• Author / Uploaded
• Haryono Yono

Citation preview

ATTACKllG CH(SS KllG'S llDIAI VOlUlf 1

DAVID VIGORITO EVERYMAN CHESS www.everymanchess.com

First published in 2010 by Gloucester Publishers plc (formerly Everyman Publishers pk), North burgh House, 10 Northburgh Street, London EC1 V OAT Copyright© 2010 David Vigorito The right of David Vigorito to be identified as the author of this work has been as­ serted in accordance with the Copyrights, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, electrostatic, magnetic tape, photocopying, recording or otherwise, without prior permission of the publisher. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN: 978 1 85744 645 6 Distributed in North America by The Globe Pequot Press, P.O Box 480, 246 Goose Lane, Guilford, CT 06437-0480. All other sales enquiries should be directed to Everyman Chess, North burgh House, 10 Northburgh Street, London EC1 V OAT tel: 020 7253 7887 fax: 020 7490 3708 email: [email protected]; website: www.everymanchess.com Everyman is the registered trade mark of Random House Inc. and is used in this work under licence from Random House Inc. This book is dedicated to two people: for my darling wife Heather, for all her love and support; and for my dear friend Joe Fang.for getting me well on my way in the King's Indian Defence with many long nights of speed chess back in the olden days...

Everyman Chess Series Chief advisor: Byron Jacobs Commissioning editor: John Emms Assistant editor: Richard Palliser Typeset and edited by First Rank Publishing, Brighton. Cover design by Horatio Monteverde. Printed and bound in Great Britain by Clays, Bungay, Suffolk.

Contents

Bibliography

5

Introduction

7

Part I: The Classical Variation 1 d4 lt:Jf6 l c4 g6 3 lt:Jc3 il.g7 4 e4 d6 S lt:Jf3 o-o 6 ii.el es 7 0-0 lt:Jc6

10

1

Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 tt:Jd3

12

2

Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 f3

38

3

Mar del Plata Variation: 9 tt:Je1 lt:Jd7 10 .ie3

47

4

Mar del Plata Variation: 9 tt:Jd2

76

s

Mar del Plata Variation: 9 b4 tt:Jhs without 10 .:tel

92

6

Mar del Plata Variation: 9 b4 tt:Jhs 10 .!:tel

105

7

Mar del Plata Variation: White's Other 9th Moves

144

8

White's Eighth Move Deviations

156

Part II: The Classical Variation 1 d4 lt:Jf6 l c4 g6 3 lt:Jc3 il.g7 4 e4 d6 S lt:Jf3 o-o 6 ii.el es without 7 o-o

167

Gligoric Variation: 7 .ie3 lt:Jg4 8 ii.gs f6 9 ii.cl tt:Jc6, 9 ilh4 tt:Jc6

168

10

Gligoric Variation: 7 .ie3 lt:Jg4 8 ii.gs f6 9 il.h4 gs

184

11

Petrosian Variation: 7 dS as without 8 ii.gs

202

9

12

Petrosian Variation: 7 dS as 8 ii.gs

222

13

Exchange Variation : 7 dxes dxes 8 '1i'xd8 .l:txd8

238

Part Ill: The Samisch Variation 1 d4 lt:Jf6 2 c4 g6 3 lt:Jc3 .ig7 4 e4 d6 S f3 o-o 6 .ie3 tt:Jc6 7 tt:Jge2 a6 8 '1i'd2 :tbs

264

14

Panno Variation: 9 h4 without 9 ... h s

266

1S

Panno Variation: 9 h4 h S 10 o-o-o

2 73

16

Panno Variation: 9 h4 hS 10 tt:J c1

293

17

Panno Variation: 9 tt:J c1

302

18

Panno Variation: 9 0-0-0, 9 .ih6 and 9 g4

320

19

Panno Variation: 9 a3, 9 !:!.bl and 9 .:tel

331

20

Panno Variation: Other lines

344

Index of Variations

3 64

B.ibliography

Books 6 ... liJc6 in the Sii.[nisch Variation, King's Indian Defense, John Watson (Chess Enter­ prises 1982) Attack ing Manual I, Jacob Aagaard (Quality Chess 2008) Beat the KID, Jan Markos (Quality Chess 2009) Beating the Anti-King's Indians, Joe Gallagher (Batsford 1996) Beating the Fianchetto Defences, Efstratios Grivas (Gambit 2006) Dangerous Weapons: The King's Indian, Richard Palliser, Glenn Flear & Yelena Dembo (Everyman Chess 2009) King's Indian Defence: 4 e4, Efim Geller (Batsford 1980) King's Indian Defence: Ma r del Plata Variation, Svetozar Gligoric (Batsford 2003) My Great Predecessors Part Ill, Garry Kasparov (Everyman Chess 2004) My Great Predecessors Pa rt IV, Garry Kasparov (Everyman Chess 2005) Play 1 d4!, Richard Palliser (Batsford 2003) Play the King's Indian, Joe Gallagher (Everyman Chess 2004) Starting Out: 1 d4!, John Cox (Everyman Chess 2006) Starting Out: The King's Indian, Joe Gallagher (Everyman Chess 2002) The Art of the King's Indian, Eduard Gufeld (Batsford 2002) The Classical King's Indian Uncovered, Krzysztof Panczyk and Jacek llczuk (Every­ man Chess 2009) The Controversial Sii.misch King's Indian, Chris Ward (Batsford 2004) The King's Indian: A Complete Black Repertoire, Victor Bologan (Chess Stars 2009) The King's Indianfor the Attacking Player, Graham Burgess (Batsford 1993) The Main Line King's Indian, John Nunn & Graham Burgess (Batsford 1996) The New Classical King's Indian, John Nunn & Graham Burgess (Batsford 1998) The Sii.misch King's Indian, Joe Gallagher (Batsford 1995) The Sii.misch King's Indian Uncovered, Alexander Cherniaev and Eduard Prokuronov (Everyman Chess 2008) Understanding The King's Indian, Mikhail Golubev (Gambit 2005)

5

A t ta c king C h e s s : The King 's In dian, V o l u m e 1 Winning with the King's Indian, Eduard Gufeld (Macmillan 1991) Winning with the King's Indian, Andrew Martin (Caissa 1989) Periodicals Chess Informant through Volume 105 New In Chess Magazine through issue 2010/4 New In Chess Yearbook through Volume 94 Electronic Resources Mega Database 2009 (ChessBase) The King's Indian, Viktor Bologan (ChessBase 2009) Chessbase.com Chesscafe.com Chess Lecture.com Chess Publishing .com The Week in Chess through issue 815

6

Introduction

The King's Indian Defence is one of the richest openings in all of chess theory. Black does not play to equalize as he does in the classical defences. Rather he seeks to unbalance the game from the outset. The last decade has seen a revitalization of the King's Indian, as even top players are often trying to win with the black pieces. Compared to the classical openings, the price of each move is quite high and a mistake by either side can easily lead to disaster. The King's Indian has always been considered a somewhat risky opening, but despite that common sentiment, the King's Indian has an impressive pedigree. While this dynamic system was pioneered in the 1950s by Russian and Yugoslav players such as David Bronstein, Efim Geller and Svetozar Gligoric, the two big names that are often attached to the King's Indian are those of its World Cham­ pion practitioners, Robert Fischer and Garry Kasparov. Whereas Fischer's retire­ ment signalled the end of his King's Indian era, Kasparov gave up our favourite opening while he was still an active player, which 'indicated' its unsoundness. At least that was the general feeling after he lost a well-known game in 1997 to Kramnik in the then dreaded 'Bayonet' system. In fact Kasparov stated something to the effect that the Sicilian and King's In­ dian were too much to keep up with at the level he was playing at, and so he stuck with the Sicilian while heading for more solid systems in the closed openings. Nowadays young players are not so worried about this; with advances in technol­ ogy many modem talents play both the Sicilian and the King's Indian, as well as other sharp defences. Opening fashions come and go. The beginning of the new millennium brought forward a great new champion of the King's Indian Defence in Teimour Radjabov. Like Kasparov, Radjabov hails from the city of Baku in Azerbaijan. Radjabov really took over where Kasparov left off, even scoring well in the aforementioned Bayo­ net (see Chapters 5 and 6). Radjabov's success influenced the younger generation as well as the old guard, and nowadays most of the top players have been found at one time or another on the black side of the King's Indian.

7

A ttac king Ch ess: Th e King 's In dian, Vol u m e 1 The King's Indian Defence has always been an opening I've felt greatly attached to. Despite the fact that I have written extensively on the Slav Defences, the King's Indian was my first real defence to 1 d4. While the King's Indian is considered to be a 'tactical' opening, I have always considered it to be very strategic in nature. It is an opening where a feeling for piece placement and pawn structure is very im­ portant. There are many thematic ideas and although the opening lends itself to frequent complications, the tactics have always seemed 'logical' to me. So, while it is true that when I 'grew up' I began to rely more on the solid Slav systems, it is always useful to have a sharp weapon available, especially when one really wants to try to win with Black. Even though the King's Indian is a complicated opening, I do not think it is so difficult to learn. For one thing, it is relatively 'move order proof'. That is, the King's Indian set-up can be employed against 1 d4, 1 c4, or 1 tt:Jf3. Also, the King's Indian lends itself to just a handful of pawn structures, so the ideas are easier to assimi­ late. In this book, as well as the second volume, I will generally focus on the main lines. The reason for this is that I think the best way to learn an opening is to study the main lines. It is easy to add other secondary systems later. The biggest exam­ ple of this is in the Classical Variation, 1 d4 tt:Jf6 2 c4 g6 3 tt:J c3 .ig7 4 e4 d6 5 lt:Jf3 o-o 6 ile2 es 7 o-o. Here I have gone for what is no doubt the main line, 7 ...tt:Jc6. This is the most uncompromising approach and it is also the most difficult. The main alternative is 7... tt:Ja6. This line is safer, easier to learn and may well be just as good as 7... tt:Jc6. In fact, I have pl ayed 7 ... tt:Ja6 myself on several occasions. Nev­ ertheless, I think it is best for the aspiring King's Indian player to learn the main lines. One great thing about the King's Indian is its flexibility - if you learn the main systems, it is easy to expand your repertoire by adding additional lines with­ out having to learn a whole new opening. In the Sa mi sch Variation, 1 d4 tt:Jf6 2 c4 g6 3 tt:Jc3 .ig7 4 e4 d6 5 f3 o-o 6 .ie3, the main line nowadays is 6... cs. This was not always the case, and 6 ...tt:J c6 and 6 ... es used to be considered Black's two main systems. I have elected not to go with 6 ... cs even though it may well be the best move. While White often steers the game into a Benoni structure, it is also possible for White to simply grab a pawn while exchanging queens as well. Modem practice has clearly shown that Black gets sufficient compen sation for the pawn, but some White players are rather well prepared in these endings. If Black is well prepared too and has a good under­ standing of these positions, he should certainly be able to draw, but I believe that heading into a pawn-down endgame straight out of the opening is probably not to everyone's taste.

8

In tro d u c tion Moreover, 6...cs against the Samisch has been very well covered in modem King's Indian literature and I did not have a lot to add to what is already out there, especially as the lack of popularity of the Samisch at high level has not produced much fresh material to examine. The classical 6 ... es must still be playable, but I always thought it was easier to play White in these lines and so decided to go for the Panno Variation with 6 ... tt:Jc6. This is an uncompromising system that still has a lot of fresh territory to explore. Note that the second volume will cover the Fianchetto Variation, the Four Pawns Attack, the Averbakh and all of White's other tries. I would like to thank a few people for their help with this book: John Emms, for suggesting the topic, allowing me to split the book into two parts and for his pa­ tience; Richard Palliser, for listening to me rant and rave about various lines that may or may not have found their way into this book; Joe Fang, for the use of his impressive library; and Vasik Rajlich, for keeping me well supplied with Rybka 3 and 4. Thanks to you all! David Vigorito Somerville, Massachusetts, October 2010

9

Part 1

The Classical Variation d4 lbf6 2 c4 g6 3 ttJc3 i..g7 4 e4 d6 s tt:Jf3 o-o 6 il.e2 es 7 o-o lbc6

1

6...tt:Ja6 and 6 ...tt:Jbd7, which aim to avoid certain lines in Part II (most nota­ bly the Exchange Variation), but in my opinion we should dive headfirst into the most critical lines. After 7 0-0 (other moves are considered in Part I I ) we will play 7 ... tt:Jc6!. Here White usually plays 8 ds (other moves are discussed in Chap­ ter Eight), and after 8 .. lt:Je7 we have the Mar del Plata Variation. .

The Classical Variation 1 d4 tt:Jf6 2 c4 0-0 6 .ie2 is one of White's most principled sys­ tems against the King's Indian Defence. White develops in a logical manner by simply developing his kingside and preparing to castle. Despite this appar­ ently modest approach, the Classical is a very complex set of systems and sub­ systems which can lead to many different types of position. We will focus on the main move, 6 ... es!. There are other moves, such as

g6 3 ltJC3 ilg7 4 e4 d6 5 tt:Jf3

10

Most of the world's top players have been found on one side or another of this position (some players like both

The

sides), because the play is very rich both strategically and tactically. In general White will play on the queen­ side and Black on the kingside, but there are many instances where one side can take the initiative on their 'weak' side. One feature that is specific to the Mar del Plata Variation is the position of Black's e7-knight. In many King's Indian lines with a blocked centre, Black's queen knight would head for the natural cs-square by ...tt:Jb8-d7-c5 or ... ltJb8-a6-cS (see, for example, the Petrosian Variation of Chapters 11 and 12). In the Mar del Plata, Black has played 7 ...tt:J c6 in order to entice White to clarify the structure in the centre with 8 dS. In general the knight on e7 is not well placed and Black's success is often connected in some way with ac­ tivating this piece. In many of the lines with 9 tt:Je1, Black brings the knight to g6. The knight is not necessarily well placed there, but it may participate in a kingside attack. Alternatively, the knight may emerge on fS (after an ex­ change of pawns on fS or e4), and sometimes Black will redeploy the

Classic a l Va ria tion

knight by playing ...'it>h8 and ...tt:Jg8. This looks funny at first, but the knight can re-emerge on f6 or even h6. There are also cases where the knight goes to c8 or even to c6. White has four basic ways to pro­ ceed in this position. He can play 9 tt:Je1, which can be considered the 'main line'. This move is considered in Chapters 1-3. A different knight rede­ ployment is 9 tt:Jd2, which is covered in Chapter 4. The dangerous 'Bayonet', 9 b4, has been White's main weapon for the last couple of decades. The play tends to be quite different here, as cen­ tral play is more prominent than it is after White's knight moves. 9 b4 is cov­ ered in Chapters 5 and 6. Other 9th moves are less common but they are not altogether harmless. These lines will be examined in Chapter 7. Some­ times too White chooses to avoid the discussion of the Mar del Plata Varia­ tion and avoids the critical 8 dS. These sidelines are not so dangerous, but Black should not ignore them, as they can be tricky to meet for the unpre­ pared. White's 8th-move deviations are covered in Chapter 8.

11

Chapt�r1· The Mar def Plata Variation .

·.

1 d4 lt:Jf6 2 c4 g 6 3 tt:Jc3 .ig7 4 e4 d6 5 tt:Jf3 0-0 6 ii.el es 7 0-0 tt:Jc6 8 ds lt:Je7 9 tt:Je1

This move quickly became White's main try when the Mar del Plata Varia­ tion first became popular and it re­ mains topical to this day. White pre­ vents Black's knight manoeuvre ...tt:Jhs, prepares to bolster the centre with f3 if necessary and avoids blocking in his c1bishop. Although the knight is passively placed at the moment, it can quickly come to d3, from where it surveys the important central squares cs, es and

12

.

.

.

f4 The knight will help force through White's important c4-cS advance and it may also drop back to the f2-square, where it will help prevent Black's ... gs­ g4 pawn break. One thing we must always be aware of is the value of Black's c8-bishop. Without this piece not only is it diffi­ cult to break with ... gS-g4, but the bishop is often needed to sacrifice itself on h3 if the kingside becomes blocked. Sometimes the bishop does not even move until it can deliver the decisive blow. So valuable is this bishop to Black's attack that we will often see White spending a number of tempi to hunt down the bishop on its original square. 9 lt:Jd7 Black clears the way for .. .f7-fS and covers the cs-square at the same time. Another plan for Black is 9 ...tt:Je8. This looks less logical, but as we shall see in the main lines following 9 ... lt:Jd7 10 tt:Jd3 fS 11 .id2 tt:Jf6 12 f3 f4, Black of...

Th e M a r de/ Plata Varia tio n : 9 tlJe1 tlJ d 7 10 tiJd3 ten drops the knight back to e8 after all in order to protect the sensitive c7square. By playing 9...t2Je8 straight­ away, Black hopes to gain a couple of tempi. Of course White does not need to be so cooperative though. For one White can continue with 10 t2Jd3 fS, but instead of 11 .id2, as played in this chapter, White may change the charac­ ter of the position with 11 f4, exploit­ ing the fact that Black's knights are far away from the es-square. Another dangerous plan is to continue along the lines of Chapter 2 by playing 10 f3 fS 11 g4, when the knight on e8 is less flexible than it would be on d7. Perhaps the greatest danger, though, comes from 10 .ie3. In this case White will have no trouble playing c4-c5 and plans involving a4-a5 become very dangerous. While 9 ... t2Je8 may be playable, the main move has always been 9 ...l2Jd7 and most experts consider it best.

From here White has three main schemes to choose between. In this chapter we look at the flexible 10 t2Jd3.

The next chapter examines the radical plan 10 f3 fS 11 g4! ?, while Chapter 3 covers White's most aggressive con­ tinuation, 10 .ie3. 10 t2Jd3 White improves the position of his knight and prepares the c4-c5 advance. This is a very straightforward continua­ tion and some of the theory goes very deep in the main lines. The fashion of such long lines comes and goes, and eventually White turned to other sys­ tems in the Classical (most notably 10 .ie3 and the Bayonet), while Black also found ways to deviate. Recently 10 t2Jd3 has become rather fashionable again. We will generally study the main lines because that will help to develop a good general King's Indian under­ standing. There are also some interest­ ing sidelines that will be mentioned that may be worthy of further research. 10 f s 11 .id2 ...

This is the main line. White simply develops his bishop and prepares !:!.cl. There are a couple of alternatives. a) First, let us take a look at 11 f3.

13

A ttacking Ch ess: The King 's In dian, Vo l u m e 1 will have to spend an additional tempo to break with c4-cS. This move does have some historical significance, how­ ever. Now 12 ... gs 13 b4

This is quite a natural reaction to Black's pawn advance, but sometimes this move is considered inaccurate. This is because Black can begin his kingside play while leaving his knight on d7 to hold up White's cs-break. Nevertheless, this move is not bad and some even consider it to be the most accurate move because it avoids the line 11 .id2 fxe4!?, which we discuss in the notes to Black's 11th move, below. White may also steer the game into independent channels by avoiding .id2 altogether. Here Black can play 11 .. .f4 right away, because there is no .ie2-g4 to worry about. 11 . ..tt:Jf6 is possible too, but Black can try to take advantage of White's move order by leaving the knight at d7 to hinder White's cs­ advance. Note that 11 ... 'it>h8?! is not so good because White can switch gears with 12 .ie3!, aiming for positions akin to those in Chapter 3 where 11 ...'it>h8 is likely to be a wasted tempo. After 11 .. .f4 White has: al) 12 .id2 looks like an inferior version of the main line because White

14

13 ...lt:Jf6 14 cs tt:Jg6 1S .l:tc1 .l:tf7 compares favourably to the main line for Black because White has played b4 (instead of tt:Jbs or cxd6, for example) which may or may not tum out to be a useful move. Black's last few moves seem rather obvious today, but this was not always so. An old plan for Black in such posi­ tions is to play ....l:tf6 and ... .rf.h6, with the idea of directly attacking down the h-file with ...'1i° e8 and ...'ifh s. If White plays h2-h3, Black will try to engineer a ....ixh3 sacrifice. Although this plan can indeed be quite dangerous to the first player, to me it seems a bit too simplistic to think that White's position can just be taken by storm like this. Often Black will have to worry about the possibility of White playing tt:Jbs after the queen has gone to e8. He can deal with this by throwing caution to the wind and allowing tt:Jbsxe7. This

Th e M a r de/ Plata Va ria tion : 9 li.Je1 li.J d 7 10 li.Jd3 sounds like fun, but if White can play li.Jf2 and h3, the h-file attack will not work. In this case, Black can change the direction of the attack by then playing ... .l:tg6 and ... hs. Another way for White to defend is by playing 'it>h1 and .ig1. Black does not really have any way to increase the pressure on h2, but if he could get a knight to g 3, it would be mate. This can be achieved either by moving the black queen from h S to threaten ...li.Jf6-hS or by a sacrifice such as ... li.Jf6xdS; exds li.Je7-fS. This all sounds very nice, but if White is alert to the possibility it is not so easy to achieve these tricks. White may also defend against mate on h2 in some positions with his bishop on f2 by playing the ugly h4. This may look fatally weakening, but it is not always easy to get through and Black's queenside will certainly disap­ pear. Black can also throw in ... a6 before moving the queen to safeguard against the knight raid to bS, but this costs a tempo and also leaves Black vulnerable to li.Ja4-b6 ideas after an exchange of pawns on d6. In general I have avoided lines where Black seeks to attack in this way, but there are exceptions as we shall see, most notably in Line C of Chapter 3. Another attacking plan, which seems more natural to me, is to play ... li.Jf6 and ... hs. Gligoric was one of the first players to develop this concept. One of his games continued (after

1s ... :tf7 above) 16 li.Jf2 h S 17 cxd6 cxd6 18 a4 .if8! 19 as .l:tg7 20 h3 li.Jh8!.

As we shall see, Black's 18th and 19th moves are extremely common­ place, but the last one requires a few words of explanation. The g6-knight often heads to h4 to both unleash the g7-rook and scare the white king. The knight manoeuvre to h8 is not so un­ common, though, and the knight will then come back to f7, from where it may hop to gs or even to h6, to further support the ... g4-break. We will even see positions where Black plays ... 'it>h8 and ... li.Je7-g8 in order to come to h6. The game continued 21 li.Jbs g4 22 fxg4 hxg4 23 hxg4 a6 24 li.Ja3 .id7 2S li.Jc4 .l:tc8 26 li.Jb6 .l:txc1 27 .ixc1 .ie8 28 .ia3 li.Jf7 29 '1i'c2 li.Jh6 30 gs .l:txgs 31 !:i.c1 .l:tg3 32 .ib2 li.Jfg4 33 li.Jxg4 li.Jxg4 34 .ixg4 .l:txg4 3S '1i'f2 .ig6 36 .l:tc4 '1i'e7 37 .ic3 '1i'h7 38 '1i'e2 .l:Ih4 39 'it>f2 f3 40 '1i'e3 .l:tf4 41 gxf3 '1i'h2+ 42 'it>e1 '1i'h1+ 43 'it>e2 .ihs 44 'it>d2 .:txf3 4S 'iY'gs+ .ig7 46 'it>c2 !:i.f2+ 47 .id2 '1i'd1+ 48 'it>c3 '1i'a1+ 0-1, M.Najdorf-5.Gligoric, Mar del Plata 19S3. Najdorf learnt a lesson

15

A tt a c king C h e s s : Th e King 's In dian, Vo l u m e 1 from this game as we shall see in Line B of Chapter 3. Going back, White probably can re­ frain from 13 b4 and play 13 !:!.cl tt:J g6 14 cs!? when instead of grabbing the pawn, Black should really just play 14 ...tt:Jf6 to tran spose to the main lines, while 13 ... hs is also possible. Another idea is 13 g4 when Black can pl ay 13 ... hs 14 h3 !:i.f6!? (Black tries to turn the delay in going ...tt:Jf6 to his benefit) 1S b4 (White should probably prefer a quick ii.el and 'it>g 2) 1S ... !:i.h 6 16 'it>g2 tt:Jg6 17 !:i.hl lt:Jh4+ 18 'it>f2 .if8 19 !:!.cl as 20 a3 c6, which gave him good play in R.Shabtai-D.Komljenovic, Biel 1989. a2) 12 g4 is quite a common counter in general - see, for instance, Line D of this chapter. Here, however, it does make a difference that Black has not played either ...tt:Jf6 or ... gs yet.

White's thrust is well met by 12 ....if6 !. Black's delay in playing ... tt:Jf6 allows this move, seizing control of the h4-square. After 13 .id2 hS 14 h3 .ih4 (alternatives are 14... �7 and

16

1 4...'it>g 7!?) 1 S ii.el .ixel 16 tt:Jxel 'it>g7 17 h4?! gs (or 17 ... hxg4 18 fxg4 .l:th8 19 tt:Jg2 gs 20 hs tt:Jf6) 18 tt:Jg2 hxg4 19 fxg4 gxh4 20 tt:Jxh4 .l:th8 Black had a strong attack brewing in S.Karp­ A.Kuzmin, Ostend 1991. a3) 12 b4 (in the main line we will see that this preparatory move usually is not necessary; the only way this move makes sense for White is if he intends to deploy the bishop on a3 or b2) 12 ... gs (Black has also tried to take advantage of White's move order with 12 ... as, but I think Black should just leave the queenside alone and get on with it) 13 cs tt:Jf6 14 a4 and now:

a31) 14...'it>h8 ! ? (here we see an­ other typical plan; Black gears up for a quick ...g4, rather than play ....l:If7-g 7, which makes some sense because White has refrained from .id2 and !:!.cl, so the c7-square does not really need additional protection) 1S .ia3 .l:tg8 16 as .if8 17 cxd6 cxd6 18 bS (or 18 h3 hS 19 tt:Jf2 when Black could try 19 ....l:I g7 20 !:!.cl tt:J eg8, with the idea ...tt:J h6 and ...g4) 18 ... g4 and Black had

Th e M a r de/ Plata Va riatio n : 9 li.J e1 li.Jdl 10 li.J d3 good play in S.Lagrotteria-B.Socko, Saint Vincent 2001. a32) The direct 14...h s is also possi­ ble and after 1s li.Jf2 li.Jg6 16 as .l::i. f7 17 c6 .if8 18 cxb7 .ixb7 19 !:i.a3 .l:tg7 20 h3 .ic8 21 bS li.Jh8 22 li.Ja2 li.Jf7 23 li.Jb4 li.Jh6 24 li.Jc6 '1i'e8 2S '1i'c2 g4 Black had good counterplay in G.Vescovi­ M.Krylov, Moscow 2010. b) The old move 11 exfs is not seen too often these days. The simplest solu­ tion is 11 ...li.Jxfs. Usually in the King's Indian Black prefers to take with the f­ pawn, but here Black immediately solves the problem of the e7-knight and has good piece play. That said, 11 ... gxfs 12 f4 li.Jg6 is also possible. Af­ ter 11 ...li.Jxfs 12 f3 li.Jf6 13 li.Jf2 Black has:

bl) 13 ....ih6 14 .ixh6 li.Jxh6 and now 1S g4 was suggested by Hort, to prevent the knight from coming to d4, but after 1s ...li.Jf7 Black looks to be fine in any case. b2) 13 ... c6 was Nunn's recommen­ dation. However, 14 li.Jfe4 li.Jxe4 (instead 14... cxds 1s li.Jxf6+ .ixf6 16 li.Jxds

.igs 17 .id3 gave White a slight edge in F.Doettling-A.Hunt, Patras 1999) 1S li.Jxe4 'ii'b6+ (1S ...cxds 16 'iY'xds+) 16 cs!? dxcs 17 .ic4 gives White danger­ ous compensation for the pawn . b3) 13 ... cs 14li.Jfe4 li.Jd4 1 s .igs .ifs looked fine for Black in T.Overbeck­ V.Rojicek, Pardubice 2009. b4) 13 ... li.Jd4 is simple enough: 14 li.Jfe4 c6 1S .ie3 li.Jxe4 16 li.Jxe4 cxds 17 cxds .ifs was equal in R.Markus­ R.Polzin, Austrian League 2006. Now we finally return to the main line, 10 .id2. 11 li.Jf6 At the cost of 'letting go' of the cs­ square, Black induces White to play f3 before advancing on the kingside. Note that the immediate 11 .. .f4? is a big strategic mistake because after 12 .ig4 Black will be hard pressed to avoid the exchange of the light-squared bishops. Black does have a couple of other ideas here, though: a) One possibility is 11 ...'it>h8!?. •••

This is Black's main alternative to the main lines. If Black wants to avoid

17

A ttac king Ch ess: The King 's In dian, Vo l u m e 1 the long main lines of 'A' and 'B', the complications of '(' and the blocked positions of 'D', this is a decent alterna­ tive. Instead of fixing the kingside, Black makes a useful, semi-waiting move. Playing 11 ... 'it>h8 does a few things. The king gets off of the a2-g8 diagonal, which may seem rather vague, but in the main lines we will see a few situations where this could be useful. Black also moves his king off of the g-file. This could be good or bad we will come across positions where after multiple exchanges on g4, a white rook or queen ends up sliding over to h3. Perhaps the biggest advantage to 11 ... 'it>h8 is that the g8-square is cleared for Black's knight. Playing ... lt:Je7-g8 creates the possibility of play­ ing ....�h6 and clears the h4-d8 diago­ nal for both Black's queen and g7bishop. This can be useful if White adopts a plan involving g4. And, lastly, by leaving the knight on d7, Black hopes to make White spend an extra move preparing the cs-advance. Perhaps the most natural continua­ tion is 12 !:!.cl. White hopes to get a favourable version of the main lines by opening the c-file. Now 12 ... tt:Jf6 13 f3 f4? just allows 14 cs and thus is a worse version of the main lines, be­ cause Black has played ... 'it>h8 rather than ... gs. In stead Black usually plays 13 ... cs to try and make the move .l:tc1 look irrelevant, while 13 ...tt:J eg8 and 13 ... c6 are altern atives. b) Another idea is 11 .. .fxe4!?. This

18

looks too simplistic, but it i s not easy for White to prove an advantage. Mar­ kos, for one, thinks enough of this reac­ tion to avoid playing 10 tt:Jd3 alto­ gether. After 12 tt:Jxe4 tt:Jfs (12 ... tt:Jf6 looks less accurate because Black would like to recapture on f6 with the queen) it seems as though Black has lost time compared to the line 11 exfs tt:Jxfs, but there White generally prefers to play f3, tt:Jd3-f2 and tt:Jfe4. White's knight moves have also left him with less control of the dS-square, so a ...c6break may be appealing.

Now 13 f3 tt:Jf6 14 tt:Jdf2 c6! (14... lt:Jd4 is also okay) looks fine for Black. One example: 1S .id3 cxds 16 cxds 'it>h8 17 tt:Jxf6 .ixf6 18 lt:Je4 '1i°b6+ 19 .l:tf2 .ih4 20 g3 .�e7 with a good game for Black in J.Sofrevski-L.Portisch, Skopje 1968. A better idea is 13 .ic3, although this is not too terrifying either: 13 ... tt:Jf6 14 .if3 lt:Jh4 (14...tt:Jxe4 15 .ixe4 '1i°h4 16 .:tel b6 17 g3 looks a little better for White) 1S tt:Jxf6+ '1i"xf6 16 .ie4 .�fs 17 '1i°e2 .ixe4 18 '1i"xe4 'iY"fs 19 .l:tael was

Th e M a r d e/ Plata Va riation: 9 li.J e 1 li.Jd7 10 li.Jd 3 agreed drawn here in P.Lukacs­ T.Radjabov, Budapest 2000. After 19 ... .l:tae8 20 f3 'iY"gs 21 '1i"e2 li.Jf5 22 li.Jf2 ii.. h6 Black should be able to hold the balance. Overall, 11 .. .fxe4 looks very solid, al­ though it is a little dull. These lines are interesting and look quite playable. I think that after learn­ ing the main lines, these alternatives would make a useful addition to a player's repertoire, but we will chiefly stick to the main line, 11 ...li.Jf6. 12 f3 f4 Here 12 ... 'it>h8?! would make little sense, as White could proceed with 13 cs immediately.

With the centre closed, the battle lines are clear - White will play on the queenside, whereas Black will try to attack the white king. Here the way divides: A:

13 cs gS 14 cxd6 cxd6 1S l:c1

B: 13 cs g s 14 cxd6 cxd6 1 s tDf2 C: 13 cs g5 14 :C1 liJg6 15 %51? D:Hg41?

Usually White just proceeds with his own play, and in Lines A and B we look at White's traditional continua­ tions where we have a typical King's Indian 'race', with both sides trying to blow the other away. Line C examines a radical idea that has become very popular of late. Finally, Line D looks at a completely different plan where White looks to slow Black down on the king­ side before pursuing his queenside ambitions. The position becomes blocked and the play greatly slows down. A) 13 cs

This is the main line. White imme­ diately initiates queenside play, trust­ ing in his chances in the ensuing race. He will decimate the black queenside while Black tries to mate the white king. Although White's play is usually faster, Black is playing for higher stakes. 13 gs There is no need for Black to be sub­ tle at this point. White will now bring a ...

19

A ttacking Chess: Th e King 's I n dian, Vo l u m e 1 rook to the c-file, but he has a choice as to how to go about it. 14 cxd6 White immediately opens the c-file. The preparatory 14 l:tcl will be seen in Line C. 14...cxd6

We will see this multi-purpose move again and again. Black defends the c7-square in advance, and prepares ... 3'.f8 and ... l:tg7. These moves will help him prepare ... g4. The bishop on f8 is often well placed - from there it pro­ tects the d6-pawn and sometimes the bishop will come to e7 from where it could join a kingside attack (with ... 3'.h4 or ...3'.gs), or emerge on the queenside ( ... 3'.d8). Note that 16 ... 'ii'b 6+?! doesn't win a piece; rather it just wastes time be­ cause White is happy to play 17 tllf2 when the bS-knight is suddenly pro­ tected by the e2-bishop. 17 'ikc2

15 !tel White's main choice, taking the c­ file immediately. Instead 15 tllf2 pre­ pares to bring the king's rook to cl. This is Line B.

1s ...tllg6 16 tllbs The knight threatens to invade on c7, from where it will go to e6. 16...l:tf7

20

White again threatens tll c 7-e6, so Black has to retreat his knight, making the ... g4 advance more difficult to achieve. Also the tactical shot tllxa7 is sometimes possible, exploiting the loose bishop on c8. 17 ...tlle8 18 a4 This move is always useful for White to include. It gives the bS-knight extra support and should Black kick the

Th e M a r def Plata Varia tio n : 9 t'De1 t'Dd 7 1 0 t'Dd3 knight back with ... a6, then t'Da3 will be played when ...bs is prevented, and White will often thr�aten l'Dc4 and 3'.as (or a4-a5), with a grip on the b6-square. If White can play t'Db6, he will be able to eliminate Black's important light­ squared bishop. Black will often feel compelled to play ... a6 to drive the knight away anyway, but usually he makes some useful moves first 1s hs Black threatens ...g4 again. 19 C'Df2

may also improve the bishop by playing ... 3'.e7. For a while 19 ... 3'.d7 supplanted 19 ...3'.f8, but I think the two moves are of approximately equal value.

...

White prevents ... g4 and increases the scope of his light-squared bishop. It is too early for ...a6 because of t'Da3-c4, followed by as and t'Db6. The e8-knight has to defend c7, so Black has to make a choice between two bishop moves. We have: A1:19..:..lf8 A2t 19.- .t.d:J'rf •

Al) 19 3'.fS This is the traditional move. Black prepares ... l:tg7, overprotects d6 and ...

20 h3 This prophylactic move is the most common. Instead 20 b4 is not terribly useful, but it does set a positional trap. After 20 ... l:tg7 (Black could consider 20 ... a6) 21 t'Dxa7 Black should avoid 21 ... l:tc7? because of 22 t'Dc6! bxc6 23 dxc6 when White had more than enough for the piece in N.Rashkovsky­ E.Gufeld, Daugavpils 1978. Instead Black can play 21 ... 3'.d7! with a good version of the pawn sacrifice discussed below. 20 'ikb3 is a tricky move order: 20...l:tg7?! (20 ... 3'.d7 transposes to Line A2, but Black should consider 20...l'Df6 too) 21 l:tc2! when White can omit the prophylactic h2-h3 and double on the c-file immediately. White does have an important al­ ternative in 20 t'Dxa7. This grabs a pawn, but Black has a counter available which regains the

21

A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1 pawn if he is willing to go into an endgame. One would think that this would favour White, but actually Black equal­ izes pretty easily. Still, most black play­ ers have preferred to sacrifice a pawn here to keep playing for an attack. Black has:

a) 20.. .lk7 forces White's reply and mass exchanges follow: 21 .Jtas l::t xc2 22 .Jtxd8 l::t x e2 23 'Llxc8 l:txa4 (worse is 23 .. Axb2 24 as! with hopes of an edge) 24 'Lld3 (24 .Jtxgs l:txb2) 24...g4 2s l:tf2 l::t e3! 26 'Llel l::t a8 27 l:tfc2 l:tb3 28 'iii> f2 l:ta2 29 l::t b1 'iii> f7 30 'ite2 .lte7 31 'Llxe7 'Llxe7 32 'Lld3 l::t a8 33 .Jtxe7 'itxe7 was completely equal in l.Novikov-1.Glek, USSR 198S. b) 20....Jtd7!? leads to a more full­ blooded struggle, but it is of course riskier: 21 'Libs l:tg7 22 h3 'Llh4 23 'ii b 3 'iii> h 8 sees Black patiently make one more little preparation for ...g4.

(seefollowing diagram) The main point of this move is to get off the a2-g8 diagonal in anticipa-

22

tion of an eventual 'Llc7-e6.

The importance of this detail can be seen after 23 ... g4?! 24 fxg4 hxg4 2S hxg4 'Llf6 26 'Lle7! (much better than 26 .Jte1 'Llh S! which gave Black coun­ terplay in L.Ftacnik-A.Sznapik, Baile Herculane 1982) 26 ... 'Llxg4 27 'Lle6 'Llxf2 28 l;Ixf2 .Jtxe6 29 dxe6 'iii> h 8 30 l:tc3 .lte7 31 l::t h 3 when White was much better in L.Ftacnik-M.Vokac, Czech Championship, Frenstat 1982. After 23 ...'ith8 White has tried both 24 l::t c 3 and 24 as. In both cases Black should sacrifice a pawn with 24... g4, when in practice he has been rather successful. This is all very nice, but I suspect that 20 ... l::t c7 is sounder. 20 l:tg7 Instead 20 ... .Jtd7 21 'ikb3 l:tg7 will likely transpose to Line A2. Also possible is 20...'Llh4. Then 21 'Llxa7 .ltd7 (note that 21...l:tc7? is not good here, because after 22 .Jtas l:txc2 23 .Jtxd8 l::t x e2 24 'Llxc8 l:txa4 2S .Jtxg s ! White wins a pawn and a tempo) 2 2 'Li b s l:tg7 23 'ii b 3 is the pawn sacrifice in variation 'b' above. ...

Th e M a r def P lata Varia tio n : 9 CDe 1 CDd 7 1 0 CDd3 21 'ikb3 White prepares to double rooks on the c-file. The queen is also useful here for defence - after mass exchanges on g4, she may slide along the third rank. Sometimes White plays l:tc3 (or l:ta3 in certain positions) to utilize a rook in this fashion too. White can also change his mind and take the pawn after all with 21 ttJxa7, when 21 ... 3'.d7 22 lLibS is again varia­ tion 'b' above. Sounder is 21 .. Jk7 22 l.as which leads to the same endgame as before with White having the extra move h3. This does not necessarily help him: for example, 22 ...l::t x c2 23 3'.xd8 :xe2 24 ttJxc8 l:txa4 2S ttJd3 g4 26 l::t f2 :e3 27 lLi el g3 was fine for Black in O.Averkin-G.Kasparov, Moscow 1979. Another idea is 21 as CDh4 (21 ... 3'.d7!? 22 'ikb3 lLih4 is probably better) 22 ttJxa7 ! ? l:tc7 (22 ...3'.d7 ! ?) 23 .'Llc6! bxc6 24 dxc6 l:tg7 2 S b4 lLic7 26 bS lLixbs ! ? 27 3'.xbs g4 and now in­ stead of 28 hxg4 CDxf3+! 29 gxf3 't!Vh4, which was unclear in A.Jakubiec­ K.Chojnacki, Lublin 2009, I suspect that 28 fxg4 hxg4 29 hxg4 3'.xg4 30 ltJxg4 l:txg4 31 l:tf2 would test Black's play. 21...ttJh4 22 l:tc2 a6 Black plays this now that he will be able to force through ... g4. Alterna­ tively: a) 22 ... g4?! (the immediate advance is premature) 23 fxg4 CDf6 24 3'.el hxg4 2S hxg4 ctJhS was played in V.lvanchuk1.Cheparinov, Sofia 2008. Here 26 CDdl! 3'.xg4 27 3'.xg4 l:txg4 28 'ikh3 'ikg s 29

ltJc7 f 3 30 3'.xh4 l:txh4 31 'ike6+ 'iii> h8 32 l:txf3 gives White a big advantage ac­ cording to Mikhalevski. b) 22 ... lLif6 ! ? is possible, though: 23 l:tfcl g4 24 fxg4 hxg4 2 S hxg4 3'.xg4 26 3'.xg4 ttJxg4 27 ttJxg4 l:txg4 has been play a few times and seems okay for Black. Then 28 3'.el f3 ! 29 3'.xh4 'ikxh4 30 'ikxf3 l::t xe4 31 l:tc3 (31 g3 'ikg4 32 'ikxg4+ l:txg4 was at least equal for Black in K.Langeweg-J.Donner, Amster­ dam 1971) 31 ... l::t el+ 32 l:txel 'ikxel+ 33 'ikfl 'ikxfl+ 34 'itxfl a6 3 S lLic7 l::t c 8 36 tLie6 l:txc3 37 bxc3 b6 38 lLic7 as 39 'ite2 3'.e7 1/2-1/2 was J.Knap-V.Tasic, cor­ respondence 2006.

23 ttJa3 lLif6 Another possibility is 23 ...3'.d7, which actually transposes to the main line of Line A2. 24 3'.e1 This move sizes up the h4-knight and clears the second rank for defence. 24...g4 25 hxg4 Instead 2S fxg4 hxg4 26 hxg4 (26 ttJxg4?! ttJxe4) 26 ...ttJxg4 should trans­ pose, but White should certainly avoid

23

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 the temptation of 27 l:txc8?.

32 l::tf3 3'.e7

He hopes to win two pieces for a rook, but Black has the trick 27 ...'Llxf2! 28 3'.xf2 (28 l:txd8 l:txg2 mate) 28 ... l:txc8 29 't!Vh3 'Llxg2 30 3'.g4 'Lle3 31 3'.xe3 'ikd7 and 0-1 in V.Talla-M.Szelag, Us­ tron 2008. 2s ...hxg4 26 'Llxg4 'Llxg4 The alternative 26 ...'Llhs may be playable, but it looks sketchy to me. Instead of 27 as 'Llg3 28 3'.xg3 fxg3 29 'ikb6 'it e7 30 l:tfc1 3'.xg4 31 fxg4 'Llg6 32 ilf3 'ikh4 33 �1 'Llf4, when Black had plenty of counterplay in L.Ftacnik­ Z.Polgar, Trencianske Teplice 1985, White should play 27 'Llc4!. Now 27 ... 'Llg3 28 3'.xg3 fxg3 29 'ikb6 'ike7 30 'Llce3 l:th7 31 l:tfc1 has been played in a couple of correspondence games. The position is messy, but I think Black is better off with the text move. 27 fxg4 3'.xg4 28 3'.xg4 l:txg4 29 'ikh3 This is a good square for defensive reasons as well as for a possible coun­ terattack. However, Black can hold the balance. 29...'ikgs 30 3'.xh4 l:txh4 31 'ike6+ 'ith8

This position has been reached a few of times in practice and Black has enough play. White's knight is far away and his pawn structure is a little loose. One game continued 33 l:th3 l:tg8 34 l:txh4+ 'ikxh4 35 l:tc3 'ikel+ 36 'ith2 l:txg2+ 1/2-1/2, l.Schuett-V.Conti, corre­ spondence 1999.

24

A2) 19...3'.d7

This is a more modem move. Black avoids any 'Llxa7 tricks and sometimes even plays on the queenside himself. Of course, the main plan is still a kingside attack

Th e M a r def Plata Varia t i o n : 9 tDe1 lDd 7 1 0 CDd3 20 'ikb3 White can also play 20 h3 3'.f8 21 'Wb3 which will transpose. 20 3'.fS ...

Black has also tried 20 ...3'.f6 ! ?, which is not without its logic. Black often plays ... 3'.f8-e7 and then sometimes either ... 3'.h4 or ...3'.d8, so it makes some sense to put the bishop on the h4-d8 diagonal in one move instead of two. White has: a) With 21 3'.b4 White decides to take advantage of Black's last move to attack d6 immediately, but he should probably wait on this, as it runs into some tactical problems: 21 ... 3'.e7 22 :te2 a6 23 CDa3 g4! 24 fxg4 hxg4 2S 3'.xg4 3'.xg4 26 ttJxg4 'ikb6+ 27 'iii> h l as 28 ctJh6+ 'iii> g7 29 ttJxf7 'ikxb4 with complications that were not unfavour­ able to Black in V.Neverov-M.Vokac, Bled Olympiad 2002. b) 21 l:tc2 'ikb8!? (consistent; Black wants to play ... 3'.d8 quickly) 22 as 3'.d8 23 l:tfcl (the prophylactic 23 ltJa3! ? is another idea when 3'.e2-bs is a possi­ bility and there are ideas like CDc4, 3'.b4

and 'ika3, laying siege to the d6-pawn) 23 ... a6 24 CDa3 bS gave Black counter­ play in S.Zilka-R.Ramondino, Vienna 2009. 21J::tc2 This is very natural. White doubles rooks on the c-file. Others: a) 21 h3 and now 21 ... l:tg7 will gen­ erally transpose elsewhere, depending where White moves his rook, but Black also has the option of making some manoeuvres on the queenside: al) 21 ... 'ikb8!? 22 as l:tg7 23 l:tc3?! (this leaves the as-pawn without sup­ port; 23 l:tc2 is more natural and could transpose to the next note) 23 ... 3'.e7 24 l:tal (not a nice move to make) 24 ... 3'.d8 2S l:tccl lLih8 ! ? (a typical idea; Black routes the knight to h6 to support the ... g4-break) 26 ttJa3 a6 27 CDc4 3'.bs 28 'iii> fl CDf7 29 3'.el ltJh6 30 lLib6 3'.xb6 31 axb6 3'.xe2+ 32 'itxe2 'ikd8 33 CDd3 g4 gave Black the initiative in M.Roeder­ M.Hebden, Bern 1992. a2) 21 ... 3'.e7!? 22 l:tc2 'ikb8 23 l:tfcl 3'.d8 24 as a6 2 s ltJa3 bS 26 3'.e1 (26 3'.b4 l:tg7 27 'iii> fl lLif6 28 'ikd3 lLih8 29 lLibl CDf7 was similar in M.Vukic­ D.Sahovic, Tuzla 1983) 26 ...lLif6 27 'ikd3 l:tg7 28 3'.d1 ctJh8! left ... g4 imminent in A.Olcayoz-E.Grivas, Mangalia 1992. b) 21 l:tc3 was played in S.Conquest­ P.Thipsay, British Championship, Southampton 1986. It guards the third rank (in the event of exchanges follow­ ing ... g4), and leaves the possibility of 'ikb3-d1 open, but compared to 21 l::t c 2, White limits his d2-bishop - an a4-aS

25

Attacking C h e s s : The K i n g 's I n dian, Vol u m e 1 advance will not be supported and .Jtb4 is not possible. c) 21 J::t c4 is a bit strange looking be­ cause there is no 'Lla3-c4 possibility: 21...a6 22 'Lla3 (again, there is no 'Llc4; perhaps White was hoping to use the rook on the fourth rank with l:tb4 to pressure the b7-pawn) 22 ...l:tg7 23 as 'Llf6 24 'ikb6 'ike8 2s h3 (2s 'ikxb7? .Jtbs) 2S ...g4 26 fxg4 hxg4 27 hxg4 'Llh4 28 l::t c7 'ikg6! 29 .Jte1 (or 29 'ikxb7 l::t d 8 30 �xa6 ii.xg4 31 l::t x g7+ .Jtxg7 32 'Llxg4 'Llxg4 33 .Jtxg4 'ikxg4 34 l:tf2 f3 3S �d3 l::t f8 with a decisive attack according to Gallagher) 29 ...l::th7 30 'ii b3 (30 'ikxb7 'ii h6! 31 'Llh3 'Llxg2 wins) 30 ... 'Llxe4 31 'ikd3? (after 31 'ikxb7 'Llg3! 32 .Jtd3 .Jtfs! 33 l::t xh7 'ikxh7 34 'ikxh7+ .Jtxh7 Black wins the exchange, but this was proba­ bly the best chance) 31 ...'Llg3 32 l::t xb7 was Z.Kozul-T. Radjabov, Sarajevo 2003.

Without White having spent a tempo on h3, 21 ...'ii b 8?! looks too slow: 22 l::t fcl a6 23 'Lla3 .lte7 24 as .Jtd8 2S 'Llc4 .Jtbs 26 .ltb4 b6? was G.Kacheish­ vili-D.Sharavdorj, Lubbock 2009, and here White could have played 27 'Llxd6! 'Llxd6 28 .Jtxd6 'ikxd6 29 .Jtxbs axbs 30 l:tc6, winning the knight on g6. 22 'Lla3 l:tg7 22 ...'Llf6 23 h3 l::t g7 is the same. 23 h3 'Llh4 Black has to time his pawn break carefully: 23 ...'Llf6 24 l:tfc1 g4?! (24...'Llh4) 2S fxg4 hxg4 26 hxg4 'Llh4 27 l::t c 7! .lte7 28 .Jte1 was much better for White in L.Ftacnik-G.Ligterink, Am­ sterdam 1977. If 28 ...'Llxg4 then 29 .Jtxg4 .Jtxg4 30 'Llxg4 l::t xg4 31 't!fh3 l::t x g2+ 32 'iii> h l wins for White. 24J::tfc1

24 'Llf6 Again 24... g4?! is premature: 2S fxg4 hxg4 26 hxg4 l:tc8 (or 26...'Llf6 27 l:tc7, transposing to 23 ...'Llf6, above) 27 l::t xc8 .Jtxc8 28 'Llc4 'Llf6 29 'ikh3 with a clear advantage in N.Rashkovsky­ A.Vitolinsh, Daugavpils 1978. .•.

Here Gallagher points out that 32 ... .Jtfs! 33 l:txh7 (33 gxfs 'Llxe2+ 34 'ikxe2 'ikxg2 mate) 33 ... .Jtxd3 34 .Jtxd3 'itxh7 3S .Jtxg6+ 'Llxg6 gives Black a clear advantage. 21 a6 ...

26

Th e M a r def Plata Va ria tio n : 9 tDe1 lDd 7 1 0 lDd3 Black also does not need to commit to 24 ... l:tb8 yet. After 2S CDC4 b6 26 as! g4 27 fxg4 lLif6 28 ttJxb6 hxg4 29 hxg4 .!Lixg2, as in V.lvanchuk-J.Timman, Hilversum 1991, White could play 30 'itixg2! ltJxg4 31 3'.xg4 3'.xg4 32 �1 3'.d7 33 'ite2 when he should win. 2S 3'.e1 Alternatively, 2S 'ikxb7? runs into 2s ... ttJxg2!, while 2S l;Ic7 can be met by 2s ... l:tb8 26 as ltJe8 27 l:t7c2 lLif6 28 l:tc7 !lJe8 29 l:t7c2 lLif6 when H.Gruenberg­ P.Hesse, Eilenburg 1984, ended in a draw. 2s ...g4 Black has made the necessary preparations and now breaks. 26 fxg4 hxg4 27 hxg4 3'.xg4 28 3'.xg4 !Llxg4 29 ttJxg4 l:txg4 30 'ikh3 'ikgs

This position was agreed drawn in F.Quiroga-E.Maggiolo, Buenos Aires 2000. White was quite a bit higher­ rated here, which shows the health of Black's position. The position is very similar to the one we saw in Line Al, but here Black has an even better ver­ sion.

B) 13 cs gs 14 cxd6 cxd6 1S ttJf2

Before playing 'ikc2 and �fcl, White shores up his defence of the g4-square. 1s ...ttJg6 Black could also begin with 1s ...h s t o try t o force White to play h3. I f 16 'ikc2 (16 h3 ttJg6 17 'ikc2 �f7 18 l:tfcl transposes the main line), then 16 ... g4! ? (16 ...l:tf7 11 ttJbs ttJg6 18 l:tfc1 ltJe8 19 a4 allows White to delay h3) 17 lLibS g3 is an untried possibility. 16 'ikc2 l:tf7 17 l:tfc1 hs After 17...ltJe8 18 a4 hS White went backwards with 19 CDcdl ! ? and won a long struggle after 19 ... 3'.f8 20 l::t a 3 in L.Aronian-H.Nakamura, World Team Championship, Bursa 2010. The text move avoids this possibility because then 18 ttJcdl could be met with 18 ... g4!. Black has also tried 17 ...a6, but after 18 a4 hS 19 h3 3'.f8 20 as we transpose to 18 ... a6 19 a4 3'.f8 in the note to Black's next move, which looks unsatis­ factory for him. 18 h3 It is important for White to hold up

27

A ttacking C h e s s : The King 's I n dian, Vo lu m e 1 ... g4 as long as he can. He can, though, also play the immediate 18 'Libs here. After 18 ...'Lle8 19 a4 3'.f8 or 19 ...3'.d7 the play is similar to the main line and may even transpose.

18 3'.fS!? Black stays calm. Practice has also seen him take immediate measures on either flank: a) 18 ... a6 stops 'Li bS, but it does weaken the queenside and White has other ways into the black position after 19 a4!. Now 19 ...b6 is considered bad after 20 l:ta3 (with the idea of 'Lla2 and l:tb3 or l:tac3, while 20 'Lla2 is also pos­ sible) 20...3'.f8 21 'Lla2 ! . The alternatives are supposed to be better, but it all just seems rather shaky to me: al) After 19 ...'Llh4 20 'ikdl g4 21 fxg4 hxg4 22 hxg4 'Llh7 23 as 'Ll gs 24 .i.f1 bS 2S axb6 'ikxb6 26 'Lla4 'ika7 27 3'.as ! l:tb7 28 l:tc6 l:tab8 29 3'.b6! White was winning in B.Lalic-L.McShane, Southend 2000. a2) Instead 19 ...3'. f8 runs into 20 as (with the idea 'Lla4-b6) 20 ... g4 21 fxg4 hxg4 22 hxg4 bS 23 axb6 'ikxb6 24 'Lla4

'ika7 2S 3'.as l:tb8, and now White has a very pleasant choice between 26 3'.c7 and 26 g s ! 'Llg4 27 3'.xg4 3'.xg4 28 'Li es! (but less good here is 28 'ikc6 3'.e2 !). b) 18 ...g4!? is a better try: 19 fxg4 hxg4 20 hxg4

...

28

20...'Lle8! (Black intends ...3'.f6-h4; a useful idea to know) 21 a4 3'.f6 22 l:ta3 3'.h4 23 'Llcdl 3'.g3 24 l:tc3 (24 'Llh3 't!fh4 2s 'Lldf2 'Llf6 26 'ikd1 3'.d7 27 as l:taf8 28 3'.el f3 29 l:txf3 'Llf4 30 l:txf4 and Yi-1/2 was G.Sosonko-F.Hellers, Wijk aan Zee 1986) 24 ... 3'.d7 2s 3'.bs?! 3'.xbs 26 axbs 'ikh4 27 'Llh3 'Llf6 28 l:tc8+? (28 gs 'Llg4 29 l:tc8+ l:txc8 30 'ikxc8+ 'Llf8! is given by Gallagher) 28 ...l:.xc8 29 'ikxc8+ 'iii> g7 3o'Lldf2 3'.xf2+ 3 1 'Llxf2 f3 ! 32 'ikfs 'Llxe4 33 'ikxf7+ (33 'ikxe4 'ikg3) 33...'itxf7 34 'Llxe4 'ikxg4 and Black had a winning attack in G.Andruet­ V.Spasov, Sofia 1990. 19 'Libs 'Li es 20 a4 The thematic 20 'Llxa7?! seems to work tactically because the c8-bishop is loose, but it appears too slow: 20 ... 3'.d7 (the counter 20 ...l:tc7? does not work well here after 21 3'.as l:txc2 22 .i.xd8

Th e M a r def Plata Va ria tio n : 9 tDe1 lDd 7 1 0 lDd3 .:.Xe2 23 ttJxc8 l:txb2 24 3'.xgs l:taxa2 2 S .:.Xa2 l:txa2 2 6 l:tb1 and White i s much better) 21 ttJbs g4! 22 fxg4 hxg4 23 hxg4 f3 ! 24 gxf3 3'.xbs 2s 3'.xbs l:txf3 gives Black excellent play: for example, 26 'ikdl l:txf2 27 'itxf2 'ikb6+ 28 3'.e3 'ikxbs M.Stean-J.Hjartarson, Lucerne Olympiad 1982, or 26 3'.xe8 't!fh4 27 �gs 'ikg3+ 28 'itf1 l:txe8 29 'ike2 (29 'Wd2 't!Vh 3+ 30 'ite2 'ikg2 31 l:tf1 lLif4+ 32 .i.xf4 exf4 is also winning for Black) 29 ... 3'.e7 0-1, J.Barkhagen-M.Tumer, Sas van Gent 1992. 20 3'.d7 With the bS-square covered, Black should not ignore White's idea: for ex­ ample, 20...lLih4 21 ttJxa7! 3'.d7 22 ttJbs g4 23 fxg4 hxg4 24 hxg4 l:tg7 2S 'ikdl .i. e7 26 l:ta3 was much better for White in Z.Kozul-A.Sznapik, Tbilisi 1988. 20...l:tg7 is playable: 21 l:ta3 a6 22 :C3 3'.d7 23 ttJa3 lLih4 24 3'.el l:tb8 2 S 'Wdl lLif6 gives Black counterplay. 21 'ikd1 This move tries to clamp down on the g4-square. Probably White should look at the alternatives, though. 21 l:ta3 a6! forces the knight back to c3, while after 21 as!? Black should play 21 ...l:tg7, which is the same as the Line A2, ex­ cept that White's rook is on al instead of fl. This probably does not change much, especially as White usually dou­ bles rooks on the c-file over the next couple of moves, but White will have the option of playing a quick 'ikdl or �el without locking his rook in on fl. 21...l:tg7 ...

Black intends ... lLih4, ... a6 and ...lLif6 to force through ... g4. White should play 22 as or double rooks with 22 l:ta3 or 22 l:tC3. Instead 22 'itfl?! 'takes ad­ vantage' of the free fl-square, but this is not a very good place for the king. After 22 ...lLih4 23 3'.el 3'.e7 24 l:tc3 'ikb8 ! ? (another idea is 24... a6 2 S ttJa3 lLif6 26 l:tacl g4 2 7 hxg4 hxg4 28 fxg4 'ike8!?, hitting the a4-pawn and prepar­ ing ... 'ikg6) 2S as 3'.d8! (here we see one of the points of 24...'ikb8; from d8 the bishop can help defend the queenside and it may even become active itself) 26 l:tb3 a6 27 CDc3 (27 ttJa3 3'.a4!) 27 ...CDC7!? (Black intends to use the bS­ square, but a more thematic plan is 27 ...lLif6, trying to force through ...g4) 28 l:tb4 'ith8 29 l:ta3 'ikc8 30 l:tc4 'ikb8 31 l:tb4 'ikc8 32 l:tab3 ttJbs 33 ttJxbs axbs 34 3'.xbs 3'.xas 3S 3'.xd7 l:txd7 36 l:ta4 l:tc7! Black suddenly took over the queenside in A.Gavrilov-A.Shomoev, Krasnoyarsk 2007. C) 14 l:tc1 White clearly will need a rook on the

29

A t ta c k i ng Ch ess: Th e King 's Indian, Vo l u m e 1 c-file, so he plays this immediately in­ stead of exchanging on d6. 14 'Llg6 1s 'Libs!? This is a very sharp move which has become popular recently. Of course lS cxd6 cxd6 brin gs us back to Line A. By delaying the exchange on d6 White is able to create some unusual problems. Black can now force White to make a positional piece sacrifice or he can play normally, although this allows White to play an original manoeuvre to weaken the black queen side. 1s l:tf7 Instead both 1S ... g4? 16 cxd6 cxd6 17 'Llc7 and 1S ...'Lle8? 16 3'.as are clearly better for White, but practice has also seen the critical 1S ... a6, trying to refute White's play. ...

...

Here White has: a) 16 'Lla3 intends 'Llc4 and 3'.as. However, this move abandon s control of e4 and Black can play 16 ... g4 imme­ diately. After 17 cxd6 cxd6 18 'Llc4 g3 White has: al) 19 h3 3'.xh3 20 gxh3 bS (better than 20 ... 'ikd7 21 'Llf2! gxf2+ 22 l:txf2)

30

2 1 l:te1? ! (better i s 2 1 'Lias 'ikd7 2 2 'Llf2 gxf2+ 23 l:txf2, although Black still has the initiative after 23 ...'LlhS) 21 ...bxc4 22 l:txc4 'Llh4 was clearly better for Black in H.Hoeksema-L.Riemersma, Dutch League 1987. a2) 19 3'.as gxh2+ (worse is 19 ...'ike7 20 'Llb6 gxh 2+ 21 'itxh2 l:tb8 22 'ikc2! ) 20 'itxh2 'ike7 2 1 l:th 1 'Llhs 22 'Llb6 'Llg3 gave Black good play in M.Yudovich ­ E.Arlind, correspondence 1974. b) 16 cxd6! (this piece sacrifice is critical) 16 ... axbs 17 dxe7 'it d 7 (worse is 17...'ike8?! 18 'ikb3 g4 19 'Lies 'Llh4 20 fxg4 and White was much better in the game F.Berkes-M.Pavlovic, Kragujevac 2009) reaches quite a complicated po­ sition.

White's play certainly seems easier, but Black is up a piece: bl) 18 'ikb3 'Lle8 with a further di­ vide: b11) 19 3'.b4 'Lle8 19 'ikb3 'Llxc7 20 3'.xf8 Lf8 looks unclear. b12) 19 'Li es 'ikd6 20 'Lle6 (after 20 3'.b4 Black can just play 20 ...'ikb6) 20 ... 3'.xe6! (much better than 20 ...'ikb6+

Th e M a r def Plata Va riatio n : 9 tDe1 lDd 7 1 0 lDd3 21 'iii> h l 3'.xe6 22 dxe6 ttJxc7 23 e7+ l:tf7 24 3'.c4 lLih8 25 l:tfdl when White had a big advantage in O.Biriukov-S.Soloviov, Saint Petersburg 1999) 21 dxe6 ttJxc7 22 3'.b4, and now both 22 ... 'ikb6+ and 22 ...'ikd4+ 23 'iii> h l l:tfc8 are possible, while earlier 21 3'.b4 'ikb6+ 22 3'.c5 'ika5 23 dxe6 ttJxc7 24 3'.xf8 'itxf8 25 l:tfdl 'ikb6+ 26 'iii> h l ttJxe6 27 'ikb4+ 'itg8 is unclear according to Soloviov. b2) 18 3'.b4 was Giri's recent try. Black has: b21) 18 ...ltJe8 19 3'.xf8 (19 'ikb3 is 'b11' above) 19 ...3'.xf8 with the idea of ...3'.d6 was given by Nunn, although Hoeksma points out that White has some initiative after 20 'ikb3, intending l:tc3 and l:tfcl. b22) 18...g4?! 19 CDc5 (19 d6! gxf3 20 gxf3 b6 21 tLif2 'it h8 22 a3 l:tg8 23 'iii> h l 'ike8 24 l:tgl gives White more than enough compensation according to Giri) 19 ... 'ikxc7 20 CDe6 and here, in­ stead of 20...'ikf7? 21 3'.xb5! when White was much better in A.Giri­ F.Nijboer, Haaksbergen 2009, Black should have tried 20 ... 'ikb6+ 21 3'.c5 'ika5. b23) 18...l:tf7 19 'ikb3 3'.f8 20 3'.xf8 'itxf8 21 'ikb4+ 'itg8 22 CDf2 CDe8 23 3'.xb5 'ikd6 24 'ikxd6 ttJxd6 25 a4 'itf8 is a line given by Giri, who thinks that Black is okay here. 1s ...l::tf1 This is a more normal continuation. Black hopes to transpose to the main lines, but White has an independent course.

16 3'.as! This is the point of White's play - he forces Black to weaken his queenside . This is somewhat annoying, but the manoeuvre costs White some time and Black is not without practical chances. Instead 16 cxd6 cxd6 17 'ikc2 would again take play back into Line A.

16 ... b6 This weakens the c6-square, but there is no choice. 17 cxd6! This is more accurate than 17 cxb6 when 17 ...axb6 may be a playable al­ ternativefor Black. 11...cxd6

31

A tt a c k i ng C h e s s : Th e K i ng 's In dian, Vo l u m e 1 Instead 17 ... bxas? is just bad after 18 dxc7 'ikf8 19 'Lies (also good is 19 'ika4 g4 20 'Lies gxf3 21 3'.xf3 'Llh4 22 'ikxas 'Lle8 23 'Lle6 and Black was mauled in V.Bhat-D.Aldama, Boca Raton 2008) 19 ... a6 20 'Lle6 'ikb4 (or 19 'Li es a6 20 'Lle6 'ike7 21 'Llxg7 'itxg7 22 d6 'ike8 23 'Lla3 with a crushing posi­ tion in M.Mchedlishvili-A.Rustamov, Tashkent 2010) 21 a3 'ikxb2 22 l:tbl 'ika2, as in D.Fridman-M.Szelag, War­ saw (rapid) 2009, and now 23 'Lld6 wins quickly. 18 3'.e1 This is better than 18 3'.b4 3'.f8 19 'Llf2 ?! (Bhat suggests 19 l:tc6 or 19 'ikc2) 19 ... a6 20 'Lla3 hS 21 CDC4 as 22 3'.el 3'.d7 23 a4 l:tg7 24 h3 'Llh8!?, which gives Black good play. After 2S 'Lla3 'Llf7 26 'Libs l:tc8 2 7 l:txc8 'ikxc8 28 'Lld3 g4 29 3'.h4 'Llgs he scored a notable upset in V.Bhat-K.Sai, New Delhi 2009.

18 g4!? This is a very principled reply and a typical reaction to the white knight's jaunt to bS. Black takes the oppqrtunity to play this advance immediately be...

32

cause the e4-pawn is not well pro­ tected. Another possibility is 18 ... a6 19 'Llc3 (after 19 'Lla3 !? Black should probably play 19 ... g4, again exploiting the fact that the e4-pawn lacks protec­ tion), and here: a) 19 ... 3'.d7?! 20 'Llb4 bS 21 'Llc6 3'.xc6 22 dxc6 was much better for White in D.Pergericht-W.Winterstein, Luxembourg 1987. b) 19 ...hs?! compares poorly with the main line: 20 'Llb4 g4 21 'Lla4 l:tb7? 22 3'.xa6 l:txa6 23 'Llxa6 bS 24 3'.as ! 'ikxas 2s l:txc8+ 'iii>h 7 26 'Ll4cS l:ta7 27 'ikel 'ikxa2 28 'Llb4 'ikc4 29 b3! 'ikd4+ 30 'iii> hl and 1-0 was G.Meier­ S.Bromberger, Zurich 2009. c) 19 ... as! is the best move. Black must cover the b4-square to prevent a knight invasion.

Here White has tried: cl) 20 'Libs g4 21 l:tc6 3'.f8 22 3'.f2 l:tb8 23 'ikc2, although Black had coun­ terplay after 23 ... g3 24 3'.xb6 gxh 2+ 2S 'itxh2 l:txb6 26 l:txc8 'ike7 27 l:tc7 'Lld7 28 'Llf2 'Llh4 in D.Cummings-B.Sam­ buev, Toronto 2010.

Th e M a r def P lata Varia t i o n : 9 tb e 1 tll d 7 1 0 tll d3 c2) 20 tllf2 3'.f8 21 tllb 5 h5 22 l:tc6 l:tg7 23 a4 3'.d7 24 'ith1! ? tllh 8 25 l:tc3 g4 26 l:tgl (winning the exchange with 26 tllc7 leads to an unclear position after 26 ... 3'.xa4 27 'ikcl 3'.d7 28 tllxa8 'ikxa8) 26 ... g3 27 tllh 3 l;Ic8 28 'ikc2 gxh2 29 'itxh2 tll g 4+! 30 fxg4 l:txc3 31 'ikxc3 hxg4 32 g3 gxh3 33 tllc7 was G.Meier­ J.Lopez Martinez, Pamplona 2009. Now Black's best looks like 33 ...tll g6, with the idea 34 tlle 6 3'.xe6 35 dxe6 d5! when Black has the initiative in a very com­ plicated position. c3) 20 3'.f2 l:tb8 21 tllb 5 3'.f8 22 l:tc6 tll e8 23 a4 3'.d7 24 l:tc3 l:tb7 25 b4 h5 26 bxa5 bxa5 27 tlla7 tllf6 28 tllc 6 'ika8 29 'ikc2 g4 30 tllb 2 l:tg7 31 tllc4 g3 32 hxg3 fxg3 33 3'.xg3 h4 34 3'.h2 tllh 5 was G.Meier-R.Polzin, Austrian League 2010. White is probably quite a bit bet­ ter here, but as usual there is counter­ play and Black went on to score a big upset against one of the main propo­ nents of White's set-up. 19 tll b4 This is critical. White pounces on the weak c6-square. Instead 19 fxg4 tll x e4 is unclear. 19 l:tc6 is also worth investigation, but after 19 ... 3'.f8 at least there is no knight coming into c6. 19...g3! A thematic and strong pawn sacri­ fice. We will now follow the game R.Rusev-A.Diamant, Malakoff 2010: 20 hxg3 20 tllc 6 just gives Black the extra option of 20...gxh2+. 20...fxg3 21 tll c 6 'ikf8 22 3'.xg3 3'.h6!

Black is a pawn down and White has made some inroads into his queen­ side, but he has very good counterplay. 2 3 l:tc3 To stop ...3'.e3+. 2 3 ...3'.f4 There are other tempting possibili­ ties as well, such as 23 .. 11g7 ! ? and 23 ... tllh 5 ! ?. 24 3'.xf4 Not 24 3'.f2? 'ikh6! 25 g3 (25 l:te1 'ikh2+ 26 'iii> f l 'ikhl+ 27 3'.gl tllh 4 mates) 25 ...'ikh3, which is winning for Black 24 ...'Llxf4 25 g3

2 s...'Llh3+ Also good is 25 ...l;Ig7!? 26 'iii> f2 'Ll4h5

33

A ttacking C h e s s : The King 's I n dian, Vol u m e 1 (or 26 ... 'Llh 3+) 27 'Itel (27 g4? 'Llxe4+) 27 ... 'Llxg3 28 l:tgl 'Llgxe4!? 29 l::t xg7+ 'ikxg7 30 fxe4 'Llxe4 31 'ikc2 'ikgl+ 32 3'.fl 'Llxc3 33 'Llxc3 3'.h3 (33...'ike3+ i s also interesting) 3 4 'ikd3 3'.g4!? which seems promising for Black. 26 'iii>g2 l:tg7 27 l:th1?? Under pressure White blunders, but after 27 'ikc2 'Llh s Black has strong pressure. 27 ...'Llxe4! 28 'Lle7+ As 28 fxe4 'ikf2 is mate, White's po­ sition falls apart. 28 ... l::txe7 29 l:txc8 'Llf4+! Black does not relax, even in a win­ ning position. 29 ...l:txc8 30 l:txh3 a6 is still much better for him, but the text is much stronger. 30 gxf4 l:tg7+ 31 '>W1 l:txc8 32 fxe4 'ikxf4+ 33 'Itel l:tcl 34 'Llxd6 l:txdl+ 35 'itxdl l:tc7 36 l:tgl+ 'ith8 3 7 3'.c4 'ikf3+ 38 'Itel l;Ixc4 0-1 In conclusion, lS 'Libs is rather dan­ gerous, but White must take risks too and not everyone wants to commit to sacrificing a piece with the white pieces right out of the opening. If prac­ tice does eventually show this line to be too dangerous for Black, it will negate several decades of opening theory! That said, although we must be pre­ pared for this line, we will not worry too much just yet. The important thing is to be aware of the possibility of l S 'Li b s and t o have something i n mind. D) 13 g4!? This is a completely different ap-

34

proach. Instead of engaging i n a straight race, White takes a stance on the kingside first. If White can com­ pletely block the kingside, he will have a free hand on the other side of the board, where he holds a spatial advan­ tage. This plan was very popular for White in the late 1980s and prompted Black to look at flexible alternatives to 11 ... 'Llf6 12 f3 f4, such as 11 ... 'ith8. Nowadays this plan is not considered to be so alarming to Black, although the play can become a bit stodgy.

13 ... gs After 13 .. .fxg3?! 14 hxg3 Black does not even have an advantage on the kingside, so he leaves the position closed and plans to open the h-file. The immediate 13 ... hs?! is bad because af­ ter 14 g s 'Llh7 lS h4 Black cannot open the kingside unless he makes an un­ sound piece sacrifice on gs. By playing 13 ...g s Black prevents h4 and will open the kingside himself with ... h s. He also prepares ... 'Llg6, aiming for the h4-square. White usually plays 'it g2, so he can contest the h-file, and a

Th e M a r def Plata Varia tio n : 9 tDe 1 CDd l 1 0 lDd3 quick 3'.e1-f2, when the bishop covers the h4-square and also takes aim at the black queenside. Admittedly, Black has a narrow field to play with, so he needs to hold things together on the queen­ side while gradually building up on the kingside, looking for tactical opportu­ nities. This often involves sacrifices on g4 or e4.

14 3'.e1 Both sides have a lot of flexibility here. Instead 14 'itg2 hS lS h3 ttJg6 16 3'.el transposes to note 'b' to White's 16th move, below. White can also play on the queenside immediately: a) 14 cs hs 1s h3 'iii> f7 (1s ... ttJg6 16 3'.el leads back to the main line) 16 l:tc1 l:th8 17 ttJbs hxg4 18 hxg4 l:th3 19 'itf2 3'.xg4! 20 cxd6 cxd6 21 fxg4 ltJxe4+ 22 'Itel 'ikb6 gave Black excellent play in G.Tallaksen-E.Lie, Gausdal 2007. b) 14 b4 hS lS h3 'iii>f7 (1S ...ltJg6 16 3'.el l:tf7 is another way) 16 3'.el l:th8 17 'it g2 ltJg6 18 cs hxg4 19 hxg4 lLihS?! (an amazing offer, which Gelfand de­ clines) 20 l:th1 (20 gxh s l:txhs 21 l:th1 l:txh l 22 'itxhl 'ilih8+ 23 'itg1 tests

Black's idea) 20 ... ttJg3 21 3'.xg3 fxg3 22 'ikd2 l:th4 23 'itxg3 3'.d7 24 l:txh4 gxh4+ 2S 'iii> h 2 3'.f6 gave Black dark-squared compensation for the pawn in B.Gelfand-G.Kasparov, Reggio Emilia 1991. Kasparov's sacrifice may not have been completely sound (it is easier to say this almost 20 years later!), but the idea is still noteworthy. 14... hs 1s h3 ttJg6

16 cs White begins his queenside play. Some alternatives: a) 16 3'.f2 l:tf7 17 a4 3'.f8 18 'itg2 3'.e7 19 as 'ikf8! 20 ttJbs 3'.d8 21 ttJxa7 hxg4 22 hxg4 3'.xg4! was a typical shot in H.Fioramonti-0.Cvitan, Geneva 199S . b) 16 'itg2 l:tf7 (I prefer this regroup­ ing to 16 ...'itf7 because Black keeps the option of a quick ...'ikf8 available) 17 3'.f2 3'.f8 and here: bl) 18 b4 l:th7 19 l:th1 3'.e7 20 'ikb3 'itg7 21 cs 3'.d7 22 l:tacl a6 (Black avoids 22 ... 'ikh8 23 c6! bxc6 24 dxc6 3'.xc6 2S 'ikc4) 23 a4 hxg4 24 hxg4 lLih4+ 2S 3'.xh4 l:txh4 26 l:txh4 gxh4 27 'ikc4 h3+ 28 'iii> hl (28 'itxh3 't!fh8+ 29

35

A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1 'itg2 'ikh4 gives Black a strong attack) 28 ... 't!fh8 29 c6 bxc6 30 dxc6 3'.c8 31 l:tg1 'Llh7 32 'ikb3 'Llgs 33 'Lids 3'.e6 34 bs axbs 3S axbs 3'.xds 36 'ikxds 'ikb8 was good for Black in P.Lukacs-E.Grivas, Budapest 1993. b2) 18 l:thl l:th7 19 'ikh3 (19 cs 'Llh4+ 20 3'.xh4 gxh4 21 b4 'ike8 22 'ith2 hxg4 23 fxg4!? l:tg7 24 3'.f3 'Llh7 2S l:tcl 3'.d7 26 l:tfl as gave Black good counterplay in S.Knott-M.Hebden, British Champion­ ship, Scarborough 2004) 19 ...'Llh4+ (Hebden is fond of this knight hop; the kingside closes up a bit, but White may miss his dark-squared bishop) 20 3'.xh4 gxh4 21 l:tagl hxg4 22 hxg4 h3+ 23 l:txh3 l;Ixh3 24 'itxh3 'Llhs! ? 2s 3'.d1 3'.e7 26 'itg2 'Llg3 27 'Lle2 'Llxe2 28 3'.xe2 3'.h4 29 l:th1 'ikg s 30 'Llf2 'itf8 31 l:th3 'iii> e7 32 'ikdl 3'.d7 33 'ikhl l:th8 gave Black enough compensation to draw in S.Knott-M.Hebden, British League 2009. 16...l:tf7

17 b4 White maintains the tension on the queenside. He may also exchange on d6 immediately with 17 cxd6 to avoid

36

the possibility of Black recapturing on d6 with his bishop. The downside to this early exchange is that White gives up some of his queenside space. He also loses flexibility, since, for example, cS-c6 is no longer an option. After 17 ... cxd6 18 l:tcl 3'.f8 19 3'.f2 l:th7 20 'ikb3 hxg4 21 hxg4 'Llh4 22 3'.xh4 l:txh4 23 'Llf2 a6 24 a4 l:th7 2S 'itg2 3'.d7 26 'Llb1 bS! 27 axbs axbs 28 'Lla3 'ikb6 29 'LlxbS (after 29 'Llc2 Black would play 29 ... 'iii> g7, with the idea of ... 3'.e7 and ... l:tah8 with an attack) 29 ... 3'.xbs 30 'ikxbs 'ikxbs 31 3'.xbs l;Ib8 32 3'.c6 l:txb2 33 l:tal, as in A.Khalifman-V.Spasov, Manila lnterzonal 1990, and now 33 ... 'Lld7 34 l:tfb1 l:txb1 3 s l:txb1 'Li es 36 l:tal 'Llb3 with the idea ... 'Lld4 equalizes according to Khalifman. 11 ...3'.fS 18 'itg2 3'.e7 19 3'.f2 'ikf8!

This is a nice set-up for Black. The d6-pawn remains well protected, from f8 the queen can go to the h-file and, more importantly, there is latent pres­ sure along the f-file, which may create opportunities for a sacrificial break­ through on g4 or e4.

Th e M a r def Plata Varia tio n : 9 t'De 1 CDd 7 1 0 CDd3 20 t'D bs .ilds! This is another typical idea. The bishop efficiently defends the c7square and an exchange of pawns on d6 will also open a route for Black's bishop to the queenside. 21 a4 3'.d7

Black has a harmonious position : a) Grabbing a pawn with 22 cxd6 cxd6 23 t'Dxa7 fails to the shot 23 ... hxg4 24 hxg4 3'.xg4!. b) Instead 22 t'Dc3 3'.e7 23 t'D bs 3'.d8

24 t'Dc3 repeats moves, but Black chose to play on with 24... a6 in Z.Gyimesi0.Cvitan, Ticino 1994. c) A good example of Black's possi­ bilities is 22 l:th l as 23 bxas l:txas 24 cxd6 cxd6 25 t'D b2? (White wants to play l'D c4, but he overlooks a tactic) 2S ... hxg4 26 hxg4 t'Dxg4! (White thought that everything was covered, but it turns out that the bS-knight is not sufficiently supported) 27 fxg4 f3+ 28 3'.xf3 3'.xbs 29 3'.g3 (the white queen is overloaded, since 29 axbs l:txal 30 'ikxal l:txf3 wins for Black) 29 ...l'D h4+! (now this check is well timed; the position has opened up and Black comes in on the dark squares) 30 3'.xh4 gxh4 31 l:th3 3'.d7 (31 ... 3'.a6! pre­ vents l'D c4 and looks even stronger) 32 l'Dc4 l:tcs 33 t'De3 l:tg 7 34 'ith1 'ikf4 when Black was still better and went on to win in l.Bedgarini-P.Popovic, Moscow Olympiad 1994.

37

Chapter 2 The Mar del Plata Variation 9 lbe1 ltJd7

10 f3

1 d4 'Llf6 2 c4 g6 3 'Llc3 3'.g7 4 e4 d6 5 'Llf3 0-0 6 3'.e2 es 7 0-0 'Llc6 8 ds 'Lle7 9 'Lle1 'Lld7 10 f3

This move can just be used for transpositional purposes, but in this chapter we will look at an independent idea. 10...fs 11 g4 Instead 11 'Lld3 transposes to Chap­ ter 1, while 11 3'.e3 transposes to Chap­ ter 3. This idea was pioneered by the Hungarian Grandmasters Benko, Pinter and Lukacs. As in Line D seen in Chap-

38

ter 1, White fights for space on the king side.

There i s a big difference here, how­ ever, because Black has not resolved the tension with ...f4. Although in gen­ eral Black should be happy about this, he must not forget that White may be able to play on the kingside too. The famous game Pinter-Nunn did a lot to dampen enthusiasm for this line from White's point of view, but what is old is new and recently there has been some revived interest in this variation. 11...'ithS

Th e M a r def Plata Variatio n : 9 tb e 1 tll d 7 1 0 f3

This flexible move is now consid­ ered the main line and will be the only continuation we examine in detail. 11 .. .f4 is considered to be a mistake because White can keep the kingside closed with 12 h4, although after 12 ... cs Black's position is very solid. Still, because Black cannot open the kingside, he is basically just trying to hold the queenside and this way of playing is too passive. 11 ...tllf6 is the other ma in continuation. After 12 tlld3 c6 Black increases the tension across the board. Then 13 .te3 'iii> h8 is similar to the main lines, except that both sides have committed themselves White with tlld3 and .te3, and Black with ... tllf6. It is logical for Black to keep the ten­ sion and the move 11 ... 'ith8 is almost always useful to him. The main idea is to improve the placement of Black's worst piece - the knight on e7. It will go to g8 from where it can re-emerge on f6 to pressure the e4- and g4-pawns. Moving the knight also opens up the h4-d8 diagonal and 11 ...'ith8 can be

seen too as preparation for an eventual ...f4. This advance will make some sense if White can be prevented from blocking the kingside. 11 ...'ith8 helps because the retreat ...tllg 8 will allow Black to fight for the h4-d8 diagonal with both his queen and dark-squared bishop. Note that the other knight re­ mains on d7 for the moment, not only to leave the f6-square free for Black's bishop in some positions, but because Black may want to play ... as and ... tll c s, increasing the pressure on the e4pawn. Because 11 g4 is becoming popular again, rather than just examine the aforementioned model game Pinter­ Nunn, we will look at all of White's pos­ sibilities in some detail. A: 12 .te3 B: 12 liJd3 C: 12 tllg2 D: 12 h4 Instead 12 .td2 tllg 8 13 'ikcl pre­ vents any ideas of ... .th6, but is a bit artificial. After 13 ...f4 Black is ready to begin his kingside play. Meanwhile 12 gs tll g 8 13 h4 transposes to Line D, as does 12 'itg2 tll g 8 13 gs f4 14 h4. Fi­ nally, 12 'iii> h l tllg 8 13 l:tgl has been played several times by the Hungarian IM Kiss. Black can play 13 ...f4, when 14 tlld3 gs transposes to Line B, or he can try 13 ... as!?. A) 12 .te3

39

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 llld 3 l:Ih7 18 .l:thl .Jth4 19 .Jtd1

Statistically this has been the most popular move, but it is not seen much nowadays. The bishop goes to an active square, but now ...f4 will come with tempo. 12 ...lllgB 13 'ii'd 2 Instead 13 lll d 3 is considered in the notes to White's 13th move in Line B. There is an alternative in 13 lll g 2 f4 14 .Jtf2 hS (Black could also try 14 ... g s with the idea 1 s h4 h S!) 1s gs (after 1 s h3 .Jtf6 16 b4 l:.f7 1 7 cs l:Ih7 18 'ii'd3 lllh 6 19 cxd6 cxd6 20 lllb s lllf7 21 h4 a6 22 lllc3 gs Black tore open the kingside in S.Mirovshchikov-S. Soloviov, St Pe­ tersburg 2002) 1s ...'ii'x gs 16 .Jih4 'ii'h6 11 lllbs gs 18 .Jtf2 llldf6 19 lllxa .Jth3 20 lllxa8 .i::tx a8 was A.Lesiege­ J.Fedorowicz, New York 1993. White has won the exchange, but Black's kingside play is still very dangerous. 13 ...f4! This advance is usually called for when White cannot play h4. Here Black has not committed to ... gs yet, so he can still use the h4-d8 diagonal. 14 .Jtf2 hS 1S h3 l:.f7 16 @g2 .Jtf6! 17

40

19 ...lllfB Often in the Mar del Plata Variation we see Black play ... llld7-f6 and ... llle7g6. Here Black has already played ... lllg 8, so he finds another way to bring a knight to g6. 20 cs .Jtxf2 21 'ii'xf2 gs 22 cxd6 cxd6 23 l:Ic1 .Jid7 24 .Jta4 .Jtxa4 2s lll xa4 lllg6 Black certainly had no problems here in D.Jacimovic-M.Vukic, Kastel Stari 1988. B) 12 llld 3

12 ...lllgB Black continues with the standard

The M a r def Plata Varia t i o n : 9 t:D e 1 t:D d 7 1 0 f3 plan. Another idea is to block the queenside and activate the d7-knight by going to cs. After 12 ... as 13 j_e3 b6 14 'ii'd 2 (14 a3 t:Dcs 1s b4? walks into 1s ... axb4 16 axb4 l:Ixa1 17 'ii'xa1 t:Dxd3 18 j_xd3 fxg4) 14 ... t:Dcs 1s t:Dxcs bxcs 16 @h1 j_d7 17 .i:Ig1 f4 18 j_f2 t:Dc8 19 h4 j_f6 20 'ii'e 1 t:Db6 Black was very solid in E.L'Ami-S. Shyam, Dieren 2009. Black can also play the immediate 12 .. .f4. Because White's 12th move did nothing to influence the h4-square, Black immediately closes the kingside, after which the e7-knight will often go to g6. Now 13 h4 t:Dg8 14 'ii'e 1 j_f6 1 S g s j_xg s (Black could also try 1s ...j_e7!? with the idea of ...h6) 16 hxgs 'ii'x gS+ is a draw. More interesting is 13 l:If2!? j_f6 14 l:Ig2 j_h4 1S b4, which was played in P.Eljanov-1.Cheparinov, Jer­ muk 2009, and here Golubev suggests 1s ... gs 16 cs t:Df6, with the idea of ... h s and ...t:Dg6. 13 @h1

White's next couple of moves indi­ cate that he may have ambitions on the kingside himself. After 13 j_e3 j_h6

(13 .. .f4 14 j_f2 hS is also possible and then 1S h3 j_f6 is similar to Line A) Black has scored quite well. Compared to 12 j_e3 t:Dg8 13 t:Dg2, White does not control the h4-square and after 14 j_f2 j_gS Black has improved the scope of his bishop. 13 ...f4 Again the plan with 13 ... as is a solid alternative. After 14 l:Ig1 f 4 1S j_d2 j_f6 16 a3 b6 17 b4 j_h4 the game K. S akaev-V.Bologan, European Club Cup, Ohrid 2009, was soon agreed drawn. 14 �g1

14...gs!? White's kingside manoeuvres indi­ cate that Black should switch plans. After 14...j_f6 1S b4 j_h4 16 cs t:Ddf6 17 j_b2 Black's pieces do not coordinate well and White can play on all parts of the board. Then 11 ... j_d7 18 a4 h S?! 19 t:Dxf4! exf4 20 g s was V.Belov­ A.Grischuk, Russian Team Champion­ ship 2009, where White won back the piece and kept the initiative. 15 j_d2 hs 16 h3 l:If6 17 I:tc1 l:Ih6 18

41

A ttacking C h e s s : Th e King 's I n d i an, Vo l u m e 1 @g2 .Jtf8 19 b4 tfJe7 20 cs tfJg6 21 cxd6 .Jtxd6!? Perhaps Black was concerned about 21 ... cxd6 22 tfJbs, although this sortie does not achieve much after 22 ... hxg4 23 hxg4 CiJh4+ 24 @f1 tfJf6 2S CiJC7 .U.b8 because White's remaining pieces are not very active. Still, the capture with the bishop is quite viable. 22 tfJbs tfJf6 23 tfJf2 .Jtd7 24 a4 This was 1.Cheparinov-A.Fedorov, Khanty Mansiysk 20os, and in this bal­ anced yet dynamic position a draw was agreed.

ered in the notes to White's 13th move in Line A. 13 ... as Because White is making prophylac­ tic kingside measures, Black goes for the queenside plan. It will not be easy for White to keep control of Black's counterplay across the whole board. 14 .Jte3 tfJcs

C) 12 tfJg2

White shores up his kingside, but this move is somewhat passive. 12 ...tfJgB Black can also play 12 ... as 13 h4 tfJcs 14 .Jte3 tfJg8 and in fact this was the actual move order of Pinter-Nunn. However 12 ...tfJg8 is both more com­ mon and more flexible, so we will con­ sider it the main line. 13 h4 The alternative 13 .Jte3 was consid-

42

15 .U.b1 This does not give White anything, but other moves have also failed to trouble Black: a) 1s gs f4 16 .Jtf2 h6 opens the kingside before White is ready for it. b) 1S Itel 'ii e7 16 a3 fxg4 17 fxg4 I:txf1+ 18 @xf1 (it is unappealing to capture this way, but it is usually nec­ essary, as here, to avoid ...tfJf6 forking White's e4- and g4-pawns) 18...tfJf6 19 .Jtf3 .Jtd7 20 b4 axb4 21 axb4 tfJa4 was pleasant for Black in T.Markowski­ S .Dolmatov, Polanica Zdroj 1993. White is beginning to look overextended. c) 1s 'ii d2 b6 16 exfs gxf s 17 gs f4 (this gives up the e4-square, but Black will get many dark squares in return)

Th e M a r def Plata Va riati o n : 9 lb e 1 lb d 7 1 0 f3 18 .Jtxcs bxcs 19 lbe4 lbe7 20 .Jtd3 lbfs gave good play in A.Chernuschevich­ J.Rowson, Bratislava 1993. d) 1S 'ii' c 2 .Jtd7 16 .U.ad1 b6 17 .Jtf2 'ii'c8 18 g s h6 19 @h2 'ii'd8 (Black's little queen moves annoy White; first the g4pawn is attacked and then when it ad­ vances, the black queen takes aim at it again) 20 exfs .Jtxfs 21 'ii'd2 @h7 22 .Jte3 'ii'd7 23 .U.g1 hS was very comfort­ able for Black in B.Rumiancevas1. Schutt, correspondence 1996. e) 1S a3 fxg4 (Black could also try 1s ... a4 16 .Jtxcs dxcs 17 lbxa4 fxg4 18 fxg4 I:txf1+ 19 @xf1 lbh6!? 20 lbc3 lbf7 21 @g1 lbd6 when White's position looks a bit airy) 16 fxg4 l:txf1+ 17 @xf1 lt:lf6 18 .Jtf3 h S!? (a notable idea; Black grabs control of g4 for his pieces) 19 gxh s gxh s 20 b4 (20 .Jtxhs lbcxe4) 20... axb4 21 axb4 I:txa1 22 'ii'xa1 lba6 23 bS lbcs 24 'ii'd1 lbg4 1/2-1/2 A.Gipslis­ D.Lapienis, Parnu 1982. White was much higher-rated here and he surely realized his position was beginning to look rather overextended. 1s... .Jtd1 16 b3 b6

17 a3?! This is a mistake, but it is not clear how White should continue. For exam­ ple, after 17 'ii'd2 fxg4 (the immediate 17 ... lbf6!? may be even stronger) 18 fxg4 lbf6 White is forced to part with his dark-squared bishop with 19 .Jtxcs, when Black can play 19 ...bxcs or even 19 ... dxcs with the idea ... lbe8-d6. 11 ... a4! This is a useful device to be ac­ quainted with - not only in this varia­ tion, but in the King's Indian in general. White's knight on g2 is a long way from controlling d4 and Black takes full ad­ vantage. 18 b4 lbb3 19 lbbs This prevents ... lbd4, but it leaves White's pawn structure looking vul­ nerable. 19...lbf6 Also good is 19 _.fxe4 20 fxe4 I:txf1+ 21 @xf1 (else 21 ...lbf6) 21 ...lbf6 22 .Jtf3 'ii' c 8! 23 g s lbQ4 24 .Jt g 1 (to cover h2) 24. ..h s and Black's firmly entrenched knights give him the advantage. 20 exfs gxfs

43

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 This is exactly what Black wants. The position begins to open up and White's kingside pawn advances just look weakening. 21 lllc 3 Instead 21 gs f4! 22 gxf6 j_xf6 23 j_f2 l:Ig8 24 @h2 l:Ixg2+! 2 s @xg2 j_xh4 26 @h1 'ii'f 6! gives Black a winning at­ tack. 21...e4! This is another typical King's Indian idea. Black sacrifices a pawn for control of the dark squares. 22 gs 22 fxe4 lllxg4 is much better for Black. 22 ...lllh s This unleashes an attack on the c3knight. 23 fxe4 f4 This further pawn sacrifice gives Black a good position, but 23 ...lllg 3 24 I:tf3 fxe4 may have been even stronger. 24 j_d2 lllxd2 25 'ii'xd2 'ii'e 8 26 j_f3 lllg 3 27 I:tfel j_e5

For the pawn Black has the bishop­ pair and a grip on the dark squares. The

44

rest of the moves of the classic game J.Pinter-J.Nunn, Thessaloniki Olympiad 1988, were: 28 llle2 lllxe4 29 j_xe4 f3 30 lllef4 fxg2 31 lllx g2 'ii'h s 32 'ii'd 3 j_g4 33 l:Ie3 'ii'f7 34 'ii'd2 'ii'g 7 35 .U.d3 I:tf7 36 l:Ie1 I:taf8 37 llle3 I:tf4 38 lllg 2 l:I4f7 39 llle3 j_h5 40 I:tf1 I:txf1+ 41 lllxf1 I:tf4 42 'ii'e 1 j_d4+ 43 @g2 'ii'e 5 44 lllg 3 j_g4 45 bS j_f2 46 'ii'xf2 j_h3+ 0-1. D) 12 h4

This committal continuation has re­ cently become popular. White is not simply trying to block the kingside; rather he has his own ambitions on that side of the board. 12 ... lllgB Black continues his plan. Again try­ ing to hold up the queenside with 12 ... as is also possible. After 13 llld3 b6 14 @g2 lllcs 15 I:th1, rather than the 1s ... j_d7 16 j_e3 lllg 8 of M.Rodshtein­ V.Bologan, Moscow 2008, when 17 h S! gives White a big advantage, Markos suggests 1s ... h S!?. With White making pawn moves on the kingside, creating more tension

Th e M a r def Plata Va ri a t i o n : 9 t:D e 1 t:D d 7 1 0 f3 with 12 ... c6!? is also logical. 13 @g2 (13 .!Z)g2 t:Df6 14 a4 stops any ...bs ideas, but the position became rather sterile after 14 ... as 1s l::t f2 j_d7 16 j_e3 l::t c 8 17 l:Ia3 cs in V.Laznicka-H.Nakamura, European Club Cup, Ohrid 2009) 13 ...t:Df6 14 t:Dd3 bS gave rise to im­ mense complications in P.Eljanov­ T.Radjabov, Elista 2008. 13 gs White initiates kingside play him­ self. Instead 13 t:Dg2 transposes to Line c.

13.. .f4 This cuts off White's kingside pawns. 13 ...h6 has also been tried, but I do not like the look of 14 exfs gxfs 1s f4 when Black's minor pieces are cramped. 14 @g2

14...h6 Black nibbles away at White's far­ advanced pawns. Instead 14...l::t f7 1S .!Z)d3 (or 1S l::t h 1 j_f8 16 t:Dd3) 1s ... j_f8 16 l:Ih1 h6 transposes to the main line. 15 l::t h l l::tf7 16 t:Dd3 j_f8 17 'ii'gll This is a strong manoeuvre. From g1

the queen may go t o the h-file and she may support the gs-pawn. With this move White can also assist his queen­ side play - the queen supports a possi­ ble cs-advance and even attacks the a7-pawn. White has done well from this position in practice, but that is mainly because Black has been ill­ prepared for a thematic sacrifice. De­ spite the closed nature of the position, Black should play concretely, paying attention to White's possibilities. 11 ...l::t h 7 Black readies himself for the open­ ing of the h-file. The downside to this move is that the f4-pawn lacks support. It is indeed protected already, but often White will sacrifice a piece on f4, when his pawn chain can become rather ominous and smother the cramped black minor pieces. There are some al­ ternatives, but they do not change the essence of the position. Black must be alert to White's sacrificial ideas. A typical example of what to avoid went 17 ...hxgs 18 hxgs+ l:Ih7 19 l::t x h7+ @xh7 20 �2 j_e7 21 t:Dxf4! (Black must always be wary of this type of sacri­ fice!) 2 1 ...exf4 22 j_xf4 .!Z)es?! 23 j_e3 t:Df7 24 f4 with a winning position for White in L.Polugaevsky-V.Arbakov, U SS R 1986. Instead after 17...@g7 18 @fl j_e7 19 j_d2 White is willing to sacrifice a pawn because he will control the open lines on the kingside and be able to play on the queenside as well. Here 19 ... hxgs 20 hxgs j_xgs 21 .!Z)bs and

45

A ttacking C h e s s : Th e King 's I n d i an, Vo l u m e 1 then both 21 ...lbf8 22 cs a6 23 lba3 �6? 24 lbc4 Ith7 2S .Jtc3 .Jih3+ 26 @e1 .U.c8 27 @d2 lbd7 28 lbf2 and 1-0 in J.Carron-J.Rosenthal, Winterthur 2008, and 21 ... .Jth6 22 cs a6 23 lba3 lbdf6 24 lbc4 lbe8 2s lbcxes! dxes 26 lbxes I:tf6 27 .Jtc3, 1 .Cheparinov-D. Stellwagen, Amsterdam 20os, are additional ex­ amples of what Black should avoid. 18 @f1

This is actually a critical position. Black must be careful or a piece sacri­ fice on f4 could blow him away. 18 ....Jte7! Black should keep the tension rather than open lines towards his own king. 18 ...hxg s 19 hxgs I:txh1 20 'ii'x h1+ @g7 21 'ii'h4 .Jte7 22 lbxf4! exf4 23 .Jtxf4 lbes 24 .Jte3 has been seen sev­ eral times with terrible results for Black The position is very similar to Polugaevsky-Arbakov, above. 19 lbxf4 If White does not play this, then Black will just take on gs, which will not only gain material, but will protect

46

the f4-pawn as well. After 19 gxh6 I:txh6 the manoeuvre ... ltJgf6-hS is in the air and Black can pick up the h4pawn at his leisure. 19...exf4 20 .Jtxf4 Itf7! 20...lbes? 21 .Jte3 hxgs 22 hxgs .Jih3+ 23 @e1 was another disaster for Black in P.Kiss-M.Ricci, Eger 1994. 21 .Jte3 hs! This move not only keeps the king­ side closed, it secures the g4-square, which in tum helps Black secure the es-square. Similar is the immediate 21 ... lbes, as played in F.Khairallah­ T.Cali stri, Cannes 2007. Black has the same idea of playing ... hs. 22 @e1 22 f4 can still be met by 22 ... lbes (Black could also try 22 ....Jtf8 23 @e1 .Jt g7) 23 @e1 (2 3 .Jtd4? I:txf4+) 23 ... lbg4. 22 ...lbes 23 'ii'g 3 .JtfB 24 lbbs

This was P.Van Hoolandt-J.Cabrera Trujillo, Cannes 2007. Here Black should play 24... c6! to open the posi­ tion when the extra piece should count for more than White's two pawns.

Chapter 3 The Mar del Plata Variation 9 lbe1 tt:Jd7 10 ..te3

1 d4 lllf6 2 c4 g6 3 lllc 3 .Jtg7 4 e4 d6 S ..'Lif3 o-o 6 .Jte2 es 7 o-o lllc6 8 ds llle7 9 lllel llld7 10 .Jte3

bishop, not only with .. .f4, but likely too from a ... g4-g3 advance in the form of a pawn sacrifice. In this variation White's queenside attack is very straightfor­ ward, but the defence of his king can be a bit more complicated. Some consider this line to be a bit of a suicidal variation for White. Indeed Black's kingside attack can be very strong, but both players must develop a feel for attack and defence to be able to enjoy success from either side of the board in these lines. 10...fs 11 f3 f4 12 .Jtf2 gs

Although the 10 llle1 fS 11 llld3 and 10 f3 fS 11 g4 lines of Chapters 1 and 2 respectively have had their own kind of resurgence in recent years, there is no doubt that the aggressive 10 .Jte3 is the main line of the 9 lll e 1 complex. From e3 the white bishop points directly at Black's queenside, not only supporting c4-c5 but also targeting the a7-pawn. This disadvantage is that Black's pawn­ storm will gain time by attacking the

47

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 The last few moves have been self­ evident, but now White faces a broad choice.

A� 13 ti4; 9tl.3.l£dJ . C;,t.J. 'llbs Q::ti l.tt ·E:U U. : � '·

"

"

'

Instead 13 g4 is rare, but not neces­ sarily bad. White tries to play alo.n g the lines of Line D of Chapter 1 and all of Chapter 2. Now should Black play 13 ... hS 14 h3, his play is slower than in those similar lines we have examined because White has got his bishop to the f2-square very quickly, where it is useful for both pressuring the queen­ side and defending the h4-square. The f4-square is weaker than before, how­ ever, and taking on g3 is now quite vi­ able. Indeed, Black can quickly initiate kingside play with 13 .. .fxg3! 14 hxg3 lllg 6.

Black takes aim at the f4-square. Now ... hS-h4 is one idea. White has:

48

a) 1 S 'ii'd2 h S 1 6 lllg 2 (16 j_e3 j_f6 with the idea of ... h4 is given by Nunn), and now rather than 16 ... as 17 llla 4!? b6 18 a3 h4 19 g4 lllf4 20 @h2 lllf6 21 llle3 with an edge for White in S . S avchenko-A.Fedorov, Nikolaev 1993, Black could play 16 ...h4 17 g4 lllf4 18 @h2 lllx g2 19 @xg2 l:.f7 20 b4 lllf8, which looked okay for him in M.Socko­ K. Kachiani Gersinska, Wuppertal 1998. Here Bologan's suggestion of 16 ...j_h6 17 j_e3 @h7 18 b4 as 19 a3 h4 with counterplay also looks sufficient. b) With 1S lllg 2 White would like to bring the knight to fs. Now: bl) 1s ... hs 16 llle3 as (or 16 ... h4 17 g4 lllf4 18 @h2 as 19 lllfs j_f6 20 'ii'd2 b6 21 b3 lllcs 22 .i::tab1 j_d7 23 j_d1 'ii'c8 24 a3 'ii'a6 with an unclear posi­ tion in D.Ruzele-R. Speckner, Regens­ burg 1998) 17 @h2 lllc s 18 llla 4 b6 19 lllxcs bxcs 20 'ii'c2 h4 21 g4 j_d7 22 lllfs j_f6 was V.Dydyshko-L.Van Wely, Yerevan Olympiad 1996. The position is messy, but looks satisfactory for Black. b2) 1s ... as (Black secures the cs­ square for his knight) 16 llle 3 (White can get the knight to fS, but this is not fatal; Black has squares too) 16 ... lllcs 17 b3?! (better was 17 'ii'c2 because now the c3-knight is loose and Black makes use of this) 17 ... lllf4! 18 j_e1 (instead 18 gxf4 exf4 19 'ii'd2 fxe3 20 j_xe3 j_es ! is nice for Black - if 21 j_xgs? j_xc3 ! ) 18.. ll.f6 ! ? 19 gxf4 gxf4 20 lllg 2 l:Ig6 21 I:.f2 @h8! and Black had a strong attack in N. Stanec-G.Timo­ shenko, Vienna 1998.

Th e M a r def Plata Varia tio n : 9 lbe1 lb d 7 1 0 iL e 3 lbg6 14 as l:If7 1s b4 lbf6 16 cs which is discussed in the notes to Black's 13th move in Line E.

A) 1 3 b4

This obvious preparation for cs has fallen out of favour since the classic game Piket-Kasparov. However, it is not completely harmless and Black should not take it too lightly. 13 lbf6 14 cs lbg6 15 cxd6 After 1S .l::t c1 .l::tf7 16 cxd6 play sim­ ply transposes, while 16 a4 gives Black the opportunity to capture on d6 with his bishop after 16 ...il..f8 17 cxd6 il..xd6! ?. This has been tried a few times, but actually I prefer 17 ...cxd6, transposing to the main line. Instead 1S a4 is important. Black has to be careful here or else he could be tricked into a line that he may not be prepared to play. ...

(seefollowing diagram) Gallagher points out that the natu­ ral 1 s ... .l::tf7 may well be unnecessary when White has not played .l::t c 1 be­ cause the c7-square does not need ad­ ditional defence. Indeed, after 16 as we have transposed to the risky line 13 a4

Black should probably prefer 1s ...h s and then: a) 16 as (16 cxd6 cxd6 17 as g4 is the same) 16 ... g4 17 cxd6 cxd6 18 lbbs g3 19 hxg3 (19 il..x a7! ?) 19 ...fxg3 20 il.. x g3 a6 21 lba3 h4 22 il..f2 lbhs 23 f4 lbhxf4 24 lbc4 'ifgs gave Black an at­ tack in L.Ftacnik-1. Smirin, Biel Inter­ zonal 1993. b) 16 h3 l:If7 17 as il..f8 18 c6 .l::th 7! (Black has plenty of support for ...g4, so the rook is better off on the-soon-to-be­ opened h-file) 19 cxb7 (19 bS b6! 20 axb6 cxb6 21 .l::ta3 'ii' c7 22 lbd3 g4 23 fxg4 hxg4 24 hxg4 'ii' g 7 2S il..xb6 'ii'h 6 gave Black a strong attack in l.Hausner­ S .Dolmatov, German League 1993. de­ spite his ugly queenside) 19 ...il..xb7 20 bS (or 20 a6 il.. c8 21 lbbs g4) 20...il.. c8 with counterplay. c) 16 c6 il..h 6 (Black could also try 16 ...aS!?) 17 bS b6 18 as g4 (unfortu­ nately 18 ... .l::tb8 19 axb6 axb6 20 'ii'a4 g4 2 1 'ii'a7 gxf3 22 'ii'xb8 fxe2 23 lbxe2

49

A ttacking C h e s s : Th e Kin g 's I n d i an, Vol u m e 1 lllxe4 24 l::t a7 does not look sufficient) 19 axb6 cxb6 20 fxg4 hxg4 21 g3 l::t f7 was unclear in V.Korchnoi-Xie Jun, Marbella 1999. 1s...cxd6 16 Itel l::tf7 17 a4

17 ...j_fB! This is Kasparov's clever idea. For the moment Black leaves the h S-square open for his knight. The alternative is the immediate 17 ... h s which also seems playable. After 18 as Black has: a) 18... g4 19 lllb s g3 20 j_xa7 lllh 7 21 @h1 (to make room for j_g1) 21 ...l::t xa7 22 .U.xc8 'ii'xc8 23 lllxa7 'ii'd 8 24 h3 lll g s (White was much better after 24 ...'ii'M 2s 'ii'c2 lt:lgs 26 j_bs lllf8 27 @g1! in L.Bass-K. Spraggett, New York 1983) 2s 'ii'a4 'ii' b 8 26 lllbs lllx h3 27 llld3 (27 gxh3? 'ii'c8 wins) 27 ...lll g s 28 l:Ic1 'ii'd8 29 'if c2 lllf8 30 'ii'c8 'ii'f6 31 @g1 l::t d7 and Black developed coun­ terplay in L.Ftacnik-T.Oral, Slovakian League 199S. b) 18...j_d7 19 lllb s j_xbs 20 j_xbs g4 21 @h1 g3 22 j_g1 gxh2 23 j_f2 a6 (23 ...h4 24 @xh2 lllh s 2s �g1 lllg 3 26 a6 bxa6 27 j_xa6 j_f8 28 llld3 l::t h 7 29

50

j_e1 h3 3 0 gxh 3 'ii'h4 31 lllf2 @h8 31 ... 'ii'h s runs into 32 j_c8! - 32 �c8! .U.xc8 33 j_xc8 j_e7 34 lllh 1! saw White defend in J.Piket-P. Paneque, Adelaide 1988) 24 j_b6?! 'ii'f8 2S j_e2 h4 26 @xh2 lllh s 27 llld3 j_f6 28 .U.g 1 lllg 3 29 'ii'd2 j_d8! gave Black good play in G.Burgess-W.Watson, British Champi­ onship, Plymouth 1989. 18 as j_d7 19 lllb s This allows Black to execute his main idea. Instead 19 j_bS j_c8 20 j_e2 repeats (if this is too off-putting then there is 17 ... h s), while White has also tried 19 @h1 l::t g 7 (19 ...'ii'e8 was sug­ gested by Kasparov), and here: a) 20 lllb s g4 21 lllxa7 g3 22 j_b6 'ii'e 8! 23 .i::te7 gxh2 24 l::t x b7 lllh s gave Black some attacking chances in G.Burgess-B.Badea, Prestwich 1990. b) 20 j_bS g4 21 j_xd7 'ii'xd7 22 fxg4 lllxg4 23 lllf3 j_e7 24 j_g1 lllh 4 was good for Black in D.Gurevich­ H.Gruenberg, New York 1991. 19...g4

20 lllc 7?! This appears to be too greedy, but

The M a r def Plata Variati o n : 9 t:De1 tiJd7 1 0 iLe3 Bareev once pointed out, the pieces move differently when one is playing against Kasparov. White can try to bail out with 20 fxg4, although Kasparov's suggestion 20 ... t:Dxe4 21 t:Dc7 il..a4 22 'ii'xa4 �xc7 looks at least equal for Black. Instead 20 t:Dxa7 g3 21 il.. b6 'ii'e 7 22 'it>h 1 t:Dh 5 23 t:Db5 'ii'h 4 (23 ...gxh2 24 fi.f2 t:Dg3+ 25 il.. x g3 fxg3 with a clear advantage for Black is given by Kas­ parov) 24 il.. g 1 I:tf6 25 t:Dc7 t:De7 26 t:Dxa8 ?! gxh 2 27 il..f2 'ii'xf2! 0-1 was K.Friesen-M.Lomineis hvili, Rotterdam 1998. Here White could look into 26 CiJd3, which was suggested by Galla­ gher. The idea is to meet the move 26 ...I:th6 with 27 'ifel or 26 ... gxh2 with 27 fi.f2. 20 ... g3! 21 t:Dxa8? This just carries things too far. In­ stead 21 hxg3 fxg3 22 il.. xg 3 il..h 6 (22 ... t:Dh 5) 23 t:Dxa8 t:Dh 5 2 4 fi.f2 t:Dgf4 25 t:Dd3 Itg7 26 t:Dxf4 fi.xf4 27 g4 il.. xcl 28 'ii'xcl t:Df4 29 'ii' e3 h5 with an attack is given by Kasparov, but this was still a better try. 21 .t:Dhs 22 @h1 22 fi.xa7 'ii'h 4 23 h3 fi.xh 3 24 gxh3 'ii'xh 3 25 .U.f2 gxf2+ 26 @xf2 t:Dh4 27 ii.fl (27 t:Dd3 'ii'g 3+ 28 @f1 t:Dg2!) 27...'ii'h 2+ 28 t:Dg2 Itg7 looks very strong. 22...gxf2 23 I:txf2 t:Dg3+ 24 @g1 'ii'xa8 Black is much better. The game J.Piket-G.Kasparov, Tilburg 1989, fin­ ished 25 il..c4 a6! 26 'ii'd 3?! 'ii' a 7 27 b5 axb5 28 il..x b5 t:Dh1 ! 0-1. as

B) 13 CiJd3

This old move seems to mix up sys­ tems, but t:Dd3 is almost always a use­ ful move, so it can hardly be that bad. 13 ... t:Df6 14 cs t:Dg6 15 a4 Instead 15 cxd6 cxd6 could easily transpose just about anywhere, but there is no benefit for White in giving up all of his flexibility - see too note 'a', below. 15 Itel is a valid alternative, how­ ever. Black should reply 15 ...Itf7.

..

Now 16 'iih 3 allows 16 ... g4 immedi­ ately, so White might prefer: a) 16 cxd6 is the most common sta­ tistically, but it is never played by

51

A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1 Grandmasters. When White gives up his queenside flexibility, it makes i t easier for Black to formulate h i s plans. After 16 ...cxd6, 17 lllb s, for example, allows 17 ... g4 with good play. b) White has also tried 16 a4 .Jtf8 17 as hS 18 h3 which can be met by 18 ...l:Ig7 or even 18 ...g4. c) The immediate 16 lllb S can, of course, be met with 16 ... g4. By now we can see a kind of relationship between the bS- and g4-squares. If White plays lllbs too early, Black can often play ... g4 without further preparation because the e4-pawn is loose. After 17 a4 g3 18 hxg3 fxg3 19 .Jtxg3 lllh s 20 .Jth2 .Jth6 21 .U.c3 a6 22 llla3 .Jte3+ 23 lllf2 .Jtf4 Black had a winning attack in Y.Du Bois-V.Bologan, Bern 1997. d) 16 .U.c2 intends to double rooks on the c-file. This move was played in one of the most famous King's Indian 'prototype' games: 16.. ..i.f8 17 cxd6 cxd6 18 'ii'd2 g4 19 .i::tfc1 g 3 ! (the the­ matic sacrifice) 20 hxg3 fxg3 21 .Jtxg3 lllh s 22 .Jth2 .Jte7 23 lll b 1?! .Jt d7 24 'ii'e 1 .Jigs 2s lll d 2 .Jte3+ 26 @h1 'ii'g s 27 .Jtf1 I:taf8 28 .U.d1 bS 29 a4 a6 30 axbs axbs 31 Ite7 Itg7 32 lllb3 lllh4 33 I:tc2 .Jth3! (every one of Black's pieces par­ ticipates in the attack)

(seefollowing diagram) 34 'ii'e2 lll xg2 3 S .Jtxg2 .Jtxg2+ 36 'ii'xg2 'ii'h 4 37 'ii'x g7+ @xg7 38 .U.g2+ @h8 39 llle 1 lllf4 40 Itg3 .Jtf2 41 l:tg4 'ii' h3 42 llld2 h s 43 I:tgs 0-1, M.Taima-

52

nov-M.Najdorf, Zurich 19S3.

e) 16 @h1 is often a useful prophy­ lactic move, but here 16 ....Jtf8 17 a4 h S 1 8 a s g4 1 9 cxd6 cxd6 2 0 lllb s g3 21 .Jtxa7 (21 hxg3 fxg3 22 .Jtxg3 is met by 22 ...h4 and 23 ... lllh s) 21 ... lllh7 (21 ...'ii'xas!?) 22 .Jtb6 (or 22 .Jt g 1 h4 23 h3 I:txas with the idea of ... .Jtxh3 and . ..'ii'd7, which gives Black an attack) 22 ... 'ii'h4 23 .Jtg1 lllg s 24 llle 1

24 ... lllh 3! 2S gxh3 was A. Shirov­ T.Radjabov, Baku (rapid) 2009, and now Golubev suggests 2s ...'ii'xh3 26 .Jtd3 lllh4 27 .i::t c2 .i::t g7 with good attacking chances. We now return to 1S a4:

Th e M a r def Plata Varia tion : 9 t:De1 t:D d 7 1 0 iL e 3

1s ... hs This is the most obvious choice, but there are other ideas: a) 1s ...I:tf7 16 as il..f8 17 cxd6 il.. xd6 (I would prefer 17 ... cxd6) 18 t:Dcs t:Df8 19 'ii b3 'ii e 7 20 t:De6?! t:Dxe6 21 dxe6 il..xe6 22 'ii xb7 'ii f 8! gave Black coun­ terplay in M.Golubev-S.Ovsejevitsch, Alushta 1999. However, 20 t:Dxb7! looks strong. b) 1s ...@h8! ? introduces a different plan and might even be Black's best. Because the move ...@h8 is useful and in this position White is not targeting C7, Black plays ...@h8, ... l:Ig8 and ... il..f8, instead of ...Itf7, ... il..f8 and ... Itg7.

After 16 as Black has: bl) Following 16 ...a6! ? 17 cxd6 (17 @h1 l:Ig8 18 cxd6 cxd6 19 b4 il.. d 7 20 .U.a3 il..f8 gave Black counterplay in Z.Kozul-Z.Lanka, Batumi 1999) 17 ...cxd6 18 t:Da4 g4 19 t:Db6 .U.b8 20 t:Dxc8 I:txc8 White is probably a bit better, although Black has counterplay. This could also come about after 13 a4 t:Dg6 14 as a6 1S CiJd3 @h8 ! ? in the notes to Black's 13th move in Line E. b2) 16 ... .U.g8 17 cxd6 cxd6 18 t:Dbs g4 19 fxg4 t:Dxe4 20 il..xa7 il.. d7 21 il..b6 'ii e7 22 t:Da I:taf8 is very messy. A cou­ ple of practical examples: b21) 23 I:ta3 t:Dgs 24 b4 e4 2s t:De1 was V.Korchnoi-Z.Lanka, Linz 1997. Now Black should play 2S ...t:Des with the idea of .. .f3 according to Lanka. b22) 23 t:De6 iLxe6 24 dxe6 t:Dgs 2s t:Db4 e4 26 t:Dds (better is 26 il..c4 with an unclear position in T.Chmielewski­ 199S) D.Lybin, correspondence 26 ...'ii xe6 27 t:Da 'fies and Black had the initiative in P.Kiriakov-Y.Zimmer­ man, Sochi 2004. 16 a 5 Instead 16 cxd6 cxd6 17 as g4 1 8 t:Dbs g3 transposes t o note 'a' to White's next move. Instructive is 16 h3 Itf7 17 c6 as!.

(seefollowing diagram) This fine move stops the advance of White's a-pawn and also slows down White's t:Db4-c6 manoeuvre:

53

A ttacking C h e s s : Th e King 's I n dia n , Vol u m e 1

18 cxb7 .i.xb7 19 b4 .i.c8 20 bxas �h6 21 lllb 4?! (better is 21 a6 .i.xa6 22 lllb4 .i.c8 with an unclear position ac­ cording to Kasparov) 21 ...g4 22 lllc6 'ii'f8 23 fxg4 hxg4 24 hxg4 .i.gs 2 5 �f3 'ii'h6 26 .i::te1 lllh4 27 .i.xh4 .i.xh4 28 g s 'ii'xgs 2 9 �e2 lllg4 30 I:tb1 .i.g3 31 �d3 'ii'h4 0-1, V.Korchnoi-G.Kasparov, Am­ sterdam 1991. 16...g4

Black has easily achieved his king­ s ide break and White must decide how to continue on the queenside. 17 c6 Others: a) 17 cxd6 cxd6 18 lllb s g3 19 .i.xa7

54

(19 hxg3 fxg3 2 0 .i.xg3 h 4 with the idea ... lllh s is the usual theme) 19 ... lllh7 (19 ...llle 8!? is also worth considering) 20 h3 'ii'h4 21 .i.b6 .i.xh3 22 gxh3 'ii'x h3 23 I:tf2 lllh4! 24 'ii'f1 gxf2+ 25 lllxf2 'ii'g 3+ 26 @h1 was B.Larsen-E. Torre, Bauang 1973. Now both 26 ... lllg s and 26 ... lllxf3 should win for Black. b) 17 a6! ? is probably the most test­ ing. Black has: bl) 17 ...bxa6 18 lllb4 g3 19 hxg3 fxg3 20 .i.xg3 h4 21 lllc6 is annoying for Black. b2) 17 ...b6 18 cxd6 cxd6 19 lllb4 g3 20 hxg 3 fxg3 21 .i.xg3 h4 and here, in­ stead of 22 .i. h2 lllh s 23 lll c6 'ifgs, the immediate 22 lll c 6! is again problem­ atic. b3) 17 ... g3 18 hxg3 fxg3 19 .i.xg3 h4 20 .i.h2 lllh s 21 axb7 .i.xb7 22 f4 lllhxf4 23 l2Jxf4 lllxf4 24 �xf4 exf4 25 �g4 was W.Harper-K.Waidyaratne, Connecticut 2007. This is a favourable version of the main line for White. 11 ... g3 18 hxg3 fxg3 19 .i.xg3 h4 20 .i.h2 lllh s 21 cxb7 .i.xb7

Black's kingside play looks ominous,

Th e M a r def Plata Varia tio n : 9 lb e 1 lb d 7 1 0 iL e 3 but White has a way to fight for space on that flank by sacrificing the pawn back 22 f4! Both sides must always keep this idea in mind. White returns the pawn in order to fight on the kingside. 22...lbhxf4 23 lbxf4 ibxf4 24 il..xf4 exf4 2s �g4 'if gs

Korchnoi's idea looks odd at first, but it is actually quite thematic. White immediately goes after the valuable c8bishop. Eventually Korchnoi gave it up, but the American GM Dmitry Gurevich still frequently plays this line. 13 ...b6 Instead 13 ...a6 falls in with White's plans: 14 lba7 l:Ixa7 (otherwise White will just take Black's precious bishop on c8) 1s il.. xa7 b6 16 b4 il.. b7 17 cs dxcs 18 Itel! and White maintained the ini­ tiative in V.Korchnoi-K.Hulak, Zagreb lnterzonal 1987. The pawn sacrifice 13 ...lbf6!? 14 lbxa7 i..d7 is interesting:

26 �e6+ @h8 27 I:tf3 il..c8 28 Ith3 �xe6 29 dxe6 �ae8 Here Black retained sufficient play in V.Korchnoi-Xie Jun, Amsterdam 2001. C) 13 lbbs a) 1s lbbs lbg6 16 'ii'c2 hs 17 cs g4 18 c6 bxc6 19 dxc6 il..c8 20 il..c4+ @h8 21 lbd3 g3 was D.Gurevich-M.Golubev, Biel 1992, when White should probably test Black's idea with 22 hxg3 fxg3 23 il..x g3 h4, although Black certainly has counterplay. b) 1s cs l:Ixa7 16 cxd6 lbc8 17 dxe7 'ii'xe7 18 il.. xa7 lbxa7 19 'ii'b3 'ii'cs+ 20 @h1 lbhs 21 lbd3 and here, instead of 21 ...'ii'e 3 22 lbf2 'ii'x e2 23 'ii'xb7 l:If7 24

55

A ttacking C h e s s : Th e King 's I n d ian, Vol u m e 1 'ii'xa7 g4 2s l:tfe1 'ii'bs 26 fxg4 with a big advantage in J.Benjamin-J.Nunn, Hastings 1987/88, Black should play 'ii'd6! (with the idea of ... lllg 3) 22 @g1 lllb s with counterplay. 14 b4 After 14 a4?! a6 1S lllc3 as White has lost two tempi and Black only one. Thus White is just a tempo down on Line E.

18 @h1 bS 19 cxd6 cxd6 20 lllb1 lllf6 21 llld2 'iVhs 22 .Jt g1 'ii'g 6 (with the idea of ...g4) 23 Itel lllh S! and White was in big trouble M.Gavilan in Diaz­ C.Matamoros Franco, Malaga 2010.

1s l:If6!? Black goes for the attack along the h-file. Usually I am sceptical about this idea, but here I think there is consider­ able justification in playing this way. This is about as good a version as Black will get of this plan because he has played ... a6 for free. This fits in with the straightforward plan to play ... .l::th 6 and ...'ii'e8 because White can no longer attack c7 with lllb s. The other logical approach is 1S ...hs. Here White's queenside play is not ob­ vious, but Dmitry Gurevich has demon­ strated a dangerous positional concept after 16 cs (White has tried other moves such as 16 Itel and 16 a4, while one high-level example continued 16 @h1 lllf6 17 cs g4 18 cxb6 cxb6 19 Itel g3 20 .Jtg1 gxh2 21 .Jtf2 h4 22 llla 4 .l:Ib8 23 bs axbs 24 .Jtxbs lllhs 2s @xh2 lllg 3 26 .l::t g 1 lllg 6 and Black had counterplay ...

14 a6 The knight has to get booted at some point. After 14... lllf6 1S cs g4?! (better is 1S ... a6 16 cxd6 cxd6 17 lllc3 hS, heading for the main line) 16 cxd6 cxd6 17 Itel White has the initiative. 1s lllc3 The alternative is 1S llla3 and here 1S ... as 16 cs axb4 17 cxd6 cxd6! ? with the idea ... lllcs is interesting, while practice has seen: a) 1S ...h s 16 cs bS 17 llla c2 lllf6 18 a4 bxa4 19 .l::txa4 lllg 6 20 bS g4 21 lllb4 g3 22 hxg3 fxg3 23 .Jtxg3 h4 24 lllc6 'ii'd7 2S .Jth2 .Jth6 was a complete mess in A.Huzman-1. S mirin, Sverdlovsk 1987. b) 1S ...l:If6 ! ? 16 llld 3 l:Ih6 17 cs 'ii'e8 ...

56

Th e M a r def Plata Varia t i o n : 9 t:De1 tiJ d 7 1 0 iL e3 in V.Korchnoi-Ye Jiangchuan, Novi Sad Olympiad 1990) 16 ... t:Df6 17 cxb6 cxb6.

It looks like White will have trouble breaching the queen side, but in fact his play is very fast The b6-pawn will come under fire and White can also exploit the weakened c6-square: 18 t:Da4! l::t b8 19 b S ! (the key move) 19 ... as 20 l:Icl g4 21 'ii'd2 (now Black has to watch out for ideas of capturing on b6 followed by 'ii'x as) 21 ...il.. d7 (after 21 ... t:Dg6 the blow 22 l::t c6 is even stronger) 22 l::t c6! .

Now instead of 22 ...il.. xc6 23 dxc6 g3 24 hxg3 fxg 3 2S il.. x g3 t:Dg6 26 'ii'c2 which was very good for White in D.Gurevich-E.Yanayt, Las Vegas 2006,

Black should play 22 ...t:Dxc6 23 dxc6 il.. e 6 24 il..xb6 (24 t:Dxb6 fails to 24... g3) 24...l:Ixb6 2s 'ii'xas CiJd7 when White has the initiative, but at least the posi­ tion remains complicated. 16 c5 After 16 t:Dd3 l::t h6 17 cs 'ii'e8 18 @h 1 bxcs 19 bxcs t:Df6, the idea of ...'ii' hs and ... g4 gave Black an attack in J.Kiltti­ V.Maki, Tampere 1998 16... l::t h 6 17 cxd6 cxd6 18 CiJd3 'ii'e 8 19 ii.el Instead 19 @h 1 'ii'h s 20 il.. g 1 t:Df6 21 t:Df2 'ii'e 8 22 t:Dg4 iLxg4 23 fxg4 l::t c8 24 Itel l::t xc3 2S .U.xc3 t:Dxe4 gave Black good compensation in D.Khismatullin­ F.Amonatov, Moscow 2009. 19...t:Df6 20 t:Df2 l:Ig6

Compared to the positions arising after lS ...h S, White's bishop has been diverted from attacking the b6-pawn. After 21 Itel hS 22 h3 Black played 22 ...il..h 6 in J.Federau-K.Klundt, German League 1988, while 22 ...il.. d7 was seen in X.Sole Fabregat-A.Pablo Marin, Cata­ lonian League 1998. Both moves look good - Black has a strong attack

57

A ttacking Chess: The King 's I n d i a n, Vo l u m e 1 D) 13 Itel

This is a very natural and obvious move, and it may transpose to other lines if White follows up with a quick b4 or llld 3. However, there is one im­ portant independent idea connected with a pawn sacrifice. 13 lllg6 This is the main line. Black has also recently tried 13 ... a6, which seems in­ sufficient to me: 14 llld3 b6 (14....l:tf6 1S cs .l:th6 transposes to 13 ....l:tf6, below) 1S b4 .l:tf6 16 cs .l:th6 17 @h1 ! ? 'ii'e 8 18 cxd6 cxd6 19 bS lllf6 20 bxa6 'ii'h s 21 .Jt g 1 'ii'g6 22 lllf2 lllh s 23 lllh 3 lll g 3+ 24 hxg3 .Jtxh3 2S gxh 3 fxg 3 ? (this is bad, but both 2S ...�xh3+ 26 @g2 and 2s ...'ii'h s 26 I:tf2 fxg3 27 .i:Ig2 'ii'xh 3+ 28 .Jth2 lll g 6 29 'ii' e1 gxh 2 30 .l:txgs also look insufficient) 26 f4! exf4 27 .Jtg4 with a clear advantage to White in J.Berkvens-E.lnarkiev, Kerner 2007. The main alternative is 13 ....l:tf6. Black intends to play ....l:th6 and ...'ii'e8 when ... 'ii'h s could be a big threat. This is a popular way of meet­ ing 13 Itel and may well be a good ...

58

practical try, but I have my doubts about its true worth. White can pre­ pare the CS-break OT play it straight­ away as a pawn sacrifice:

a) 14 b4 has scored badly for White overall, but matters are not so clear. 14... .l:th6 1S cS! 1S ... a6 (1S ...'ii'e 8 16 lllbS 'ii'h S 17 h4 lllg6 18 g4 fxg3 19 .Jt xg3 lllf4 20 lllx c7 may look risky for White, but it is hard to find anything convinc­ ing for Black) 16 cxd6 cxd6 17 llld3 lllf6 18 llla4! 'ii'e8 19 lllb6 �S 20 h4 was S.Atalik-1. Nikolaidis, Halkida 1997.

This is a good example of White's 'desperate' h4 defence, leaving Black's attack looking too speculative. Now

Th e M a r de/ Plata Va ria t io n : 9 lbe1 lb d 7 1 0 iL e 3 2 0 ... l::t b8 21 lbxc8 .U.xc8 22 .U.xc8+ lbxc8 23 'ii'c2 lbe7 24 'ii'c7 lbg6 2S 'ii'xd6 lbxh4 26 'ii'e6+ @h8 27 il..xh4 'ii'xh4 28 'ii' h3 'ii'xh3 29 gxh3 l::t xh3 30 Itel hS was suggested by Atalik, but this looks in sufficient for Black. b) 14 lbd3 also has scored badly overall, but looks promising for White as well after 14 ...l::t h 6 1S cs when again Black has to decide whether or not to prevent lbbS: bl) 1S ...'ii'e 8 still seems too optimis­ tic after 16 lbbs 'ii'h s 17 h4 lbg6 18 g4 fxg 3 19 il.. x g3, which looks ugly, but Black's queen side is melting away and 19 ...lbf4 20 lbxc7 is almost identical to variation 'a' above. Even though White's king position looks bad, it is not clear how Black should continue as his pieces are awkwardly placed. b2) 1S ... a6 and now White has two promising moves: b21) 16 cxd6 cxd6 17 a4 'ii'e8 18 ii.el! lbf6 19 ibf2 l::t g6 20 as il..d7 21 lbb1 (with the idea lba3-c4-b6) 21 ...hS 22 h3 g4 23 fxg4 hxg4 24 hxg4 il..h 6 2S l::t c7 l::t c8 26 .i::txc8 lbxc8 27 lbd2 f3 28 gxf3 lbhS 29 lbh3 gave White a win­ ning position in M.Dambacher­ R.Ponomariov, European Club Cup, Kal­ lithea 2008, which was a game with a large rating disparity. b22) 16 c6 bxc6 17 dxc6 lbf8 18 lbb4 'ii'e8 19 @h1 lbe6 20 lbcdS @h8 (20 ... lbxdS 21 lbxdS is similar) 21 lbxe7 'ii'xe7 22 lbds 'ii'f7 23 il..g 1 lbd4 24 il..c4 il.. e6 2S 'ii'd3 as 26 b3 l::t b8 27 'ii'd2 l::t a8 28 a3 l::t h4 29 b4 axb4 30 axb4 il..xdS 31

il.. xds 'ii'h s 32 .U.a1 l::tf8 33 l::t a7 g4 34 fxg4 'ii'xg4 3S il..xd4 exd4 36 l::t f3 with a clear advantage for White in A.Lenderman-R.Barcenilla, Mesa 2009. c) With 14 cs White continues his plan and sacrifices a pawn to open the c-file. Then 14 ... lbxcs 1S b4 lba6 reach es a key theoretical position for the 1 3 - Af6 sub-variation .

White has: cl) 16 lbbs l::t h 6 17 lbxa7 il..d7 18 il.. xa6 (White goes after the c-pawn; instead 18 il.. bs lbxb4 19 'ii'a4 il.. xbS 20 'ii'xbS b6 21 a3 lba2 ?! 22 lbc6 lbxc6 23 .U.xc6 .U.xa3 24 'ii'b2 l::t a7 2S lbd3 g4 26 fxg4 'ii'g S was l.Rajlich-M.Al Sayed, Bu­ dapest 2001, and here 27 h3 would give White a clear advantage) 18 ...bxa6 19 'ii' c2 g4 20 'ii'xc7 'ii'e8 and now: c11) 21 g3? (this turns out badly) 21... @h8 22 .U.c2 lbg8 23 fxg4 il..xg4 24 il..b6 'ii'g 6 2S lbc6 fxg 3 26 lbd8 l::t xd8 27 'ii'xd8 gxh2+ and White had seen enough in V.Korchnoi-1.Cheparinov, Amsterdam 2008. After 28 l::t xh2 il.. f3+ 29 .U.g2 il.. xg2 30 lbxg2 'ii'xe4 Black has the initiative and is up material.

59

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 c12) 21 fxg4 il.xg4 and Black is not threatened by either 22 tllc6 tllxc6 23 dxc6 'it'h5 24 h4 il.f6 (24..Jk8!? was suggested by Mikhalevski) 25 'it'b7 nf8 26 nC3 il.xh4 27 il.xh4 'it'xh4 28 tllf3 hf3 29 ncxf3 'it'h2+ 30 �2 ng6 3 1 ng1 ng3 32 �f1 �h8 33 'it'e7 nfg8 34 'it'f6+ n8g7 35 'it'f8+ ng8 36 'iVf6+ n8g7 37 'iVf8+ V2-V2, V.lotov-V.Saravanan, Kalamaria 2008, or 22 il.b6 'i!Vh5 (if Black wants to play for a win then there is Mikhalevski's suggestion of 22 ...tllc 8! 23 tllc6 tllxb6 24 'it'xb6 'i!Vh 5 25 tllf3 iLxf3 26 nxf3 'it'xh2+ 27 �fl il.f6 and Black retains chances to attack) 23 tllf3 il.xf3 24 nxf3 'it'xh2+ 2 5 �fl ng6 26 nc2 ng3 ! 27 il.gl nxf3+ 28 gxf3 'ilr'h3+ 29 �e2 'it'g2+ (or 29 ...tllf5 immediately) 30 il.f2 tllf5 3 1 exf5 e4, which leads to perpetual check, S.Arounopoulos­ M.Ghinda, correspondence 2007. c2) The assessment of 16 tlld3! ? has gone back and forth a bit. 16...nh6 17 a4 'it'e8 18 �hl and now:

c21) 18... 'it'h5 19 il.g1 il.d7 20 tll a2 nc8 21 tllf 2! 'it'e8 22 tllg4 ng6 23 b5 tll c 5 24 il.xc5 dxc5 25 nxc5 gave White

60

the advantage i n R.Vera Gonzalez1.Nataf, Havana 2002. c22) 18 ...il.d7 was Natafs improve­ ment against the same opponent, but as Golubev points out, this may still be insufficient: 19 'it'b3 �h8 20 tll b 2?! (better is 20 il.gl, as there is no ...tllh 5 to worry about and 20 ... 'it'h5 is met by 21 tllf2) 20 ... 'it'h5 21 il.g1 g4 22 fxg4 il.xg4 23 il.xg4 'it'xg4 24 tll b 5 il.f6 2 5 tll x e7? tllxe7 26 nxa 'it'e2 27 nd1 'it'xe4 28 tllc4 tllf5 29 il.f2 il.h4 30 il.xh4 was R.Vera Gonzalez-1.Nataf, Montreal 2003, and now the easiest win is 30...tllxh4 31 'it'h3 ng8 32 nc8 tllf5, as pointed out by Gallagher. Instead 19 b5?! 'it'h5 20 il.g1 tll c 5 is good for Black, but Golubev points out that 19 il.gl! threatens b5 when 19...'ilr'h 5 just trans­ poses to variation 'c21'. Now we finally return to 13 ...tllg 6.

Here 14 b4 and 14 tlld3 will trans­ pose to Lines A and B respectively, but White has another idea. 14 cs!? Kozul's pawn sacrifice rekindled in­ terest in 13 nc1. 1t is clear that 14... dxc5

Th e M a r def Plata Va riatio n : 9 tD e1 tb d 7 1 0 ii. e3 lS b 4 is no fun after either 1S...cxb4 16 tDbs or 1S ...b6 16 t:Dd3, so Black has to take the pawn with his knight. 14 tDxcs 15 b4 tDa6 Not 1S ...tDd7? 16 tDbs. This is the point of White's sacrifice - the black knight has been sidelined for the foreseeable future. Now White has a choice: he can immediately regain the pawn, or he can play slowly and count on gradual pressure against Black's queenside. 16 tDbs •••

This is the most forcing continua­ tion. White can also play 16 tDd3. Then 16 ...l:tf7 17 t:Dbs 1'..d 7 (17 ... hs 18 a4 ii..h6 19 t:Dxa7 1'..d 7 would be the same) 18 a4 hS (18 ...'it'e8 has also been played, but it is probably too artificial) 19 t:Dxa7 ii..h6 20 l:tc3 �g7 21 t:Dbs tDf8 ! ? 22 h3 tDh7 23 ii.el tDf6 2 4 tDf2 (otherwise ... g4 comes) 24...tDxb4 2S �xc7 tDa6! 26 �xb7 tDcs 27 �c7 g4! 28 hxg4 hxg4 29 fxg4 tDfxe4 30 tDxe4 tDxe4 31 'it'b3 tDg3 32 ii..x g3 fxg3 33 'it'xg3 ii..xbs 34 �xg7+ �xg7 3S gs (3S ii..xbs? 'it'b6+ 36 �f2 ii..e3 wins for Black) 3 S ... ii..x gs 36 ii..x bs

'it'b6+ 37 �f2 'it'e3 was level in V.Korchnoi-0.Cvitan, Pula 1997. 16 ii..d 7 17 tDxa7 Instead 17 a4 l:tf7 18 t:Dd3 would lead us back to 16 t:Dd3, above, but Black could also try 17 ... h s or 17 ... 'it'b8. 11 h s This looks like the most accurate move. Instead 17 ...'it'b8! ? 18 tDbs (or 18 ii..bs l:tf7 19 ii..xd7 �xd7 20 'it'a4 �f7) 18 ....l:tf7 19 a4 h S 20 tbd3 ii..h6 21 �c4 �g7 22 �hl 'it'd8 was the actual move order of Atalik-Gufeld below, which leads us back to the main line. Black has also tried 17....l:tf7 and here: a) 18 ii..xa6 ! ? looks a bit better for White: 18 ... bxa6 19 'it'C2 (19 tDc6 ! ?) 19 ... ii..e8 20 'it'e2 h S 21 'it'xa6 g4 22 tDc6 ii..xc6 23 'it'xc6 �xa2 24 �c2 �a3 2 S bS ii..f8 26 b6 g3 27 ii.cs (27 hxg3 fxg3 28 ii..x g3 cxb6 29 ii..f2 tDf4 30 ii..xb6 'ifgs 31 'it'c8 was suggested in New in Chess and looks strong) 27 ...�b3 28 'it'a4 �bl 29 'it'a2 �bS 30 'it'a6 �bl and Yi-Y2 was D.Gurevich-J.Becerra Rivero, Las Vegas 2007. b) 18 a4 h s 19 tDbs ii..f 8 (19 ...ii..h6 20 �c4 is the main line after 17...h s) 20 t:Dd3 (also good is 20 tDc3 �g7 21 tDd3 �h8 22 �hl 'it'f6 23 as 1'..e 7 24 bS tDb8 2S �al with some advantage in A.Yermolinsky-J.Fedorowicz, North Bay 1998) 20 ... �g7 21 �hl lDM (insuffi­ cient is 21 ... tDh8 22 g3 fxg3 23 ii..xg3 tDf7 24 �gl �c8 2S ii..f2 'iVf6 26 ii..e 3 �g6 27 'it'd2 1'..e 7 28 tDc3 �h7 29 bS tDcs 30 t:Dxcs dxcs which occurred in •••

•••

61

A ttacking C h e s s : The King 's I n dia n, Vol u m e 1 A.Yermolinsky-H.Yoshiharu, Chicago 2000, as 31 as looks much better for White) 22 g3 fxg3 23 il.xg3 tllg 6 24 l:tgl 'it'f6 25 'it'd2 l:tc8 26 il.f2 il.e7 27 il.e3 and here, instead of the 27 ...tllh 4? 28 il.xg S ! of A.Yermolinsky-J.Becerra Rivero, Chicago 2008, Black should play 21 ...tllf4 28 tllxf4 gxf4 29 il.f2 l:txg1+ 30 l:txgl+ �h8 31 tlla7 l:tg8 32 .llx g8+ �xg8 33 il.xa6 bxa6 34 tllc6 when White is better, but Black is still in the game. 18 a4 Capturing on a6 is still an option: 18 il.xa6 bxa6 19 tllc6 'it'f6 20 tlla s 'it'd8 21 tll c 6 'iVf6 22 tlla7 'it'd8 23 tll c 6 was drawn in D.Gurevich-J.Becerra Rivero, US Championship, Stillwater 2007. Here too White might have tried 19 'i:Vc2!?.

18...il.h6 Instead 18 ....llf7 brings us back to 17 ... .llf7, but an interesting idea of Golubev's is 18 ...tllx b4!? with the idea 19 'it'b3 g4!. This has been seen in prac­ tice and the complications led to a draw following 20 fxg4 hxg4 21 'i:Vxb4

62

g3 22 'i:Vxb7 gxf2+ 23 .llxf2 f3 24 tllxf3 il.h6 2 5 l:tc3 'it'b8 26 'it'xb8 .llfxb8 27 tllc6 l:tbl+ 28 l:tfl l:txfl+ 29 �xfl l:txa4 30 il.c4 il.e3 in V.Erdos-P.Acs, Hungar­ ian League 2005. 19 .llc4 l::t f7 20 tllbs l:tg7 21 �h 1 tllf8

Black intends to play ... g4 with an attack However, because one of Black's knights is so far away, White can afford to fight on the kingside directly. 22 g3! fxg3 23 il.xg3 tllg6 24 tlld 3 tllf4 This is better than 24...�h7 2 5 l:tgl .llf7 26 il.f2 with an edge in A.Yermolinsky-S.Kindermann, Gronin­ gen 1997.

2 5 il.xf4

Th e M a r def Plata Va riati o n : 9 llJe1 lLl d 7 1 0 ii.e3 This is the most common, but prac­ tice has also seen 2S ll:lxf4 gxf4 26 ii..f2 �h8 (Black could try 26 ... M with the idea of ... h3) 27 l:tg1 ii.gs 28 1'.. f l nc8 29 'it'e1 'it'e7 30 ii.. g 2 ii..f6 31 'if cl 1'.. h 4 32 ii..xh4 'it'xh4 33 ll:lxd6! cxd6? (a better chance was 33 ... ncg8 34 ll:lfs ii..xfs 3 S exfs 'it'f2 3 6 nc2 'it'b6) 3 4 nxc8+ ii..xc8 3S 'it'xc8+ �h7 36 ii.fl and White was wining in l.Rajlich-A.Brustman, Ostrava 1999. 2s ...gxf4 26 ng1 ii.gs!? This looks better than 26 ...nxgl+ 27 "i' xgl+ �h7 28 ii.fl 'it'e7 (Black could try 28 ... ii..g s or 28 ... nc8) 29 ll:lxc7! l::tc8 30 'it'b6 ii..xa4 31 ll:lb2 (this leads no­ where, so 31 b S ! is better) 31 ...ii..b3 32 ll:lxa6 ii..xc4 33 ll:lxc4 bxa6 34 'it'xa6 ng8 3S 'it'b6? (3S 'it'xd6 'it'xd6 36 ll:lxd6 ii..f8 is about even) 3 S ... 'it'g7? (3S ... 'it'h4! with the idea of ...'if el wins) 36 'it'f2 (36 ll:lxd6 ! ?) 36 ...'it'b7 was drawn in S.Atalik-E.Gufeld, Honolulu 1997. 27 ii.fl 1'.. h 4 28 nxg7+ �xg7

29 'it'c2 Instead Atalik suggests 29 'if cl nc8 30 nc2, while here Timoshenko gives

30 ii.. g2 �h6 31 'it'f1 'it'e7 32 ii..h 3 l:tg8 33 ii..x d7 'it'xd7 34 lLlC 3. Both feel White is a little better, but the position is still very complicated. 29... �h6! 30 ll:lc3 'it'e7 31 ll:le2 b6 32 nc6 We've followed the game S.Atalik­ G.Timoshen ko, Timisu de Sus 1998. Here Atalik suggests 32 ... ng8! with an attack E) 13 a4

This thoroughly modern continua­ tion is the main line and has been played more than all the other 13th moves combined. The move a4 is al­ ways useful and White may be able to play the cs-break without any prepara­ tion, either through tactical means or after Black moves the d7-knight to support ... g4. The a-pawn may advance to as, which often helps White by con­ trolling the b6-square if he manages to play ll:lbs followed by capturing on a7, especially because the f2-bishop will have the b6-square to go to in case of an eventual ... g3 by Black. The possibil-

63

A ttacking C h e s s : Th e K i n g 's I n dian, Volu m e 1 ity of a quick tllbs remains, and this sortie could give White an improved version of Line C. 13 as Eventually this positional move has come around to establishing itself as the main line. Black prevents White's further advance of the a-pawn and fights for control of the dark squares on the queenside. White will still break through of course, but Black hopes that the time White spends breaking down his queenside pawn structure can be used to create counterplay. It is too early to play 13 ... h s because, as we shall see, sometimes Black can do without this move. There is another plan with 13 ...�f6 14 as a6, with the familiar idea of ... �h6 and ...'it'e8-hS. The play is similar to that discussed in the notes to Black's 13th move in Line D, but here Black has even scored worse, so we will not spend time look­ ing at it. Black's most principled alternative is 13 ... tll g 6, which is the move perhaps most in the King's Indian spirit. ...

64

Here Black ignores White's play for the moment and continues his own campaign. It is very risky, but it is excit­ ing. After 14 as (White can also lunge forth with 14 tllb s ! ?) Black has: a) 14... tllf6 is too compliant, as it al­ lows 15 cs immediately. Following 1s ... hs 16 cxd6 cxd6 17 tllbs g4 both 18 tllxa7 and 18 ii..x a7 have scored heavily for White. b) 14...a6!? is not so bad and could use further tests: 15 tlld3 tllf6 16 cs �h8 17 cxd6 cxd6 18 tlla4 g4 19 tllb6 �b8 20 tllxc8 �xc8 is probably better for White, although the position is not so clear. We also considered this posi­ tion via a different move order in the notes to Black's 15th move in Line B. c) 14...h S is usually a signal for White to play tllb s. This is because if White plays tllb s too early (as in Piket­ Kasparov), Black may be able to get ... g4 in without preparing it with ...h s be­ cause the e4-pawn is not defended. This is usually an achievement for Black, not only because he saves a tempo on ... h s, but also because the move ...tllh s remains a possibility. However, with Black spending time on ...h s, White should always consider this lunge. Indeed, here 15 tllb S! tllf6 (or 1s ...a6 16 tlla7) 16 tllxa7 has scored heavily in White's favour. d) 14...�h8 intends ...�g8 and ... ii..f8. Markos considers this to be Black's best try, before recommending for White 15 tlld3 �g8 16 cs tllf6 17 cxd6 cxd6 18 tll b s g4 19 fxg4! tllxe4 20 'it'c2. After

Th e M a r def Plata Va riatio n : 9 tDe1 tbd 7 1 0 ii. e 3 Golubev's 20...t:Dxf2, 21 t:Dxf2 controls e4 and gives White an edge according. Instead 20.. Ji.d7? runs into 21 t:Dxes and 20 ...'it'e7 21 tDc7 tDxf2 22 t:Dxf2 nb8 23 tDe4 1'..d 7 24 'it'b3 ! (with the idea tDe6) 24 ...ii..xg4 2S ii..xg4 'it'xc7 26 ii.. e 6 is also good for White according to Markos. The best try may be giving up the exchange with 20... tDf6 !? 21 tDc7 ii.xg4 22 t:Dxa8 ii..x e2 23 'it'xe2 e4 24 lDb4 'it'xa8 2S 1'..d4, but again Markos thinks White is somewhat better and I tend to agree. e) The main line has always been 14...nt1.

Now White has: el) Just like in Line D, Kozul has gone lS cs here. I have actually tried this several times with White, but now I suspect lS b4 is just stronger. Follow­ ing lS ...tDxcs 16 ii..x cs dxcs 17 1'..c4 (af­ ter 17 a6 b6 18 1'..c4 ii..f 8! 19 d6 c6! 20 'it'b3 tDh8!, as in Z.Kozul-A.Kuzmin, Oberwart 199S, Black will pick up the d6-pawn and have two good pawns for the exchange) 17 ... �h8 (17...ii..f8 18 d6 c6 19 tDd3 ii..xd6 20 tDa4 is much better

for White) 18 a6 a critical position is reached.

Black has: ell) 18...b6?! 19 d6 nd7 (or 19 ...nf8 20 ii..d s nb8 21 t:Dbs cxd6 22 t:Dxa7) 20 'it'ds c6 (forced, because 20 ... nb8 21 dxe7 'it'xc7 22 'it'g8 is mate!) 21 'it'xc6 nb8 22 lDbS ii..f8 was D.Vigorito­ E.Yanayt, Philadelphia 2003. Now 23 'it'c7! ii..xd6 24 'it'xd8+ nxd8 2S t:Dxa7 gives White a big advantage. e12) 18...nf6?! 19 axb7 ii..xb7 20 tDd3 ii..f8 21 nt2 (also possible is 21 nas ii..c 8 22 t:Dxcs, as in D.Gurevich­ A.Sherzer, Durango 1992) 21 ... ii..d6 22 lDa4 'it'e7 23 nc2 ng8 24 tDf2 hS 2S ii..e 2 ii..c 8 occurred in D.Vigorito­ J.Fedorowicz, US Championship, San Diego 2006, and here 26 nacl! looks good for White. e13) 18 ... bxa6 19 tDd3 ii..f8 20 tDa4 (20 Ji.xa6 Ji.xa6 21 nxa6 C4 22 lDf2 Ji.CS 23 'it'e2 g4! gives Black counterplay ac­ cording to Bologan) 20 ... g4 (20 .. Ab8 has played by Popovic, and after 21 nt2 or 21 b3 I would suggest 21 ... c6! ?) 21 fxg4 'if gs 22 tDf2 ii..d6 23 na3 nb8 24

65

A ttacking C h e s s : The King 's I n dia n, Vo l u m e 1 nc3 nb4 was M.Llopis de Aysa-K.Volke, Biel 1993, which is unclear according to Bologan. e2) The calm move lS b4 may be best.

Here Black has: e21) 1s ...il.f8 16 cs dxcs! was sug­ gested by Markos, leading to 17 d6 il.xd6 18 il.c4 cxb4 and now we have the following: e211) 19 'llt'b 3 tllh 8! 20 tllb s il.cs! 21 il.xcs tll x cs 22 nd1 'iVf8 23 il.xf7+ tllxf7 24 'it'xb4 tlle6 2S 'it'xf8+ �xf8 26 l:tf2 gives Black compensation according to Markos. e212) 19 tllb s tllcs (the 19 ... tlldf8 20 'it'b3 'it'e7 of P.H.Nielsen-M.Solleveld, German League 2002, can be met with 21 tllxd6) 20 tllx d6! (20 il.xcs il.xcs+ is check!) 20...'it'xd6 21 'it'xd6 cxd6 22 il.xcs dxcs 23 tll d 3 looks a little better for White. e22) 1s ...tllf6 is queried by Markos, but this is the only move Bologan con­ siders: 16 cs il.f8 17 c6! (17 cxd6 is also dangerous) 17 ... h s (Markos suggests 11 ... ng1 18 cxb7 il.xb7 19 a6 il.c8 20

66

�hl! il.d7 2 1 il.bs g4 2 2 il.xd7 'it'xd7 23 fxg4 tll x g4 24 tllf3 tllxf2+ 2S nxf2 il.e7, as in P.H.Nielsen-Y.Kruppa, Minsk 1996, although White looks somewhat better here as well, and Markos's suggestion of 22 il.c6 may be even better) 18 �hl and White looks faster, but Bologan has some ideas.

Black has: e221) 18...nh1 19 tll b s g4 20 tllxa7 g3 21 tllxc8 bxc6 22 il.g1 gxh2 23 il.f2 'it'xc8 24 dxc6 h4 2 S il.c4+ �h8 26 'llt'b 3 was good for White in P.Kiriakov-M.Al Sayed, Dubai 2002. e222) Bologan's idea is to play on the g-file after 18...ng1!?. His main line goes 19 cxb7 il.xb7 20 a6 il.c8 21 tll bs g4 22 tll x a7 g3 23 il.g1 l:txa7 24 il.xa7 cs 2s il.xcs dxcs 26 d6 tllh 7 27 il.c4+ �h8 28 na2 'i(h4 29 h3 il.xh3 30 gxh3 'it'xh3+ 31 �gl tllh4 32 'it'e2 cxb4 33 a7 nxa7 34 nxa7 il.xd6 3S na8+ �g7 36 nas 'it'c8 when Black is two rooks for a piece down, but has tremendous coun­ terplay. Now we return to the relatively­ safer 13 ... as:

Th e M a r def Pla ta Varia tio n : 9 tDe1 tb d 7 1 0 ii.e3 and cs-breaks. 15 ...tDf6 16 tDf2 The immediate 16 b4 axb4 17 t:Db s transposes to Line E21, below. 16... h 5 17 h3

14 lDd3 b6 Black must not allow cs. Now White can regroup his pieces for both attack and defence, or he can strike at once on the queenside. El: 15 .i.e1 E2: 15 b4 El) 15 ii.el

This is a flexible, yet somewhat slow continuation. White wants to bring the d3-knight to f2 in order to hold up Black's ...g4-break. He also prepares to recapture with his bishop after playing b4 which would then support the as-

11 ...�hS! This is a very important move. The natural 17... tDg6?! 18 t:Dbs leaves Black struggling to achieve ... g4, whereas White is ready to break open the queenside with b4. Thus the text in­ tends to regroup with ...tDeg8, ...tDh6, ....llf7, ... ii..f8 and ....ll g7 when ... g4 will come with great force. 18 t:Db5 Both sides can employ a few differ­ ent move orders here. One example is 18 b4 tDeg8 19 bxas bxas 20 cs tDh6 21 t:Dbs llf7 22 cxd6 cxd6 23 .ll c l ii..f8, transposing to the main line. 18...tDeg8 19 b4 .llf7! It is important for Black to avoid capturing on b4, because then White's dark-squared bishop would spring to life. After 19 ... axb4?! 20 ii..xb4 the aS­ break is coming and Black's play is too slow. Instead 19 ...tDh6 is possible,

67

A ttacking Chess: Th e Kin g 's I n dian, Vo l u m e 1 though, and will lead to the main line after 20 bxas (or 20 cs bxcs 21 bxcs) 20...bxas 21 cs nf7 22 cxd6 cxd6. 20 bxas Again 20 cs bxcs 21 bxcs makes no difference. 20...bxas 21 cs il.f8 22 cxd6 White should not become too fancy: a) 22 tlla 3?! tllh 6 23 tllc4 ng1 24 na3 �g8 2S tllx as g4 26 fxg4 hxg4 27 hxg4 tllh xg4 28 tllx g4 tllxg4 29 tllc4 'it'g s 30 il.xg4 il.xg4 31 'it'c2 il.e7 32 nf2 'ikh S 33 g3 il.dl! 34 'it'b2 nf8 and Black had a strong attack in S.Ghane Gardeh­ V.Spasov, Izmir 2003. b) 22 c6?! tllh 6 23 na3 ng1 24 il.d2 g4 2S fxg4 hxg4 26 hxg4 tllh xg4 2 7 tllxg4 tllxg4 28 'it' el (this prevents ... 'it'h4 and also eyes the a7-pawn) 28 ... �g8! (after breaking with ... g4, this is a typical move to avoid any na3-h3+ business) 29 nff3 (29 il.xg4 il.xg4 30 il.xas 'it'g s 31 nf2 nh1 32 tllx c7 nxas 33 'it'xas 'it'h4 34 nf1 'it'h2+ 3S �f2 f3 gives Black a winning attack according to Mikhalevski) 29 ... il.e7 30 nh3 tllf6 (or 30 ... tlle 3 31 nhxe3 il.h4 32 g3 fxe3 33 nxe3 'it'g s 34 nf3 'it'g6 with a clear ad­ vantage - Mikhalevski) 3 1 iLfl ! ? il.xh3 32 nxh3 'it'c8 33 'it'h4?! �f7 was much better for Black in S.Krivoshey­ G.Schebler, German League 2004. 22 ... cxd6 23 nc1 tllh 6 24 nc4 After 24 'it'c2 tlle 8 2 S 'it'b3 (instead 2S 'it'd3 ng1 26 nc2 �g8 27 'iVdl looked very strange in A.Bachmann-A.Frolov, Berlin 1994, and here Black should probably continue 27 ... g4 28 hxg4 hxg4

68

2 9 fxg4 tllf6) 2 s ...ng1 2 6 nc6 il.d7 27 il.d2 il.xc6?! 28 dxc6 'it'b6 29 nc1 g4 30 fxg4 il.e7 31 'it'e6 was good for White in P.H.Nielsen-M.Golubev, Internet (blitz) 2004. Instead Golubev suggests 27 ... g4, while I think Black could also investigate 27 ...tllf6!?. The text reaches a critical position.

24...g4! This is better than 24...ng7?! 2s 'it'c2 g4 which runs into 26 il.xas!. After 26 ... 'it'e8 (even worse is 26 ...'it'xas 27 nxc8 gxh 3 28 tllxh3, J.Fang-B.Dean Kawamura, Parsippany 2008), White should play 27 fxg4! hxg4 28 tllc 7 'it'g6 (or 28 ... nxa 29 il.xc7 gxh 3 30 tllx h3 tllh g4 31 il.b6) 29 tllx a8 gxh 3 30 il.f3! when he survives the attack and keeps a decisive material advantage. Instead 27 h4? gxf3 28 il.xf3 il.g4 29 il.xg4 tllfxg4! 30 il.b6 'it'e7 31 tllxg4 tllx g4 32 il.f2 f3 33 g3? (this loses quickly, but Black is still much better after 3 3 'JJ..C7 fxg2 34 ne1 'iVf6 3 S nxg7 il.xg7 36 �Xg2 il.h6 OT 33 nc3 fxg2 34 na1 ng8 3S il.g3 il.h6 36 naa3 il.f4) 33 ... tlle 3 ! 34 il.xe3 nxg3+ 3 S �2 ng2+

The M a r def Plata Va ria tio n : 9 tDe1 tb d 7 1 0 ii.e3 36 �Xf3 nxc2 37 nxc2 'ifxh4 38 ngl nxa4 39 tDc3 nc4 40 �e2 ii..h 6! 41 ii..b6 nxe4+ and 0-1 was the well-known game P.H.Nielsen-V.Kotronias, Hastings 2003/04. 25 fxg4 After 2S 'it'c2 1'..d 7 26 tDC7 Black has the shot 26 ...ii..x a4!. 25 ...hxg4 26 hxg4 ng1

21 gs!? Worse is 2 7 'it'c2 t:Dfxg4 28 t:Dxg4 tDxg4 29 ii..x g4 ii.xg4 when Black has good attacking chances. 21.. Jbgs 28 'it'c2 After 28 tDc7 na1 29 'it'c2 (or 29 tDe6 ii.xe6 30 dxe6 nag7 31 ii..f3 tDhg4 with counterplay) 29 ...1'..d 7 30 tDe6 ii.xe6 31 dxe6 ng8 both sides have chances ac­ cording to Avrukh. 28...tDhs All of 28...t:Dfg4, 28...tDhg4 and 28...ng8 deserve attention as well. 29 ii..f3? 29 ii..xhs nxhs 30 na 1'..d 7 is un­ clear. 29... tDg3 30 tDc7 This was S.Krivoshey-N .Ortiz

Aguirre, Sort 2006. Black should now play 30 ... t:Dxfl 31 �xfl ng1 with good chances, such as after 32 t:Dxa8 ii.. a 6. 82) 15 b4

This is a more direct approach from White. 1s ...axb4 Black should certainly take now, be­ cause otherwise White will either cap­ ture on as himself or else even break with cs immediately, when the f2bishop remains very active. Now White has a further choice: E21: 16 "llbS E22:

16 "bxb4

E21) 16 t:Dbs White is hoping to obtain a favour­ able version of Line El by playing ii.el and ii..xb4, but the early knight hop gives Black an opportunity to acceler­ ate his attack and this method of play for White has a bit of a suicidal reputation. In this variation both sides must try to break through quickly.

69

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1

16...tllf6! 17 il.e1 Instead 17 tllxb4 looks misguided, but matters are not so clear after 17 ... g4 18 il.h4 g3 19 hxg3 (after the 19 h3 of P.H.Nielsen-H.Harestad, Gausdal 1996, Black could consider 19 ...'it'd7! ?, with the idea of ...tll g 6, or Gallagher's 19 ... il.d7) 19 ... tllg6 20 tllc6 �d7, which is murky, E.Kobylkin-S.Soloviov, Alushta 2001.

17 ...g4! Because the e4-pawn is loose, Black does not need any further preparation for his key kingside break and can ad­ vance at once. 18 il.xb4

70

18 fxg4 smacks of desperation, but really it is just unclear: 18 ...tllxe4 19 il.xb4 il.d7 (or 19 ...tll g6 20 as tllcs 21 tllx cs bxcs 22 il.c3 e4 23 il.xg 7 �xg7 24 'it'c2 'it'e7 2S 'it'c3+ tll e s with the initia­ tive, A.Nikitin-S.Soloviov, Alushta 2002) 20 'it'c2 tll gs 21 h4 tlle4 22 il.e1 tllg 6 23 h s tll g 3 24 hxg6 'it'h4 2s tllf2 il.xbs 26 gxh7+ (26 cxbs e4 with an attack) 26 ... �h8 (26 ...'it'xh7 wins immediately after 27 �xh7+ �xh7 28 il.d3+ e4 29 tllxe4 tllxfl, but the text is certainly good enough) 27 cxbs e4 28 nd1 f3 29 gxf3 tllx e2+ 30 'it'xe2 exf3 0-1, J.Chabanon-J.Degraeve, French League 1999. 18...g3 Another idea is 18 ...tll g 6 as sug­ gested by Golubev. Indeed, the further 19 as bxas 20 nxas l:tb8 21 l:ta7 g3 22 'it'c2 nf7 (better than 22 ...tllh s 23 nxa gxh2+ 24 �xh2 'it'h4+ 2S �gl tll g 3 26 il.xd6) 23 nb1 gxh2+ (23 ... tllxds 24 exds e4 2S fxe4 nxbs 26 na8!) 24 �xh2 tllxdS !? 2S exds e4 certainly deserves attention. 19 h3 Instead 19 as should be met with 19 ...bxas! (19 ...gxh2+ 20 �xh2 bxas 21 .l:txas nxas 22 il.xas tlle8 23 cs is better for White according to Markos) 20 nxas (20 il.xas nxas 21 nxas c6) 20 ...l:txas 21 il.xas tlle 8 (with the idea of ...tll g 6) 22 cs tll g6 23 'it'c2 'i!fh4 24 h 3 il.xh3 2S gxh3 'it'xh3 2 6 il.d1 tllh4 2 7 ne1 dxcs, which i s a little better for Black according to lkonnikov in New in Chess.

Th e M a r def Plata Varia t i o n : 9 tDe 1 tb d 7 1 0 ii.e3

The text reaches a controversial po­ sition. Black's play is very far advanced, but with the kingside closed he will have to sacrifice to break through. 19 ii..x h3! The assessment of this move has gone back and forth from '?!' to '!'. Black will clearly obtain an attack, but will White be able to defend? For a while I believed that this sacrifice was not quite good enough, but upon fur­ ther investigation I believe it is indeed the best, both from a practical and an analytical point of view. We will see, however, that Black's order of moves is extremely important. If Black does not sacrifice immedi­ ately, White may be able to hold the kingside together: for example, 19 ... tDg6 20 ne1 tDe8 21 ii.fl tDh4 22 'it'e2 nf6 23 nebl ng6 24 tDel ii..f8 25 as gave White all the play in D.Vigorito-S.Muhammad, Chicago 2007. Instead 19 ... tDe8 was praised by Markos, but 20 'it'd2 (20 ne1 with the idea of j,.f1, ne2, tDel and nea2 can be met with 20 ...cs!?, trying to hold White

up on the queenside) 20 ... tDg6 (20 ... cs was suggested by Markos) 21 ntc1 lDh4 22 tDel nf6 23 ii.fl should be some­ what better for White, even though Black eventually made a sacrificial breakthrough in N.Sulava-V.Neved­ nichy, Bad Worishofen 2000. 20 gxh3 'it'd7 21 'it'c2 This is the best defence. White will try to defend along the second rank. Instead 21 �g2? gets crushed: 21 ... tDg6 22 nh1 tDh4+ 23 �g1

...

23 ...t:Dxe4! 24 fxe4 f3 25 ii.. d 2 (or 25 ii..xf3 nxf3 26 tDel 'it'f7 27 t:Dxf3 t:Dxf3+ 0-1, S.Ghane Gardeh-S.Krivoshey, Dubai 2006) 2s .. .f2+ 26 t:Dxf2 nxf2 27 l:th2 gxh2+! 28 �xf2 'it'xh3 29 'ikh1 nf8+ 30 �el 'it'g2 31 ii..f3 t:Dxf3+ and 0-1 in R.Pogorelov-C.Matamoros Franco, Dos Hermanas 2003. 21 'it'd2 is better, but unimpressive: 21 ... t:Dg6 (or perhaps even 21 ... 'it'xh3) 22 ii..dl ii..h6 23 tDel �h8 24 na2 ng8 25 'it'd3? (25 'it'g2 tDh4 26 'it'h1 is unclear ­ Gallagher) 2s ... 'it'xh3 26 .i:tg2 tDh4 27 'it'e2 ngs was winning for Black in S.Krivoshey-Xie Jun, Linares 1997.

71

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 21...tllg 6! This move is very important. The obvious 21 ...'it'xh3 is not as good, but it is worth looking at it to see some typi­ cal ideas for both sides. After 22 il.d1 tllg 6 23 'it'g2 'it'h6 practice has seen: a) 24 'iVhl tllh4 25 tlle l?! (25 ne1!) 25...tllxe4! 26 tllg 2 was S.Koutsin­ A.Frolov, Kiev 1995, and now 26 ...tllxf3+! 27 il.xf3 'it'xhl+ 28 �xhl tllf2+ 29 nxf2 (29 �gl e4) 29 ... gxf2 30 nt1 (30 il.e4 nxa4!) 30 ...nxa4 31 il.a3 nta8! (this seems even better than Gal­ lagher's 31 ..Axc4 32 nxf2) 32 il.b2 nxc4 33 nxf2 na2 should win for Black. b) 24 tll x c7! tllh4 25 'it'd2! (but los­ ing is 25 'it'e2? tllxe4! 26 'it'xe4 tllg 2! as given by Gallagher) is a key idea White lines up against Black's queen to avoid the ... tllxe4 and .. .f3 bomb.

'it'xf3+ 38 �c2 was winning for White in C.Persson-G.Calzolari, correspon­ dence 1998. Now back to 21...tllg 6!:

22 ntb1 This move is often recommended. Instead: a) 22 il.d1? loses to 22 ...tllM 23 tll e1 'it'xh3 24 tllxe7 (24 na2 tllh 5) 24...tllh 5! 25 tllxa8 g2 2 6 tllx g2 tllg 3 and 0-1 in M.Dziuba-J.Czakon, Koszalin 2005. b) 22 nfdl is similar to OUT main line: 22 ...'it'xh3 ! (22 ...tllh4 23 tll e1 'it'xh3 24 iLfl at least staves off the first wave of the attack) 23 il.f1 'it'h5 24 tllxa (or 24 tlle1 tllxe4 25 fxe4 f3) 24...tllxe4! 25 tllxa8 tllf2 wins, since after, for exam­ ple, 26 il.g2 'it'h2+ 27 �fl tllh4 28 tll e1 e4 White will be crushed. 22 .. 'it'xh3! This move was mentioned by Matamoros back in 2003 in New in Chess, but it seems to have gone unno­ ticed. Weaker is 22 ...tllM 23 tll e 1 with a clear advantage for White as given by Markos. 23 il.f1 .

Indeed, 25 ... g2 (after 25 ... nac8 26 tlle6 g2 27 ne1 nxc4 both 28 tllxf8 and 28 tllf2 give White a big advantage) 26 ne1 tllxf3+ 2 7 il.xf3 'it'hl+ 28 �2 'it'h4+ 29 �e2 tllxe4 30 il.xe4 f3+ 31 �dl f2 32 il.xg2 fxel'iV+ 33 tllx el 'it'g4+ 34 il.f3 �f3 3 5 tllxf3 nt8 36 na3 �f3 37 nxf3

72

Th e M a r def Plata Vari a t i o n : 9 llJ e 1 lLl d 7 1 0 ii. e 3 After 23 ii..d 1 Black has 23...ll:lh4! (23 ... ll:lxe4 24 'it'g2) 24 ll:le1 ll:lxe4! with familiar devastation.

23... 'it'hs 24 ll:le1 No better is 24 ll:lxc7 ll:lxe4! 2S ii..g 2 (otherwise, 2S ll:lxa8 ll:lgs 26 ll:lel e4 wins, but 2 s ... ll:lf2 may be even better: for example, 26 ii..g 2 'it'h2+ 27 �fl ll:lh4 28 ll:le1 e4) 2s ... ll:lgs 26 ll:le1 e4 27 l:ta3 ll:lh4 and White cannot defend.

24...ll:lxe41 25 fxe4 2 S ii..g 2 ll:lf2 26 ll:lxe7 ll:lh4 winning. 2s ...f3 26 ll:lxf3 �xf3 27 'it'g2 ll:lf4 28 'it'h1 'it'g4 29 �b2 �f8 30 ll:lxc7 �f2 Black has a decisive attack - Mata­ moros.

E22) 16 ll:lxb4

While both lS ii.el and lS b4 axb4 16 ll:lbs are very ambitious, the simple recapture with 16 ll:lxb4, while very obvious, almost looks naive. The knight cannot defend the kingside and it is also seems harder for White to support the advances a4-aS and c4-cS. How­ ever, as is often still playable and the b4-knight may hop to c6. As always, a lot will come down to specifics. 16...ll:lf6 Instead 16 ...ll:lcs 17 as! bxas (17 ...�xas 18 �xas bxas 19 ll:ld3 is similar) 18 ll:ld3 ll:lxd3 19 'it'xd3 gives White a slight but clear edge according to Markos. White will double rooks on the a-file to win back the pawn and Black's counterplay is not so easy. 17 ll:ld3 White has several alternatives: a) 17 ll:lc6 is the most direct: 17 ...ll:lxc6 18 dxc6 'it'e8 (18 ...ii..e 6 is also possible) 19 ll:lds �f7 20 as bxas 21 'it'a4 g4 22 'iYbs (or 22 1'..h4 ll:lxds 23 cxds g3 24 hxg3 fxg3 2S ii..x g3. as in P.Kiriakov­ D.Lobzhanidze, Groningen 1996, and

73

A ttacking Chess: Th e K i n g 's I n dian, Vo l u m e 1 now Black should go 2S ... il.h6) 22 ...tllxds 23 cxds g3 24 hxg3 fxg3 2s il.xg3 il.h6 gives Black counterplay. b) 17 as bxas 18 tllc6 tllxc6 19 dxc6 'it'e8 20 cs dxcs 21 il.xcs nf7 22 'it'b3?! 'it'xc6 23 'it'bs 'it'xbs 24 il.xbs il.f8 was better for Black in O.Wilgenhof­ F.Nijboer, Dieren 2009. c) 17 il.el and then: cl) 17... hs 18 h3 nf7 (worse is 18 ...�h8?! 19 tlld 3! tlle g8 20 tllf2 tll h6 21 tllbs .l:tf7 22 as bxas 23 nxas nxas 24 il.xas il.f8 2S cs!) 19 tll d3 il.f8 20 tllf2 is a little better for White. c2) 17 ...'it'd7 18 tlld3 tllg 6 19 tllf2 h S 2 0 h3 tllh4 21 a s ! nf7 (or 21 ...nxas 2 2 nxas bxas 23 tllb s) 22 axb6 nxal 23 'it'xal cxb6 24 'it'b2 'it'd8 2S tlla4 looks shaky for Black, although he drummed up some counterplay after 2S ...g4 26 fxg4 f3 27 gxf3 tll h7 28 'it'xb6 'iVf6 in O.Gladyszev-S.Soloviov, Soc hi 200S. d) 17 na3 il.d7 18 tll b s was Yusu­ pov's choice and has been endorsed by Markos.

Now Black has: dl) 18 ...�h8 19 'it'al! (19 il.el ng8

74

gave Black counterplay in A.Yusupov­ G.Kasparov, Yerevan Olympiad 1996) 19 ...nas (or 19 ... g4 20 as gxf3 21 il.xf3 bxas 22 nxas nb8 23 l:ta7) 20 tlld3 g4 21 il.el na6 22 as is a line given by Markos. White's play looks faster. d2) 18 ... g4!? and then: d21) 19 fxg4 tllxe4 20 il.d3 tllxf2 21 l:txf2 il.f6 is level according to Yusupov. d22) 19 il.el gxf3 20 il.xf3 �h8 21 tlld3 tllg 6 22 as bxas 23 il.xas il.xbs 24 cxbs 'it'd7 2S 'it'b3 nfb8 26 nb1 tllh4 gave Black counterplay in M.Bley1.Novak, correspondence 200S. d23) 19 il.h4 g 3 ! ? (Black could also try 19 ... gxf3 20 il.xf3 tll g 6 21 il.f2 'it'c8 or 19 ...nf7!? with the idea 20 il.xf6 gxf3 21 il.xe7 fxe2 22 il.xd8 exdl'it' 23 nxdl nxd8 24 tllxe7 nc8) 20 hxg3 tllg 6 was suggested by Yusupov. Then 21 tllc6 il.xc6 22 dxc6 tllxh4 23 gxh4 tllh s 24 as is given by Markos, but after 24 ... bxas 2S cs tllg3 I think I would prefer Black. 11 hs consider Black should also 17 ... tllg 6!? with the idea 18 tllbs g4!. ...

1s tt::i bs tllg6 After 18 ...il.d7 19 il.el 'it'b8 20 �hl g4 21 as bxas 22 il.xas 'it'b7, as in S.Ghane Gardeh-V.Kotronias, Cappelle la Grande 2008, Markos suggests 23 'it'c2 ntc8 24 nfbl when White is way ahead. Black could, though, consider 18 ...g4 19 il.h4 gxf3 (after 19 ... g3 20 hxg3 tllg6 21 as tllxh4 22 gxh4 Black does not have ...tllh s available, so it is not so easy to attack) 20 il.xf3 tllg4 with counterplay.

Th e M a r def Plata Va ria t i o n : 9 tDe1 tb d 7 1 0 ii.e3 Dembo also mentions 24...h 3 ! ?. 2 s nxas?! After 2S cxd6! cxd6 26 tDc7 nb8 27 tDe6 ii.xe6 28 dxe6 tDg3 the position is very messy. 2s nxas 26 ii.. x as ii..h 6! •••

19 a 5 Instead 19 'if el is interesting, taking a look at both the as- and h4-squares: 19...g4 20 as h4! ? 21 ii..xh4 (or 21 axb6 nxal 22 'it'xal g3 23 ii.el h3 21 fxg4 tDxe4 22 axb6 .l:txal 23 'it'xal cxb6 with counterplay - Golubev) 21... tDxM 22 'it'xh4 nxas 23 lDb4 was E.Guseinov­ M.Golubev, Moscow 2006. Now 23 ...1'..d7 24 t:Dc6 ii..xc6 2S dxc6 'it'a8 26 tDxe7 (or 26 nxas bxas with the idea of 27 cs?! 'it'xc6 28 1'..c4+? dS 29 exds 'it'xcs+, win­ ning) 26 ...'it'xc6 27 tDe6 ne8! (E.Guseinov) 28 t:Dxg7 �xg7 29 fxg4?! t:Dxe4 gives Black the initiative. The threat is 30 ... 'it'cs 31 �hl tDg3!. 19. bxas 20 cs 20 ii.el g4 21 nxas .l:tb8 22 lDb4 nb7 23 tDc6 'it'e8 24 na8 gave White a big advantage in Z.Kozul-D.Rogic, Bled 1997, but Black should try 20...a4 or 21 ...1'..d 7. 20 g4 21 'it'c2 g3! 22 hxg3 After 22 ii.el t:De8 23 hxg3 fxg3 24 ii..x g3 h4 2S ii..f2 'it'gs Black has attack­ ing chances. 22 fxg3 23 ii..x g3 h4 24 ii.el tDhs ••

•••

•••

Black has excellent attacking chances. 27 ii..d 2 Dembo also gives the lines 27 t:Df2 h3, 27 t:Dxd6 h3!, 27 tDxe7 ii..e 3+ 28 �h2 'it'gs, 27 cxd6 ii.. e 3+ 28 �h2 'it'gs and 27 ii.xc7 ii..e 3+ 28 nf2 (or 28 tDf2 'it'gs 29 ii.. x d6 tDgf4 30 g4 hxg3) 28...'it'gs. In all cases Black retains a strong attack 27 ii..x d2 28 'it'xd2 t:Dgf4 29 tDxf4 If 29 cxd6? 'it'gs 30 t:Dxf4 t:Dxf4 threat­ ens both mate on g2 and 31...tDh3+. After 31 �2 'it'xg2+ 32 �e3 (or 32 �el ii..h 3, opening the way for Black's rook to a8) 32 ...cxd6 Black's attack is decisive. 29 tDxf4 Black is much better. Indeed, after 30 .l:tdl 'it'gs 31 ii.fl 'it'g3 32 nc1 tDh3+ 33 �hl t:Df2+ 34 �gl h3 White re­ signed in K.Nikolaidis-Y.Dembo, Corfu 2007, in view of 3S 'ikh6 hxg2 36 ii..x g2 Ji.h3 OT 3S 'ifxf2 h2+. •••

•••

75

Chapter 4 The Mar del Plata Variation 9 '2Jd2 1 d4 ll:lf6 2 c4 g6 3 ll:lc3 il.g7 4 e4 d6 5 ll:lf3 o-o 6 il.e2 es 7 o-o ll:lc6 8 ds ll:le7 9 ll:ld2

By playing 9 ll:ld2 White takes a very optimistic view towards his queenside play. This line was the main battle­ ground in the early 1990s until 9 b4 experienced a tremendous boom in popularity. Like 9 ll:lel, 9 ll:ld2 prevents Black's ... ll:lhs and at the same time prepares to bolster his centre with f2f3. In this case, however, instead of re­ routing the knight to help support the c4-c5 break (for example with ll:le1-d3),

76

White intends to force through c4-c5, even as a pawn sacrifice, after which the d2-knight will take up a tremen­ dous post at c4. From there it will pres­ sure d6, while its control of b6 often ends up being important as well. At first 9 ll:ld2 looks more active than 9 ll:lel, but if White cannot effec­ tively reposition the knight, White's development can become a bit clogged up. Logically Black should look to pre­ vent White from playing cs and ll:lc4 effectively. In that sense 9 ... cs might be considered Black's most natural re­ sponse, blocking the position. At one point this was, indeed, the main line, but after 10 �bl ll:le8 11 b4 b6 12 bxcs bxcs 13 ll:lb3 White's knight finds use­ ful employment via another route. For example, 13 .. .fs 14 il.g s! h6 15 il.xe7 'Wlxe7 16 ll:las and the knight becomes a menace on c6. Another idea is to play a 'straight race' with 9 ...ll:ld7 (or 9 ...ll:le8 - the two moves will generally transpose after

The M a r def Plata Vari a t i o n : 9 tll d 2 the knight returns to f6) 10 b4 fS 11 cs �f6 12 f3 f4 13 tllc4 gs 14 a4 ( 14 il.a3 is also possible) 14...tll g 6 lS il.a3 l:tf7. Here both Beliavsky's 16 as and the main line, 16 bS, are theoretically very dangerous, although Black has had some recent successes, especially in the later variation after 16 ... dxcs! 17 il.xcs hS 18 as g4 when he has dangerous counterplay. 9 as ...

This is established as the main line. Black prevents b4 and in a sense he gains a tempo, because White will have to play both a3 and l:tbl in order to achieve the b4-advance. In the 9 tllel line, this move would be ineffective be­ cause White could play something like a3 and il.d2 when the al-rook is pro­ tected by very natural means. After 9 tlld2, however, the d2-knight prevents the cl-bishop from moving, so White does not have this plan available. 10 a3 White can also play 10 l::tb l, when we can still play 10 ... il.d7. After 11 b3 (11 a3 a4) Black can play 11 ...tll c 8, 11 ...tll e 8 or

11...c6 when in each case White does not really have anything better than 12 a3, which will just transpose back into our main lines, below. After 10 a3 Black faces a very impor­ tant strategic decision. First of all, he must avoid the race. Even though 9 ... as has gained a tempo of sorts, it has weakened Black's queenside and after the analogous 10 a3 tlld7 11 l::tb l fS 12 b4 tllf6 13 f3 f4 14 cs axb4 lS axb4 gs 16 tllc4 tllg 6 17 bS! (worse is 17 il.a3 �f7) 17...dxcs 18 b6! Black's queenside is destroyed before he could even dream of a kingside attack. The block­ ing plan with 10 ... cs is more respect­ able, although after 11 l:tbl tlle 8 12 b4 axb4 13 axb4 b6 14 bxcs bxcs lS tllb3 White again does pretty well. Black should still consider this structure in some cases, though, and we will allow it sometimes if we can obtain a rela­ tively favourable version. Traditionally the big choice has been between 10...tlld7 and 10 ... il.d7. Originally I had hoped to include both lines, but then decided that, with 9 tlld2 being considered less critical than 9 tllel and 9 b4 these days, one main line should suffice. I've always kind of liked the move 10...il.d7, because Black may actually play on either side of the board. Instead 10 ...tlld7 actually mir­ rors White's developmental logjam, but because 10 ...tlld7 is twice as popular and a more traditional 'attacking' move, I began to delve into its intrica­ cies first.

77

A ttacking C h e s s : Th e King 's I n dia n , Vo l u m e 1 I decided to avoid 10 ... ll:ld7 because of 11 nb1 fS 12 b4 �h8

13 f3 (after 13 'it'c2, I do not like 13 ... tllg 8 14 exfs gxfs 1s f4, but 13 ... tllf6! 14 f3 axb4 l S axb4 c6! is an effective way for Black to deal with White's threat to play cs and tllc4, and 16 dxc6 ll:lxc6 has proved to be satisfac­ tory for him) 13 ...f4 (after 13 ... axb4 14 axb4 c6 lS dxc6! ? ll:lxc6 White has gained as his queen has not gone to c2 and Black's knight is still on d7, so he is quicker to attack the d6-pawn) 13 .. .f4 (after 13 ...tll g 8 lvanchuk's 14 'it'c2 ll:lgf6 l S il.d3 ! forces 1S ...f4 when Black's knights are suddenly misplaced).

78

This line was Gallagher's recom­ mendation back in 2004 and it was played successfully by Kasparov. How­ ever, the more I delved into White's possibilities, the more disheartened I became. White can prepare the cs­ break in various ways, such as with 14 tllb3, 14 tlla4 and 14 tll bs, and in fact even the immediate 14 cs!? is danger­ ous. As it turns out, even if Black tries to prevent this break with ...b6, White can almost always sacrifice a pawn with c4-cS anyway. Once that knight gets to c4, White obtains excellent play. Although Black has not done so badly in practice with 10 ... tlld7, while studying the relevant games I found it much easier to find improvements for White than for Black Black's pieces all seem to be a bit jumbled up. I once came across a comment by Golubev that said something to the effect that he was not convinced that combining ... as and ... tlld7 had to be logical, and I have come to understand this reason­ ing. After all, the 'logical' 9 ... cs and 9 ... as 10 a3 cs do not stop White's ini­ tiative even though they prevent the c4-cS advance. While 9 ... as and 10 ...tlld7 seems to discourage White's breakthrough, they do not prevent it. So after a lot of time spent on 10 ... tlld7, I felt I had to abandon it, al­ though I think an understanding of these positions is still beneficial, espe­ cially the idea of playing a pre-emptive strike with ... c6. However, when I came around to investigating 10 ... il.d7, I was

Th e M a r def Pla ta Varia tio n : 9 tll d 2 quite pleased with the play that arose. 10 il.d7 ...

This move not only develops a piece, but in effect it gains another tempo because White usually chooses to pre­ vent Black from damaging his pawn structure with ... a4. 11 b3 This by far the most common move, preventing Black from locking down the queenside with ... a4. It is also pos­ sible to ignore Black's idea: a) 11 .l:tbl a4 12 b4 axb3 13 tllxb3 (after 13 .llxb3 b6 14 a4 Black should not bother with 14 ... il.xa4 lS tllxa4 .llxa4 16 .l:txb6!, but instead just play 14 ... tlle8 lS .lla3 fS with counterplay) 13 ...b6 (alternatives are 13 ...il.a4, 13 ... c6 and 13 ... cs 14 dxc6 il.xc6 lS 'it'd3 tll e 8! 16 .lld l fS) 14 .ll al (White is ready now to push the a-pawn, but he has wasted a lot of time) 14 ... tlle 8 (stopping the a­ pawn with 14...'it'e8!? is also an attrac­ tive idea) lS a4 fS and now: al) 16 as bxas 17 tllxas (after 17 .llxas .llx as 18 tll x as 'it'a8 19 tll b 3 tllf6 20 il.d3 .llb8 Black already had a nice

initiative i n L.Brunner-J.Nunn, Nurem­ berg (rapid) 1990) 17 ...tllf6 is fine for Black. White's queenside play has not really gone anywhere. a2) 16 f3 tllf6 (alternatives are 16 ...�h8 and 16 ... cs) 17 il.e3 cs!? (or 17 ... tllh s ! ?) 18 dxc6 il.xc6 was fairly level in E.Gleizerov-1.Nataf, Montecatini Terme 1997. b) 11 .ll a2 hopes to obtain an im­ proved version of variation 'a' by main­ taining the rook on the a-file: 11 ... cs (I still think 11 ...a4 is playable and after 12 b4 axb3 13 tllxb3 Black could inves­ tigate 13 ... c6, 13 ... cs. and 13 ... b6) 12 dxc6 (after 12 b3 tlle 8 13 .llb2 fS 14 b4 axb4 lS axb4 b6 White has wasted a lot of time, so Black should be fine) 12 ...bxc6! (this is an ideal circumstance for this recapture, as White won't re­ gain control of the dS-square) 13 b4 axb4 14 axb4 .llxa2 lS tllxa2 'it'a8 (also possible are 1s ...'it'b6!? and 1s .....lfi.e6!?) 16 tllc3 dS and Black had the initiative in G.Flear-M.Gurevich, Tel Aviv 1989. Instead after 11 b3 White is all set to play 12 llbl and 13 b4.

79

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 One thing I like about the 10 ... 1'.d7 line is the fact that Black's play is flexi­ ble and there are a few decent options here:

2t)c8 8: 11-.lhel

A: 11

...

C: 11 c6 ...

It is also possible to play 11 ... cs, be­ cause Black will end up a tempo ahead of the line 10...as 11 a3 cs as White will have to spend two moves with his b­ pawn. Still, after 12 hi.bl ll:ie8 13 b4 axb4 14 axb4 b6 lS bxcs bxcs 16 ll:ib3 (16 'i¥b3 ll:ic8! exploits Black's extra tempo by covering b6) 16 ...fs 17 f3 lLif6 18 ii.d2 f4 19 hi.al hi.xal 20 'ifxal g s 2 1 Va s White has the initiative. I n general Black should be willing to accept this structure, but only on his own terms. Here even the extra tempo does not guarantee Black equality. A) 11 ll:icS ...

I have always been attracted to this funny-looking move. The idea is to play

80

...ll:ib6 which will allow Black to meet b4 with ...ll:ia4. Exchanging this knight not only provides Black with some re­ lief from his spatial disadvantage, but he may sometimes even take the initia­ tive on the queenside. We will see a similar plan frequently in the Petrosian Variation, although there Black plays ...ll:ia6-cS-a4. Black may also initiate play with ... c6. If White is not careful he can very easily lose the initiative on the queen side. 12 hi.bl This is the natural move, but there are others: a) 12 l:ta2 always looks strange to me. If White plays b4 soon he will end up with a knight on a2, which hardly seems ideal for him. After 12 ...ii.h6 (this is a typical idea for Black in the 9 ll:id2 variation) 13 'ifc2 ll:ib6 14 htdl 'i¥e7 lS 'ifd3 ll:ie8 16 h!.c2 fs Black is al­ ready comfortable. White felt the need to complicate with 17 cs dxcs 18 ll:ic4 ii.xcl 19 d6 ll:ixd6 20 ll:ixb6 fxe4 21 ll:ixe4 ii.ts 22 'ifds+ �g7 23 l:tcxcl ii.xe4 24 'ifxcs cxb6 2S 'i¥xd6 'i¥xd6 26 h!.xd6 ii.c6, but then Black was just up a pawn in S.Skembris-V.Kotronias, Xanthi 1991. b) 12 ii.b2 ll:ib6 (12 ... cs and 12 ...ii.h6 have also been tried) 13 'ifc2 'ife7 (again 13 ... ii.h6 is possible) 14 b4 ii.h6 (or 14... axb4 lS axb4 ii.h6 16 hi.fdl c6!?) 1S ll:ib3 axb4 16 axb4 ll:ia4 17 ll:ias (17 ll:ixa4 ii.xa4 is fine for Black) 17...ll:ixb2 18 'i¥xb2 h!.ab8 19 bs ii.e8! (Black is ready to block the queenside with

Th e M a r d e f Plata Va ria t io n : 9 lLi d2 ...ll:id7 and this causes White to panic) 20 b6?! cxb6 21 ll:ib3 (21 'i¥xb6 ii.d2 wins the e-pawn) 21 ... 'ifc7 left Black with an extra pawn and the bishop­ pair in S.Kishnev-B.Damljanovic Sibenik 1990. c) 12 'i¥c2 ll:ib6 13 a4! ? sees White switch plans, hoping to make the b6knight look silly. He may continue with ii.a3 and b4. Still, this is time­ consuming and freezing the queenside looks strange.

Black has: cl) 13 .. lte8 14 ii.a3 ii.f8 1s ll:i bs cs?! (1s...ii.xbs 16 cxbs ll:ibd7 is better, but after 17 b4 White still has an edge) 16 dxc6 ii.xc6 17 Itfdl 'i¥b8 18 ll:ic3 ll:ibd7 19 ll:if1 ll:ics 20 ii.xcs! dxcs 21 ll:ie3 gave White a positional advantage in D.Komarov-Kr.Georgiev, Lyon 199S. c2) With 13 ...ll:ic8 Black is satisfied that the knight has done its job and now prepares to block the queenside: 14 ii.a3 b6 lS b4 axb4 16 ii.xb4 cs 17 dxc6 ii.xc6 18 ll:ids ll:id 7 19 1'.g4 ll:ics 20 ii.xcs bxcs 21 ll:ib1 ll:ia7 (or 21 ...ll:ib6! ? when 22 ll:ibc3 fails to 22 ...ll:ixc4 and 2 2

ll:ixb6 'ifxb6 23 ll:ic3 Vas is fine for Black) 22 ll:ibc3 ii.e8 23 ll:ibs ll:ic6 24 ll:idc7 Itb8 2S ll:ixe8 Itxe8 26 'i¥d3 ll:id4 27 ll:ixd4 exd4 and Black was at least equal in l.Cosma-B.Damljanovic, Bel­ grade 199S. c3) 13 ...ii.h6 ! ? is untried, but 14 1'.a3 'i¥e7 lS b4? would run into 1S ... ll:ixa4!, exploiting the loose knight on d2. 12 ll:ib6 ...

13 'i¥c2 After 13 b4 axb4 14 axb4 ll:ia4 lS 'i¥c2 Black has: a) 1s ...ii.h6 16 ll:ibs ii.xbs (16...'ife7?! 17 cs! is variation 'b21' to Black's 13th, below) 17 cxbs ll:ib6 18 ll:ib3 ii.xcl 19 Itbxcl ll:ie8 20 ll:ias Itb8 was perhaps slightly better for White in D.A nagnostopoulos-1.Nikolaidis, A no Liosia 199S, but Black is very solid. b) 1s ... ll:ixc3 16 'ifxc3 ii.h6 (this is logical, but Black could consider 16 ...c6, 16 ...h!.a2 or 16 ...'ife7) 17 ii.b2 ! ? (White intends f4, so Black is obliged to grab a pawn) 17 ... ii.xd2 18 'i¥xd2 ll:ixe4 19 'i¥e3 fs 20 f4 exf4 21 h!.xf4 gives White com­ pensation for t he pawn.

81

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1 13 'i¥e7 Black connects his rooks and covers the cs-square. He also has: a) 13 ... ll:ie8!? 14 b4 axb4 lS axb4 ll:ia4 16 ll:ixa4 ii.xa4 17 'if c3 ii.d7 18 cs fs 19 f3 ii.h6 (better is 19 ...ll:if6) 20 c6 and White was ahead in K.Lerner­ G.Pieterse, Amsterdam 1988. b) 13 ... ..ih6 and here: bl) 14 b4 axb4 lS axb4 ll:ia4 (1s ... 'i¥e7 16 c s ! ) 16 ll:ibs leads back to 13 b4, above, while avoiding 1S ... ll:ixc3, although this is no special accom­ plishment for White. b2) 14 ll:ibs 'i¥e7 lS b4 (White can­ not take on c7 because of ... l:tfc8, so he just continues his queenside play) 1S ... axb4 16 axb4 and here: b21) 16 ... ll:ia4?! 17 cs! is a trick Black must watch out for. After 17 ... dxcs 18 ll:ixc7 l:tac8 19 ll:ic4! he was clearly worse in G.Kamsky-V.Spasov, Tilburg 1992. b22) 16 ...h!.fc8 17 cs ll:ia4 was L.Psakhis-V.Spasov, European Team Championship, Debrecen 1992, and now the simple 18 lLif3 ii.xcl 19 l:tfxcl gives White the initiative. b23) Best is 16 ... c6!, a typical pre­ emptive strike: 17 dxc6 ii.xc6 18 ll:ic3 l:tfc8 19 ll:ib3 ii.xcl 20 hi.fxcl ll:ia4 was fine for Black in Wu Shaobin-M.Al Mo­ diahki, Beijing 199S. 14 b4 This is consistent. Instead 14 ii.b2 ii.h6 lS l:tbel 1'.g4!? is possible, which is an idea we will see too in the Petro­ sian Variation. After 16 ii.xg4 ll:ixg4 17 •..

82

a4!? ll:id7 18 ll:if3 ll:ics 19 ii.a3 b6 20 ii.xcs bxcs 21 g3 ll:if6 the position was fairly level in M.Gurevich-1.Smirin, Haifa 199S. 14...axb4 15 axb4

1s ...ll:ia4 Or 1S ... l:tfc8 16 ll:ib3 ll:ia4 17 ii.d2 ll:ixc3 18 ii.xc3 c6 19 dxc6 ii.xc6 (19 ...bxc6 !?) 20 f3 dS! 21 cxds ii.xds 22 'i¥b2 ii.e6 23 hi.al ll:id7 with equality in A.Beliavsky-Ki.Georgiev, Groningen 1994. 16 ll:ixa4 White acquiesces to the exchange of knights. Instead 16 ll:id1?! is too pas­ sive. After 16 ...ll:ig4!? (16 ... c6 also looks good) 17 f3 (or 17 h3 ll:ih6 with the idea of ... fs) 17 ...ll:if6! 18 ll:ib3 ll:ihs (the point of Black's manoeuvres) 19 ll:ias ll:if4! Black had the initiative in R.Kasimdzh anov-1.Smirin, Elista Olym­ piad 1998. After 16 cs Black should avoid 16 ... ii.h6 17 ll:ibs! and play either 16...h!.fc8 or 16 ... ll:ixc3 17 'it'xc3 c6!?, with the idea of 18 cxd6 'it'xd6 19 lLic4?!

Th e M a r def Plata Va riati o n : 9 lLid2 16 ... ii.xa4 17 'ifc3 After 17 'it'd3 the moves 17 ... c6, 17 ...l:tfc8, and 17 ...1'.d7 all look reason­ able. 11...ii.h6 Pressuring the e4-pawn. 17... c6!? is another idea. 18 'i¥d3 Instead 18 f3 is met by 18 ...ll:ih5!. White can again consider sacrificing a pawn with 18 ii.b2 ii.xd2 19 'i¥xd2 ll:ixe4 20 'it'e3 f5 21 f4, although this is somewhat speculative. 18 ... c6 19 l2Jf3

26 �g2 'ifd7 21 ll:ig1 ii.xe2 28 ll:ixe2 fs Trading light-squared bishops is not always fatal in the King's Indian if Black can generate counterplay. Here Black could also consider 28 ... 'ifb5 29 ll:ic3 'i¥c4 with equal chances. 29 ll:ic3 ll:if6 30 f3 l:tfc8 31 'i¥d3 l:ic7 32 ll:ie2 h!.xcl 33 h!.xcl h!.a2 34 h!.c2 After 34 ll:ic3 fxe4! 35 fxe4 l:tb2 36 l:tbl l:txbl 37 'ii' x bl 'i¥c7 38 'it'd3 'ifb6 Black has counterplay. His queen is ac­ tive and the e4-pawn will always need attention. Given time, Black can also consider something like ...h 5, ...�h7 and ...ii.h6. 34...fxe4 35 fxe4 'if a4 Black had sufficient play in M.Gurevich-1.Smirin, Elenite 1994. B) 11...ll:ieS

19 ...1'.g7!? Black keeps pieces on the board. Af­ ter 19 ...ii.xcl 20 l:tfxcl Mikhail Gure­ vich believes White is better, but Black looks okay to me. 20 ii.gs cxds 21 cxds h6 22 ii.d2 1'.d7 Black could also consider playing 22 .. JUc8 immediately. 23 h!.bcl 'ifdB 24 'it'b1 24 b5 can be met by 24...h!.a4! when the e4-pawn is in trouble. 24... ll:ihs 2s g3 1'.g4 Not 25 ... f5? 26 ll:ih4.

This typical King's Indian move has recently garnered some attention. Black simply prepares .. .f5. 12 hi.bl f5 13 b4 Instead 13 f3 is well met by 13 ... ii.h6, while Black could also con­ sider 13 .. .f4.

83

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1 (the complications begin) 18...fxe4 19 fxe4 dxcs 20 d6 cxb4 21 dxe7 'ifxe7 22 ll:ia4 ll:ig4 23 ll:ib6 ll:if2+ 24 h!.xf2 h!.xf2 2 s 'it'e1 (2s ll:ixa8 'ii'g s!) 2 s ...l:taf8 26 ll:ixd7 'ii'g s ! 27 g3 hi.8f7 28 ll:ics 'it'e3 29 1'.c4 was R.Nolte-W.So, Tagaytay City 2010. Now 29 ... 'ifxcs 30 ii.xf7 hi.xf7 would be unclear.

13 ...axb4 14 axb4 ll:if6 This is by far the commonest move, but Black could also consider 14...�h8 1s cs ll:ig8 16 f3 ll:igf6 17 ll:ic4 ll:ihs or 14... ii.h6 lS cs fxe4 16 ll:idxe4 ii.xcl 17 'ifxcl ll:ifs. 15 cs The most direct. Instead lS f3 c6!? (or just 1 s...ii.h6), and here: a) 16 cs!? cxds 17 exds ll:iexds 18 ll:ixds ll:ixds 19 1'.c4 ii.e6 20 ll:ie4 (this is the point of White's play, but it does not lead to an advantage) 20...fxe4 21 ii.xds 'ifd7 22 'tlfb3 ii.xds 23 'ifxds+ �h8 24 'ifxe4 ds 2 S htdl l:tad8 was level in C.Bemard-A.David, Andorra 1996. b) 16 dxc6 ll:ixc6 17 ll:ibs (17 ll:ib3 ii.e6 18 ll:ibs fxe4 19 ll:ixd6 exf3 was unclear in E.Gleizerov-1.Zaitsev, Podolsk 1992) 17 ... ii.e6 18 'it'c2 'ife7 (or 18 ... l:tc8, as in R.Scherbakov-J. Nunn, H astings 1993/94) 19 ii.d3 ll:ihs (both 19 .. .f4!? and 19 ...ll:ie8 ! ? are possible too) 20 cs dxcs 21 'ifxcs was V.Chuchelov-J.Nunn, Leeuwarden 199S, and now 21 ...ll:if4 would give Black a slight initiative. c) 16 �h1 �h8 17 ii.b2 h s 18 c s ! ?

84

15 ...fxe4 Black gives up the e4-square in or­ der to generate active piece play him­ self. There are a couple of alternatives: a) 1s ...�h8 16 f3 ll:ihs 17 ll:ic4 ll:if4 18 ii.e3 ll:ic8 (Black covers the d6-pawn and intends ...'if gs with counterplay) 19 �hl! g s 20 ll:ias (20 b s also looks good) 20...'ife8 21 ll:ixb7 h!.a3 was M.Gurevich-B.Gelfand, Munich 1993. Now 22 l:tcl seems to refute Black's play because after 22 ... h!.xc3 23 h!.xc3 ll:ixe2 24 'ifxe2 ii.bs 2S h!.c4 Black can­ not exploit the pin and White will re­ main a good pawn up. b) 1s ... ii.h6!? is Black's latest at­ tempt to fight for the initiative: 16 ll:ic4 (16 ii.b2 and 16 f3 are possible as well) 16 ... ii.xcl (perhaps Black should prefer

Th e M a r def Plata Va ria t io n : 9 lLid2 16 ... ll:ixe4!? 17 ll:ixe4 1'.a4 18 'ifel ii.xcl 19 l:txcl fxe4) 17 'ifxcl (after 17 h!.xcl ll:ixe4 18 ll:ixe4 fxe4 19 cxd6 cxd6 20 ll:ixd6? is not possible because of 20 ... ii.a4) 17 _.fxe4 18 cxd6 and here: bl) 18 ...ll:iexds 19 ll:ixds ll:ixds 20 ll:ixes ii.e6 21 dxe7 'ifd6 (no better is 21...'ifxe7 22 'i¥xc7 ll:ixc7 23 l:tbcl) 22 'i¥b2 and White is more comfortable. b2) 18 ...ll:ifs! ? and then: b21) 19 ll:ixes?! cxd6 20 ll:ic4 l:tc8 (20... bs!?) 21 'if d2 hi.e8 22 l:tfel and now 22 ... e3! 23 ll:ixe3 ll:ixe3 24 fxe3 'ifb6 2 S ii.f3 hi.c4 gave Black excellent play for the pawn in A.Khalifman1.Cheparinov, Amsterdam 2007. b22) 19 dxe7 'i¥xc7 20 d6! ? (20 l:tdl e3! 21 ll:ixe3 hi.fc8 22 ll:ic4 bS is not clear, but both 20 'i¥d2 and 20 'i¥b2 look somewhat more pleasant for White) 20 ... ll:ixd6 21 ll:ixes ii.ts? (21 ... hi.ac8) 22 ll:ixe4! won a pawn in V.Chuchelov­ J.Degraeve, German League 2008. 16 ll:idxe4 ll:ifs 17 ii.gs White is willing give up his dark­ squared bishop to consolidate his grip on e4.

11...h6 This natural move has always been played, even though 17 ... 'i¥e7 and 17 ... 'i¥e8 look reasonable. In fact I be­ lieve the strongest move is 17 ... ll:ixe4! 18 ll:ixe4 (18 ii.xd8 ll:ixc3 19 'if el ll:ixe2+ 20 'i¥xe2 l:taxd8 gives Black three won­ derful minor pieces for the queen) 18 ... 'i¥e8, which looks fine for Black, while here 18 ... 1'.a4!? may be possible as well. 1s ii.xf6 ii.xf6 19 bs White's most recent try was 19 hi.al and 19 ... l:txal 20 'ifxal ii.gs?! 21 'ifa7 was much better for him in A.Beliavsky­ Z.Lanka, Austrian League 2007. Black can improve with 20 ... ll:id4 21 ii.d3 ii.el when White has a small edge at best after 22 'i¥a2. 19...b6 20 c6 ii.cs 21 ii.d3 ii.gs 22 l:ta1 l:txal 23 'ifxal

23 ... hs This is Nunn's recommendation. In­ stead 23 ...1'.f4?! 24 ll:ie2 'ifh4 was G.Kamsky-J.Nunn, Monaco (blindfold) 199S, when Nunn gives 2S g 3 ! 'ifh3 26 ll:ixf4 exf4 27 hi.el! ll:ih4 28 gxh4 'ifxd3

85

A ttacking Chess: Th e Kin g 's I n dian, Vo l u m e 1 29 ll:if6+ �f7 30 'i¥a7! and White is winning. After 23 ... hs Lukacs and Hazai give 24 ll:ixgs 'ifxgs 2s f4 (?!) exf4 26 ll:ie4 'i¥e7 27 h!.xf4 as slightly better for White, but Black can play the more ac­ tive 26 ...'t!Vh4! 27 'ii'a1 ll:ie3 28 'if xe7 ii.ts with the idea ... 'ifg4. After 29 g3 fxg3 30 hxg3 'i¥g4 Black has the initiative.

allows Black t o show another idea be­ hind 11 ... c6.

C) 11...c6

This move was once rather uncom­ mon, but has become established as the main line. Black creates more tension in the centre and keeps his options open on both sides of the board. White now has a choice: C1: U .:.%b1 C2: 12 lla2

C3: 12 1'.b2 Cl) 12 l:tb1

(seefollowing diagram) This is the most natural move, but it

86

12 ... bs1 This opportunistic move gives Black good counterplay. He can also play more quietly with 12 ...'ifb8 13 b4 axb4 (after 13 ... cxds 14 cxds hi.c8 White can play lS ll:ic4! targeting b6 and as, as in Metz D.Giacomazzi-A.Shchekachev, 199S) 14 axb4 (White may be better off with 14 l:txb4 as in M.Notkin­ V.Bologan, Saint Petersburg 199S, when Black should probably play 14 ... l:ta7) 14... cxds 1s cxds l:tc8 (or the immediate 1s ... bs) 16 ii.b2 bs (16 ... ii.h6!?) 17 ii.d3 'ifb6 18 ll:ib3 ii.h6 19 ll:ias hi.Cl gave Black good play in l.Nemet-J.Gallagher, Swiss League 1994. 13 dxc6 Taking the pawn with 13 cxbs gives Black excellent chances after 13 ...cxds 14 exds ii.ts! (but not 14... ll:iexds? 1s ll:ixds ll:ixds 16 ll:ie4) 1s l:tb2 ll:ifxds 16 ll:ixds ll:ixds. Ignoring Black's play with 13 b4 also comes to nothing after 13 ... axb4 14

Th e M a r def Plata Va ria tio n : 9 l2Jd2 axb4 bxc4 15 l2Jxc4 cxds 16 exds ii.ts 17 hi.b3, which was V.Chuchelov­ J.Filipek, Leuven 1999, and here Black has 17 ... l2Je4!, with the idea of 18 ltJas 'if c8! when he enjoys excellent play. 13 . b4! ..

White's jumbled queenside pieces allow Black to fight for the initiative. 14 ctJdS Instead 14 cxd7 bxc3 15 l2Jf3 l2Jxe4 is the main tactical justification of Black's play. White does have a few other tries: a) 14 C7 looks too fancy after 14 ... 'ifxe7 1 5 ltJds ltJexds 16 cxds a4! (Black's better development allows him to take over the initiative on the queenside) 17 axb4 h!.fb8 18 hi.al (or 18 bxa4 ii.xa4 19 'ifel ii.c2 20 h!.b2 l2Jxe4 when Black is much better) 18 ... 'i¥c3 19 ii.a3 axb3 20 l:tcl 'i¥d4 and White posi­ tion fell apart in G.Grigore­ V.Nevednichy, Calarasi 1995. b) 14 ltJbs is fairly safe. After 14 ... ii.xc6 15 ii.f3 'tlfb8 16 a4 l2Jd7 17 Itel ltJcs 18 l2Jf1 l:td8 the position was level in G.Grigore-J.Baron Rodriguez,

Benasque 1999. c) 14 axb4 axb4 and now: cl) 15 cxd7 fails again after 1 5 ... bxc3 16 l2Jf3 l2Jxe4. c2) 15 ltJds l2Jxc6 16 l2Jxf6+ ii.xf6 is certainly fine for Black c3) 15 C7 'ifxe7 16 ltJbs (after 16 ltJds ltJexds 17 cxds l:ta2 18 ii.d3 hi.b8 19 Itb2 Itxb2 20 ii.xb2 ii.e8! 21 l2Jc4 l2Jd7 22 ii.cl ltJcs Black was totally equal in J.Gokhale-P.Mahmoud, Cal­ cutta 1996), and here 16 ... ii.xbs 17 cxbs ds gives Black counterplay accord­ ing to Bologan, but 16...'ifcs ! looks even better. 14...l2Jxc6 1s l2Jxf6+ Instead 1 5 ii.b2 ii.e6 16 l:tal l2Jd7 17 f4?! exf4 18 ii.xg7 �xg7 19 hi.xf4 ltJcs 20 l2Jf1 'ifgs was much better for Black in L.Ftacnik-R.Gadjily, Moscow Olym­ piad 1994, and Black indeed managed to score a notable upset. 15 ...ii.xf6 16 l2Jf3 'i¥e 7 17 a4

This structure is fine for Black He has good squares for his pieces and can easily cover his only weakness on d6. 17 ...l2Jd8!

87

A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1 The knight belongs on cs. 18 ii.h6 l:te8 19 'ifd2 ii.c6 20 ii.d3 ll:ie6 Black had an excellent position in the game P.Lukacs-W.Uhlmann, Vienna 199S. He intends moves like ...'ifb7, ...ii.d8 and ...ll:ics. C2) 12 l:ta2

This move always looks strange, but it does prepare the b4-advance without allowing Black the counterplay we saw in Line Cl. 12...'ifbS Here 12 ... bs? fails to 13 dxc6 b4 14 cxd7 bxc3 because White has lS ll:ibl! when 1S ... ll:ixe4 loses to 16 ii.f3 fs 17 ii.xe4 fxe4 18 'i¥xd6. Instead 12 ... cs is as usual solid but stodgy after 13 h!.b2 ll:ie8 14 b4 axb4 lS axb4 b6 16 bxcs bxcs. Others: a) 12 ... ii.h6 13 h!.c2 cs 14 ll:idbl! 1'.g7 (14...ii.xcl lS l:txcl with the idea of f4 gives White the initiative) 1 S l:tb2 ll:ie8 16 b4 was M.Gurevich-R.Kasim­ dzhanov, Hoogeveen 1999. and here 16 ...axb4 17 axb4 b6 18 bxcs bxcs 19 ll:ibs fs 20 ll:i1c3 looks better for White.

88

b ) 12 .. JWb6 13 hi.c2 cs 14 l:tb2 l:tfb8 lS ll:idbl 'i¥d8 16 b4 axb4 17 axb4 b6 was very solid for Black in M.Gurevich­ A.Shirov, Prague (rapid) 2002. Black has at least managed to connect his rooks here and can hold firm on the queen­ side. c) 12 ...ll:ie8 is very sensible and looks fine. After 13 b4 axb4 14 axb4 l:txa2 l S ll:ixa2 cxds 16 cxds f s Black had good play in M.Gurevich-V.Babula, German League 2000.

13 hi.c2 hi.cs 14 ll:idb1 A good example of Black's play was 14 ii.d3 cxds 1 S cxds bs 16 b4 axb4 17 axb4 ll:ihs 18 hi.el ii.h6 19 ll:idbl ii.xcl 20 'ifxcl hi.c7 21 'i¥e3 fs 22 f3 f 4 23 'i¥e2 gs 24 l:tec1 l:tb7 2s ll:id1 ll:if6 26 g4 h s 27 h 3 h!.a4 2 8 l:tb2 'i¥a7+ 2 9 'iff2 'i¥d4 with good play in V.Chuchelov­ V.Bologan, Istanbul 2003. 14 ... cxds Black can also play 14 ... bs when lS cxbs cxbs leads to the same position. 15 cxds bs Black has a piece formation that is very typical of the Ruy Lopez or the Old

Th e M a r def Plata Varia tio n : 9 lLid2 Indian Defence. With such good devel­ opment, Black is not afraid to initiate play on the queenside. 16 b4

Black had good counterplay in A.Beliavsky-D.Stellwagen, Amsterdam 2006. C3) 12 ii.b2

16...ll:ieS This is a very flexible move, not rushing to determine the queenside pawn structure. Black has also resolved the tension immediately: a) 16 ... a4 is probably a little prema­ ture, as White is not badly placed to fight on the kingside: for example, 17 ii.e3 l:tc7 18 l:tc1 'i¥e8 19 'if d3 l:tb7 20 f4 and White had the initiative in M.Gurevich-Y.Vovk, Cappelle la Grande 2010. b) 16 ... axb4 is not so bad. After 17 axb4 l:ta1 18 ii.d3 ll:ihs 19 g3 l:ta7 20 ii.gs f6 21 ii.e3 hi.ac7 22 f3 fs Black had counterplay in D.Sharavdorj-M.Al Mo­ diahki, Yangon 1999. 17 ii.e3 f5 18 f3 ll:if6 19 'ifd2 a4 Now Black decides to close off the queenside because he will have good play on the other side of the board. 20 h!.fcl h!.c7 21 'if d3 htb7 22 ll:id1 f4 23 ii.f2 gs

This move makes a lot of sense. White completes his development and does not declare his intentions just yet. 12 ...ii.h6 This is a typical deployment, al­ though Black must mind the long di­ agonal that White has just occupied. Others: a) 12 ... 'ifb8 is the thematic move, but White is better developed here and after 13 b4 axb4 14 axb4 h!.xa1 1s 'if xal cxds (Bologan recommends Black try to draw by sacrificing a pawn starting with 1s ... bs!?) 16 cxds hi.c8 17 Vas White has maintained control of the queenside and enjoys a slight edge. b) 12 ... 'ifb6 13 dxc6 (13 l:tcl cs is solid) 13 ... bxc6 (Black could consider recapturing with a piece) 14 ll:ia4 'ifc7 1s cs ds 16 ll:ib6 l:tad8 17 ii.c3 with complications in V.Malakhov­ V.Zvjaginsev, Poikovsky 2004. Here

89

A tt a c k i n g C h e s s : The King 's In dian, Vo l u m e 1 Black could consider 11 ... ii.e6 or 17 ... a4. c) 12 ... cs is common, but after 13 ll:ibs ll:ie8 14 b4 axb4 lS axb4 l:txal 16 'ii'xal ii.h6 (16 ... b6 17 bxcs bxcS 18 ii.c3 gives White a pleasant edge), 17 ll:ixd6 ! is a strong sacrifice, M.Ulibin-A.Galkin, Russian Championship, Elista 1996. 13 'it>hl White takes a moment to make a prophylactic move and again does not disclose his intentions. This is a patient approach, but many other moves have been tried too. a) 13 dxc6 ll:ixc6 (this is fine, but both 13 ... ii.xc6 and 13 ... bxc6 are possi­ ble too) 14 ll:ibs ii.e6 1s ii.d3 ll:id7 16 ll:ib1 ll:ics 17 ll:i1c3 ll:ia7 18 ll:ixa7 l:txa7 19 ii.c2 h!.a6 20 'ii'e2 h!.c6 21 l:tadl 'ii'b6 22 ll:ibs a4! 23 b4 ll:ib3 gave Black good counterplay against the c4-pawn in Shen Yang-L.Van Wely, Moscow 2009. b) 13 .rf.bl ll:ic8!? (again we see this manoeuvre) 14 dxc6 ii.xc6 lS ii.d3 ll:ib6 16 b4 axb4 17 axb4 ll:ihs (after 17 ... 1'.d7 18 ll:ib3 ll:ia4 19 ll:ixa4 .Jha4 20 ii.c3 bS!? 21 'ii' e2 - or 21 cxbs 'ii' b6 21 ...bxc4 22 ii.xc4 ii.d7 23 hi.fdl 'ii'b6 Black was also doing well in Li Wenli­ ang-F.Nijboer, Groningen 2003) 18 g3 ii.d7 19 'ii'e2 l:tc8 20 ll:ib3 ll:ia4 21 ll:ixa4 ii.xa4 22 ii.cl ii.xcl 23 l:tfxcl ii.xb3! 24 Itxb3 ll:ig 7 and with the knight bound for e6, Black had good play in L.Ftacnik­ V.Topalov, Polanica Zdroj 199S. c) 13 cs!? is a sharp try: 13 ...dxcs (better than 13 ... cxds?! 14 cxd6 ll:ic8 l S ll:ixds ll:ixds 16 exds f6 17 a4 ll:ixd6 1 8 ii.a3 ii.ts 19 ll:ic4 ll:ixc4 20 ii.xc4 when

90

White had the initiative in R.Sherbakov-V.Syrtlanov, Koszalin 1996) 14 ll:ic4 cxds 1s exds e4! (after 1s ... ll:ifs 16 ll:ixes ll:id4 17 1'.c4 ii.ts 18 ll:ie2 ii.c2 19 'ifel 1'.g 7 20 ll:ixd4 cxd4 21 ii.xd4 ll:ixds 22 ll:ixf7! White was doing well in R.Sherbakov-A. Korotylev, Par­ dubice 1996) 16 ll:ia4 (after 16 ll:id6 Black played 16 ... ii.g7 in F.Atakis­ T.Remmel, correspondence 2003, but 16 ... ii.fs, with the idea of 17 ll:ixb7? 'ii' b6, looks even better) 16 ... ll:iexds 17 ll:ixcs ii.c6 18 'it'd4 ll:if4 19 'it'xd8 ll:ixe2+ 20 �hl l:tfxd8 21 ii.xf6 .rf.e8! ? 22 ll:id6 ii.f8 23 ll:ixe8 ii.xcs gave Black compen­ sation for the exchange in R.Jansen­ F.Nijboer, Dutch Championship, Rot­ terdam 1999, and 21 ... l:tds 22 ll:ixe4 l:tbs with a strong initiative is a good alternative. 13 cs ...

White's slow play finally convinces Black to acquiesce to the blocking of the queenside. 14 ii.d3 After 14 'it'c2 ll:ie8 1s ll:ibs fs 16 ii.d3 fxe4 the game was agreed drawn in

Th e M a r def Plata Va ria tio n : 9 lLid2 K.Sakaev-F.Amonatov, Dagomys 2009. If we continue with 17 ll:ixe4 ii.xb5 18 cxb5 both 18 ...ll:if5 and 18...ll:if6 look fine for Black. White can also swing back his knight with 14 ll:if3, but 14...ll:ih5 (more active than 14 ... ll:ie8 15 ii.cl ii.xcl 16 'it'xcl ll:ic8 17 ll:ig1 ll:ig7 18 f4! with the initiative in A.Beliavsky-M.Erdogdu, European Championship, Budva 2009) 15 g3 f5 gave Black good play in Zhao Xue-Wenjun Ju, Jiangsu Wuxi 2008. 14...ll:ihs 15 g3 1'.h3 16 hi.el fs 11 exfs ll:ixf5 18 ll:ide4 Of course not 18 g4? ii.xd2. 18 ...ll:id4 Black has active pieces and no struc­ tural problems. White decides to elimi-

nate the d4-knight, but this allows Black to contest White's domination of the e4-square.

19 ll:ibs ll:ixbs 20 cxbs ll:if6 By now Black already had a slight initiative in V.Kramnik-L.Van Wely, Nice (blindfold) 2008.

91

Ch,a,pte'(c 5 Tfte Mar· def Plata Variation

1 d4 ll:if6 2 c4 g6 3 ll:ic3 1'.g7 4 e4 d6 5 l2Jf3 0-0 6 ii.e2 es 7 0-0 ll:ic6 8 ds ll:ie7 9 b4

The Bayonet. White had tremen­ dous success in the second half of the 1990s with this move which forced Black to really shore up his defences. White's king is usually (but not al­ ways!) pretty safe and despite the rela­ tively closed nature of the position, there is often a lot of play in the centre. 9 ll:ihs This is the challenging main line, which will be the variation to which we ...

92

will devote our attention. Black takes advantage of the possibility to jump to a more active square than d7 or e8, eyes the f4-square and prepares .. .fs.

It is easy to forget that the critical 10 hi.el of our next chapter is a rela­ tively new move. Even though it only came to prominence as late as 1995, it has been played more than all the other moves combined and by quite a wide margin. Nevertheless, White has other continuations on move 10 that Black cannot ignore and it is these lines thatare examined in this chapter.

Th e M a r def Plata Vari a t i o n : 9 b4 tl'i h s with o u t 1 0 hi. e l

A: 10 �2 8.: 10 tlt'c2 C: 10 C5

0; 10 g3 A) 10 tl'id2

White forces Black's knight to f4. This old move has been considered in­ ferior since the well-known game Pet­ rosian-Gligoric, Zagreb 1970, but it is likely that this attractive game does not represent best play for either col­ our. Nowadays this move is very rare and this is understandable because Black should not really experience any problems. 10...tl'if4 11 ..if3 White can also play 11 a4 fS 12 ii.f3, transposing to the main line. This avoids the possibility given in the note to Black's 11th move, but gives Black the option of playing 11 ... tl'ixe2+ 12 'ifxe2 fs. 11...fs Black can also play 11 ...tl'id3 12 ii.a3 as 13 bxas hi.xas, as in V.Anand-

B.Gelfand, Dortmund 1997. As the main line is fine for Black, we will focus our attention there. 12 a4 gs This is commonest and quite sound, but Black can already consider experi­ menting: a) 12 ...�h8!? 13 tl'ib3 g s 14 exfs tl'ixfs 1s g3 ll'ih3+ 16 �g2 tl'ih6!? 17 1'.e4 ll'ig4 18 h!.a2 'ife8 19 f3 ll'if6 20 ii.bl 'ifhs was unclear in G.Sosonko­ F.Nijboer, Dutch Championship 1996. b) 12 ...ii.d7 13 cs (13 ii.a3 gs 14 exfs tl'ixfs 1s g3?! ll'ih3+ 16 �g2 'ifc8! is good for Black) 13 ...gs 14 exfs tl'ixfs 1s g3 ll'ih3+ 16 �g2 'ifc8 17 1'.e4 g4 was also messy in D.Zagorskis-G.Beckhuis, Muenster 1994. 13 exf5 ll'ixfs 14 g3

14... ll'ih3+ This looks sounder than 14 ... ll'id4!? lS gxf4 tl'ixf3+ (1s ...exf4!? has been suggested as a better move order, be­ cause 16 tl'ide4 tl'ixf3+ 17 'ifxf3 g4 leads to variation 'a', below, but White could try 16 l:ta3 ! ?) and here: a) 16 'ifxf3 g4 17 't!fh1? exf4 18 ii.b2

93

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 ii.ts 19 l:tfel f3 20 ll:ide4 'ifh4 21 h3 ii.es gave Black a strong attack in T.Petrosian-5.Gligoric, Zagreb 1970. b) However, 16 ll:ixf3 ! gives Black problems: 16 ... e4? (this loses, but 16 ... exf4 17 ii.b2 g4 18 �hl! gxf3 19 l:tg l also gives Black big problems) 17 ll:ixg s ! ii.xc3 18 l:ta3 1'.g7 19 htg3 and, facing an overwhelming attack, Black resigned in A.Martin-R.Britton, British Championship, Eastbourne 1991. Black could, however, throw in the move 14 ... as!? to disrupt White's queenside. In fact, I think this move would improve Black's cause in almost any position that follows. 15 '>t>g2 'it'd7! Instead 1 S...g4?! 16 ii.xg4 ll:ixf2 backfired after 17 'it>xf2 ll:ih6+ 18 ii.f3 ll:ig4+ 19 '>t>gl ll:ie3 20 'it'e2 ll:ixfl 21 ll:ide4 1'.h3 22 ii.gs 'i¥d7 23 ii.g2 with a clear advantage for White in J.Parker­ J.Gallagher, Hastings 1991/92, but here too 1s ... as!? is possible.

16 1'.e4 Other moves are likely to bring White trouble:

94

a) 16 1'.g4? is bad because 16 ...ll:ixf2! does work now. After 17 �xf2 ll:id4+ 18 ii.f3 g4 Black is winning. b) 16 ll:ide4 is natural, but White's light-squared bishop becomes a target after 16 ... l2Jd4 when 17 ii.hs ? ll:if4+ 18 gxf4 'ifh3+ 19 �hl g4 gives Black a winning position. c) 16 ll:ib3 ll:id4 (16 ... as!?) 17 ll:ixd4 exd4 18 ll:ibs c6 19 ll:ia3 (instead 19 ll:ixd4 l:txf3 20 �xf3 'if g4+ 21 �e3 looks crazy, but 21...cS 22 bxcs dxcs 23 ii.b2 is actually not so clear) 19 ... h!.xf3 (sim­ pler is 19 ...ii.es with an excellent posi­ tion) 20 'ifxf3 g4 with compensation in R.Keene-L.Kavalek, Teeside 197S.

16...g4 Yet again 16 ... as looks good, while 16 ... 'iff7 is also interesting. Instead 16 ...ll:ixf2 17 �xf2 ll:ixg3+ 18 '>t>gl ll:ixfl 19 ll:ixfl was unclear in D.Ruzele­ D.Lapienis, Lithuanian Championship 1993, but I would rather not trade away Black's active minor pieces, and 16 ...ll:id4 17 f3 Itf6 (of course 17 ... as! is again a likely improvement) 18 ll:ib3 l:th6 19 l:ta2 ll:ixb3 20 'i¥xb3 'iff7 21 cs

Th e M a r def Plata Va riatio n : 9 b4 tl'i h s with o u t 1 0 l:t e1 was a little better for White in Vilnius V.Grabliauskas-R.Zukauskas, 199S. 17 ll'ib3'i¥e7 18 'ifd3 hS 19 cs

19...ll'if4+ This is a bit speculative. Black could consider 19 ... as or 19 ... tl'igs with coun­ terplay. 20 gxf4 exf4 21 f3 Panczyk and llczuk suggest 21 'it>hl ii.xc3 22 i.xfs i.xal 23 ii.xc8 l:taxc8 24 tl'ixal h!.ce8 2S 1'.xf4 'i¥e4+ 26 'ifxe4 h!.xe4 27 ii.g3 when White may be a little bit better in the endgame. 21... ii.es 22 hi.a2 'i¥g7 23 tl'id1 Instead Smirin gives 23 ii.xfs ii.xfs 24 ll'ie4 l:tae8 2S �hl 'i¥g6 with com­ pensation in Info rmant 69. 23 ... g3 24 htgl gxh2+ 25 �xh2 ll'ig3 Black had some compensation for the piece and managed to overwhelm his opponent in J.Manion-1.Smirin, Las Vegas 1997. B) 10 'i¥c2 Epishin's move looks funny but it should be taken seriously. Black pro-

tects the c3-knight in anticipation of ...ll'if4; ilxf4 exf4, after which he will try to overrun Black in the centre. One downside to this move is that with his queen off the d-file, White will lack control of dS and perhaps d6, and this could give Black some additional possi­ bilities to undermine the centre. A similar idea for White is 10 'ifb3 ll'if4 (Black can also consider 10.. .fs! ? or 10 ... h6) 11 ii.xf4 exf4 12 l:tadl h6 13 cs gs 14 es! (this is basically forced, as Black would otherwise play 14 ...tl'ig6 with a great position) 14... dxes lS d6 cxd6 (after lS ...tl'ifs both 16 dxc7 'ifxe7 17 tl'ids and 16 h 3 ! ? cxd6 17 ll'ie4 ds 18 l:txds 'ife7 19 .�c4, as in S.Skembris­ A.David, Cannes 1991, give White com­ pensation for the pawn) 16 l:txd6 and here both 16 ...'ifC7 and 16 ...'i¥e8 have been tried with unclear play.

10.. .fs I like this straightforward move. In­ stead 10...tl'if4 11 1'.xf4 exf4 12 h!.adl seems to play into White's hands: for example, 12 ... as 13 bs h6 14 ll'id4 gs 1s ii.h s ii.es 16 h3 tl'ig6 17 tl'if3 'i¥e7 18

95

A t ta c k i ng C h e s s : Th e King 's Indian, Vo l u m e 1 b6!? cxb6 19 ll:ia4 hta6 20 'it'b3 and White had the initiative in A.Zhigalko­ V.Kovalev, Minsk 2010. A decent alternative, however, is 10 ... as 11 bxas hi.xas 12 hi.el cs 13 a4 h!.a6!? with a position similar to the line 10 hi.el as 11 bxas hi.xas 12 ll:id2 ll:if4 13 ii.fl cs 14 a4 h!.a6, but in this case Black will not have to waste as much time playing the typical retreat ... lLif4hS-f6. As that line is not part of our repertoire after 10 l:tel, I will stick with the text move, which I think gives Black good chances. 11 ll:igs After 11 cs I would propose 11 ...ll:if6!? 12 ll:igs fxe4 (this is logical because playing 'ii'c2 and cs has left the dS-pawn needing White's atten­ tion), with the idea of ... ll:ifs when the queen on c2 is a target.. 11...ll:if4 Also good is 11 ... ll:if6 when 12 ii.f3 (possible too is 12 f3, but I cannot see how White's queen on c2 can be an improvement on the normal hi.el in this position) 12 ... h6 (again 12 .. .fxe4 with the idea ... lLifS-d4 is also a reason­ able idea) 13 ll:ie6 ii.xe6 14 dxe6 was played in A.Yermolinsky-D.Sharavdorj, Berkley 200S, and here Panczyk and llczuk suggest 14...'it'c8 ! lS lLids ll:ifxds. This is rarely this a good idea, but it works here: 16 cxds c6 and Black has good counterplay. 12 ii.xf4 This is always played, but 12 ii.f3 must also be considered: 12 .. .fxe4 in-

96

tending ...ll:ifs looks like a good answer, while Black could consider throwing in 12 ... as!? too. 12 ...exf4 13 h!.ael

White is loading up in the centre, but Bl ack has a good tactical answer. 13 ...ll:ic6! 14 dxc6 After 14 ll:ie6 Black has a couple of good responses: a) 14...ii.xe6 lS dxe6 ii.xc3! 16 'ii' x c3 fxe4 17 1'.g4 hs 18 1'.h3 (or 18 ii.d1 'ii'f6) 18 ...gs with good play. b) 14 ... ll:id4 lS ll:ixd4 ii.xd4 16 ii.f3 ii.es 1 7 cs as 18 'it'b3?! (18 a3 is more solid) 18 ... axb4 19 'ii'xb4 was V.Epishin­ V.Menoni, Bratto 1999, and here 19 ... dxcs 20 'ii'xcs 'ii'd6 looks good for Black. 14...'ii' xgs 15 ii.f3 After lS cxb7?! ii.xb7 White has problems on the long diagonal. 1s ... bxc6 16 bs 1'.d7 A decent alternative is 16 .. .fxe4 17 ll:ixe4 (17 ii.xe4 f3!) 17...'it'fs 18 bxc6 h!.b8 with good play for Black. 17 exfs 'ii'xfs 18 1'.e4 'ifgs 19 bxc6 1'.h3 20 'ii'd 3 ii.xg2! 21 ii.xg2 ii.xc3 22 h4

Th e M a r def Plata Varia tio n: 9 b4 tl'i h s with o u t 1 0 hi. e l 'it'f6 23 hi.e6 'ifg7 24 ii.ds Here the game V.Epishin-F.Nijboer, Apeldoorn (rapid) 2001, was agreed drawn, although after the forced 24...�h8 I prefer Black. C) 10 cs

This move is the most direct. Al­ though it is not bad, this line is not seen very much nowadays. 10...tl'if4 Black can also play 10 .. .fs, but I pre­ fer the text move, which is considered the main line. 11 ii.xf4 exf4

12 hi.cl

This is by far the most common move. White logically breaks the pin along the long diagonal, but he can also protect the loose c3-knight with his queen: a) 12 'it'd2 h6 (oth erwise, 12 ... 1'.g4 13 ll'id4 is annoying, but 12 ... as!? is worth considerin g) 13 .rf.adl gs 14 es g4!? lS exd6 cxd6 16 ll'iel and here Black can play 16 ... tl'ig6 or 16 ...dxcs!? 17 d6 tl'ig6 18 bxcs (not 18 d7 ii.xd7 when the c3-knight is loose) 18 ... ii.d7 with an unclear position. b) 12 'tlfb3 1'.g4 (12 ... h6 13 h!.adl was considered via the move order 10 'ii'b3 in Line B) 13 hi.adl ii.xf3 14 ii.xf3 gs 1s ii.h s (or 1s 1'.g4 tl'ig6 16 tl'ie2 lLle s) 1 S ...tl'ig6 16 ii.xg6 hxg6 was pretty level in A.Schneider-E.Gufeld, Helsinki 1992. 12... as This move scores well and was played by Kasparov, which is quite a good endorsement. Again 12 ... 1'.g4 is met by 13 ll'id4, but Black has a couple of other lines worth considering: a) 12 .. .fs is worth looking at, espe­ cially to compare it to 12 ... as 13 a3 axb4 14 axb4 fs below. After 13 es dxes 14 d6 cxd6 1s cxd6 tl'ic6 16 'ii' ds+ �h8 17 bs ll'id4 18 tl'ixes 'it'e8! 19 'it'xd4 'ii' x es 20 'ii'xes ii.xes 21 hi.fd1 hi.d8 22 hi.ds 1'.g7 23 ii.f3 ii.e6 24 hi.d3 the posi­ tion was unclear in J.Vilela-L.Perez, Villa Clara 1998; Black has the bishop-pair, but White has a passed d-pawn and some queenside pressure. b) 12 ... h6 is the main line. White has some choice here:

97

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 bl) 13 ll:id2 g s 14 ll:ic4 ll:ig6 (14 ... a6 l S ll:ia4! ?) lS 1'.g4 ll:ies 16 ll:ixes ii.xes 17 ii.xc8 h!.xc8 18 ll:ibs and here, rather than 18 ...'it'f6 19 c6!, Black should play 18 ... dxcs 19 bxcs c6! with counterplay. b2) 13 lLid4 g s 14 ii.hs as lS a3 axb4 16 axb4 c6! 17 ll:ide2 would transpose to A.Beliavsky-V.Spasov, Manila Olympiad 1992, and here the simplest move is 17 ... ii.es with equality according to Beliavsky. b2) 13 h 3 ! ? g s 14 a4 ll:ig6 l S as l:te8 (both 1S ...'ife7 and 1S ... a6 are possible too) 16 ll:id2 ll:ies 17 cxd6 cxd6 18 ll:ibs l:te7 19 'i¥c2 a6 20 ll:id4 was M.Diesen­ L.Day, Lone Pine 1977, and here Black should play the simple 20 ... ii.d7! when the position is fairly level.

13 cxd6 Other moves are also possible: a) 13 a3 is the obvious move, but opening the a-file should favour Black compared to the lines above. After 13 ... axb4 14 axb4 fS! ? (of course, 14 ... h6 is also possible) lS es dxes 16 d6 cxd6 17 cxd6 ll:ic6 18 'it'ds+ �h8 19 bs the idea of 19 ... Itas!? becomes a new pos-

98

sibility: 2 o 'ifd2 (20 l:tfdl 1'.d7) 2 o...ll:id4 21 ll:ixd4 exd4 22 ll:ids f3 ! (better than 22 ... ii.e6 23 ll:ixf4) 23 ii.xf3 (after 23 gxf3 ii.e6! 24 ll:if4 'it'gs+ wins for Black and here 24 ll:ic7 Ita2 2s ll:ixe6 Itxd2 26 ll:ixd8 h!.xd8, with the idea of ...ii.es, is also much better for him) 23 ...l:txbs 24 'i¥e2 1'.d7 and Black has an extra pawn. b) 13 ll:ibs could also be played im­ mediately. Then 13 ... axb4 14 cxd6 cxd6 l S 'it'b3 1'.g4 16 htc7 ll:ic8! is a very typi­ cal move. The knight heads to a7 to challenge White's own knight and un­ dermine the Cl-rook. Now White has: bl) 17 ll:ifd4 ii.xe2 (Black could also consider 17 ... ii.xd4 18 ii.xg4 ii.cs, with the idea 19 l:txb7 'it'as!) 18 lLixe2 f3 19 gxf3 ll:ia7 20 hi.xb7 ll:ixbs 21 Itxbs 'ifgs+ 22 �hl hi.a3 23 'it'xb4 l:txa2 gave Black compensation for the pawn in F.Schirm-G.Schmid, German League 1991. b2) 17 h!.xb7 Vas 18 1'.c4 and here:

b21) 18...ii.xf3 19 gxf3 ! ll:ib6 20 l:tbl ll:ixc4 21 'ifxc4 ii.es (Grivas suggests 21 ... h!.fc8 22 'it'xb4 l:tcl+ 23 �g2 'i¥xb4 24 Itxb4 Itxa2 2S htb8+ ii.f8 26 ll:id4

Th e M a r def Plata Variati o n : 9 b4 tl'i h s with o u t 1 0 hi. e 1 Itdl, but after 2 7 tl'ie6! fxe6 2 8 dxe6 �g7 - 28 ... l:ta7 29 l:ta4! is an important point - 29 l:t4b7+ �f6 30 l:txf8+ �xe6 31 l:tff7 �es 32 l:tbe7+ �d4 33 l:tf6 White can still try to press) 22 'it'xb4 'it'xb4 23 l:txb4 l:txa2 24 J:ic4 was a little better for White in E.Grivas-J.Cooper, Novi Sad Olympiad 1990. b22) 18...tl'ib6! is very strong: 19 l:ic7 tl'ixc4 20 'ii'xc4 l:tfc8! 21 l:tc6 l:txc6 22 'ii'xc6 (22 dxc6 ii.e6!) 22 ...'ii' xa2 (or 22 ...ii.xf3 23 gxf3 'ii'xa2 24 ll'ixd6 b3) 23 tl'ixd6 was E.Grivas-J.Murey, Tel Aviv 1991. Here Black shouldn't play 23 ...ii.xf3 24 gxf3 b3, transposing to the previous note, but rath er the immedi­ ate 23 ... b3! with a clear advantage. 13 ... cxd6

14 tl'ibs After 14 'ii'd 2 axb4 1s tl'ibs Black could play 1S ... 1'.g4 16 h!.c7, which is our main line below, or lS ...fS !? 16 ii.d3 fxe4 17 ii.xe4 h!.a4 with counterplay. In stead 14 bS limits White's possi­ bilities on the queenside and 14... a4!? 1s 'it'd3 (1s tl'ixa4 'it'as) 1s...ii.d7 16 tl'id2 fs 17 hi.fe 1 tl'ic8 18 ii.f3 tl'i b6 was

unclear in N.Petre-L.Vajda, Sovata 1999. 14...1'.g4 Also possible is 14... axb4 which leads us to note 'b' to White's 13th, above. 15 l'.!c7 axb4 16 'ii' d2

16...ii.xf3 Th is was Kasparov's choice, but I think that Black may be better off with the older 16 ... tl'ic8!? and then: a) 17 l:tfcl ii.xf3 18 ii.xf3 ii.c3 !? (I think Black can improve with 18 ...tl'ia7 19 'it'xb4 tl'ixbs 20 'it'xbs l:txa2 21 'ii'xb7 1'.d4 when he is obviously better - f2 is weaker than f7 and Black has much the better bishop) 19 'it'e2 (better is 19 'ifxf4 tl'ia7 20 h!.7xc3 tl'ixbs - after 20 ...bxc3 21 tl'ixd6 White has compen­ sation - 21 l:tb3 when White is at least equal) 19 ...ll'ia7 20 l:txb7 tl'ixbs 21 l:txbs was B.Malich-L.Vogt, East German Championship, Erfurt 1973, and here 21 ... l:ta3 would be better for Black. b) 17 ll'ifd4 ii.xd4!? (17 ...ii.xe2 18 tl'ixe2 was agreed drawn in K.Grigorian-E.Geller, USSR Champion­ ship, Leningrad 1971) 18 ii.xg4 ii.cs

99

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 (with the idea of ... ll'la7) 19 es (19 .i::!.xb7 �as is better for Black) 19 ... ll)a7 20 exd6 tllxbs 21 .i::!.x cs tllxd6 22 'ii'xf4 .i::txa2 23 'ii'xb4 'ii'f6 with equality. 17 �xf3 �es Again 17 ...tllc8! ? looks good: 18 'ii'xb4 (18 .i::!.x b7 'ii'a s) 18 ...ll'la7 19 .i::!. x b7 tllxbs 20 'ii'xbs .i::tx a2 and if anyone is better it's Black.

'ii'xb7 21 'ii'x as .i::!.a8 22 'ii'd2 when he has an extra pawn, although it will be difficult to convert. 18 'ii'as 19 ll'ld4 'ii'x a2 20 'ii'xa2 .i::tx a2 21 .i::txb4 •••

21 .i::!.fa8 Black has also tried 21 ... 3'.xd4 22 .i::!.xd4 g s 23 .l::!.b4 (White should really prevent Black's next by playing 23 �hS with equality) 23 ...tll g 6 24 .i.g4 tll e s 2s h3 .i::!.fa8 with an edge in H.Herndl­ S.Kindermann, Austrian League 1996. 22 �g4 Gallagher suggests that White should at least activate his bishop with 22 tllc6 tllxc6 23 dxc6 .i::tc2 24 .i.d1 .i::!.xc6 2S �b3, although he will clearly be grovelling here as well. 22 .i.xd4 23 I!xd4 gs 24 h4?! After 24 �hS .l::!.b8 Black can try to improve his position with ... '>t>g7 (Gal­ lagher), but White should hold here. 24 gxh4 25 'it>h2 tll g6 26 '>t>h3 .i::t b2 27 �fs tll es 28 'it>xh4 h6 29 �h3 '>t>g7 Now White's king is in some trouble and Black went on to win in G.Kamsky­ G.Kasparov, New York (rapid) 1994. •••

18 .i::!.xb7?! This may be good enough to main­ tain equality, but White will have to work for it. Instead after 18 .i::!.fcl Galla­ gher gives 18 ....i::t a s! and then: a) 19 'ii'xb4 .i::txa2 20 .i::!.xb7 .i::!.b2 21 'ii'a3 'ii'a 8! 22 .i::!.x e7 (or 22 'ii'x a8 .i::txa8 23 tllxd6 .i::tc2! ) 22 ....i::t x bs when Black is better. b) 19 .i::!.xb7 �a8 20 �xb4 (not 20 .l::!.ce7 .i::txa2 winning, while 20 .i::!. x e7 .i::!.xbs also favours Black) 20...'ii'x b7 21 'ii'xas .i::!.a 8 22 'ii'b4 ( 2 2 .l::!.e7 'ii'b 8!) 22 .. 1Ixa2, again with a slight advan­ tage for Black. White has an improvement on all of this, however, as he can play the clever 18 .i::tb 1 ! ? .i::t a s 19 .i::!.xb7 'ii'a 8 20 'ii'xb4

1 00

•••

•••

Th e M a r de/ Plata Var i a t i o n : 9 b4 li'i h 5 with o u t 1 0 I! e 1 D ) 10 g3 This favourite of Van Wely is White's most dan gerous alternative to 10 I!el and sometimes the play is simi­ lar. Rather than discouraging ...li'if4, White prevents it altogether. 10 fs 11 li'igs li'if6 12 f3 f4 ...

This is Black's most popular choice and is favoured by the top players. In­ stead 12 ...c6 is also popular, but White has done well after 13 �e3. The imme­ diate 12 ... h6 13 li'ie6 �xe6 14 dxe6 c6 lS �e3 also looks good for White, and after 12 ...'it>h 8 both 13 �e3 and Van Wely's 13 'it>g2 score well for White, so we will stick with 12 .. .f4, which looks to take advantage of White's 10th move. 13 bs This has been Van Wely's latest choice, but we may have seen the end of it. White has a couple of other op­ tions: a) 13 cs has fallen out of favour be­ cause of 13 ... dxcs!? 14 �c4 (14 bxcs h6 lS li'ie6 �xe6 16 dxe6 'ii'd4+! is Black's idea) 14 ... cxb4 lS d6+ 'it>h8 16 li'ibs (af­ ter 16 dxe7 'ii'x e7 Black threaten s both

...bxc3 and ...'ii' c s+) 16 ... h6 17 li'if7+ (Nunn gives 17 dxe7 'ii'xe7 18 li'if7+ I!xf7 19 �xf7 'ii'xf7 20 'ii'd8+ 'ii'e8, win­ ning, and 17 li'ixc7 hxgs 18 li'ixa8 li'ic6 19 li'ic7 !i'id4 as good for Black, because 20 li'ibs? li'ixbs 21 �xbs 'ii'b6+ wins) 11 ...I!xf7 18 �xf7 cxd6 19 li'ixd6 �h3 gave Black good compensation for the exchange in Cu.Hansen-J.Nunn, Wijk aan Zee 1991. b) 13 'it>g2 is actuallythe most popu­ lar move.

A good part of the theory of this line is based on Van Wely's games, but more recently he has preferred 13 bS. Here Black has a wide choice: bl) 13 ... as 14 bxas (instead 14 �a3 axb4 lS �xb4 �h6 16 li'ie6? loses to 16 ... .i.xe6 17 dxe6 cs, but 14 bS!? should be considered) 14 ..l!xas lS 'ii'b3 li'ie8 16 �d2 I!a8 17 cs was a little bet­ ter for White in E.Lobron-B.Gelfand, Munich 1992. b2) 13 ...h6 14 li'ie6 �xe6 lS dxe6 c6 (not 1s ...li'ic6?! 16 li'ids) 16 bS 'ii' c 7 17 bxc6 bxc6 18 �a3 I!fd8 19 'ii'a4 'ii'c8 20 I!abl 'it>h 8 21 .i::!.fdl 'ii'xe6 (21 ... as? has

101

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 been suggested, but 22 .i::!.xd6 just wins) 22 'ii'a 6 hS 23 .l::!.b 7 gave White good compensation for the pawn in L.Van Wely-Ye Jiangchuan, Biel lnterzonal 1993, although Black has some coun­ terplay on the kingside. b3) 13 ... 'it>h8 14 cs h6 lS cxd6 (after lS ll:le6 �xe6 16 dxe6 dS 17 exds ll:lexds 18 ll:lxds ll:lxds 19 �c4 Black has 19 ... c6 with the idea of ....i::!.e 8) 1s ...�xd6 (now after 1s ...cxd6 16 ll:le6 �xe6 17 dxe6 Black cannot support the dS-square with ... c6) 16 ll:lbs �b6 17 a4 ll:lfxds 18 exds hxgs 19 as 'ii'f6 20 ll:lxe7 .l::!.b8 21 g4 and now 21 ... �d7? 22 bS .i::!.fc8 23 d6 left White much better in L.Van Wely-J.Nunn, Wijk aan Zee 1992, but both 21 ...�d6 21 ...e4 improve and look satisfactory for Black. b4) 13 ... ll:lhs!? and here: b41) 14 cs fxg3 lS hxg3 ll:lf4+ 16 gxf4 exf4 17 �el (after 17 lLlxh7 Black can play the simple 17 ... 'it>xh7 18 .l:!.hl+ 'it>g8 19 �b2 gs OT 17 ...�XC3 18 ll:lxf8 �xf8 19 .l::!.b l �h3+! 20 '>t>gl �xfl 21 �xfl 'ii'f6 22 �h3 '>t>g7 with a good po­ sition) 17 ...ll:lfs 18 �xf4 (18 lLlh3 ! ?) 18 ... �XC3 19 �XC3 ll:lh4+ 20 '>t>f2 (not 20 '>t>g3 .i::!.xf4 21 'it>xf4 ll:lg2+ 22 '>t>g3 �xgs+) 20....i::txf4 21 ll:le6 �xe6 22 dxe6 'ii'e7 23 cxd6 cxd6 24 .i::ta cl .i::taf8 with a messy position, L.Van Wely-0.Cvitan, Moscow Olympiad 1994. b42) 14 g4 �f6! lS ll:le6 �xe6 16 dxe6 ll:lg7 17 cs 'it>h8 (17 ...ll:lc6!?) 18 cxd6 cxd6 19 ll:lbs ll:lc8 (19 ...ll:lxe6!?) 20 �c4 .l:!.e8 21 �dS .l::!.e 7 (also after 21 ... �e7 22 �b2 ll:lb6 23 a4 a6 24 ll:lc3

1 02

ll:lxe6 2 S as ll:lc8 White has compensa­ tion) 22 'ii'b 3 'ii'e 8 23 a4 a6 24 ll:lc3 ll:lxe6 2s �xe6 .i::!.xe6 26 ll:lds �d8 27 �b2 with compensation for the pawn in L.Van Wely-A.Zapata, Matanzas 199S. bS) 13 ...c6 increases the tension and has scored very well.

White has: bSl) 14 bS cs! lS 'ii'd3 ll:le8 16 ll:le6 �xe6 17 dxe6 was B.Finegold­ G.Michelakis, Groningen 1993. Here 17 ... 'ii'c8 18 ll:lds 'ii'xe6 should be somewhat better for Black bS2) 14 cs looks logical enough, but remains untried. Gallagher gives 14 ... h6 (14 ... cxds! ? 1s exds ll:lfs is an­ other idea) lS �c4 (or lS ll:le6 �xe6 16 dxe6 dS!) 1s ... hxgs 16 cxd6 'it>h7 17 dxe7 �xe7 18 bS (both 18 g4 'ii'xb4 and 18 d6 'ii'd7 should be at least okay for Black) 18 ... g4 with counterplay. bS3) 14 'ii'b3 h6 lS ll:le6 �xe6 16 dxe6 �c8 17 .i::!.d1 .l:!.d8 (not 17...'ii'xe6? 18 .i::!.xd6!), with a final divide: bs31) 18 ll:lds ! ? cxds 19 cxds gs 20 �d2 .i::!.f 8 21 g4 hS 22 h3 ll:lg6 23 .i::!.dcl

Th e M a r de/ Plata Varia t i o n : 9 b4 lLJ h 5 with o u t 1 0 .i::!. e 1 � 8 2 4 ..i.el hxg4 2S hxg4 lLJe8 2 6 a4 �f6 27 as �d8 was unclear in L.Van Wely-A.Fedorov, Batumi 1999. bs32) 18 cs ds 19 exds lL:ifxds 20 lLJe4 'ii'xe6 (instead 20 ...lL:ifs was sug­ gested by Mikhalevski, but after 21 e7!? lLJfxe7 22 lLJd6 'ii'd7 23 bS White has the initiative) 21 g4 'it>h8 (probably too slow; Mikhalevski suggests 21 ... b6, while 21...bS!? is also possible) 22 '>t>hl lLJg8 (again 22 ... b6 looks better) 23 �b2 lLJgf6 24 gs lLJxe4 2s fxe4 lLJC7 26 gxh6 �f6 27 �c4 'ii'g4 28 .i::!. d 6! With an ini­ tiative in L.Van Wely-R.Pruijssers, Dutch League 2006. We now return to 13 bS.

At first this move seems illogical from a positional standpoint, but White's play is based on his light­ square control. 13 ... h6 After 13 .. .fxg3 14 hxg3 lL:ih s (it is probably better to head back to the main line with 14 ...h6 lS lLJe6 �xe6 16 dxe6 'ii'c 8) White has: a) lS 'ii'e 1 h6 (better is 1s ...�f6 16 lLJe6 �xe6 17 dxe6 lLJg7 18 �h6 lLJxe6

19 �xf8 �xf8 with some compensa­ tion) 16 lLJe6 �xe6 17 dxe6 'ii'c 8 18 lL:ids 'ii'xe6 19 g4 lLJf4 20 lLJxc7 'ii'c 8 21 lLJxa8 'ii' cs+ 22 'it>h2 I!xa8 23 'ii'f2 was better for White in G.Prakash­ K.Sasikiran, New Delhi 2001. b) lS 'it>f2! lLJf4? (this fails, but after 1S ... ..i.f6 16 lLJe6 �xe6 17 dxe6 both 17 ... lLJg7 18 cs and 17 ... 'ii'c8 18 cs 'ii'xe6 19 cxd6 cxd6 20 �3 leave White with the initiative) 16 gxf4 exf4 17 'ii'd3 h6 18 lLJe6 ..i.xe6 19 dxe6 lLJc6 20 .l:thl lL:ies 21 'ii'd2 'ii' gs 22 .i::!.h 3 'ii'f6 23 lL:ids �xe6 24 'it>g2 �f7 2S 'ii'xf4 'ii'd7 26 'ii'g 3 and 1-0 was L.Van Wely-S.Dyachkov, Rus­ sian Team Championship 2008. 14 lLJe6 �xe6 15 dxe6 fxg3 Black only plays this because he has a concrete idea in mind. 16 hxg3 �cS 17 lLJds

17 ...'ii'xe6 18 lLJxc1 'ii'h 3 This is the point of Black's play. 19 l!f2 After 19 lLJxa8 'ii'xg3+ 20 '>t>h l Black can, of course, take the perpetual check, but he could also try 20 ... lLJhS ! ? 21 'ii' el .i::tx a8 22 .i::tf2 'ii'h4+ 23 'it>g2

1 03

A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1 ll:lf4+ 24 �xf4 exf4 2S .i::!.d1. Then 2s ...�d4 26 .i::txd4 'ii'g 3+ 27 'it>h1 'ii'h4+ is still a draw, but Black could also go in for either 2s ... �es or 2s ....i::!.c 8! ? with the idea of ... .i::tc s.

19 ll:lxe4! Correct, whereas after 19 ... .i::tac8 20 .i::th 2 'ii'xg3+ 21 .i::!.g 2 'ii' h3 22 'ii'xd6 .l::!.f7 23 cs �f8 24 ll:le6 White was winning in L.Van Wely-M.Golubev, Romanian Team Championship 2000. 20 fxe4 This leads to a very delicate posi­ tion. Instead 20 .i::th 2 'ii'd7 21 ll:lxa8 ll:lxg3 22 �xh6 �xh6 23 .i::!.xh6 '>t>g7 24 .i::th 2 ll:lefS! gave Black compensation for the exchange in L.Van Wely­ J.Degraeve, Mondariz Balneario Zonal 2000. 20 ....i::txf2 21 'it>xf2 .i::!.f8+ 22 '>t>e3 'ii'xg3+ 23 'it>d2 .i::tf2 24 ll:leS After 24 'ii'b3 'ii'g 2 2S 'ii'e 3 h S ! (Gal­ lagher) Black gets his h-pawn moving and threatens ... �h6. 24...'ii'f3! A tremendous improvement on

24 ... h s (24 ... ds has also been consid­ ered), although this is playable too: 2S ll:lxg7 'it>xg7 26 'ii'b3 'ii'g 2 27 'ii'e 3 ll:lg8! 28 cs dxcs 29 � b2 and now, instead of 29 ... 'ii'h 2?! 30 .i::!.e l ll:lf6 31 '>t>dl b6 32 �c3 'it>h7 33 'ii'g s when White was much better in L.Van Wely-T.Radjabov, Dresden Olympiad 2008, Black should have played 29 ... ll:lf6! 30 .i::!.el (not 30 �xes? .i::!.x e2+ 31 'ii'xe2 'ii' g s+) 30 ... ll:lg4 when he has counterplay with his ac­ tive pieces and dangerous h-pawn.

.••

1 04

2 s ll:lxd6 .i.f6! 26 cs After 26 '>t>el 'ii'g 3! 26 'ii'e l .i::!.h 2 27 cs Black has 27...ll:lc8! 28 '>t>dl (or 28 ll:lxc8 'ii'xe4) 28...'ii'g 2 29 'ii d2 ll:lxd6 30 cxd6 'ii'h l+ 31 'ii'e l 'ii'xe4 when White is in trouble. 26...ll:lds! 27 exds e4! 28 'it>e1 Now 28 ...'ii'g 2 29 'it>d2 �xal was good enough to win in L.Van Wely­ D.Stellwagen, Dutch Championship, Amsterdam 2009, but even stronger was 28 ... �c3+! with the idea of 29 �d2 'ii'g 3 30 �xc3 e3 31 'ii'd3 .i::tfl+ 32 'it>xfl 'ii'f2 mate!

Chapter 6 The Mar del Plata Variation 9 b4 tbhs

10 l::.e 1

1 d4 ll:lf6 2 c4 g6 3 ll:lc3 �g7 4 e4 d6 5 ll:lf3 o-o 6 �e2 es 7 o-o ll:lc6 8 ds ll:le7 9 b4 ll:lhs 10 .i::te1

Truth be told, Kasparov never 'blamed' this loss to the Bayonet for the nixing of the opening from his reper­ toire. Since then, however, not only has Teimour Radjabov resurrected the King's Indian, but he has had great suc­ cess in this particular 'death knell' variation from the black side. Kramnik himself has faced Radjabov twice in the Classical King's Indian, and both times he has avoided the Bayonet... 10 fs The main alternative is 10 ... as 11 bxas .i::txas 12 ll:ld2 ll:lf4 13 �fl cs 14 a4 .i::ta6 15 .i::ta 3 with a stodgy game ahead. This has never appealed to me and the fact that nowadays the retreat 15 ... ll:lhs, with the idea of a further re­ treat with ...ll:lf6-e8, has become con­ sidered Black's most reliable method of play furthers my suspicions. 11 ll:lgs ll:lf6 The alternative 11 ...ll:lf4 is not so bad. After 12 �xf4 exf4 13 .i::!. cl Black can choose between 13 ... �f6, 13 .. .fxe4 ...

This move was introduced into modem practice by Ivan Sokolov. Soon hordes of top GMs, headed by Vladimir Kramnik, started playing this way. The game Kramnik-Kasparov, Novgorod 1997, sounded the alarm and Kasparov quickly stopped playing the King's In­ dian. The world took notice and the King's Indian became a 'dubious' open­ ing and virtually vanished from the upper echelon of the chess world.

105

A ttacking Chess: Th e King 's I n dian, Vo l u m e 1 and 13 ... j.h6!?. The position tends to simplify greatly, but Black is rarely in any 'danger' of winning the game.

After 11...lL'if6, White's e4-pawn is attacked, so his choice is limited. White has two main lines here and they are both quite complicated. A: 12

.i.f3

B: 12 f3 A) 12 j.f3

and this variation tends to lead to forc­ ing play. 12 c6 Black shuts down the f3-bishop's diagonal and increases the tension in the centre. Black may also initiate the sequence 13 ... h6 14 lL'ie6 j.xe6 15 dxe6 followed by either .. .fxe4 and ... d5 or ... lL'ie8-c7 depending on the circum­ stances. The only other sensible continua­ tion is really 12 .. .fxe4 with the idea of ... lL'if5. After 13 lLicxe4 lL'if5 14 �b2 or 13 lLigxe4 lL'if5 14 .i.g5 lLid4 15 lL'ib5! White is somewhat better, although Black has a solid position. In general I do not like this structure when the c­ pawns are still on the board, because the advance c4-c5 gains space for White and Black always has to consider the possibility of both cxd6 and c5-c6. After 12 ...c6 White has several possibilities: ...

Al: 13 'W'b3 A2: 13 b5 A3: 13 .i.b2 A4: 13 .i.e3

White increases his piece play on the light squares. The diagonal for the light-squared bishop can easily be opened if White plays lL'ig5-e6 and exf5,

1 06

Other moves are less dangerous: a) 13 lL'ie6 looks premature. After 13 ... j.xe6 14 dxe6 fxe4 15 lLixe4 lLixe4 16 j.xe4 d5 17 cxd5 cxd5 18 .i.c2 e4 19 .l::!.bl 'ii'b6 (but not 19 ...'ii'd 6? 20 j.xe4 j.e5 21 .i.f3 j.xh2+ 22 'it>h1) Black is already better. b) 13 .l::!.bl is Bareev's favourite move in various positions, but here it looks

Th e M a r de/ Pla ta Va ria tio n : 9 b4 li'i h 5 1 0 I! e 1 irrelevant after 13. . .h6 1 4 li'ie6 �xe6 15 dxe6 fxe4 16 li'ixe4 li'ixe4 17 �xe4 d5 18 .i.c2 'ii'd6 19 'ii'g4 (E.Bareev­ T. Radjabov, Sarajevo 2003), when both 19 ...e4 and 19 ... .i::!.f6 look good for Black. c) 13 .i.a3 is not bad and will gener­ ally transpose to Line A2, although there are a couple of independent pos­ sibilities:

cl) 13 ...h6 is the most forcing: 14 li'ie6 .i.xe6 15 dxe6 fxe4 16 �xe4 (16 li'ixe4 li'ixe4 17 �xe4 d5 18 cxd5 cxd5 19 b5 reaches the main line of 'A2') 16...d5 17 �c2 ! ? (again, both 17 cxd5 cxd5 18 b5 li'ixe4 19 li'ixe4 dxe4 and 17 b5 li'ixe4 18 li'ixe4 dxe4 19 'ii'xd8 .i::!.fxd8 20 .i.xe7 I!e8 21 .i.c5 cxb5 22 cxb5 bring us to Line A2) 11 ... 'ii'd6 18 cxd5 cxd5 19 �b3 'it>h7 20 'ii'e2 li'ih5 21 I!ad1 li'if4 22 'ii'f1 and now, rather than 22 ...'ii'xe6 23 b5! .i::!.ad8 24 �xe7 'ii'xe7 25 li'ixd5 li'ixd5 26 .i.xd5 with an edge for White in J.Chab­ anon-0.Touzane, Marseille 2010, Black should try 22 ... .i::ta c8! with good play. c2) 13 ... cxd5 gives White the possi­ bility to deviate, but Black should still be fine:

c21) 14 cxd5 h6 15 li'ie6 �xe6 16 dxe6 fxe4 (also possible is 16 .. 1Ic8 17 'ii'b3 d5, as in A.Shariyazdanov­ M.Cebalo, Rabac 2003) 17 .i.xe4 d5 18 b5 li'ixe4 19 li'ixe4 dxe4 and once more we have transposed to Line A2. c22) 14 exd5!? e4 15 .i.e2 and then: c221) 15 ...li'ifxd5 16 li'ixd5 .i.xal 17 li'ixe7+ 'ii'xe7 18 'ii d5+ �g7 19 .i::!.x al 'ii'xg5 20 .i.b2+ 'it>h6 21 'ii'x d6 gave White very good compen sation for the exchange in L.Portisch-P.Acs, Rethym­ non 2003. C222) 15 ... !i'Jh5 ! ? 16 'ii'd 2 li'if4! 17 'ii'xf4 .i.xc3 18 'ii'h4 h5 19 .i.xh5 was J.Lautier-T.Moriuchi, Tokyo (simul) 2002, and now 19 ... gxh 5 20 �xh5 .i::!.f6 21 'ii'h 7+ �8 22 'ii'h 8+ li'ig8 23 li'ih7+ �7 24 li'ig5+ is a draw. c223) 15 ... h6 looks the simplest: 16 li'ie6 .i.xe6 17 dxe6 d5 18 cxd5 li'ifxd5 19 li'ixd5 �xd5 is equal according to Portisch. Al) 13 'ii'b3

This is a tricky move. White creates pressure on the a2-g8 diagonal, but

107

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 now there is no pressure on the d6pa wn, a consideration that Black should logically try to take advantage of. 13 h6 This is the main move, but 13 ... 'it>h8 is also possible. 14 lL'ie6 �xe6 15 dxe6 'ii'c 81 Because nothing is attacking the d6-pawn, Black can attack the e6-pawn right away. He has a good alternative in 1s ...fxe4 16 lLixe4 lLixe4 17 .i.xe4 �c8 and now after 18 'ii' h3 dS 19 .i.c2 lL'ifs Black has counterplay, while 18 .i::!.d 1 .i::!.d8 19 bS (after 19 'ii' h 3 dS 20 .i.c2 e4 Black looks to be doing well) 19 ...'ii'xe6 leaves White with some compensation for the pawn, but probably no more than that. ...

16 bs White plays on the light squares. There are several alternatives: a) 16 cs?! looks panicky and back­ fires quickly: 16 ...fxe4 17 cxd6 exf3 18 dxe7 .i::!.e 8 19 �b2?! (better is 19 gxf3 I!xe7 20 .i::!.x es lL'ihs 21 .i::t e 3 lLJf4 22 .i.b2 'ii'f8 23 .i::t a el 'ii'f6, but Black is doing

1 08

well here too) 19 ... .i::tx e7 20 .i::!.x es 'ii'C7 21 .i::!. e el fxg2 (White's kingside is a wreck and Black threatens ...lL'ig4) 22 h3 .i::!.f8 23 .i::t e 3 lL'ihs gave Black a strong attack in Y.Pelletier-E.lnarkiev, Istanbul 2003. b) 16 �a3 is an interesting move order.

After 16 ... 'ii'xe6 17 .i::!.adl .i::!.fd8 18 bS 'it> h7 (note that compared to the main line, 18...cs?! is not so good here be­ cause after 19 exfs gxfS 20 �xb7 .i::!.ab8 White has better piece control to sup­ port 21 ..i.dS!) 19 .i::t d2 .l::!.d7 20 bxc6 bxc6 21 .i::!.edl .i::!.ad8 22 'ii'a4 f4 White had some compensation for the pawn in P.Eljanov-T.Radjabov, Moscow 200S, although Black has chances here as well. c) 16 .i::!. d1 .i::!.d8 17 bS (instead 17 cs fxe4 18 cxd6 exf3 19 dxe7 .i::!.x dl+ 20 'ii'xdl 'ii'xe6 21 'ii'd8+ 'it>h7 was H.Kallio­ V.Kotronias, Batumi 2002, when White could not take the rook because 22 'ii'x a8? 'ii'g4 wins for Black) 17...'ii'xe6 18 �a3 (after 18 bxc6 both 18 ... lLixc6 19 exfs gxfs 20 'ii'xb7 lL'id4, E.Bacrot-

The M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1 T.Radjabov, F I D E World Championship, Tripoli 2004, and 18 ...bxc6 19 �a3 'it>h7 20 'ii'a4 .i::!.ab8 21 �a6 fxe4 22 li'ixe4 li'ifs 23 li'ixf6+ 'ii'xf6 24 �xc6 .l::!.b6 25 'ii'a4 li'id4, Y.Pelletier-A.David, French League 2007, gave Black enough piece play) and here:

cl) 18 ...'it>h8 19 bxc6 li'ixc6?! 20 exfs gxfs (after 20 ... 'ii'xfs 21 'ii'xb7 li'id4 22 �e4 li'ixe4 23 li'ixe4 .i::!.ab8 24 'ii'ds li'ic2 25 li'ixd6 'ii'f6 White has 26 li'if7+!) 21 �ds 'ii'e7 22 li'ibs li'ie8 23 .i::tab1 and White had the initiative in E.Bacrot­ T.Radjabov, Bled Olympiad 2002. c2) 18 ...'>t>h7 is given by Bologan. This is a subtle improvement that I had to figure out and 19 bxc6 (or 19 .i::!.a bl .l::!.d7 20 bxc6 li'ixc6 21 exfs 'ii'xfs 22 �xc6 bxc6 23 .i::!.xd6 .i::!.ad8 24 cs e4 Bologan) 19 ...li'ixc6!? 20 exfs 'ii'xfs 21 'ii'xb7 li'id4 22 �xd6 (it transpires that this time after 22 �e4 li'ixe4 23 li'ixe4 .i::tab8 24 'ii'ds li'ic2 25 li'ixd6 'ii'f6 there is no check on f7 and so Black wins ma­ terial ! ) 22 ....i::tac8 23 li'ibs li'ixf3+ 24 'ii'xf3 'ii'xf3 25 gxf3 a6 is equal accord­ ing to Bologan.

16...'ii'x e6 17 �a3 White keeps the tension, but now Black can close the position. The im­ mediate 17 bxc6 has not been played or mentioned, probably because after 17 ... bxc6 18 �a3 .i::!.ab8 19 'ii'a4 li'id7! White cannot take the a7-pawn and ...li'ib6 is threatened.

11 . .cs!? Black can also play 17...'>t>h7 when 18 .i::!.adl .i::!.fd8 19 bxc6 bxc6 (after 19 ... li'ixc6?! 20 exfs both 20...gxfs 21 'ii'xb7 li'id4? 22 .i::!.xd4 and 20...'ii'xfs 21 'ii'xb7 li'id4, as in V.Mikhalevski­ C.Matamoros Franco, Drammen 2005, and then 22 �xd6! are bad for Black) 20 'ii'a4 .i::!.a c8 21 'ii'a6! gave White com­ pensation for the pawn in P.Eljanov1.Nataf, Mallorca Olympiad 2004. Here Black could try 20...f4, with play similar to Eljanov-Radjabov in note 'b' to White's 16th move, above. 18 exfs This wins back the pawn, but Black will get good piece play. Instead 18 .i::!.adl f4 19 .i::t d3 .i::!.ad8 20 h3 was S.Porat-R.Djurhuus, Drammen 2005, .

1 09

A ttacking C h e s s : Th e King 's I n dia n , Vol u m e 1 and here Mikhalevski suggests 20 ... g5 21 .i::!.edl b6 with the idea of ...h 5 and ... g4. 18 ...gxf5 19 �xb7 I!ab8

20 .i.ds After 20 �f3 e4 Black intends ... lLic6 or ... lLig6: for example, 21 �e2 lLic6! 22 �b2 lL'id4 23 'ii'dl lL'id7 with excellent piece play. 20 ...lLifxds 21 cxds �f7 22 �b2 e4 23 .l:!.adl ll'ig6 Black stood well in A.Korobov­ D.Yevseev, Sochi 2008. A2) 13 bS

This forcing continuation has ex-

110

perienced certain bouts of popularity. At first this move looks very strange positionally - usually White is aiming for c4-c5, but in the Bayonet the ad­ vance b4-b5 is a common theme be­ cause White's initiative is largely based on his light-square play, especially with lL'ig5-e6 always in the air. These days it is known that the main line basically leads to a drawish ending. Black does have one main deviation, but it entails some real risks. 13 ...cxds Instead 13 ... h6 14 lLie6 �xe6 15 dxe6 fxe4 16 lLixe4 lLixe4 17 �xe4 d5 18 cxd5 cxd5 19 �a3 leads to the same thing. This move order would deny Black of the option given at move 16, however. 14 cxds Sometimes the recapture 14 exd5 has to be taken into account when Black plays ...cxd5 before ...h6, but in this case White has problems on the long diagonal after 14 ... e4. 14...h6 If Black really wants to avoid the forcing lines that follow, the rare 14 .. .fxe4 could be played. After 15 lLigxe4 lLixe4 16 lLixe4 lL'if5 White is a little better, but these positions are not so bad for Black when the c-pawns have been exchanged. 1s lL'ie6 �xe6 16 dxe6 fxe4 Thi s is the theoretically approved line, but if Black wants to take a chance there is 16 ....i::!.c 8!?. This looks attractive (at least in the

Th e M a r de/ Plata Va ri a t i o n : 9 b4 li'i h 5 1 0 .i::!. e1 sense that i t does not just lead t o a level endgame), but I am not so sure it is worth the trouble ... Here White has:

a) 17 �b2 fxe4 18 li'ixe4 li'ixe4 19 .l:txe4 (19 .i.xe4 dS 20 �a3 dxe4 21 'ii'xd8 .i::!.fxd8 22 �xe7 .i::!.e 8 is a familiar ending, but Black has the extra move ..i::tac8) 19 ... ds 20 .i::!. x es �xes 21 .i.xes is very similar to Line A3. After 21 ...�b6 (21 ... .i::tc4 is also possible) 22 �b2 .l::!.c4 (not 22 ...'>t>h7 23 �e2 d4 24 h4 .i::!.f6 2S .l:!.e1 d3 26 �es when White was win­ ning in T.Nyback-D.Stellwagen, Ger­ man League 2oos) 23 'ii'e2 'ii'a 24 �e3 �f4 2S 'ii'xa7 (after 2S �a3 'ii'd2 26 .l:!.dl l:te4! 27 '>t>fl 'ii' c 2 28 �xe7 'ii'xb2 29 'ii'd7 .i::!.xf3 30 gxf3 'ii'e2+ 31 'it>g2 .l::!.g 4+! it's a draw) 2s ....i::t c2 26 �a3 'ii' e s 27 .i::tf1 .i::tx a2 28 'ii' c s .i::!.xf3 29 gxf3 �gs+ 30 'it>h1 'ii'fs 31 'it>g2 �gs+ 32 'it>h1 �fs White could not avoid a draw in Z.Peng-Li Shilong, Wijk aan Zee 2008. b) 17 �b3 ds 18 li'ixds li'ifxds 19 exds e4 20 d6 and here: _

(seefollowing diagram)

bl) 20 ... .i::t c3 ?! 21 dxe7 (much stronger than the 21 �a3? .i::!.xb3 22 dxe7 .i::txa3 23 exd8'ii' .i::!.x d8 24 .i::!.adl of Ki.Georgiev-R.Ponomariov, Istanbul Olympiad 2000, when both 24...l:te8 and 24....i::!.c 8 are good for Black) 21 ...'ii'xe7 22 'ii'a4 exf3 23 �b2 f4 (or 23 ... �gs 24 g3 'ii'g4 2s 'ii'xg4 fxg4 26 .i::tadl .i::ta 27 e7 .i::!.e 8 28 .i::!.d8 1-0, S.Savchenko-F.Guilleux, Le Touquet 2007) 24 'ii'e4 'ii'g s 2s g3 .i::t e 3 26 �c4 was drawn here in R.Szuhanek­ N.Grigore, Eforie Nord 2008, but White is much better. b2) 20 ... �xal 21 dxe7 'ii'xe7 22 �a3 'ii'f6 23 e7+ .l::!.f7 24 �e2! (Zakhartsov's idea; instead 24 �dS .i::!.e 8 2S �dl 'ii'c 3 was good for Black in S.Savchenko­ M.Mozharov, Sochi 2008) 24 ... �es (also bad are 24... '>t>g7 2S .i::t xal 'ii'xal+ 26 .i.fl and 24 ... '>t>h7 2S b6! a6 26 'ii'a4) 2 S b 6 ! a 6 ( 2 s...�xb6 26 'ii'xb6 axb6 2 7 .i.bs) 26 �c4 '>t>g7 21 �xf7 'ii'xf7 28 .i::!. d l and the e-pawn will give Black big problems. b3) 20...'ii'xd6 21 �a3 �f4 (worse are 21 ...�dS? ! 22 .l:!.adl 'ii'xb3 23 axb3

111

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 exf3 24 �xe7 .i::!.fe8 2 S .l::!.d7 �f6, as in Xu Jun-Ye Jiangchuan, Taiyuan 2004, and then 26 �b4!, and 21 ... 'ii'b6 22 �xe7 .i::!.fe8 23 �b4 �xal 24 .i::txal exf3 2 S e7+ 'l;g7 26 �a3 ! with a strong ini­ tiative) 22 .i::!.a dl (not 22 �xe7 .i::!.fe8 23 �a3 .i::tc 3 24 'ii'ds exf3 2s �d6 'ii'g s 26 �g3 f4 when Black is better) 22 ....l::!.C 3 23 �4 .i::txa3 (better than 23 ... .i::!.e8 24 �e2) 24 'ii'xa3 exf3 2S .l::!.d7 �f6 26 'ii'xf3 'ii'b4 27 'ii'e2 with an unclear po­ sition according to Zakhartsov. 11 ll:lxe4 17 �xe4 ll:lxe4 18 ll:lxe4 dS 19 �a3 makes no difference 17 ...ll:lxe4 18 �xe4 ds 19 �a3 This leads by force to an ending where White hopes to be marginally better. Instead he must avoid 19 �c2 e4 20 .l::!.bl 'ii'b 6 when Black is hitting both f2 and e6. 19...dxe4 20 'ii'x d8 .i::!.fxd8 21 �xe7

21 ....i::te s The alternative 21 ....i::td s!? has only seen one outing, although it was suc­ cessful: 22 .i::!.adl .i::!.x bs 23 �d8 .i::!.c 8 24 .i::!.d6 .i::!.c6 2S .i::!.xc6 bxc6 26 h4 �8 27

112

.i::!. cl 'l;e8 28 e7 V2-V2 L.Fressinet­ M.Hebden, Lausanne 2001. 22 �cs .i::!.x e6 23 �e3 This is White's latest attempt to keep a little something in the position. He hopes to develop some pressure against Black's queenside. In general, taking on e4 does not lead to anything because Black's king can become active very easily, which makes up for the weak es-pawn. My first experience with this line was back in 1996 when GM Igor Nataf asked me to play the white side of this endgame in some blitz games on line to see if it was ten­ able for Black. It was. A more serious example: 23 .i::!.xe4 a6 24 b6 .i::!.c 8 2S �e3 .l:!.d6! and Black was at least equal in M.Brodsky-A. Shimanov, St Petersburg 2008; Black's king will come to e6. Instead 23 .i::!.a dl a6 24 bxa6 .i::!.exa6 2S a3 �f8 was also level in E.Bacrot­ T.Radjabov, Dresden Olympiad 2008. 23 ...a6 Black removes the a7-pawn from the scope of White's bishop. The care­ less 23 ...b6?! 24 a4 a6? runs into the breakthrough 2s as! axbs 26 axb6 when Black is in big trouble. 24 b6 The only try to eke anything out of the position is to advance the pawn, but it could become weak here. 24....i::!. d S This looks like the simplest move to me. Black should certainly not go to sleep just yet, though, as these exam­ ples show:

Th e M a r de/ Plata Varia t i o n : 9 b4 li'i h 5 1 o .i::!. e 1 a) 24...�f6 intending ... �gs has al­ ways been considered a simple solu­ tion, but White may still push a little bit here: 2S g3 (2S a4 has also been tried) 2s ... �gs 26 �xgs hxgs 27 .i::!.xe4 .i::!.xb6 28 .i::!.x es .i::t bs 29 .i::t ael (after 29 .i::!.xbs axbs 30 .i::t d 1 '>t>g7 31 .l::!. d7+ 'it>h6 32 .i::!.xb7 .i::txa2 33 .i::txbs g4 Black drew easily in J .Lautier-V.lvanchuk, Mallorca Olympiad 2004) 29 ... g4 30 .i::t1e4 .i::!.x es 31 .i::!.x es bS 32 .i::tg s �7 33 .i::!.xg4 and White went on to win in S.Savchenko­ V.Talla, Legnica 2007. b) 24....i::!.c 8?! 2s .i::tacl .i::!.ec6 (2s ....i::tcc6 26 .i::!.edl �f8 27 'it>fl also gives White some hope) 26 .i::!.c dl! .i::!.f6 27 .l::!. d7 .l::!.f7 28 .i::!.edl .i::!.e8 29 .l::!.c 7 and White main­ tains some pressure. c) 24 ... '>t>f7!? leads to the main line after 2s .i::ta cl .i::!.d 8 26 .l::!. e7+ .l::!. e 7 27 �cs .i::!.e d7. 25 .i::t a cl '>t>f7 Or 2s ... .i::!.ed6 with the idea of .. 1I8d7 and ... '>t>f7-e6. 26 .l::!. c 7+ .l::!. e7 27 �cs .i::!.e d7

�d6 32 'it>g2 'it>e6 33 .l::!.g4 gs 34 h4 the game was agreed drawn immediately in N .Rashkovsky-F.Florian, Biel 2001, although Black could have tried 34 ... �xe7 with the idea of 3S �xc7 (better is 3S .l::!. c 4) 3S .. 1Ixc7 36 hxgs 'it>f S! when he is suddenly much better. Another game went 28 g4 'it>e6 29 'it>g2 .l::!.xe7 30 bxe7 .i::!.c8 31 �b6 �f8 32 .i::!.xe4 �d6 and Black even went on to win in A.Van Beek-G.Kodentsov, Gron­ ingen 2007. A3) 13 �b2

This move became popular after Shirov used it to defeat Radjabov. The bishop does not look very effective on b2, but White plans a dangerous ex­ change sacrifice to justify it. 13 h6 Black heads down the rabbit hole. The alternative is the immediate 13 ... cxds 14 cxds (Black must also worry about 14 exds) 14...fxe4 (instead 14 ... h6 lS li'ie6 �xe6 16 dxe6 fxe4 17 li'ixe4 li'ixe4 18 I!xe4 dS brings us back to the main line) lS li'igxe4 li'ifs 16 •••

Black is totally fine and after 28 g3 �f8 29 .i::!.xe4 .l::!. xe7 30 bxe7 .l::!. c 8 31 �b6

113

A ttacking C h e s s : Th e King 's I n dia n, Vo l u m e 1 ll:lxf6+ and now both 16 ... �xf6 17 lLle4 �g7 18 .i::!.cl �d7 (V.Laznicka-G.Jones, European Team Championship, Herak­ lion 2007), and 16 ... .i::txf6 17 lLle4 .l::!.f7 18 g3 (D.Vigorito-J.Rihel, Somerville 2009) 18...�d7 are only a little bit better for White. 14 ll:le6 �xe6 15 dxe6 fxe4 Black can try to avoid the coming complications with 1S ...'ii'C7 16 'ii'b 3 (16 cs ds 17 exds ll:lfxds 18 �xds .i::ta d8 looks okay for Black), and here 16 .. .fxe4!? looks worth a try after either 17 i.. xe4 .i::!.a e8 or 17 ll:lxe4 ll:lxe4 18 �xe4 .i::!.f6. In practice Black has pre­ ferred 16 ... .i::tad8, but 17 .i::!.adl fxe4 18 ll:lxe4 ll:lxe4 19 �xe4 �c8 20 'ii' h3 I!f6 21 ..tc2! .i::!.xe6 22 f4 I!f8 23 fxes (worse is 23 cs dS 24 fxes, as in W.Paschall­ D.Karatorossian, Budapest 2004, be­ cause of 24...h s with the idea of ...ll:lfs) 23 ... dxes 24 cs ll:lds 2s �b3 .l::!.e 7 26 �xc8 .i::!.xc8 27 �xds+ cxds 28 I!xds e4 29 ..txg7 'it>xg7 30 'it>f2 was clearly bet­ ter for White in L.Vrbica-T.Pupak, corre­ spondence 2004. 16 ll:lxe4 Instead 16 �xe4 gives Black a rather pleasant choice: a) 16 ...'ii'b6 17 �b3 dS! 18 cxds cxds 19 �xds ll:lg4 gave Black counterplay in S.lvanov-B.Avrukh, Beersheba 1998. b) 16 ...ll:lxe4 17 ll:lxe4 ll:lfs 18 'ii'g4 'it>h7 (18 ... 'ii'e8 19 e7!) 19 .i::!.adl 'ii'e 7 20 Ild3 .i::tae8! (not 20...'ii'xe6 21 .i::!.xd6) was satisfactory for Black in O.Averkin­ V.Zakharstov, Krasnodar 1998. c) 16 ... ds 17 cxds (Black was better

114

after 1 7 �d3 e4 1 8 �fl 'ii'b6 19 .l::!.bl ll:lhs in E.Bareev-T.Radjabov, Enghien­ les-Bains 2003) 17 ...cxdS 18 �c2 and: cl) 18...'ii'd6 19 �b3 .i::!.ad8!? (this looks better than 19 ... 'ii'xe6 or 19 ...'ii'xb4) 20 ll:la4 ll:le4 21 f3 ll:lgs 22 ll:lcs ll:lxe6 looks okay for Black. c2) 18 ... e4 19 ll:la4 (it may be better to just play 19 �b3 'ii'b6 20 .l:!.cl 'it>h7 21 bS 'ii'xe6 22 �a3 .i::!.ad8 with an unclear position) 19 ... b6 20 f3 exf3 21 'ii'xf3 �d6! 22 .i::!.adl .i::tac8 23 �bl .l::!.c4 and the strong threat of ... ll:lg4 gave Black the initiative in A.Mista-R.Antoniewski, Trzebinia 1998. 16...ll:lxe4 17 .i::!.x e4 Worse is 17 �xe4 dS 18 cxds cxds 19 �c2 �b6 20 'ii'd2 e4! 21 �xg7 'it>xg7 22 �b3 .i::tfs and Black was better in M.Peek-A.David, Amsterdam 2000.

11 ...ds The only real alternative is Galla­ gher's old suggestion of 17 ...ll:lfs. After 18 bS! .i::!.c8 19 .i::te 2 .i::!.e8! ? (worse is 19 ...�e8 20 bxc6 bxc6 21 'ii'a4, V.Golod­ T.Nedev, European Championship, Plovdiv 2008) White has:

Th e M a r de/ Plata Varia tio n : 9 b4 li'i h 5 1 o .i::!. e 1 a ) 20 bxc6 bxc6 21 cs (21 �a4 'ilia 22 .l::!.bl .i::!.xe6 23 �c3 .l::!.e 7 24 �e4 'it>h7 2S g4?! li'ih4 26 �el .l::!.f7 27 .i::!.b3 .l::!.f4! gave Black the upper hand in A.Wirig­ N.Pokazanjev, Winterthur 2008) 21 ... ds 22 �xes �xes 23 .i::!. x es �f6 24 .i::t e 1 was agreed drawn here in R.Dautov­ S.Kindermann, Nussloch 1996. After 24 ... .i::txe6 the position is pretty level. b) 20 �a4 'fib6 21 .i::!.d1 (or 21 bxc6 bxc6 22 �e4 I!xe6 23 .l:tb1 �c7 24 �c3 li'id4 2S .l:teb2, which gives White com­ pensation for the pawn) 21 ... .i::txe6 22 �b3 �d8 23 cs ds 24 .l:txds! cxds 2 s �xds .i::!. x cs 2 6 �xe6+ '>t>h7 27 g3 li'id4 28 �xd4 �xd4 29 'it>g2 and White kept an edge in R.Buhmann-M.Klenburg, Neustadt an der Weinstrasse 2009. 18 cxds Retreating with 18 .i::!. el gives Black no problems after 18 ... 'fid6. 18...cxds

19 Ilxes! The point of White's play. 19...�xes 20 �xes 'fib6 21 �b2! This was Shirov's innovation. White opens the e-file to protect the e6-pawn

and introduces the threat of a battery on the a1-h8 diagonal. other moves are less dangerous: a) 21 �d4 �xe6 22 �cs was played in J.Vilela-R.Vazquez, Cuba 1996. Here Markos suggests 22 ... .i::tfd8 23 �d2 li'ifs when White probably does not have quite enough compensation for the exchange. b) 21 'fid2 'illx e6 (21 ... '>t>h7 22 .i::!.e l as was played by a young Radjabov, but is not as trustworthy) 22 .i::!.e l (or 22 �d4 I!xf3) 22 ....i::txf3 ! is a thematic idea. Black gives back the exchange to ruin White's pawn structure and give his knight a strong outpost. After 23 gxf3 li'ifs the knight is well placed to both attack and defend. One example, which was a serious upset, T.Nyback­ M.Matthiesen, Copenhagen 2003: 24 �g3 �f7 2s .i::t c1 d4 26 .i::t a 'fids 27 'fid3 .i::!.e8 28 a4 hS 29 'it>g2 �e6 30 .i::!.xb7 h4 31 .l:txa7 h3+ 32 '>t>gl 'flel+ 33 'fifl li'ixg 3 34 fxg3 'fle3+ 0-1.

21....i::ta d8 This is considered best by Markos. Black creates the possibility of shutting

115

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1 the long diagonal with ...d4. others: a) 21 ...'>t>h7 was Black's original try. Then 22 'ii'e2 d4 23 h4 .i::!.f6 24 .l:tel 'ii'xb4 2S a3 'ii'd6 26 h S .i::!.af8 27 'ii'e4 ll:lc6 28 hxg6+ '>t>g7 29 �cl! 'ii'e 7 30 �xh6+ 'it>xh6 31 'tlfli4+ 'it>xg6 32 �xc6 bxc6 33 .l:te S leads to a further theoreti­ cal divide: al) 33 ... 'ii'xe6? 34 .i::!.xe6 .i::!.xe6 3 S 'ii'g4+ '>t>f7 3 6 'ii'xd4 was winning for White in A. Shirov-T.Radjabov, Linares 2004. a2) 33 ... .i::txe6 is a better try: 34 'tlfli s+ '>t>g7 3 S .i:tgs+ 'ii'x gs 36 'ii'x gs+ .i::!. g 6 3 7 'ii'e s+ 'it>g8 38 'ii'xd4 .l::!.f7 39 'ii'd8+ (after 39 g3 .i::!.f g7! White will have trouble advancing his pawns) 39 ....l:tf8? (39 ...'>t>h7 has to be a better try, with the idea of 40 g3 .l:tfg7!, al­ though White will still certainly try to probe with 40 'ii'd3 !) 40 'ii'd7 .l::!.f7 41 'ii'e 8+ '>t>g7 42 g3 and White won the ending in M.Casella-1.Zenyuk, U S Championship, S an Diego 2004. I can still remember preparing this line with Casella before the game ... b) 21 ...'ii'xb4 is Bologan's recom­ mendation. However, after 22 .l:tbl .i::!.ac8 (instead 22 ...'ii'c s 23 �d4 'ii'c 7 24 �b2 'ii'c s 2s �d4 'ii'c 7 26 �b2 was drawn by repetition in W.Fademrecht­ F.Gerhardt, correspondence 2004, but White could also consider 24 'ii'd 2 'it>h7 2 S h4 with compensation) 23 �g7 (23 �es 'ii'a 3 eyes cl and Bologan gives 24 'ii'd2 'it>h7 2S .i::!.xb7 'ii cl+ 26 �dl 'ii'xd2 27 .l:txe7+ 'it>g8 28 .l::!. g 7+ with perpetual check) 23 ...'ii'xbl 24 'ii'xbl 'it>xg7 2 S

116

'ii'b 2+ (Bologan only gives 2 S h4 b6 as unclear) 2s ... .i::!.f6 26 h4 b6 27 g4 g s 28 �g2 .l::!.c4 29 'ii'e s White kept some pressure in M.Boccia-S .Giuliani, corre­ spondence 2004.

22 'ii'e 1 This is an interesting try. White pro­ tects the b4-pawn and may still con­ tinue with 'ii'e s at some point. a) 22 'ii'e 2 'ii'xb4 23 'ii'e S allows Black to execute his idea with 23 ... d4 when he stands better. b) 22 'ii'd3 'ii'x e6 23 'ii'd4 transposes to variation 'd2', below. c) 22 bS 'ii'x e6 23 'ii'd4 ll:lfs 24 'tlfli8+ '>t>f7 2 S 'ii'h 7+ 'it>e8 26 �g4 .l::!.f 7?! gave White excellent compensation after 27 'tlfli8+ 'it'd 7 28 'ii'c 3 .i::!. c 8 29 'ii'd2 .l::!.c4 30 �f3 in B.Lalith-A.Lahiri, Bhubaneswar 2010, but Black can improve here with 26 ... 'ii'f 7! 27 .l:tel+ ll:le7 28 'ii'xf7+ (28 'ii'xh6 'ii'xf2+) 28 ....i::txf7 when he is bet­ ter. d) 22 'ii'd2 has been White's most common choice. Then following 22 ... 'ii'x e6! (after 22 ... d4 23 .i::!. el .i::!.f6 24 ..lli.g4 'it>h7 2S h4 h S 26 �h3 Black could

Th e M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1 hardly move in R.Buhmann-T.Nedev, Hersonissos 2007) we have: dl) 23 'ii'x h6 d4 24 'ii'd 2 .i::!.xf3 ! 2 5 gxf3 li'if5 is given by Markos. This is a typical set-up for Black. White's king position is also loose and the strong d­ pawn and knight give Black compensa­ tion for the pawn. d2) 23 'ii'd4 looks scary, but the checks do not lead to anything: 23 ...li'if5 24 'tlfli8+ 'l;f7 25 �7+ 'l;e8 26 'ii'xb7 .l::!.f7 27 'ii'h 5+ �8 when Black is able to get 'castled' again and his pieces coor­ dinate well. Still, White retains com­ pensation for the exchange. d3) 23 I!el could lead to perpetual check after 23 ... 'ii'f7 (forced) 24 'ii'x h6 (if 24 'ii'd4 'l;h7 25 'ii'xa7 li'ic6 Black can fight for the initiative after 26 'ii'b6 .i::!.fe8 or 26 'ii'c 5 d4 27 b5 J:tfe8! ) 24 ... d4 and then:

d32) 25 .i::!.x e7 (forcing the draw) 2 5 ... 'ii'xe7 26 'ii'xg6+ 'ii'g7 27 'ii'e6+ 'ii'f7 (27... .l::!.f7 28 �h5 does not help Black, while 27 ... 'l;h8 28 �h3+ �7 29 'ii'g4 'ii'b 1+ 30 �dl 'ii'xb2 31 'ii'h 5+ is also a draw) 28 'ii'g4+ 'ii'g 7 29 'ii'e6+ with per­ petual. 22 .i::!.d 6 Black could also try 22 ...d4 23 .i::!.d1 (after 23 �g4 h5 24 �h3 .i::!.de8! Black intends ...li'if 5) 23 ... h 5! ?. 23 I!dl Putting pressure on the d5-pawn makes it more difficult for Black to play ... li'if5. Instead 23 'ii'c 3 d4 24 'ii'b3 .i::!.f6 brings White's initiative to a standstill, while 23 'ii' d2 .l:!.xe6 24 �d4 (after 24 'ii'x h6 d4 25 'ii'd2 Black plays the usual 2 5 ... .i::txf3 ! 26 gxf3 - not 26 �xd4? .i::!.d6 winning - 26 ...li'if5 with thematic play) 24 ... 'ii'd6 25 'ii'xh6 is again not as scary as it looks after 25 ...li'if5 26 �8+ 'l;f7 27 �7+ 'l;e8 28 �c3 J:te7 29 'ii'h4 d4!? with unclear play. 23 .i::txe6 24 'ii'c3 .i::!.ef6 2 5 'ii'd 2 •.•

.•.

d32) 25 'ii'd2 li'if5 is unclear, but 26 �dl? (26 �e4) 26 ...'ii'xa2 27 h4 .i::!. de8 28 .i::!.xe8 .i::!.xe8 29 g4 'ii'e6 30 'l;h2 'ii'd6+ 31 rt;g2 'ii' d5+ 32 �h2 li'ixh4 33 rt;g3 'ii'g 2+ and 0-1 was a significant upset in A.Fier-K.Sai, Bhubaneswar 2010.

2 s....i::txt3 This is certainly safer than 2 5 ... l!f4!?

117

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 26 �es .i::!.xb4 27 'ii'xh6 when White has the initiative. After 27 ...'l;f7 28 'ii'd2 'l;e8 29 �d4 .i::!.xd4?! (29 ... 'ii'd6 is better, but White has compensation for the exchange) 30 'ii'xd4 'ii'xd4 3 1 .i::!.xd4 White had the better ending in V.Golod-N.Huschenbeth, Hamburg 2008. 26 gxf3 .i::!.xf3 27 .i::te1 .i::tf7 28 �d4 �d6 29 'ii'xh6 Or 29 ..i.es 'ii'b6 30 �d4 'ii'd6 with a repetition. 29 ... .l:!.h7 30 �cs .i::!.xh6 31 �xd6 lllfs The endgame is equal (Golod). A4) 13 �e3

This move used to be the main line and it has been played more than any other move. Nowadays it is not consid­ ered so dangerous for Black. 13... h6 Instead 13 ... 'l;h8 14 exfs! was pointed out by Markos and this looks quite good for White. Bologan states that Black should play 13 ... cxds immediately, although he gives no explanation why this is so. I

118

do not like this move order because it allows the extra option of 14 exdS! ? which scores rather well for White. In­ stead after the more common 14 cxds h6 lS tlle6 �xe6 16 dxe6 fxe4 17 tllxe4 tllxe4 18 �xe4 dS we just transpose back into the main line with 13 ...h6. A more serious option is 13 .. .f4!?. After 14 �cl it seems as though White has just wasted time, but the centre has been stabilized and now the bishop will head to a3. Following 14... h6 lS tll e6 �xe6 16 dxe6 tllc 8 17 bS 'ii' e 7 (worse is 17 ...'ii'e 8?! 18 bxc6 bxc6 19 cs 'ii'xe6 20 �a3 dxcs 21 ll'la4 tll b 6 22 tllxcs 'ii'f7 23 'ii' c l �c4 24 �dl! with the better chances for White in V.Kramnik-B.Gelfand, Belgrade 1997) 18 bxc6 bxc6 19 cs dxcs 20 �a3 tllb6 21 ll'la4 .i::tfd8 22 'ii'b 3 tllxa4 23 �xa4 �xe6 24 �xcs ll'ld7 2s .i::!.e cl tllxcs 26 I!xcs .i::!.d6 the position is equal. 14 tlle6 �xe6 15 dxe6

1s ...fxe4 Instead 1S .. .f4 16 �cl takes us back to 13 .. .f4 in the previous note. Black could also consider l S ... gs.

Th e M a r de/ Plata Va ria tio n : 9 b4 li'i h 5 1 o .i::!. e 1 This has not had much success, but it does not look too bad. After 16 exfs li'ixfs 17 �d3 li'ixe3 18 .i::!.x e3 "ille7 19 .i::!.d1 .i::!.ad8 (19 ... flixe6 20 "illxd6 �xc4 21 bS gives White the initiative) 20 li'ie4 Black has: a) 20... g4?! 21 li'ixf6+ flixf6 22 �xg4 �xf2+ 23 '>t>hl gave White a big advan­ tage in V.Kramnik-F.Nijboer, Wijk aan Zee 1998. After this game nobody wanted to play 1s ... gs anymore. b) 20 ... ds 21 li'ixf6+ flixf6 22 cxds cxds 23 e7 (23 �xds? .i::!.xds) 23 ... �xe7 24 �xds+ 'it>h8 is given as slightly bet­ ter for White by Kramnik. This certainly looks pretty miserable for Black. c) 20 ... li'ixe4 21 �xe4 fllxe6 22 bS gives White good compensation for the pawn. d) 20 ... �xe6 and now Kramnik claimed that White has the upper hand after 21 li'ixd6, but 21 ... li'ie8 22 cs li'ixd6 23 cxd6 I!f6 looks tenable, since Black will win the d-pawn and have an extra pawn. White certainly has some com­ pensation, but Black does not really appear to stand worse. 16 li'ixe4 li'ixe4 17 �xe4 dS 18 cxdS White can also refrain from this ex­ change with 18 �c2, but this does not look too dangerous: 18 ... b6 (Black plays as if the exchange had already oc­ curred on dS, but 18 ... d4 19 �d2 and now 19 ... .i::!.f6 or 19 ... "illd6 could seek to punish White's omission as the a2-g8 diagonal is not yet open) 19 "illg 4 (19 cxds would transpose to the main line) 19 ....i::!.f6 20 bS!? (White tries to break

apart Black's centre) 20 ... fllc8 (taking immediate aim at the e6-pawn; 20... d4 21 .i::!.adl "illd6 is an alternative) 21 bxc6 (2 1 �b3 d4 22 �d2 "ilia 23 c s ! ?, main­ taining the e6-pawn, is also possible) 21 ... d4 22 �d2 fllxe6 23 fllxe6+ .i::!.xe6 (Black is likely to pick off the c6-pawn, but White's bishop-pair will be fair compensation) 24 �e4 was Huang Qian-Ju Wenjun, Hefei (rapid) 2010. Now Black could play 24....i::!.c8 because 2s .i.ds? fails to 2s ...li'ixds 26 cxds .i::!.d6 when the pawns begin to fall. 18...cxds

19 �C2 This leads to a complicated middle­ game. White can also head for an end­ ing with 19 �cs dxe4 20 flixd8 .i::!.fxd8 21 �xe7 .i::!.e 8 22 �cs. The position is identical to the main line of Line A2, but here White's pawn is still on b4. If anything, this should help Black be­ cause his queenside will not be under any pressure. Both 22 ... b6 23 �e3 I!xe6 24 a4 �f8 2S bS (N.lbrayev­ S.Mamedyarov, Khanty Mansiysk 200S) 2s ...a6! and 22 ...I!xe6 23 I!xe4 .i::!.d8 24

119

A ttacking C h e s s : Th e K i n g 's I n dian, Vo l u m e 1 h4 a6 2 s a4 .l:!.ds 26 ii.e3 .l::!.c6 27 .l:tb1 bS (Z.Gyimesi-V.Baklan, Baile Tusnad 2oos) are equal. 19 b6 ...

This is a very important move. Black must prevent ii.e3-cS or his centre will fall apart. He has a strong centre and White's e6-pawn is likely to fall, but Black must be careful here because White has the bishop-pair and the safer king. Indeed, Black must try to keep his pieces well coordinated and should not necessarily be in a hurry to win the e6-pawn. Any misstep could result in a serious tactical accident. As long as Black is not too careless, his chances are good, though, and in prac­ tice he has done pretty well from this position. 20 'lli'g4 White must play actively or the e6pawn will fall. Black is likely to win it anyway, but he must not allow White too much activity in return. Some other tries: a) 20 ii.b3 immediately is inaccu­ rate because White will not be attack-

120

ing the g6-pawn after 'lli'g4. After 20 ... 'lli'd6 21 'lli'g4 h S 22 'lli'h 3 (E.Ovod­ V.Rajlich, Budapest 2002) 22 ... e4, in­ tending ....:tfs, the b4- and e6-pawns are weak. b) 20 ii.a4 'lli'd6 21 ii.d7 is mis­ guided. The e6-pawn is secure, but White's bishop is totally out of play. After 21 ...'l!Vxb4 22 .l::!.b1 'l!Vh4 23 f3 J:tfs 24 ii.f2 'lli'f6 2 S ii.g3 h S 26 h3 .l::!.f8 27 �hl? 'l!Vg s 28 ii.h2 e4! Black broke through to White's king in V.Malakhatko-F.Jenni, Istanbul Olym­ piad 2000. c) 20 bS 1'6d6 21 ii.cl is not an un­ common idea, but here it is painfully slow. Now 21 ... .l:tad8 22 ii.b3 �h7 23 a4 'l!Vxe6 24 ii.a3 was V.Malakhatko­ S.Sivokho, Polanica Zdroj 1999, and here 24 ... e4! 2S .:tel ii.es 26 ii.xe7 (26 .l::!.x e4 ii.xh2+) 26 ...'l!Vxe7 is good for Black because 27 ii.xds? loses to 27 ...'l!Vg s, intending ....l::!.xd6 or ...ii.g7.

20 e4 The alternative 20....l:tf6 is not so bad theoretically, but Black must work for a draw after 21 .l:tadl 'lli'd6 22 ii.b3 .l:td8 ...

Th e M a r de/ Plata Varia tio n : 9 b4 l'i::i h 5 1 o .l:. e 1 (22 ... .l:r.xe6 can be met by 23 J:txds l'i::ixds 24 .l:tdl when White will recoup his ma­ terial with a slight edge) 23 bS (other­ wise, 23 .l:txds oversimplifies: 23 ...l'i::ix ds 24 .l:r.dl 'l!Vxe6 2s 'l!Vxe6+ J:txe6 26 .l:txds .l:txds 27 ii.xds �f7 and the ending is drawn) 23 ... 'l!Vxe6 24 'l!Vxe6+ .l:txe6 2S ii.cl! �7 26 ii.a3 e4.

Here White can try to improve his position or he can win back the pawn immediately: a) 27 ii.xe7 seems to be the most ac­ curate: 27 ... �xe7 28 .l::!.x ds .l:txds 29 ii.xds .l::!.e s (instead 29 ... ii.c3 30 .l:te2 .l::!.e s 31 ii.xe4 .l::!.xbs 32 ii.d3+ .l:tes 33 .l:r.xes+ ii.xes 34 ii.xg6 is Kramnik-Shirov, be­ low) 30 ii.xe4 (White must play this, because 30 .l:txe4 �d6 31 .l:txes �xes 32 ii.f7 gs 33 a4 was already agreed drawn in V.Mikhalevski-M.Klinova, Gi­ braltar 2008). Now 30 ... .l:txbs 31 ii.d3+ .l::!. e s 32 J:txes+ ii.xes 33 ii.xg6 is Kram­ nik-Shirov again, so Black should play 30...�6 31 a4 which is Mikhalevski­ Melekhina below, but with a couple of tempi less for each side. b) 27 g3 and here:

bl) 27 ... .l:td7 28 ii.xe7 �xe7 29 J:txds .l:txds 30 ii.xds ii.c3 (also better for White is 30 ....l::!.e s 31 ii.xe4 .l:txbs 32 ii.xg6+ �6 33 ii.d3 .l:tb2 34 a4) 31 J:te2 J:tes 32 ii.xe4 and White was much bet­ ter in V.Kramnik-A.Shirov, Linares 1998. b2) 27....:tes ! 28 .l::!.d4 with a final di­ vide: b21) 28 ....l:te6 29 ii.xe7 �xe7 30 .l:txds .!:txds 31 ii.xds J:tes 32 ii.xe4 �f6 33 a4 was V.Mikhalevski-A.Melekhina, Philadelphia 2008. White can certainly play on for a fair while, although Black did manage to draw this game. b22) 28 ... t'i::ifs ! is best. Then Mik­ halevski gives 29 .l:r.dxe4 l'i::id4 30 .l::!.x es ii.xes 31 �g2 ii.f6 with equality. 21 .l:r.ad1 White threatens ii.xe4, so Black must move his queen off the d-file. 21 'l!Vc7 22 ii.b3 Instead 22 ii.a4 still looks like the wrong idea. After 22 ....l::!.fs (Black could look into 22 ... 'l!Vc3 or 22 ...'l!Vc4) 23 ii.d7 .l:taf8 24 bS, thematic is 24 ... hS when White must commit his queen: 2S 'lli'h4 (2S 'l!Ve2 could be met by 2s ... ii.es or 2s ... ii.c3 26 .l:r.c1 'l!Ves) 2s ...ii.f6 26 'lli'g 3 (26 'l!Vh3 is possible, but the queen does not seem well placed here) 26 ... ii.es 27 'lli'h4 ii.f6 repeats. Instead 24 ..'lli'c3 2 S 'l!Ve2 'l!Ves 26 'lli'd2 'lli'd6 (now 26 ...h s can be met with 27 'l!Vb4! ii.f6 28 'lli'a4 'lli'b8 29 .l::!.e 2!, preparing to double rooks on the d-file) 27 h3?! (not 27 ii.xh6 ii.xh6 28 'lli'xh6 .:ths, but 27 a4 gs 28 'l!Vc1 is possible) 27...g S ! 28 a4 was B.Lalith­ V.A.Rajesh, Delhi 2010, and here Black •••

121

A tt a c k i n g C h e s s : Th e King 's I n d i a n, Vo l u m e 1 should play 28 ...d4! 29 ii.xd4 (not 29 'l!Vc2 dxe3 30 .l::!.xd6 exf2+) 29 ....l::!.ds 30 .l:!.xe4 .l:tf4! when White has some prob­ lems. 22 .l::!.fs! ...

This move protects dS, cuts off the e6-pawn and Black may also double rooks on the f-file or play ..J::t e s. White has tried many things here, but has been unable to prove much. 23 .l:td2 White prepares to double rooks on the d-file and leaves the dl-hS diagonal open for his queen. There are many other moves: a) 23 'l!Ve2 .l::!.af8 24 'lli'a6 (24 .l::!.fl and 1/2 -Yi was T.Radjabov-J.Moreno Carnero, Pamplona 2002) 24 ... ii.c3 2S .!:tel 'l!Ves 26 .l:tedl ii.xb4 27 ii.d4 (27 'l!Vxa7!?) 27 ... 'lli'b8 28 'lli'a4 ii.d6 29 g3 hS gave Black a good position in A.Janusonis­ D.Chocenka, correspondence 2oos. b) 23 ii.d4 ii.xd4 (23 ... .l:taf8!?) 24 l:i'.xd4 'l!Vc3 2S 'lli'dl .!::taf8,, 26 .!::t e 2 'l!Vc7 27 g3 'l!Vc6 28 'l!Vc2 1/2-Yi was S.l skusnyh­ A.Motylev, Moscow 1999. c) 23 .:tel 'Ii' d6 24 .l::!. e dl .l::!.af8?! (bet-

122

ter is 24...'l!Vxb4 2 S .l::!.c7 h S , leading to variation 'e2' below with Black having avoided the note to White's 24th move there) was L.Gerzhoy-V.Rajlich, Buda­ pest 2002, and White could play 2S 'l!Vxe4! here. d) 23 bS is logical. The pawn is not hanging anymore, c6 is controlled and ii.c1-a3 becomes a possibility.

Here Black has: dl) 23 ...'l!Ves?! 24 ii.cl! .l::!.af8 2s ii.a3 hS 26 'lli'g 3 'l!Vxe6 27 ii.xe7 'l!Vxe7 28 ii.xds+ �h8 29 ii.xe4 .l:tg s 30 'lli'd3 and White was up a pawn with better posi­ tion in L.Lenic-A.Jankovic, Sibenik 2006. d2) 23 ....l:taf8 24 .l:te2 'l!Vc8 2S ii.cl (White should consider 2S .l:tc2 'l!Vxe6 26 .!::tc7) 2s ... hs 26 'i!Vh4 ii.f6 27 'l!Vh3 �h7 28 ii.a3 'l!Vxe6 29 g3 .l::td8 30 f3 'lli'c8 gave Black a good position in M.Bosiocic1.Saric, Split 20os. d3) 23 ...�h7 24 .l:te2 (after 24 ii.cl .!::taf8 2S ii.a3 .l:txf2 26 ii.xe7 'l!Vxe7 Black is much better following 27 .l:txds e3 or 27 ii.xds 'l!Vcs) 24 ... .l:td8 2s ii.cl ii.es 26 g3 hS 27 'l!Vh3 'l!Vcs is comfortable for Black, who is well coordinated.

Th e M a r de/ Plata Varia t i o n : 9 b4 l'i::i h s 1 0 .l:. e 1 e ) 23 .l:te2 is similar t o the main line, but here White may bring the rook to c2 as well.

Black has: el) 23 ....l:td8 24 bS hS 2S 'lli'h 3 �h7 26 ii.cl 'l!Vcs 27 .l:tc2 'l!Vxbs? (better is 21...'l!Vd6) 28 .l:te7 'l!Vb4 29 J:txa7 'l!Vcs 30 'l!Ve3 gave White a winning position in N.Alfred-S.Nagle, Budapest 2004. e2) 23 ...'lli'c 3 24 .l:tc2 (it is probably better to play 24 bS .l:taf8 2S .l:tc2 with some initiative for White, V.Mikha­ Vlissingen levski-A.David, 2000) 24 ... 'l!Vxb4 2S l:te7 h S 26 'l!Vh3 (and here better is 26 'l!Ve2 'lli'd6 27 l:txe7 'l!Vxe7 28 .l:txds .l:txds 29 ii.xds .l:tc8 30 g3 ii.f6, which is pretty even) 26 ... as! 27 g4! (forced, because 27 a4 is good for Black and after 27 .l:tb7? a4 28 ii.c2, as in J.Sales-Lim Chuing Hoong, Kuala Lum­ pur 20os, Black has 28 ... .l::!.c 8! 29 ii.bl d4 winning, as pointed out by Mik­ halevski) 27 ...hxg4 28 'lli'h4 .l:te8 29 ii.gs e3! and Black has good counterplay according to Bologan. e3) 23 ...h S ! takes immediate advan­ tage of the fact that the queen cannot

retreat along the dl-hS diagonal. After 24 'l!Vh4 ii.f6 2 s 'lli'g3 ii.es (2s ...'lli'c 6!? could be tried) 26 'lli'h4 .�f6 27 'lli'g 3 ii.es 28 'lli'h4 ii.f6 the game was drawn by repetition in Xu Jun-R.Ponomariov, Shenyang 2000. Returning to 23 .l:td2:

23 'l!Vc3!? This is the most enterprising move. Others: a) 23 ... ii.c3? 24 .:tel 'l!Ves loses to 2S .l:txds ! l'i::ixds 26 'l!Vxg6+ �h8 2 7 'lli'x h6+ �g8 28 'lli'g 6+ �h8 29 'l!Vh6+ �g8 30 e7!, R.Sherbakov-S.lskusnyh, Tula 1999. b) 23 ...J:taf8 24 .l:tedl 'lli'c6 was S.Savchen ko-R.Ponomariov, Alushta 2000. Now 2S .l:tc2 'lli'd6 (worse is 2s. . .'l!Vxe6 26 l:te7 - Ponomariov) 26 'l!Vxe4 ii.es ! 27 .l:txds ii.xh2+ 28 �hl .l:txds 29 ii.xds ii.f4 was level in A.Koz lowicz-F.Schwarz, correspondence 2003. c) 23 ...'l!Vc6 24 bS!? 'l!Vxe6 2S f3 exf3 26 .�f2 'lli'f6 27 .l:txe7 'l!Vxe7 28 ii.xds+ .l::!.x ds 29 .l:txds 'l!Ve2 30 gxf3 .l:tf8 31 'l!Vxg6 'lli'xf3 led t o a draw in Z.Straka­ D.Fridman, correspondence 2004. .•.

123

A tt a c k i n g C h e s s : Th e King 's I n d i a n, Vo l u m e 1 d) 23 ... �h7 24 .l:tedl 'Ii'c6 25 .l:tc2 (25 'l!Vh3 .l:td8) 2s ...'l!Vxe6 (2s ...'lli'd6 26 'l!Vxe4!) 26 .l:tc7 was B.Lalith-A.Lahiri, New Delhi 2009. Now 26 ... hs with the idea of ...'lli'd6 would be unclear. e) 23 ... .l:td8!? is untried but looks sensible.

24 'l!Vd1 Instead after 24 .l:tedl 'l!Vxb4 25 ii.d4?! (better is Panczyk and llczuk's 25 'l!Ve2 when 2s ....l:td8 26 'l!Va6 would give White counterplay) 2s .. Jk8 26 'l!Ve2 ii.xd4 27 .l:txd4 'l!Vcs 28 g4 .:tes 29 'lli'd2 .l:tf8! 30 'lli'x h6 .l:txe6 Black stood well in J.Werle-A.David, Amsterdam 2000. 24 ....:tds Black has some other options: a) 24... .l:taf8 25 bS (25 g4 .l:tes) 2s ...�h 7 (2s ...'lli'b 4!? has been proposed by Panczyk and llczuk) 26 J:tc2 'l!Ves 27 ii.cl d4 28 ii.a3 d3 29 ii.xe7 (29 J:td2 ii.f6 30 .l::!.xd3 .l:txf2 31 .l:td7 ii.h4! is un­ clear) 29 ... dxc2 30 ii.xc2 .l:te8 31 .l:txe4 'l!Vxbs 32 ii.a3 'lli'ds 33 'lli'xds 1/2-1/2, A.Eriksson-G.Glatt, correspondence 2002. After 33 ... .l:txds 34 e7 ii.f6 35 ii.b3 .l:td7 36 ii.a4 .l::!.e xe7 the position is to-

124

tally equal. b) 24 ...'l!Vxb4 25 g4 .l:tes (instead 2s ... .l:tf6 26 .l:td4 'l!Vd6 27 .l:txds lllxds 28 'lli'xds 'lli'xds 29 ii.xds .l:te8 30 .l::!. c 1 gives White compensation for the exchange) 26 .l:td4 'l!Vc3 ! (not 26 ... 'l!Vas 27 ii.d2 'lli'bs 28 a4 'lli'c6 29 ii.b4 with a big advan­ tage for White - Tsesarsky) 27 ii.f4 .l:tf8 28 ii.xes ii.xes 29 .l::!.xds (forced) 29 ... ii.xh2+ 30 �g2 .l:txf2+ 31 �xf2 lllxds 32 ii.xds 'lli'g 3+ 33 �1 (not 33 �e2? 'IJVf3+ 34 �d2 'IJVd3+ 35 �Cl Ji.f4+ 36 �b2 ii.es+ 37 �cl 'lli'c3+ 38 'l!Vc2 'l!Vxel+) 33 ...'lli'h 3+ with a draw. 2s bs 'lli'b4 Black can also play 2s ... 'lli'c8 when 26 ii.d4 ii.xd4 27 .l:txd4 'l!Vcs transposes to the main text, and in fact this was the actual move order of Xu Jun-Ye Ji­ angchuan, below. However, White has an extra option here in 26 g4!? .l:tff8 27 ii.xds ii.c3 28 ii.xe4 ii.xd2 29 ii.xd2 'l!Vxe6 30 'l!Ve2 with good compensation. 26 ii.d4 But not 26 a4? ii.c3. 26....:tdfB 27 a4 ii.xd4 28 .l:txd4 'l!Vcs 29 .l:te2

Th e M a r de/ Pla ta Va r i a t i o n : 9 b4 tb h s 1 0 .l:. e 1 White covers the vulnerable f2pawn. 29 .l::!. es 30 h3 Instead 30 .!:tdd2 'lli'd6 31 .l:tc2 .l::!.xe6 32 'lli'c1 .l:tf7 33 .l:te7 (not 33 'lli'xh6? .l:th7) 33 ...tt:ifs 34 .l:txf7 �xf7 3S 'lli'd2 �f6!? 36 ii.xds .:tes 37 ii.b3 'lli'xd2 38 .l:txd2 .:tcs gave Black a slightly better ending due to his active pieces in J.Skeels-S.Boyd, correspondence 2006. After 30 h3, the game Xu Jun-Ye Ji­ angchuan, Shanghai 2001, was agreed drawn, but Black could play on after 30 ...J:tffS OT 30...�g7. ...

is not looking very attractive these days because of Bareev's idea 13 cs tl:if4 14 ii.c4 �h8 lS .l:tbl!. Bl) 12 c6 ...

B) 12 f3 Black increases the tension as in Line A while also threatening ... 'lli'b6+. This choice is somewhat more risky than Line B2, but it is quite playable and leads to a double-edged game. White has two fundamental lines: 811: 13 'ili>h1 812: 13 .i.eJ

This move is less forcing than 12 ii.f3 and it is for that very reason that it has become very popular. Indeed, a lot of tension remains in the position, al­ though Black has a broader choice here:

B1: '12 c6 •..

82: 12 ¢>hl •••

Black's other main move, 12 ... tl:ihs,

Other moves do not mind the g1-a7 diagonal and are rare: for example, 13 bS h6 14 tl:ie6 ii.xe6 lS dxe6 fxe4 16 fxe4 'lli'b6+ 17 �hl ex bS 18 ex bS .l:tac8 when White must be careful, because 19 tl:ids loses to 19 ...tl:iexds 20 exds tl:ie4 and 19 ii.b2 fails to 19 ....l::!.xc3 20 ii.xc3 tl:ixe4. Instead 13 'l!Vb3 is reasonable, al­ though it takes pressure off the d-file and 13 ...h6 14 tl:ie6 ii.xe6 lS dxe6 'lli'c 8 16 .l:tdl .l:td8 looks fine for Black.

125

A ttacking C h e s s : T h e King 's I n d i a n , Vo l u m e 1 811) 13 �h1

White safeguards his king by re­ moving it from the g1-a7 diagonal. 13 ... h6 Black forces the play. 14 l2Je6 ii.xe6 15 dxe6 lbe8 Threatening to play ... lbe7xe6-d4. Bologan suggest the rarer 1s ... 'l!Vc7 when 16 bS (instead 16 ii.b2 .l:tad8 17 'l!Vb3 .l:tfe8 18 .l:tadl �h7 19 ii.d3 fxe4 20 lbxe4 lbxe4 21 J:txe4 l2Jfs 22 f4 J:txe6 23 fxes was V.Epishin-1.Nemet, Biel 1996, and here Epishin suggests 23 ...ii.xes) 16 ....l:tfd8 17 bxc6 bxc6 18 .l:tbl 'lli'c 8 19 'lli'a4 'l!Vxe6 20 .l:tb7 as 21 ii.e3 .l:td7 22 .l:tb6 .l:te7 23 .l:tebl lbd7 is fairly level in his view. 16 'l!Vb3 After 16 bS Black should play 16 ... 'lli'c8, rather than 16 ... cs?! 17 lbds lbc7 18 l2Jxc7 'lli'xe7 19 exf s gxfs (19 ... lbxfs may be better, but even if Black wins the e6-pawn White will al­ ways have light-square compensation) 20 g4 when White had the initiative in D.Lima-A.Romero Holmes, Leon 1996. 16...l2Jc7

126

17 c5 This is almost always played, but Black should also be ready for 17 .l:td1! ? which looks natural. Here he can con­ sider: a) 17 ... l2Jxe6 18 cs dS 19 exds l2Jd4 20 .l:Ixd4 exd4 21 dxc6+ (not 21 d6+ �h7 22 dxe7 'l!Vxe7 and Black wins ma­ terial) 21 ... �h7 22 cxb7 .l:tb8 23 ii.f4 is very messy, but looks roughly balanced. b) 17 ...�h7 18 cs dS 19 exds cxds 20 lbxds lbexds (worse is 20...lbcxds 21 ii.c4 e4 22 .l:tb1) 21 ii.c4 e4 22 ii.xds (after 22 .l:tb1 'lli'h 4! 23 g3 'l!Vh3 24 ii.xds lbxds 2s 'lli'xds .l:tad8 26 'l!Vb3 J:td3 ! 27 'l!Vc2 exf3 Black has good play) 22 ...lbxds (not 22 ... ii.xal 23 ii.xb7 with a big advantage) 23 .l::!.x ds 'lli'f6 24 e7 'l!Vxe7 2S ii.b2 ii.xb2 26 'l!Vxb2 .l:tad8 and the position is level. 11... ds 18 exds cxds After 18...lbexds 19 ii.c4 White maintains pressure. 19 ii.b2 Black has a tremendous centre, but it is not easy to hold it together if White plays aggressively.

Th e M a r de/ Plata Varia tio n : 9 b4 l'i::i h s 1 0 .l:. e 1

19 b6!? Black seeks to break up White's queenside pawn chain. Other moves are risky: a) 19 ... 'lli'c8 20 t'i::ib 5! l'i::ixb5 (this is probably a mistake; instead 20 ... a6 21 l'i::id6 'l!Vxe6 22 l'i::i xb7 was E.Bareev­ R.Polzin, Rethymnon 2003, and here 22 ... e4! ? would be unclear, while 20 ... a5! ? is another option) 21 ii.xb5 'l!Vxe6 22 ii.xe5! ii.xe5 23 f4 gives White a comfortable advantage. b) 19 ...'l!Ve8 is the most popular move. Black removes his queen from the d-file and keeps an eye on the b5square as well. Markos stops here, pointing out that Black has scored well, but the position is actually not so easy for him. White has: bl) 20 .l:tadl .l::!. d 8 21 a4 a6 22 b5 'it>h7 23 ii.a3 l'i::i xe6 24 bxa6 l'i::id4 (probably better is 24 ... bxa6 2 5 ii.xa6 l'i::id4 26 'l!Va2 'lli'd7 with an unclear posi­ tion) 25 'l!Vxb7 .l::!.b8 26 'lli'e7 l'i::ic2 27 l'i::ib5 was E.Bareev-V.Topalov, Dortmund 2002, and now 27 ... .l:tc8 is probably okay for Black. However, 27 ii.b5 looks ...

stronger, since after 27...'lli'f7 28 c6! l'i::i x a3 29 l'i::i x d5 l'i::i xd5 30 'lli'xf7 .l::!.xf7 31 .l:txd5 White is much better. b2) 20 a4 a6 21 .l:tadl (the sacrifice 21 l'i::i xd5 does not work as well here: 21 ...l'i::ic xd5 22 l:!.ad1 'lli'c6 23 ilc4 l:!.fd8 24 ii.xe5 ii.xe5 2 5 .l::!.x e5 t'i::if6 26 .l:teel b5 27 'l!Vc3 l'i::ifd5 28 ii.xd5 .l:txd5 29 axb5 axb5 30 'lli'f6 'lli'e8 gave Black good chances in Van Wely-Kotronias, Euro­ pean Championship, Warsaw 2005) 21 ....l:td8 22 l'i::i x d5 .l:txd5 (worse is 22 ...l'i::ic xd5 23 ii.c4 'lli'c 6 24 b5 axb5 2 5 axb5 'l!Vxe6 26 ii.xe5 ii.xe5 27 f4) 23 ii.c4 .l:txdl 24 .l:txdl 'l!Vb8 25 l:!.d7 and even though White has only a pawn, he has good compensation for the piece, L.Van Wely-D.Stellwagen, Dutch Cham­ pionship, Leeuwarden 2005. b3) 2o l'i::ix d5!

This immediate sacrifice is more dangerous than in Van Wely-Kotronias above. Leaving the respective a-pawns on their home squares helps White. After 20 ...l'i::i c xd5 21 .l:tadl there is: b31) 21 ...'l!Vc6 22 ii.c4 .l:tfd8 23 b5! 'l!Vxc5 24 ii.a3 ! is the difference because

127

A tt a c k i n g C h e s s : The King 's I n d i a n, Vo l u m e 1 the a-file is not open for Black's rook and 24 ... 'l!Vc7 2 5 ii.xd5 seems better for White, although I suspect Black can hold after 25 ... .l::!.xd5 26 .l:txd5 l2Jxd5 27 'lli'xd5 'l!Vc3! 28 e7+ �h7 29 .l::!.g l 'l!Vxa3 30 'lli'd8 .l::!.xd8 31 exd8'1li' 'l!Vxa2. b32) 21 ... �h7 22 ii.c4 l2Jf6 23 ii.xe5 b5 24 ii.fl gave White good compensa­ tion for the piece in L.Van Wely­ V .Kotronias, European Team Champi­ onship, Gothenburg 2005. After 24... a6, Bologan suggests 25 �gl (instead of the game's 25 g3) 2S ....l::!.a7 26 a3 with some advantage. 20 cxb6 Worse is 20 c6 �h7 21 .l::!.adl l2Jxc6 22 l2Jxd5 lbxe6 23 lbxb6 'l!Vxb6 24 'l!Vxe6 .l::!.fe8, which was drawn here in L.Van Wely-F.Nijboer, Dutch Championship, Hilversum 2006. However, instead of 24....l:tfe8 Black could play 24 ...'l!Vxb4 25 .l:t bl (25 'l!Vxc6 'l!Vxb2 leaves Black a pawn up) 25 ... .l:tf6 26 'Ii'c4 'l!Vxc4 27 ii.xc4 .l:td6 and despite White's bishop­ pair, the extra pawn gives Black reason to play on. 20 ...axb6 21 .:tad1 'lli'b8

128

Worse is 21 ...'lli'd6 because of Mik­ halevski's suggestion 22 lbb 5! lbxb5 23 ii.xb5 d4 (23 ...'l!Vxe6 allows the familiar 24 ii.xe5 ii.xe5 25 f4) 24 a4 .l::!.fc8 25 ii.d7 .l:tc7 26 b5 with the idea of ii.a3 when White has pressure. 22 lbxds After 22 lbb5 lbxe6 23 ii.c4!? dxc4 (Black should avoid 23 ...l2Jf4 24 ii.xd5+ l2Jfxd5 25 l2Jc3) 24 'l!Vxc4 'lli'c8 Mik­ halevski gives 25 lbc7 'l!Vxc7 26 'l!Vxe6+ .l::!.f 7 27 .l::!.d 7 'lli'c6 28 .l:txe7 'l!Vxe6 29 .l:txe6 .l:txa2 30 ii.xe5 ii.xe5 31 .l:t6xe5 �g7 32 .l:te6 l:!.f6 33 .l:te7+ .l:tf7 with a draw. 22 lbcxds 23 .l:txds lbxds 24 'lli'xds •••

.:tds 2 s 'lli'b3 �h7 26 ii.bs White's advanced d-pawn gives him enough compensation for the ex­ change, but no more. Black soon sacri­ fices back the exchange to equalise. 26 'lli' d 6 27 ii.d7 .l:!.a7 28 .l:td1 'lli' e 7 29 a4 .l:taxd7 30 exd7 Here a draw was agreed in E.L'Ami­ F.Nijboer, Dutch Championship, Hilver­ sum 2008. .••

812) 13 ii.e3

The M a r de/ Plata Va r i a t i o n : 9 b4 lb h s 1 0 .l:. e 1 White covers the diagonal towards his king by developing. 13 ...ii.h6 This is the main line. Black tries to exploit the pin along the c1-h6 diago­ nal and threatens .. .f4. Instead 13 ... h6 is possible as well. After 14 lbe6 ii.xe6 lS dxe6 -.a (Markos does not mention this possibility, giving only 1s ...lbeS 16 -.b3 lbc7 17 .l::!.adl when White has the upper hand) 16 .l:!.b1 (16 �3 .l:tfdS 17 bS -.cs is equal according to Bologan) 16 ... .l:!.fdS 17 bS -.cs 1S bxc6 bxc6 19 -.a4 �h7 (19 ...-.xe6 and 19 ...f4 are also possible) 20 .l:tedl -.xe6 21 .l:tb7 .l:td7 22 -.a6, as in D.Komljenovic-A.Kuzmin, Benasque 1999, then 22 ... hs 23 .l:txd7 -.xd7 24 ii.cs lbcS 2s ii.xd6 lbxd6 26 cs ..tfS is equal (Bologan).

14 h4 Too tame is 14 ii.d2. After 14.. .f4 lS lbh3 (1s lbe6 ii.xe6 16 dxe6 -.cs wins the e6-pawn right away) 1s ...ii.xh3 (not 1s ... gS?l 16 lbf2 when White was better in E.Lobron-R.Barcenilla, Yerevan Olympiad 1996) 16 gxh3 ii.gs Black has a good game.

Instead 14 cs seems to ignore Black's threat, but after 14 ...f4 (14 ...cxdSl? l S cxd6 cxd6 could be an improvement) l S cxd6 fxe3 16 dxe7 -.xe7 17 d6 -.ds 1S ii.c4+ �g7 19 lbe6+ ii.xe6 20 ii.xe6 �6 (or 20 ... lbh s l ? with the idea of ...lbf4 and ....l:tf6) 21 lbe2 .:tads it was unclear in E.Bareev­ J.Balcerak, German League 2001. 14...cxds 15 cxds White has also tried lS exds. This gives him access to the e4-square, but it hands fs to Black. After 1s _.f4 16 ii.f2 ii.xgs 17 hxgs lb hs 1s cs lLifs 19 lbe4 (following 19 ii.bs lbhg3 20 lbe4 lbxe4 21 .l:txe4 -.xg s 22 -.d2 the position is unclear according to Mikhalevski) 19 ... lbfg3 20 ii.xg3 (after the 20 ii.c4 lbxe4 21 .l:txe4 of M.Bosboom-F.Nijboer, Wijk aan Zee 1997, Black should simply play 21 ...-.xgs) 20 .. .fxg3 21 ii.c4 ii.fs 22 lbxd6 -.xgs 23 lbxfs .l:txfs 24 d6+ �g7 Black had a strong attack in T.Kotanjian-V.Kotronias, Thessaloniki 2007. 1s ...ii.d1 Black will have to play this move at some point, both to cover the bS­ square and to allow the e7-knight to move off the queen's path to gs. Black has also tried the immediate 1 S .. .f4 16 ii.f2 ii.xgs 17 hxgs lbhs 1s .l:tc1 and then : a) 1S ... lbg3 19 -.d2 ii.d7 (no better is 19 ...lbxe2+ 20 -.xe2 ii.d7 21 lbbs) 20 ii.bs and White was better in T.Radjabov-F.Nijboer, Wijk aan Zee 2001.

129

A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1 b) 18 ... ii.d7 19 ii.bs ! gave White some advantage in J.Berkvens-P.Smir­ nov, Patras 2001, but 19 lbbs ii.xbs 20 ii.xbs l2Jc8 gave Black counterplay in A.Huzman-1.Caspi, Israeli League 2008.

hxg s .l:ta4! was suddenly much better for Black in S.Brynell-V.Kotronias, Stockholm 2006. Correct, though, is 20...ii.xgs 21 hxgs l2Jxe4 22 l2Jc7 l2Jfs 23 l2Jxa8 'l!Vxa8 with some compensation for the exchange in an unclear posi­ tion. 16 f4 Instead 16 ... a6 is a sneaky move be­ cause Black will have to play this at some point anyway. After 17 a4?! (a waste of a tempo; it would be much more testing to just bring a rook to cl) 11...f4 18 ii.f2 ii.xgs 19 hxgs lbh s 20 bs l2Jc8 21 b6 'l!Vxgs 22 .l:!.ecl 'lli' d 8 23 'l!Ve1 lbg3 ! 24 ii.xg3 fxg3 2S 'l!Vxg3 'l!Vxb6+ 26 'lli'f2 'l!Vas left Black a pawn up in L.Murzin-V.Kotronias, Linares 2003. 11 ii.f2 ii.xgs 18 hxgs lbhs ...

16 'lli' d2 This move was recommended by Markos and it is White's latest try. White connects his rooks and there is also a subtle point in lining up White's queen against the gs-square, as we shall see. The alternative is 16 .l:!.cl a6 17 bS and then: a) 11 ...f4 18 ii.f2 ii.xgs 19 hxgs lbhs 20 'l!Vb3 as (not 20 ... l2Jc8 21 bxa6 bxa6 22 'l!Vb7) 21 b6 l2Jc8 22 ii.bs .l:tf7 23 ii.xd7 .l:txd7 24 lbbs was S.lvanov­ H. Rudolf, Panormo 2001, and here Black should take the pawn with 24 ... lbxb6, although White has a strong initiative after the simple 2S .l:tc2, in­ tending to double rooks on the c-file. b) 11... ii.xbs! ? 18 ii.xbs axbs 19 lbxbs fxe4 20 fxe4 l2Jxe4?! 21 'lli'g 4?! (White could play 21 lbxe4 ii.xe3+ 22 .l:txe3 'lli'b6 23 'l!Vb3! lbfs 24 l:tcel with an obvious advantage) 21 ...ii.xgs 22

130

This is the typical structure for the 13 ii.e3 line. Black will win the g s-pawn and have some kingside pressure, but White will obtain a strong initiative on the other wing. 19 .l:tec1 Instead 19 ii.bs looks premature. After 19 ... ii.xbs 20 lbxbs a6 21 l2Jc3

Th e M a r de/ Plata Varia tio n : 9 b4 lb h s 1 0 .l:. e 1 �h8 ( 21 . .lbc8 looks better) 22 .l::!. ecl lbg8 23 lbe2, as in l.lkonnikov­ V.Kotronias, Gausdal 2008, Black should play 23 ....l:tf7 to cover C7, with the idea of ...'l!Vxgs with an unclear po­ sition. 19 ...a6 Now this is necessary because after 19 ... lbc8 both 20 lbbs and 20 ii.bs look strong. 2o bs White consistently plays to open the queenside. Markos also mentions the strange 20 'l!Ve1?!, but this is playing on the wrong flank and Black gets good counterplay after his 20 ...lbc8 21 ii.h4 h6! 22 gxh6 gs 23 ii.f2 .l:tf6. 20 ... lbcB 21 bxa6 bxa6 22 .l:tab1 'l!Vxgs 23 .l:tb7 .l:tf7

Instead 24 .l:tcbl lbg3 (24 ...'lli'd8 2 S 'l!Ve1 g s 26 l:i'.b8 .l:txb8 27 .l:txb8 was bet­ ter for White in Xu Jun-S.Safin, Bled Olympiad 2002, because the a6-pawn is so weak) 2 S ii.xg3 'l!Vxg3 26 ii.d3 is unclear according to Markos. 24 ...'lli'dB Black wants to play ... lbg3 and able to meet ii.xg3 with .. .fxg3. Still, the immediate 24 ... lbg3 2S ii.xg3 'l!Vxg3 looks playable as well. 25 'l!Ve1 gs 26 lbd1 lbf6 27 J:tcc7 .l:tb8!? Black gives back the pawn to relieve some of White's queenside pressure. 27 ... ii.e8 was an alternative. 28 ii.xa6 .l::!. x b7 29 .l:txb7 g4 Here Black had counterplay in L.Van Wely-F.Nijboer, Dutch Championship, Leeuwarden 2001. 82) 12 ...�hB

Now Black threatens ... ii.h3. Note that Black has trouble creating direct threats with his queen and knight be­ cause ...lbg3 will be met by ii.xg3 when Black cannot recapture with the pawn - this is one point of White's queen being on the d2-square. 24 ii.fl

This is considered the most reliable move nowadays, mostly due to the ef­ forts of Radjabov. Because 12 f3 is a bit slow, Black makes a useful waiting move without committing his pawn structure. White has tried many ideas

131

A ttacking Chess: Th e King 's I n dian, Vo l u m e 1 here but Black's defences continue to hold. B21: 13 .i.e3 B22: 13 lUe6 823: 13 cs 824: 13 .:.b1

White has also tried 13 bS and here: a) 13 ... lbhs 14 ii.fl (14 l2Je6 ii.xe6 1 S dxe6 l2Jf4 16 hf4 exf4 17 .:tel c 6 18 ii.d3 fxe4 19 ii.xe4 ii.es was about equal in P.Ricardi-D.Lemos, Berazategui 2007) 14 ...ii.f6 (14 ... l2Jf4!?) lS lbe6 ii.xe6 16 dxe6 lbg7 (16 ... l2Jf4!?) 17 ii.h6 .l:te8 18 l2Jds l2Jg8 19 ii.xg7+ ii.xg7 20 cs (20 exfs looks like a better try) 20.. .fxe4 21 fxe4 dxcs 22 'l!Vb3 c6 23 l2Je3 'lli'e7 24 bxc6 bxc6 2S ii.c4 l2Jf6 and Black took over the initiative in F.Abbasov­ N.Mamedov, Baku 2009. b) 13 ... l2Je8 14 ii.e3 ii.f6 (after 14.. .f4 lS ii.f2 lbxds 16 lbxds 'l!Vxgs 17 cs White has some initiative for the pawn) lS l2Je6 ii.xe6 16 dxe6 lbg7 (16 .. .f4 17 ii.f2 lbg7 18 cs is given as unclear by Carlsen, although after 18 ... l2Jxe6 19 cxd6 'lli'xd6 20 'lli'xd6 cxd6 2 1 ii.c4 lbd4 22 .l::!. a cl .l:tfc8 Black can hardly complain with the result of the opening) 17 ii.h6 l2Jxe6! 18 ii.xf8 'lli'xf8 19 cs!? lbxcs (19... dxcs!? 20 'lli'd7 'lli'c8) 20 ii.c4 ii.gs 21 'l!Ve2 'l!Vh6 22 .l:tad1 .l:tf8 gave Black sufficient compensation for the exchange in M.Carlsen-V.lvanchuk, Foros 2008. 821) 13 ii.e 3

132

This has often been played, but is not very popular nowadays. 13 ...ii.h6 Black plays along the lines of 'B12', but 13 ...lbhs ! ? is also interesting. After 14 cs (perhaps White will look for im­ provements here) 14...l2Jf4 practice has seen: a) lS ii.c4 fxe4 (Wells gives the line lS ... lbexds! 16 lbxds 'l!Vxgs 17 hf4 exf4 18 lbxe7 fxe4 19 l2Jxa8 ii.h3 20 ii.fl exf3 21 'lli'x f3 ii.g4 22 'l!Vxb7 f3 when Black indeed has a strong attack) 16 lbgxe4 lbfs 17 ii.f2 l2Jd4 has scored well for Black; one example being 18 .:tel ii.fs 19 �hl 'lli'd 7 when Black had a reasonable position in A.Rychagov­ N.Pokazanjev, Novokuznetsk 2008. b) lS .:tel fxe4 (Wells's idea works here too, although not quite as well: lS ... lbexds!? 16 lbxds 'l!Vxgs 17 hf4 exf4 18 lbxe7 fxe4 19 l2Jxa8 ii.h3 20 ii.fl exf3 21 'lli'xf3 ii.g4 22 'l!Vxb7 f3 and here, with White's rook on cl instead of al, he can hold: for example, 23 .l:tc2 f2+ 24 .l:txf2 ii.d4 2S �hl ii.xf2 26 .l::!.e 7 'l!Vh6 27 .l:tf7 dxcs 28 'l!Ve7 .l:txf7 29 'lli'xf7

Th e M a r de/ Plata Va riat io n : 9 b4 lb h s 1 o .l:.e 1 id4 30 'lli'e 8+ �g7 31 'l!Ve7+ with a draw) 16 lbgxe4 lLifs 17 ii.f2 lbd4 18 ifl was slightly better for White in E.Bareev-1.Cheparinov, Amsterdam 2008. 14 lbf7+ The slow 14 ii.d2 should not give Black any trouble after 14.. .f4 lS lbe6 (1s h4 ii.xg s 16 hxgs lbh s) 1s ... ii.xe6 16 dxe6 c6 with ideas like ...'lli'b 6+, ... 'lli'c8 and maybe ... lbe8-c7. Instead 14 h4 f4 lS ii.f2 ii.xg S 16 hxg s lbh s 17 cs lbg8 looks like a better version of Line B12 for Black because ... �h8 is a useful move as it clears g8 for his knight. After 18 'lli'd2 (M.Thejkumar-K.Ramu, Hyderabad 2006), Black can just play 18 ... 'l!Vxgs (18 ... ii.d7 and 18 ... a6 are reasonable too) 19 lbbs 'lli'd8 with the idea of ... lbg3 and ... 'lli'h4. 14....:txf7 15 ii.xh6 f4

White has obtained the bishop-pair, but it has taken time and Black has a clear plan on the kingside. 16 ii.gs White wants to keep his dark-

squared bishop. Instead 16 cs makes a lot of sense because 16 ... lbeg8 is rather forced. Then 17 ii.g s h6 18 ii.h4 gs 19 ii.f2 .l:tg7 20 g4 was R. Pogorelov-1.Nataf, Reykjavik 2004, and here Black should probably just continue with 20 ... hs. White could also play 16 g4 imme­ diately. After 16 ...lbeg8 17 ii.gs hS (Black can also play 17 ...h6 18 ii.h4 gs 19 ii.f2 hS, heading into the main line) 18 gxhs (18 h3 hxg4 19 hxg4 .l:th7 20 �g2 fbh6 21 :J:!.h l fbf7 22 J:txh 7+ �Xh7 23 ii.h4 �g7 should give Black enough play) 18 ... .l:th 7 we have: a) 19 hxg6 .:th S 20 h4 was given as much better for White by Gofshtein, but Black has 20 ....i::!.x gs+ 21 hxgs lbxds 22 �f2 lbxc3 23 .l:th1+ �g7 24 l:th7+ �xg 6 2s 'l!Vh1 'l!Vxgs 26 .l:tg1 'l!Vxgl+ 27 �xgl lbxe2+ 28 �2 lbg 3. This looks quite risky, but there is no mate: 29 'lli'h4 ii.e6 30 .l:txc7 lbf6 just looks very unclear. b) 19 h4 .!:txhs 20 �2 was Z.Gyimesi-P.Acs, Budapest 2004. Now Black has 20 ....l:txh4! 21 ii.xh4 lbg4+ 22 fxg4 'lli'xh4+ 23 �g2 'lli'g 3+ 24 �1 wh en he can take the draw or play 24...'l!Vxc3 with good compensation for the ex­ change. 16...lbegB 16 ...h6!? looks even stronger be­ cause 17 ii.xh6?! lbh7 18 h4 lbg8 19 ii.g s lbxgs 20 hxgs 'l!Vxg s is nice for Black. Instead 17 ii.h4 gs 18 ii.f2 .l:tg7 is an improved version of the main line because Black has not spent a tempo on ... lbeg8. Indeed, after 19 cs g4! 20

133

Attackin g C h e s s : Th e King 's I n dian, Vo l u m e 1 ii.h4 l'Llg6 21 ii.xf6 'lli'xf6 22 fxg4 h S ! Black had good play i n A.Vlasov­ V.Tukhvatullin, Sterlitamak 2008. 11 g4 h6 1s ii.h4 gs 19 ii.f2 hs 20 h3

Van Wely won the first one, but then lost the next three, at which point he headed towards other variations ... 13...ii.xe6 14 dxe6

20...hxg4! This looks better than 20 ....:th7 21 cs l'Lle7. In this position Black has lost two tempi compared to 16 ...h6 and after 22 'it>g2 l'Llg6 23 .l:thl White was somewhat better in B.Gelfand-T.Radjabov, Sochi 2008, although Black managed to win anyway. 21 hxg4 l'Llxg4 22 fxg4 f3! 2 3 ii.xf3 'lli'f6 24 .l:tf1! After 24 'it>g2? ii.xg4! 2S J:te3 .l:taf8 White's position falls apart - Mik­ halevski. 24...'l!Vxf3 25 'lli'xf3 .l:txf3 26 l'Llbs! ii.xg4 21 l'Llxc1 .l:tc8 28 l'Llbs .l:txc4 Black has equal chances in this double-edged endgame.

14 ...l'Llhs This has been played the most, but it may not be best. Others: a) 14.. .fxe4 lS fxe4 l'Llc6 is a typical idea, but it does not fare well here: 16 l'LldS! l'Llxe4 (compared to the position where .l:tbl and ... h6 are thrown in, 16 ...l'Llg8 is not so good because after 17 bS l'Lld4 18 e7 l'Llxe7 White has 19 ii.gs! .l:te8 20 ii.g4 h6 21 ii.xe7 .l:txe7 22 l'Llxe7 'l!Vxe7 23 'lli'd 3 and Black does not have enough) 17 ii.f3 l'Llf6 18 bS! was very good for White in L.Van Wely­ T.Radjabov, Khanty-Mansiysk 200S. b) Van Wely suggests that 14 ....l::!.e 8!? is Black's best in New in Chess.

(seefollowing diagram) 822) 13 t'Lie6 The most direct move. White does not even wait for ...h6. Van Wely liked this approach for a while, playing it in four of his games against Radjabov.

134

The game A.Grischuk-T.Radjabov, Elista 2008, continued lS l'Llds fxe4 16 l'Llxf6 (after 16 fxe4 Black could try 16 ...l'Lleg8!?; the alternative is 16 ...l'Llxe4

The M a r de/ Plata Varia tio n : 9 b4 lb h s 1 0 .l:.e1 1 7 ii.f3 lbf6 when both 1 8 ii.gs lbeg8 and 18 lbxf6 ii.xf6 19 ii.xb7 .l:tb8 seem satisfactory for Black)

16 ... ii.xf6 17 fxe4 lbc6 18 ii.g4 lbd4 19 1i.e3 c6 (Black should also consider both 19 ... lbxe6 20 ii.xe6 .l:txe6 21 'lli'ds ii"c8 and 19 ...'l!Ve7 20 1i.xd4 exd4 21 .l:tfl :ts) 20 1i.xd4 exd4 21 J:tf1 .l:tf8 22 'lli'd3 -W-e7 when Black was better, but the position was drawish. Indeed, after 23 :n 1i.g7 24 .l::!.af1 .l:txf 3 2s .l:txf3 .l:tf8 26 :xf8+ 'lli'xf8 27 'lli'f3 hS 28 'lli'xf8+ hf8 29 1i.e2 'itig 7 30 bS �6 31 bxc6 bxc6 32 .i.d3 'it>xe6 the game was drawn. This method of play looks like a simple solu­ tion to 13 lbe6, which will probably not be seen so much anymore. Neverthe­ less, we will still examine the sharp 14 ...lbh s. 15 g3 White has a serious alternative in lS cs. After 1s ...lbf4 16 ii.c4 fxe4 17 .l:txe4 (17 fxe4 lbc6 with the idea of ... lbd4 is at least equal for Black) 11 ... lbfs and here: a) 18 ii.xf4 exf4 19 lbds .l::!.e8 20 lbxf4 c6 (not 20 ... ii.xal? 21 'l!Vxal+ lbg 7

2 2 e 7 .l:txe7 23 lbxg6+ hxg6 2 4 .l:th4 mate) 21 e7 .l::!.xe7 22 lbe6 'l!Vb8 (after 22 ...'lli'd7 23 lbxg7 .l:txe4 24 fxe4 lbe3 2S 'l!Ve2 'l!Vxg7 26 .:tel lbxc4 27 'l!Vxc4 dxcs 28 bxcs White is a little better accord­ ing to Mikhalevski) 23 lbxg7 .l:txg7 24 cxd6 (Mikhalevski suggests 24 1i.e6 !? dS 2s .l:te2 'lli'f4 26 ii.xfs gxfs 27 'l!Ve1 .l:tag8 28 'l!Vc3) 24 ...lbxd6 2s .l::!. d 4 lLifs (playing for a win; 2s ... lbxc4 26 .l:txc4 is just equal) 26 J:td3 .l:te7 27 'lli'd2 'l!Ve S 28 .l:tdl .l::!. a e8 was level in L.Van Wely­ T.Radjabov, Monaco (blindfold) 2007, although Black did indeed manage to win. b) 18 g3 lLih 3+ 19 'it>g2 lLigs 20 .:tg4 (Radjabov gives both 20 ii.xg s 'l!Vxgs 21 lbds c6 22 cxd6 lbxd6 23 .l:tg4 'l!Vh6 24 e1 cxds 2s exf8'1li'+ .l:txf8 26 'lli' x ds lLifs 27 .l:te4 lbe3+ 28 J:txe3 'l!Vxe3 29 .l:tdl b6 and 20 e1 'l!Vxe1 21 lbds 'lli'd8 22 ii.xgs 'l!Vxg s 23 lbxc7 lbe3+ 24 .l:txe3 'l!Vxe3 2S lbxa8 e4! 26 f4 .l:txa8 27 cxd6 .l:td8 as unclear) 20...lbxf3 ! 21 'it>xf3 e4+ 22 lbxe4 1i.xa1 23 ii.gs lbe1+ 24 'it>g2 ii.es was rather unclear in L.Van Wely­ T.Radjabov, Biel 2007.

135

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 1s ...ii.f6 A typical idea. Black wants to play ...lbg7xe6. 16 c5 This does not lead to much for White. More interesting is 16 exfs lbxfs 17 g4 (else ... l2Jd4) 17 ... e4 18 l2Jxe4 ii.xa1 19 ii.gs and here: a) 19 ...ii.f6 20 l2Jxf6 l2Jxf6 21 gxfs (Mikhalevski prefers 21 ii.d3, but after 21 ... l2Je7 22 'l!Va1 �g7 23 ii.xf6+ .l:txf6 24 gs 'lli'f8 the position is unclear) 21...gxfs 22 f4 .l:tg 8 23 'lli'd4 cs 24 'l!Va1 'lli'e7 2S �2 .l:taf8 26 bxcs dxcs 27 ii.f3 .l:tg6 28 'l!Ves when White has excellent com­ pensation and went on to win in V.Mamonovas-C.Gatto, correspondence 2007. b) 19 ..id4+ 20 �hl 'lli'e8 21 e1 lbxe1 (21 ..Af7 22 gxfs ii.es 23 f4 l2Jxf4 24 f6 lbe6 2S ii.h4 lbd4 is messy, but looks better for White) 22 'lli'xd4+ lbg7 23 bS l2Jg8 (L'Ami relates how at this point both players were playing very quickly!) 24 ii.d2 'lli'f7 (after 24 ... 'l!Ves White has the simple 2s 'l!Vxes dxes 26 ii.c3 or can play 2S 'l!Ve3 .l:tae8 26 'l!Vxa7! .l::!.a8 27 'l!Vxb7 .l::!.xa2 28 'lli'xe7 which looks very strong) 2s ii.c3 h6 (2s ... .l:tae8 26 'l!Vxa7 b6 27 'l!Va3 h6 28 ii.b2 .l:ta8 29 'l!Vb3 �h7 30 ii.d1 .l:tae8 31 'l!Vc3 also looked good for White in V.Mamonovas-G.Staf, cor­ respondence 2007) 26 f4 .l:!.ae8 27 .:tfl l2Jf6 28 ii.f3 was 1.Cheparinov-T.Radja­ bov, Sofia 2008. Here 28...l2Jxe4 29 ii.xe4 �h7! 30 fs (not 30 ii.xb7?! lbe6 31 'l!Vxa7 lbxf4 32 'lli'd4 cs 33 bxc6 'lli'e7 34 �xf4 .l:Ixf4 3S 'lli'xf4 .l:tf8 with a strong attack -

136

L'Ami) 30....l:txe4! 31 'l!Vxe4 gxfs 32 'lli'd3 a6 33 a4 'lli'e6 34 ii.xg7 �xg7 3S gxfs 'l!Ves is level according to L'Ami. 16 ...f4

17 �g2 Worse is 17 g4?! lbg7 18 ii.c4 lbc6 19 cxd6 cxd6 20 lbe2 (after 20 lbbs lbxe6 21 'lli'xd6, as in T.Bakre-A.Kuzmin, Dubai 2002, Black has 21 ...lbg s! 22 'lli'xd8 .:taxd8 23 l:tfl a6 24 l2Jc3 l2Jd4 with the threats of ... .l:tc8 and ...l2Jgxf3) 20....l:!.c8 21 ii.ds lbxb4 22 .l::!.b1 l2Jc2 23 .l:tfl b6 was much better for Black in L.Van Wely-T.Radjabov, Wijk aan Zee 2007. 17 lbc6 18 cxd6 cxd6 19 lbds •••

Th e M a r de/ Plata Vari a t i o n : 9 b4 lb h s 1 0 .l::!. e 1 19 ...lbd4 19 ... lbg7 also looks good. 20 ii.b2 lbxe6 21 g4 lbhg7 Chances were fairly balanced in this murky position in A.Shirov-T.Radjabov, Wijk aan Zee 2001.

16 exds lbfxds 11 lbxds 'lli'x ds! Worse is 17...lbxd5 18 'l!Vb3 (both 18 ii.b2 and 18 ii.c4 lbxb4 19 .l:tbl look dangerous too) 18 ... c6 19 ii.c4 with pressure.

823) 13 cs

This obvious and direct move was considered harmless for a long time, but lately White has been trying to squeeze something out of the end­ game that arises. 13...h6 Black has also tried 13 ...dxc5 14 bxc5 h6 1 5 lbe6?! ii.xe6 16 dxe6 'lli'd4+! 17 'lli'xd4 exd4 18 lbb5 fxe4 with good counterplay in Z.Gyimesi-T.Radjabov, Moscow 2005, but Avrukh's suggestion 15 lbh3 ! fxe4 16 fxe4 ii.xh3 17 gxh3 indeed loo ks good for White. Black could check out 13 ... a5!?. 14 lbe6 ii.xe6 15 dxe6 ds Black cannot sneak back into Gy­ imesi-Radjabov with 15 ... dxc5, because of 16 'l!Vxd8! (not 16 bxc5 'lli'd4+) 16 ....l:tfxd8 17 bxc5 with a big plus.

18 1'6b3 White's hopes for an edge have be­ come associated with this finesse. By threatening ii.c4 he induces Black to exchange queens, after which the a-file will be opened for the al-rook. The immediate 18 'l!Vxd5 should also not be taken too lightly. After 18 ... lbxd5 19 ii.c4 lbxb4 20 .l:tb1 lbc6! White has: a) 21 J:txb7 lba5 22 .l:tb4 lbxc4 23 J:txc4 .!:tfe8 is no problem for Black. b) 21 ii.d5 .!:tfd8 22 ii.xc6 bxc6 23 .l:tb7 .l:te8 24 .l:txe7 .l:txe6 25 .l:td1 ii.f8 was fine for Black in G.Rechlis-B.Avrukh, Israeli League 2003. White has com­ pensation for the pawn, but no more than that. c) 21 ii.e3 .l:tab8 22 .l:tbdl .!:tfd8 23 .l:td7 .l:txd7 24 exd7 ii.f8 25 f4 .l:td8! (worse is 25 ... e4 26 .l:tbl!, A.Korobov-N.Huschen­ beth, Pardubice 2008) 26 fxe5 .l:txd7 27 e6 .l:td8 and if anything Black is better.

137

A t ta c k i ng C h e s s : Th e K i n g 's Indian, Vo l u m e 1 d) 21 e7! ? looks like White's best try, intending 21 ....l:tfe8 (not 21 ...l2Jxe7 22 .l:txb7) 22 i..f7 .l::!. x e7 23 i..x g6.

20 .l:tfeB! Radjabov's move looks like Black's best road to equality. Black has also pushed his a-pawn: a) 20 ... as?! is loosening. Instead of 21 bS?! l2Jd4 22 i..c4 .l:tfe8 23 i..e 3?! l2Jc2 24 J:te2 l2Jxe3 25 .l:txe3 e4 26 i..d s .l::!.ad8 27 fxe4 fxe4 28 i..xe4 .l:!.xe6 when Black was better in Z.Gyimesi-A.Volokitin, German League 2007, the simple 21 bxas! just looks much better for White. b) 20... a6 is a little bit better: 21 i..b2 (21 bS axbs 22 i..xbs .l::!.fe8 23 J:txa8 .l:txa8 24 i..b2 .l:te8 25 i..xe 5 i..x es 26 .l:txes �g7 27 J:tds .l:te7 should hold) 21 ....l:tae8 22 i..c4 .l:te7 23 i..d s .l:td8? 24 i..xc6 bxc6 25 i..xe s i..x es 26 .l:!.xes with a healthy extra pawn for White in Z.Gyimesi-E.Hagara, Jenbach 2009. A better try is 23 ...l2Jd8 24 .:ta2 (24 i..x es i..x es 25 .l:txes c6 26 i..c4 �g7) 24 ... c6 25 i..c4 lbxe6 26 i..xe 5 l2Jc7 27 i..x g7+ .l:txg7, as given by Golubev, but White is somewhat better here as well. 21 i..c4 This does not lead to much, but White has trouble getting anywhere: ...

Black has: dl) 23 ... e4 is considered Black's best by Avrukh, but Black will suffer here: 24 i..xfs!? (24 fxe4 fxe4 25 .l:txe4 i.. d4+ 26 .l:txd4 l2Jxd4 27 i..b 2 .l:td8 28 .l:tdl �g7 29 i..c2 .l:te2 30 .l:txd4 .l:txc2 is given as equal by Avrukh) 24 ...exf3 25 .l:txe7 lbxe7 26 i..e4 i.. d4+ 27 �fl fxg2+ 28 �xg2 .l::!. g 8+ 29 �h3 b6 30 cxb6 cxb6 31 i..x h6 should be a draw, but White can certainly play on for a while with his bishop-pair. d2) 23 ... .l:tf8 24 .l:txb7 is given as slightly better for White by Avrukh , but 24... e4 gives Black counterplay. d3) 23 ... l2Jd4 also looks okay after 24 .l:txb7 .l:te6 or 24 i..b 2 e4!. 18 'lli'x b3 19 axb3 lbc6 20 .l:ta4 White protects the b4-pawn and may double on the a-file. Instead 20 i..bs .l::!.fe8 21 i..xc6 bxc6 22 bS .l:txe6 (22 ...cxbs 23 .l:ta6 gives White compen­ sation) 23 b6 cxb6 24 cxb6 cs 25 bxa7 .l:te7 is equal according to Giri. ...

1 38

Th e M a r de/ Pla ta Va ria tio n : 9 b4 lb h s 1 0 .l:. e 1 a) 21 ii.b2 .l:txe6 22 bS lbd4 23 ii.c4 J:tee8 24 ii.f7 .l:tf8 25 ii.xg6 lbxb3 26 ii.xes lbxcs is level. b) 21 ii.e3 has been used success­ fully by Arun Prasad. Now Black has: bl) 21 ...lbd8 22 .:teal lbxe6 23 .l::!.x a7 .l:tab8 24 ii.c4 (or 24 .l:tdl lbd4 25 ii.c4) 24.. .f4 2 5 ii.d2 e4 26 fxe4 ii.xal 27 J:txal .l:tbd8 28 ii.c3+ �h7 29 h4 and White had good compensation in 5.Arun Prasad-A.Lahiri, Gurgaon 2010. b2) 21 ... a6 22 bS axbs 23 .l:txa8 .l:txa8 24 ii.xbs .l:te8 25 .l:td1 .l::!.xe6 26 .l:td7 was clearly better for White in 5.Arun Prasad-V. Rajesh, Chennai 2010, as 26 ....l:te7? loses to 27 ii.xc6. b3) 21 ... .l:txe6 is the simplest. Now 22 ii.c4 J:tee8 23 bS lbd4 transposes to variation 'c'. c) 21 bS lbd4 22 ii.c4 lbxe6 23 ii.e3 lbd4 and now: cl) 24 ii.xd4 exd4 25 J:txe8+ .l:txe8 26 .l::!.xa7 d3 27 ii.xd3 ii.d4+ 28 �fl ii.xcs 29 l:txb7 ii.b6 and with White's rook trapped on b7, Black cannot be worse. c2) 24 ii.f7 lbxb S ! ? (Black could also play 24.. 1H8 25 ii.xd4 exd4 26 ii.xg6 d3) 25 ii.xe8 (25 ii.xg6 J:tf8 is fine for Black) 2s ....l:txe8 26 ii.d4 .l:te7 27 ii.b2 c6 28 .l:taal �g8 29 .l::!.adl �f7 and Black had two pawns for the exchange in M.Kozakov-R.Vidonyak, Lvov 2009. 21... a6! 22 bs axbs 23 .l:txa8 .l::!. xa8 24 ii.xbs .l::!.e8! Black finally rounds up the e6pawn. 2 s .l::!. d 1 25 ii.c4 lbas ! does not help White,

so he activates his rook. 2s ....l:txe6 26 .l:td7 ii.f8! 27 .l:txc7 ii.xcs+ 28 �fl .l:te7 29 .l:txe7 Perhaps simpler was 29 .l:tc8+ �g7 30 ii.xc6 bxc6 31 J:txc6 with equality. 29 ... ii.xe7 30 ii.b2 After 30 ii.xh6 gS! White will have to give up a pawn to save his bishop. 30...lbd4 31 ii.xd4 exd4 Despite Black's extra pawn, the endgame is a draw, A.Grischuk­ T.Radjabov, Linares 2009. 824) 13 l:tb1

This flexible move has become White's latest attempt to get some­ thing from the position. The rook gets off the long diagonal and takes a far­ sighted look at the b-file which, as we have seen, can open up in some varia­ tions. This line was also recommended by Markos in Beat the KID. Moreover, after a few reversals with 12 lbe6, Van Wely turned to this variation to beat Radjabov. He was not going to aban­ don the Bayonet and he praised Rad­ jabov for sticking to his guns as well,

139

A ttacking C h e s s : Th e King 's I n dian, V o l u m e 1 expressing that he would rather 'lose a game than lose face'. It will certainly be interesting to see what their future battles bring. 13 ... h6 Forcing the pace. After 13 ... tllh s 14 cs tllf4 lS il.c4 we have transposed to Bareev's 12 f3 tllh s 13 �bl! which was discussed at the beginning of Line B. The main alternative is 13 ... c6 14 il.e3 and now: a) 14 ... il.h6 lS h4 cxds 16 cxds f4 17 il.f2 tllh s 18 'lli'd2 il.xgs 19 hxgs is very similar to Line B21. As Markos points out, the additional moves �bl and ... 'iii>h 8 help both sides to some degree, because White will often open the b­ file with b4-bS (Black usually has to play ... a6 to prevent a white piece from hopping to bS), while ...'iii>h 8 could be useful too because Black can play a quick ... tll g 8 to open his queen's path towards capturing the gs-pawn. b) Black can also play 14 ... h6 lS tlle6 il.xe6 16 dxe6, but this should compare unfavourably for him to the line 12 ... c6 13 il.e3 h6, because of the extra moves �bl (which White plays anyway) and ...'iii>h8 (which Black may or may not play). However, maybe it is not so bad after 16 ...tll e 8 17 bS tll c 7 18 bxc6 bxc6 19 �b7 tllx e6 20 'lli' a4 (not 20 il.xa7?? �xa7 0-1, S.Kojima-G.Timoshenko, Cappelle la Grande 2009) 20 ...tll d4 with an unclear position in R.Markus­ Y.Dembo, Budapest 2002. 14 tlle6 il.xe6 15 dxe6 fxe4 Instead lS ... tllhs 16 exfs! opens the

140

position favourably for White after both 16 ... tllxfs 17 il.d3 and 16 ... gxfs 17 g3, intending f4 when Black is weak on the light-squares. 16 fxe4 Exchanging pieces helps Black. After 16 tllxe4 tllxe4 17 fxe4 tllc6 Black is fine because White has no knight coming to ds. 16...tllc6 17 tlld s

17 tllg S! Black cannot play 17 ...tlld4? 18 e7 and 17...tllx e4? 18 'lli'd3 tllf6 19 'l!Vxg6 must also be avoided. 18 il.d3 Opening a path for the white queen. Other moves do not lead to much: a) 18 l:tfl �xfl+ 19 il.xfl tlld 4 20 'lli'g 4 c6 (with the idea of ... 'lli'f6) virtu­ ally forces White to sacrifice a piece with 21 'l!Vxg6 cxds 22 exds, but this looks insufficient after 22 ...'l!Vh4!. b) 18 cs tlld4 19 il.c4 tllxe6 20 'lli'g4 'lli'e 8 21 cxd6 cxd6 (with the idea ... tlld4) 22 il.bs ! ? 'l!Vxbs 23 'l!Vxe6 'l!Vd3 24 il.e3 tllf6 2S �bdl 'lli'c 4! is also satisfactory for Black - Markos. ...

The M a r de/ Plata Va ria tio n : 9 b4 ll'i h s 1 0 .U. e 1 c ) 18 'lli'd3 ll'id4 19 'l!Vh3 .U.e8! 20 �g4!? (an interesting pawn sacrifice; instead 20 e7 ll'ixe2+ 21 l:txe2 'lli'c8 22 il.e3 ll'ixe7 23 'l!Vxc8 .U.axc8 24 ll'ixe7 :.Xe7 2S il.xa7 is equal according to Markos) 20 ... c6 21 ll'ic3 'lli'e7 22 bS ll'ixe6 (similar is 22 ... 'iii>h7 23 bxc6 bxc6 24 ll'ie2 !Lixe6 2S il.xe6 'l!Vxe6 26 'l!Vxe6 .U.xe6, which has led to draws in some corre­ spondence games) 23 il.xe6 'l!Vxe6 24 'l!Vxe6 .U.xe6 2S bxc6 bxc6 26 il.e3 l:te7 27 .U.edl �d7 gives White compensation for the pawn, although probably no advan­ tage. Then 28 a4 ll'if6 29 as �c8 30 a6 dS!? (30 ... �ce7 is solid, but obviously rather passive) 31 exds cxds 32 ll'ixds was A.Korobov-K.Maslak, Pardubice 2010, and now 32 ... ll'ig4! would keep the game within the bounds of equality. 18 ... ll'id4 19 'lli'g4 gs This move is forced. 19 ... 'iii>h 7?! 20 �e3 ll'if6 21 'l!Vh3 with the idea of e6-e7 is a problem, and 19 ... c6?! 20 'l!Vxg6! is now very strong: for example, 20...'lli'e 8 (20 ...cxds 21 exds ll'if6 22 il.xh6 �g8 23 �gs is crushing) 21 'l!Vxe8 �axe8 22 !Lie3 �xe6 23 bS gives a good endgame.

20 h4 This move was Van Wely's choice in his defeat of Radjabov. There are some othertries, though: a) 20 il.e3 c6 21 .U.fl! ? is speculative, but White was . also successful after 21 ... cxds 22 exds ll'if6 23 'l!Vh3 e4? 24 il.xd4 exd3 2S 'lli'xd3 'lli'e7 26 cs in S.Arun Prasad-J.Ashwin, Kavala 2009. The calm 23 ...'iii> g 8 would have been more testing, but the idea is interest­ ing. b) 20 bS hopes to open the b-file in the event of ... c6 by Black, but Markos points out that Black can play 20 ... 'lli'c 8! 21 h4 'l!Vxe6 22 'lli'hs 'lli'f7 23 'lli'xf7 .U.xf7 24 hxgs hxgs 2 s il.xgs il.h6 with good prospects. c) 20 'l!Vh3 was Ponomariov's choice when this position first occurred and this move is considered best by Markos. After 20 ... c6 White can try: cl) 21 e7 ll'ixe7 22 il.xgs cxds 23 il.xh6 is a sharp piece sacrifice, but Markos illustrates how Black can neu­ tralize it: 23 ... 'lli'c 8! (not 23 ... 'iii>g8 24 il.xg7 'iii>x g7 2s 'lli'g4+ ll'ig6 26 exds �f6 27 l:tfl with a strong initiative) 24 'lli'h4 dxc4! (again holding on to the piece is dangerous after 24 ... 'lli'd7 2 S il.e3+ 'iii>g 8 26 il.xd4 exd4 27 exds) 2 s il.xg7+ 'iii>xg7 26 'l!Vxe7+ �f7 27 'l!Vgs+ 'iii>h8 28 'l!Vh6+ �h7 29 'lli'xd6 ifa 30 'lli'xe7 �xe7 and Black has excellent compensation for the pawn in the endgame. c2) 21 ll'ie3 'lli'f6 and now: c21) 22 ll'ifs ll'ixe6 23 �fl ll'if4 is al­ ready better for Black.

141

A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1 c22) 22 tllg4 'l!Ve7! 23 il.e3 (better is 23 tll e 3 'l!Vxe6 24 'l!Vxe6 tll xe6 2s tllfs .U.ad8 26 il.e3 with some compensation for the pawn) 23 ... tllxe6 24 g3 'lli'd7! ? (perhaps even better is 24...tll d 4!? 2 S �fl 'lli' e 6) 2S �fl was R.Ponomariov­ T.Radjabov, Wijk aan Zee 2003. Here the simplest is 2s ... tll d4 with the idea of 26 il.xg s? tll f 3+. c23) 22 �fl 'l!Vxe6 23 �xf8 (Markos suggests 23 tllf s tllx fs 24 exfs 'lli'f6 2s il.b2 dS 26 bS with compensation for the pawn) 23 ...�xf8 24 'l!Vxe6 tllxe6 2S tllfs .U.d8 26 il.e3 il.f8 27 g3 tll g 7 28 tllx g7 'iii>xg7 29 bS cs 30 �b3 il.e7 31 .U.a3 .U.a8 32 'iii>f2 and White had enough compensation for the pawn in A.Klimov-A.Shomoev, Kemerovo 2007. 20...tllf6 21 'lli'g3 After 21 'l!Vh 3 g4 Black keeps the kingside closed.

21...tllxe6! This is much better than 21 ... gxh4?! 22 'lli'xh4 tllxe6 23 il.xh6 'iii>g 8? (this loses, although after 23 ...tll h7 White has a choice between an endgame edge, with 24 'l!Vxd8 �axd8 2S il.e3, or a

142

continuing initiative after 24 il.xg7+ 'iii>x g7 2S 'lli'g4+ tllh gS 26 l:tfl OT 26 tll e 3) 24 'l!Vh3 il.xh6 2S 'lli'xh6 c6 26 J:te3 ! with a winning attack in L.Van Wely­ T.Radjabov, Wijk aan Zee 2009. 22 hxgs tllx ds Worse is 22 ...tllxg S 23 il.xg S hxgs 24 'l!Vxgs when White has all the play. 23 gxh6 After 23 exds tllxg s 24 'lli'g4 'iii>g8 (24... �f6 is also possible) 2S il.e3 'lli'c8 26 'lli'h4 'lli'd8 27 'lli'g4 'lli'c8 28 'lli'h4 'lli'd8 29 'lli'g4 the game was drawn in K.Landa-A.Shomoev, Ulan Ude 2009. 23 ... il.f6 This is fine, but 23 ...tlldf4 24 hxg7+ tll x g7 2S 'l!Vh2+ 'iii>g 8 26 hf4 �xf4 27 cs .U.h4 28 'lli'g 3 'lli'f6 also looks okay for Black. 24 exds il.h4

An important resource. The position remains very complicated. 25 'lli' h 3 After 2S 'lli' g 6? il.f2+ 26 Wfl il.e3+ 27 il.fs l:txfs+! 28 'lli'xfs tll d 4 29 'lli'f7 'lli'f8 Black wins, while 2S 'lli'g4 il.xel 26 dxe6 (26 il.e3 'lli'h4 27 'lli'x h4 il.xh4 28 dxe6

Th e M a r de/ Plata Va ria t i o n : 9 b4 ll'i h s 1 0 .U. e 1 :ae8 is much better for Black) 26 ...il.f2+ leads to: a) 27 �1 il.h4+ 28 'iii>e 2 (28 'iii> g l ..if2+ is a draw) 28.. ltf2+ 29 'iii> d l -.e7 intending ....U.g8 is unclear. b) 27 'iii>h 2 .U.g8 28 -.e4 -.e7! (28 ... �4+ 29 -.xh4 il.xh4 was given as much better for Black by Velickovic and Sasa in the Inform ant, but White has two dangerous passed pawns and looks to be doing well after 30 'iii>h 3 ! ) 29 i.d2 .U.af8 with counterplay. 25 ... il.xel 26 dxe6 il.f2+ 27 'iii>h 1

21 ...-.f6I This is more ambitious than the sim­ plifying 27 ...°ii'h4.

28 il.d2 l:tae8 29 .U.f1 -.xe6 30 -.xe6 �xe6 31 .il.e4

White looks to have sufficient com­ pensation for the exchange, especially if the kingside pawns get moving, but Black finds a clever resource to take over the initiative. 31 ...1If4!? Instead 3 1 .. ltef6 32 g4 il.g3 33 �xf6 .U.xf6 34 il.gs is not so easy for Black. 32 il.xf4 exf4 33 il.f5 Better was 33 J:txf2 l:txf2 34 g4!. 33 ...1Ixh6+ 34 iLh3 il.g3 Black had a big advantage in Shen Yang-Ju Wenjun, Xinghua Jiangsu 2009.

143

Chapter 7 The Mar del Plata Variation White's other Ninth Moves

1 d4 tllf6 2 c4 g6 3 tllc3 il.g7 4 e4 d6 5 tllf3 o-o 6 il.e2 es 7 o-o tllc6 8 ds tlle7

In this chapter we look at moves other than 9 tlle l, 9 tlld2 and 9 b4. The first of these, 9 il.gs, seems illogical, but it has had its bouts of popularity. 9 il.d2 was once a main line, but it seem s too passive for modern tastes. Finally, both 9 a4 and 9 'iii>h l look a bit indul­ gent, but they should not be underes­ timated. Although these various lines are not considered to be theoretically dangerous, they all have their ideas and Black should still be well prepared.

144

A: 9 �gS B: 9 .il.d2 C: 9 34 D: 9 �h1 There are a few other moves to con­ sider, although they are rather experi­ mental: a) 9 'l!Vc2 looks a bit pointless, but it cannot be too bad: 9 ... tllh s (9 ... tlld7 and 9 ... tlle8 are possible as well) 10 g3 (10 b4 is Epishin's line considered at the beginning of Line B in Chapter 5) 10 ...fs 11 tllgs was P.Eljanov-M.Al Sayed, Khanty-Mansiysk 2009. Here 11 ...tllf6 looks sensible and best. b) 9 il.e3 rather invites 9 ...tllg4 (oth­ erwise White will play 10 tlld2), but it is not so bad: 10 il.gs (after 10 il.d2 fS 11 tllg s tllf6 the position is the same as the main line of the Bayonet, but White has played il.d2 rather than b4 and .U.e1, and after 12 exfs tllxfs! Black is fine) 10 .. .f6 (worse is 10 ... h6 11 il.h4 fS 12 tlld2 tllf6

Th e M a r de/ Pla ta Va ria t i o n : White 's O t h e r N i n t h M o ves 13 f3 gs 14 il.f2 f4 lS cs with the initia­ tive) 11 il.h4 (11 il.d2 fS is the same as 10 il.d2) and now after 11 ... gs 12 il.g3 �h6 we reach a line of the Gligoric where White has castled rather prema­ turely - see note 'c' to White's 12th move in Line Bl of Chapter 9. c) 9 �bl is undoubtedly playable, but it should not be too scary for Black theoretically.

following 12 .. ..i:!xfs 13 il.g4 or 12 ...gxfs 13 f4) 10 g3 (another point to Black's 9th move is that it covers the Cl-pawn, so 10 il.e3 can be met by 10 ... tllx dS!) and now, rather than 10...ilh3 11 �el fs 12 exfs tllxfs 13 tllf3 ! ? h6 14 tlle4 when the h3-bishop was a bit mis­ placed in G.Hertneck-W.Watson, Ger­ man League 1994, I would prefer 10.. .fs 11 exfs tllxfs 12 tllf3 tllf6 13 il.d3 h6 with an equal position in C.Bauer­ F.Libiszewski, Montpellier 2004. A) 9 iLgS

When faced with a move like 9 �bl Black must mainly rely on his under­ standing of the Mar del Plata in gen­ eral, as he did in E.Lobron-L.Brunner, German League 1990: 9 ... tllh s (this is very logical; instead 9 ... tlld7 10 b4 as 11 a3 'iii>h 8 12 il.e3 ! tllg 8 13 cs! axb4 14 axb4 fS lS tllg s favoured White in l.Nikolaidis-A.Jankovic, Leros 2010) 10 tt:Je1 tllf4 11 tlld 3 tllx e2+ 12 'l!Vxe2 fS and there was a more or less normal­ looking position similar to some found in Line A, below. d) 9 tllh4 looks quite extravagant, but it is playable too: 9 ... tlle 8 (after 9 ... tlld7 10 g3 fs 11 exfs tllxfs 12 tllxfs we can see some point to White's idea

This move was brought to promi­ nence by the creative Ukrainian Grand­ master Vereslav Eingom and was favoured by the Swiss Grandmaster Yannick Pelletier for a while. It does not seem logical for White to trade off his dark-squared bishop, but if Black forces the exchange then White hopes to gain time for a queen side assault. 9 tll h s Black has a decent alternative in the popular 9 ... h6. After 10 il.xf6 il.xf6 11 b4 Black has: ...

145

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 il.xb7 or the untried 14... as!? with counterplay.

a) 11 ...c6!? was recommended by Bologan. We have seen this pre­ emptive strike before and after the 12 .U.cl (12 a4 as! 13 bxas cs is comfort­ able for Black) 12 ... as 13 a3 il.g7 14 'l!Vb3 cxds l S exds!? b6 16 �fdl axb4 17 axb4 il.d7 18 bS of Y.Pelletier­ V.Bologan, Cap d'Agde 2002, Bologan suggests 18 ... g S ! ?. b) 11 ...il.g7 has been the most common choice: 12 cs (12 tlld2 fS transposes) 12 .. .fs 13 tlld2 fxe4 (this is usually played, but it scores badly; still, 13 .. .f4? is bad because of 14 il.g4 and other moves like 13 ... 'iii>h 7 look silly) 14 tlldxe4 tllfs looks like a better version of the 9 tlld 2 variation for White after lS a4, 1 S �Cl OT 1 S il.g4. c) 11 ...'iii> g 7 has always looked stur­ diest to me: 12 cs tllg 8 13 a4 (13 cxd6?! cxd6 releases the tension prematurely and 13 tlld 2 il.e7 14 c6 bxc6 lS dxc6 fS gives Black good play) 13 ... il.e7 14 c6 (after 14 cxd6 both 14 ... il.xd6! ? and 14...cxd6 lS as fS 16 tlld2 tllf6 give Black counterplay), and now Black can play 14 ... bxc6 lS dxc6 fS, 14 .. .fS l S cxb7

146

10 tlle1 White attacks the hS-knight and prepares to bring his own knight to d3. Other moves are less dangerous: a) 10 .U.el h6 11 il.d2 tllf4 12 il.xf4 exf4 gives Black a position similar to those in the Bayonet with 10 cs, but here Black is better placed. After 13 �d2 gs 14 h3 tllg 6 lS tllh 2 tlle s 16 .U.acl a6 Black had no problems in J.Gonzalez Zamora-V.Bologan, Turin Olympiad 2006. b) 10 g3 h6! (this move will prove useful because it controls gS; instead 10 .. .f6 11 il.d2 fS transposes to a line which can arise from 9 il.d2 too, but Black should avoid it as White has 12 tllg S!) 11 il.d2 (there is not much point now in exchanging the bishop with 11 il.xe7?! 'l!Vxe7) and here: bl) 11 ... il.h3 12 .U.el fS 13 tllh4 tllf6 14 exfs g s ! l S tllg 6 tllx g6 16 fxg6 il.fs 17 il.e3 'lli'e8 18 cs a6 19 .U.cl 'l!Vxg6 20 cxd6 cxd6 was fine for Black in l.Bem­ W.Watson, Gausdal 1991.

Th e M a r de/ Plata Varia tio n : White 's O t h e r N i n t h M o v e s b2) 11 .. .fs 12 tllh 4!? (after 12 exfs Black should play 12 ... tllxfs, because 13 g4? fails to 13 ...tlld4 14 gxhs tll x e2+ lS 'l!Vxe2 il.g4) 12...tllf6 13 exfs g s 14 tll g2 (14 tll g 6 tllx g6 lS fxg6 il.fs is similar to variation 'bl', but White's rook is not on el) 14 ... tllxfs l S il.d3 tlld4 was level in M.Krasenkow-W.Watson, Cappelle la Grande 1990. 10 tllf4 11 tll d3 tllxe2+ Black secures the bishop-pair. Prac­ tice has show that White keeps some edge after 11 ... tllx d3 with both 12 'l!Vxd3 and 12 il.xd3. 12 'l!Vxe2 ...

12 h6 Again this move will prove useful. Instead 12 ...f6 13 il.d2 (less incisive is 13 il.e3 fS 14 f3 f4) 13 .. .fs 14 f4! com­ pares favourably (for White!) with the note to Black's 13th move, below, be­ cause Black does not have control of gs. 13 il.d2 Instead 13 il.xe7 'l!Vxe7 is harmless, while 13 il.e3 fS 14 f3 (after 14 f4 exf4 1s tllxf4 g s 16 tllh s il.es 17 exfs tllxfs Black hits the e3-bishop - compare

with variation 'c' in the note to Black's next move) 14... g S ! (14.. .f4 lS il.f2 g s 16 cs tll g6 i s also possible) lS c s tll g 6 16 .U.fcl �f7 17 .U.C2 (better is 17 exfs il.xfs according to Pelletier) 17 ... tllf4 18 'lli'd2 tllxd3 19 'lli'xd3 g4 20 fxg4 f4 21 il.f2 il.xg4 22 cxd6 cxd6 23 .U.acl a6 24 tl:Ja4 'l!Vg s 2s tllb6 .U.af8 and Black had an excellent attacking position along with the bishop-pair in Y.Pelletier­ A.Fedorov, Plovdiv 2003. 13 gs Black clamps down on the kingside. The natural 13 ...fs leads to sharp play, but is risky after 14 f4! (instead 14 f3 is well met by 14.. .f4 or 14...c6!?) 14... exf4 1s tllxf4 gs. ...

•••

Here White has: a) 16 tlld3 f4 is fine for Black, as 17 es dxes 18 tllx es fails to 18....U.e8!. b ) 16 tlle6 il.xe6 17 dxe6 f4 (intend­ ing ... tllg6) 18 g3 tll g6 (but not 18.. .fxg3 19 hxg3 tll g 6 20 tll d s! c6 21 'l!Vh s with the initiative in A.Yermolinsky­ V.Topalov, Wijk aan Zee 1999) 19 gxf4 il.xc3! 20 bxc3 gxf4 21 'iii>h l 'lli'f6 gave Black promising counterplay in A.Blees-

147

A ttacking Chess: Th e Kin g 's I n d i a n, Vo l u m e 1 H.Klarenbeek, Heraklio 1993. c) 16 tllh s ! il.d4+ (worse is 16...il.es 17 exfs tllxfs when, instead of 18 g4 tllg 7 19 tllx g7 'iii>x g7 20 �xf8 'lli'xf8, as seen in D.Berczes-L.Valdes, Budapest 2006, White can play 18 tlle 4! il.xb2 19 il.c3 il.xc3 20 tllx c3 with a strong initia­ tive for the pawn) 17 'iii>h l f4 sees Black threatening to consolidate with ...tllg 6.

Thus White must act quickly with 18 g3! and here: cl) 18 ... il.h 3 19 gxf4! looks promis­ ing: 19 ...il.xfl 20 �xfl tllg 6 21 es! (after 21 'lli'd3 Black should avoid 21 ...il.h8 22 es 'lli'e8 23 e6 as in M.Galyas-S.Husari, Balatonlelle 2003, and play 21...il.xc3! 22 il.xc3 tllxf4 23 tllxf4 �xf4 24 �xf4 gxf4 2s 'lli'd4 'l!Vgs 26 'lli'h8+ 'iii>f7 when White does not have more than a per­ petual check with 27 'lli'h 7+ because 27 'l!Vxa8? gets mated after 27.. .f3) 21...dxes 22 'lli' d 3! tllxf4 (White wins after 22 ...'iii>h 7 23 fxg s hxgs 24 �fS! !Ixfs 2s 'lli'xfs with a decisive attack) 23 l:.xf4 (not 23 il.xf4 'lli'e 8!, turning the tables) 23 ...exf4 24 'lli'xd4 'l!Ve7 2S tlle4 and White has a huge advantage.

148

c2) 1 9. ..fxg3 1 9 �xf8+ 'lli'xf8 20 tllb s! (after 20 �fl iLh3 ! 21 �xf8+ !Ixf8 Black will win back the queen with ...�f2, as in the game G.Prakash-P.Konguvel, Nagpur 1999) 20 ...'lli'f2 21 'lli'xf2 il.xf2 22 tllx c7! (worse is 22 hxg3 il.b6 23 il.c3 a6 24 tlld4 'iii>f7 2S 'iii> g 2 'iii> g 6 26 �hl il.d7 when Black was at least equal in G.Prakash-P.Konguvel, Chennai 2000) 22 ... il.h3 23 hxg3 �c8 was P. Lukacs­ V.Rajlich, Budapest 2001, and now 24 tllb s �xc4 2S 'iii>h 2! would give White a comfortable plus. 14 g4 White blocks the kingside. Instead 14 f3 fS gives Black counterplay, while 14 h4?! g4 lS f4 gxf3 16 'lli'xf3 fS gave him the initiative in L.Oll-A.Shirov, Til­ burg 1992. 14...tll g6 The knight heads for f4. 15 f3 tllf4 16 tllxf4 White could also try 16 il.xf4!? exf4 17 .U.acl with the idea of opening the c­ file, but Black's strong bishops should certainly hold the balance. 16...exf4

The M a r de/ Plata Va ria t i o n : White 's O t h e r N i n th M o ves Black is already quite comfortable with the bishop-pair and th e possibility of opening the h-file with ... hs. 17 �fdl More solid is 17 tlldl with the idea of exch anging Black's dark-squared bishop immediately. However, after 17 ...c6 18 il.c3 cxds 19 il.xg7 'iii> x g7 20 c:xds il.d7 21 tllf2 �c8 22 'Ii' d2 'lli'b6 Black was still very comfortable in L.Oll­ S.Dolmatov, Rostov 1993. 11 hs! 18 h3 il.es 19 il.e1 Safer was 19 'iii>g2 with the idea of defending with �hl at some point. 19. 'iii>g7 Strengthening the long diagonal with 19 ...'lli'f6 was also possible. 20 J:td3?! �h8 21 tlld 1 hxg4 22 fxg4 After 22 hxg4 both 22 ...'lli'f6 and 22 ... �h3 with the idea of ... 'lli'h 8 are very good for Black. 22 iff6 23 il.c3 il.d7 24 tllf2 �ae8 Black had a clear advantage in V.Pelletier-T.Radjabov, Biel 2006. ...

..

...

days, but it is not completely harmless. White prepares to bring his rook to cl and remains very flexible. 9 ...tlle8 Black can also play 9 ...tlld7 when 10 �cl fS 11 tll g s tllf6 leads us back to the main line. However, White has an extra option in 10 b4 fS 11 tllgs tllf6 12 f3, so 9 ... tll e8 seems more accurate. Instead 9 ... tllh s was for a long time considered the main line, but after 10 g3 fS 11 tllg s ! (more common is 11 exfs tllxfs 12 tlle4, but Black is fine after 12 ...tllf6) 11 ...tllf6 12 f3 White has scored well. This position is similar to the Bayonet with 12 f3, but here White has played il.d2 instead of b4. This may not favour him, but it could well intro­ duce a new world of complications that is not really worth delving into consid­ ering how rare 9 il.d2 is nowadays. 1o �c1 fs

B) 9 il.d2

This old move is rarely seen nowa-

After 10 tll el fS 11 tlld3 tllf6 we have transposed back to Chapter 1. In­ stead 10 b4 fS 11 'lli'b 3 (11 tll g s makes less sense here because the e6-square is defended) 11 ...tllf6 12 exfs should be

149

A ttacking Chess: Th e King 's I n dian, Vo l u m e 1 met by 12 ... tllx fs! when 13 tllg s? runs into 13 ...tlld4. 11 tll gs Others: a) 11 exfs gxfs (11 ...tllxfs! ?) 12 tll g s h6 13 tlle6 il.xe6 14 dxe6 'lli'c8 lS 'l!Vb3 c6 16 il.hs 'l!Vxe6 17 'l!Vxb7 tllf6 18 il.e2 �fb8 19 'lli'a6 .U.xb2 gave Black good play in M.Taimanov-R.Fischer, 1st matchgame, Vancouver 1971. b) 11 'l!Vb3 b6!? 12 exfs gxfs 13 tll g s tl\f6 (after 13 ...h6 14 tll e6 Si.xe6 lS dxe6 'lli'c8 16 tllds 'l!Vxe6 17 tllxe7+ 'l!Vxe7 18 cs+ 'iii>h8 19 cxd6 tllxd6 20 .U.c6 White had decent compensation for the pawn in M.Taimanov-Ma.Tseitlin, USSR 1973) 14 f4 h6 1s fxes dxes 16 cs tllfxds 17 tllxds tll x ds 18 cxb6 axb6 19 .U.c6 'iii>h8 20 tl\f3? (better is 20 'lli' h3 when Black can play either 20.. J:tf6! ? or 20 ... tllf6 21 il.c3 f4! ? 22 'lli'h4 il.b7) 20 ... il.b7 21 �g6 tl\f4 22 Si.xf4 exf4 23 .U.dl 'lli'e7 gave Black a big advantage in M.Taimanov­ R.Fischer, 3rd matchgame, Vancouver 1971. 11...tllf6

This looks better than 11 ... h6 12

150

tlle6 Si.xe6 13 dxe6 'Ii'c8 14 'l!Vb3 c6, as in L.Ftacnik-Kr.Georgiev, Groningen 1976. Here lS cs!? would give White the initiative. After the text, again we have a posi­ tion resembling the Bayonet. Here White has played the rather slow il.d2 and l:tcl, though, instead of b4 and l:tel. 12 f3 12 exfs should be met by 12 ... tllxfs!, heading for the d4-square. 12 ...c6 13 'l!Vb3 h6 14 tll e6 il.xe6 15 dxe6 'lli'c8 16 .il.e3 '1li'xe6 This is fine, but Black has decent al­ ternatives in 16 ... .U.d8 and 16 .. .f4. 17 l:tfdl Instead 17 'l!Vxb7 is met by 17 ... .U.fb8 18 'lli'a6 �xb2 with an unclear position. 11 ... 'iii>h7 18 'l!Vxb7 .U.fb8 19 'lli'a6

19 .. .fxe4! An important move. Here 19 .. ..i:!xb2 is bad because of 20 'l!Va3! which also hits the d6-pawn. Likewise, 19 ... hS?! 20 cs ds 21 exds tllfxds 22 tllx ds tllxds 23 il.c4 gave White the upper hand in J.Plachetka-V.Babula, Stare Mesto 1992.

Th e M a r de/ Pla ta Varia tio n : Wh i t e's O t h e r N i n t h M o ves 2o fxe4 If 20 tllxe4 tllxe4 21 fxe4 then 21 ... �xb2 hits the bishop on e2. 20...�xb2 21 'l!Va3 �xe2 22 tllxe2 Not 22 �xd6 .U.xg2+! 23 'iii>xg2 'lli'g4+ 24 'iii>f2 �f8 when Black has a strong attack. 22 ...tllxe4 With two pawn for the exchange and squares for his pieces, Black has a good position. C) 9 a4

This move is rare but is hardly bad. Moreover, it has certain similarities with and may even transpose to 9 tll e l tt:ld7 10 il.e3 fs 11 f3 f4 12 il.f2 gs 13 a4 (Line E of Chapter 3). 9 .as This is a sound reply. Black can also play 9 ...tllh s as he would against 9 b4. White chooses between 10 as and 10 g3 with rather unexplored play. 10 tlle1 Instead 10 b3 looks slow, but White can still play this way: 10 ...tlld 7 (10...tllh s ! ?) 11 il.a3 il.h6 (worse is ..

11 ... fs 12 tll g s tllf6 13 cs!, but 11 ...b6 and 11 ...tllcs are sensible alternatives) 12 b4 axb4 13 il.xb4 fS was M.Carlsen­ A.Morozevich, Biel 2006, and Black was a tempo up on a line of the Bayonet (9 b4 as 10 il.a3 axb4 11 il.xb4 b6). 10...tll d 7 11 tll d 3 This is more dangerous than 11 il.e3 fS 12 f3 when Black can keep the ten­ sion with 12 ...tllc s!? (instead 12 .. .f4 13 il.f2 gs leads to the main lines of Chap­ ter 3) 13 tlld3 b6 and now: a) 14 b4 tllxd3 lS 'lli'xd3 axb4 16 tllb s 'iii>h8 17 'l!Vb3 (after 17 il.d2 Black can consider 17 .. .fxe4 18 fxe4 .U.xfl+ 19 il.xfl tl:Jg8 20 il.xb4 tl:Jf6 21 aS il.h6 OT 17... cs!? 18 dxc6 tll x c6) 17 ... tll g 8 18 'l!Vxb4 and here 18 ... tllf6 19 exfs gxfs 20 il.gs h6 21 il.h4 il.d7 22 .U.a3 'l!Vb8 23 il.f2 .U.g8 24 �bl tllh s 2s 'iii>h l 'l!Vd8 gave Black counterplay in V.Korchnoi­ G.Kasparov, Barcelona 1989, but later Kasparov preferred 18 .. .fxe4! 19 fxe4 �xfl+ 20 �xfl il.h6 21 il.xh6 tllxh6 22 'l!Vd2 'iii> g 7 when Black is certainly bet­ ter. b) 14 �a3 ! ? 'iii>h 8 lS 'lli'd2 f4 16 il.f2 h S (after 16 ... g s Black must consider 17 g4! ?) 17 tllb s gs 18 b4 axb4 19 'l!Vxb4 g4 20 �fal tll g 6 21 'lli'el (or 21 as �xas 22 �xas bxas 23 �xas? gxf3 24 gxf3 tllx d3 2S il.xd3 iLh3 when White cannot de­ fend against 26 ... 'l!Vgs+), and now 21 ...tllb7! ? was unusual but held up White long enough on the queenside for Black to develop good kingside play in A.Evdokimov-D.Jakovenko, Russian Team Championship 2010.

151

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 11...fs 12 f3 After 12 exfs tllxfs 13 tlle4 tllf6 Black is fine.

12 ...f4 This is rather committal, but it is thematic and logical. Instead 12 ...b6 13 il.d2 tllf6 14 b4 axb4 lS tllxb4 il.d7 16 tlld3 cs!? 17 �a3 �f7 was unclear in A.Evdokimov-L.Nisipeanu, European Championship, Budva 2009. Another possibility is 12 ...'iii> h8 13 il.e3 b6 14 b4 (if 14 'l!Vb1 il.a6! ? then lS b3 tllg 8 is equal according to Bologan and lS tll b s can be met by the reply 1 s ... tllf6, pressuring e4) 14 ... axb4 l S tllxb4 tllf6 16 tlld3 tlleg8!? 17 tllf2 tllh s 18 �el il.h6 19 il.xh6 tllx h6 20 iLfl f4 21 tllb s g s gave Black good counterplay in S.Khmelevsky-G.Beckhuis, Vienna 2006. 13 b3! This is the point of White's play - he will be able to open the queenside. This may look slow, but it is the right plan and if we compare it to Line E of Chap­ ter 3, White is actually getting his bishop to b4 more quickly by playing

152

b3, il.a3, b4 and il.xb4, rather than il.e3-f2-e1, b4 and il.xb4. 13 ...gs 14 il.a3 b6 Instead 14...tllf 6?! allows the imme­ diate lS cs, as in S.Atalik-V.Bologan, Manila Olympiad 1992. 15 b4 axb4 16 il.xb4 tllc s 17 as

This is a critical position. Black is too far behind to just play for mate: 17 ... il.d7?! 18 tllb s gave White the ini­ tiative in V.Popov-F.Amonatov, Sochi 2007, and 17 ... il.a6 18 axb6 (better than 18 tllbs c6!) 18 ... cxb6 19 tllx cs bxcs 20 il.as 'lli'd7 21 'lli'a4 looks better for White as well. Black's best looks like 17 ...c6!? 18 axb6 (instead 18 tll a4 tllxd3 19 'lli'xd3 bxas 20 il.c3 cxds 21 cxds il.a6!? 22 'lli'd2 il.xe2 23 'l!Vxe2 'lli'e7 24 �fbl �fb8 was fine for Black in A.Klimov­ E.Bondarenko, Tomsk 2008) 18 ... tll x d3! (but not 18 ....U.xal? 19 'l!Vxal tllxd3 20 il.as!) 19 �xa8 (19 il.xd3? loses to 19 ...�xal 20 'l!Vxal 'l!Vxb6+) 19 ... 'l!Vxb6+ 20 cs! tllxcs 21 il.as 'lli'b7 22 �xc8 .U.xc8 23 dxc6 tllxc6, which looks okay for him.

Th e M a r de/ Pla ta Va ria t i o n : Wh ite 's O t h e r N i n t h M o ves D ) 9 �h1

Tony Miles's old favourite looks pointless, but it is not completely with­ out merit. 9 tlle 8! One point of White's play is that 9 ... tllh s can be met by 10 tllg 1 tllf4 11 l.f3, although i t is not clear that this is so dangerous. Instead 9...�h8 can be met by 10 a4 or 10 tll e l, when it is hardly clear which player the king moves will favour. As I found out for myself, 9 ... tlld7 is less accurate than 9 ...tlle 8, though, be­ cause it allows 10 g4!?. With this move White discourages Black from opening the kingside, after which he may tum his attention back to the queen side. a) 10...fs 11 gxfs gxfs 12 exfs (White can use both the e4-square and the g-file) 12 ...tllf6 13 �gl tllxfs 14 il.d3 �h8 lS tllg s left White with an edge in R.Kaufman-D.Vigorito, Chicago 2008. b) 10...�h8 11 �gl! as 12 il.e3 tll cs gives White an active version of some similar plans we have seen with the g4advance. ...

A few possibilities: bl) 13 a3 ! ? looks promising. White will play b4 next unless Black sacrifices a pawn with 13 ...a4 14 il.xcs dxcs lS tllxa4. b2) 13 l:tcl is a bit slow as the rook does not have much purpose on this square:

13 ... il.d7 14 b3 tllg 8 lS tlld2 f6 (Black intends ... il.h6) 16 h4 fS! (now that White has weakened his kingside, Black tries to open things up) 17 gs tllxe4 18 tlldxe4 fxe4 19 tllxe4 tl\e7 20 il.g4 tllfs gave Black good counterplay in M.Carlsen-T.Radjabov, Biel 2006. b3) 13 'lli'd2 (with the idea of �g3

153

A ttacking Chess: Th e King 's I n d i a n, Vo l u m e 1 and .U.agl) 13 ...tll g 8 14 l:tg3 b6 lS l:tag l tllf6 16 'l!Vc2 il.d7 17 tlld2 c6 18 f3 gave White an edge in M.Tosic-A.Kolev, Vrnjacka Banja 1990. b4) 13 tlle l fS 14 f3 fxe4 lS fxe4 c6 16 tlld3 cxds 17 cxds b6 18 tllxcs bxcs 19 �d2 �b8 20 !Iafl and again White could claim some advantage in A.Miles­ F.Gheorghiu, Cran s Montana 2001. 10 tlle1 Instead White can also play 10 a4 as 11 tlle l fS 12 tll d3 tllf6 13 f3 b6, which is similar to Line C, although here White has played 'iii>h l rather prema­ turely. It looks a bit late for 10 l:tgl because Black can play 10...fs 11 exfs when 11 ...gxfs 12 tll g s tllf6 has been played with some success, but I prefer 11...tllxfS ! when White's kingside shuf­ fling looks rather silly. 10 ... fs 11 exfs tllxfs! Again I prefer activating the knight rather than allow 11 ...gxfs 12 f4!?.

12 tllf3 Covering the d4-square, but this is obviously time-consuming. White's

1 54

king move looks out of place here. White can also play 12 il.d3 tllf6 13 tllf3 which transposes to the main line. Instead 12 tllc2 tllf6 13 g4? just weak­ ens the kingside too much. After 13 ...tll d4 14 h3 tllx g4! ? lS hxg4 'lli'h4+ 16 'iii>g l tllx e2+ 17 'l!Vxe2 il.xg4 18 f3 a draw was agreed in A.Miles­ D.Gormally, British Championship, Hove 1997, but Black should really have played on with 18 ... il.xf3! 19 .U.xf3 'lli'g4+ 20 'l!Vg2 'lli'xf3 21 'lli'xf3 .U.xf3 22 'iii>g 2 l:taf8 when he is certainly better. Here a clearer refutation is available even sooner with 14 ... tllx c2! lS 'l!Vxc2 il.xg4 16 il.xg4 (White gets mated after 16 hxg4 tllxg4 17 il.xg4 'lli'h4+ 18 'iii>g2 'l!Vxg4+ 19 'iii>h2 .U.f3) 16 ... tllx g4 with a winning position. 12 ...tllf6 13 il.d3 tlld4 14 tllxd4 exd4 15 tlle4 tll xe4 16 il.xe4

16...'lli' h 41 This is more aggressive than 16... il.d7 17 'l!Vd3 cs 18 dxc6 bxc6 19 il.d2 .U.b8, which was also satisfactory for Black in A.Miles-A.Beliavsky, Nova Gorica 1999.

Th e M a r de/ Plata Varia tio n : White 's O t h e r N i n t h M o ves 17 'l!Vc2 il.fs! Black develops very quickly. 18 il.xfs �xfs 19 'iii>g1 This is a bit humbling. White 'takes back' his 9th move. However, he has trouble developing his queenside and a good alternative is hard to find, be­ cause 19 il.d2 loses to 19 ...�xf2. 19 �es Black has a fantastic position. All of his major pieces are active, he has a passed d-pawn and White has no ac­ tive possibilities.

This is forced, because 24 'lli'fl loses to 24 ... il.xf2+ 25 'lli'xf2 .U.el+.

••.

20 a4 il.es Black could also play 20...c5!?. 21 h3 d3!? This forces the pace. Again 21...c5 was possible too. 22 'l!Vxd3 .U.xf2 23 �xf2 il.d4 24 g3

Now after 24 ... 'lli'xh3 ? White man­ aged to hold on with 25 il.f4 il.xf2+ 26 'iii>xf2 'lli'h2+ 27 'iii>t3 'l!Vh5+ 28 'iii>g 2 .U.e2+ 29 'iii>fl .U.xb2 30 'lli'e4! (this covers the hl-square and prepares to check the black king) 30 ... 'lli'h 3+ 31 'iii> el �g2 32 'lli'e8+ 'iii>g 7 3 3 'lli'e7+ 'iii> g 8 34 'lli'e8+ and V2-V2 in A.Miles-H.Tirard, Cappelle la Grande 2000. However, Black could have crowned his excellent play with 24 ...'lli'f6! 25 'lli'fl (or 25 il.f4 il.xf2+ 26 'iii>xf2 'l!Vxb2+) 25 ... 'lli'f3 !, paralyzing White and threat­ ening ...�e2: for example, 26 il.d2 �e2 27 il.el 'l!Vxg3+ 28 'l!Vg2 (28 'iii>h l il.xf2 29 'l!Vxe2 'l!Vg1 mate) 28 ... .U.xel+ 29 J:txel il.xf2+.

155

Chapter 8 The Mar del Plata Va riation White's Eighth Move Deviations

1 d4 tl:Jf6 2 C4 g6 3 tl:Jc3 J.g7 4 e4 d6 5 tl:Jf3 0-0 6 J.e2 es 7 0-0 tllc6

cause it i s not easy to create counter­ play unless one is familiar with certain ideas. A: 8 dxes B: 8 J.e3 .

Usually when White plays 7 o-o he is willing to enter into a sharp fight. However, that is not always the case and in this chapter we look at moves other than 8 dS. It seems that with the moves 8 dxes and 8 J.e3 White is changing his mind about what kind of game he wants and these variations are certainly less popular than 7 dxes and 7 J.e3. While the lines considered here are not terribly dangerous, they should not be ignored, especially be-

156

Note th at 8 J.g 5 is not very good because of 8... tllxd4 9 tllxd4 exd4 10 'lli'xd4 tllxe4! 11 J.xd8 J.xd4 12 J.xc7 J.xc3 ! (this is better than 12 ... tllxc3 13 bxc3 J.xc3 14 �abl which is just equal) 13 bxc3 J.e6 14 J.f3 l:tac8! when Black is much better. A) S dxes White plays a kind of delayed ex­ change variation. This should not trou­ ble Black, but it is not as innocent as one might think and Black should not go to sleep just yet. 8 dxes 9 J.gs Instead 9 'lli'xd8 tllxd8 is solid, but Black can also just play 9...�xd8 10 J.gs l:tf8! (White may keep a pull after ...

Th e M a r de/ Plata Varia tio n : W h i t e's Eig h t h M o ve D e viatio n s 1 1...l:te8 1 2 tlld s tllx ds 1 3 cxds tlld4 14 ..'!Jxd4 exd4 lS il.d3 or 11 ...�d7 12 .i..d l!) 11 �fdl il.g4 which actually transposes back into the main line. 9 'lli'xdl Black has an interesting alternative in 9 ... il.g4!? which immediately fights for the d4-square. Then 10 'lli' x d8 and: a) 10...�fxd8 11 tlld s tll x ds 12 exds f6 13 dxc6 fxgs 14 cxb7 l:tab8 lS h3 .i..fs was l.Rausis-1.Nataf, Enghien les Bains 1997, and here 16 cs! would give White some initiative. b) 10 ... �axd8 11 il.xf6 (better is 11 ..'!::l d s tllxe4 12 il.xd8 �xd8 13 �fdl when Black probably lacks sufficient compensation for the exchange) 11 ... il.xf6 12 tlld s il.xf3 13 il.xf3 il.g s was fine for Black in R.Ponomariov­ V.Topalov, Vitoria Gasteiz 2007; 14 ..'!::lxe7 l:td2 gives Black enough activity for the pawn. c) 10...tllxd8 11 il.xf6 il.xf6 12 tlld s .i..xf3 13 gxf3 ! (after 13 tllxf6+?! 'iii> g 7 14 .i..xf3 'iii>xf6, with the idea of ...tllc 6-d4, Black is already better, while 13 il.xf3 �g s 14 tllx c7 l:tc8 is certainly not worse for the second player) 13 ... il.g s 14 tllxe7 .:.c8 lS tll d s tlle6 gives Black some compensation, but it is a bit specula­ tive - White's extra pawn is 'real' on the queenside. 10 .U.fxd1 White can also play 10 �axdl, but then the fl-rook would be a bit stuck. 10 il.g4 Instead 10 ... h6 11 il.e3 il.g4 looks fine too. ...

...

11 h3 Black is certainly not troubled by 11 �acl h6 (11 ... tlld4!?) 12 il.e3 .U.fd8, but 11 �d3 is interesting. Black could try 11 ... hf3 12 il.xf3 tlld4 13 tlld s tllxds 14 cxds fS or even the immediate 11 ... tllxe4!? 12 tllxe4 fS, but practice has focused on 11 ... h6 12 il.e3 tllxe4! (White is a little better after 12 ...�fd8 13 l:tadl �xd3 14 �xd3) 13 tllxe4 fS 14 tll c s e4 lS �b3 and now: a) 1S ...f4 16 il.d4 (Black is certainly okay after 16 tllxe4 fxe3 17 l:txe3 il.xf3 18 il.xf3 il.xb2 19 l:tbl il.d4) 16 ... tllxd4 17 tllxd4 il.xd4 18 il.xg4 il.xcs 19 �xb7 il.d6! (but not 19 ... il.b6 20 b4!) is equal but unbalanced. b) 1S ... exf3 16 gxf3 f4 17 fxg4 fxe3 18 fxe3 tlld4!? 19 exd4 il.xd4+ 20 'iii>hl il.xcs 21 �fl?! �xfl+ 22 il.xfl b6 was drawish but still a little bit better for Black in B.ltkis-M.Golubev, Sovata 2000. However, 21 �xb7 il.b6 22 cs! is a bet­ ter try, so I would prefer to deviate ear­ lier with 18 ... tlla s 19 �bs b6 or just 18 ... �ae8. In both cases Black has good compensation for a pawn.

157

A tt a c k i ng Chess: Th e King 's Indian, Vo l u m e 1 11...il.xf3 12 il.xf3 tlld4 13 tlld s tllx ds 14 cxds

Even though the d4-knight looks nice, White has the bishop-pair and potential pressure on the c-file, so Black must act quickly. 14 ...fs Worse is 14...c6 1 5 dxc6 bxc6 16 .U.acl �fb8 17 b3 as 18 l:tc4 when White has some pressure. 15 .U.acl .U.f7 16 il.e3 After 16 'iii>fl Black should avoid 16 ...f4?! 17 il.g4! and play 16 ... il.f8 or 16 ...l:taf8. 16 ...fxe4 17 il.g4 Wisely avoiding 17 il.xe4? tlle 2+ and 17 il.xd4?! exf3 18 il.e3 fxg 2. After 17 il.g4 tllfs 18 l:tc3 hs 19 il.e2 il.f8 20 l:tc4 tlld6 21 l:tc3 tllfs 22 l:tc4 tlld6 23 l:tc3 tllfs the game was drawn in l. Krush­ A.Melekhina, US Women's Champion­ ship, Saint Louis 2009. 8) 8 .il.e3 White tries to keep the tension in the centre. Black has the traditional equalizer 8 ... �e8 or he can play along

158

the lines of the Gligoric Variation with the sharper 8 ... tllg4.

81: 8...�e8 82: 8 tllg4 •..

81) 8 ...l:te8

This move has long been known to equalize. 9 dxes The point of Black's rook move is seen after 9 dS tlld 4! 10 tllxd4 exd4 11 il.xd4 tllxe4 12 il.xg 7 'iii>x g7 13 tll xe4 (or 13 'lli'd4+ 'lli'f6) 13 ...�xe4 when the posi­ tion is equal, although Black does have slightly the better long-term chances

The M a r de/ Plata Varia tio n : W h i te's Eig h t h M o ve D e viations because of his better bishop. 9 dxes ...

1o 'lli'xd8 This is not very ambitious, of course, but other moves do not give White much either: a) 10 cs it.g4! (fighting for the d4square) 11 .tbs 'l!Vc8 (Smirin's idea; 11 ... tlld7 12 it.xc6 bxc6 13 h3 it.xf3 14 ·tvxf3 'l!Vb8 is another option) and now: al) 12 h3 �d8 (it is probably better to just play 12 ...it.hS! 13 it.xc6 bxc6, transposing to variation 'a2') 13 tlld s !Dxds 14 exds it.e6 (instead 14...it.xf3 lS 'lli'xf3 tlld4 16 it.xd4 exd4 17 �fel ·tvfs 18 'lli'xfs gxfs 19 it.c4 it.f8 20 b4 as 21 a3 d3 is possible, but Black is really just playing to draw here) lS dxe6! :xdl 16 exf7+ 'iii>xf7 17 it.c4+ �8 18 :taxdl gave White good compensation for the queen in P.Lukacs-A.Stummer, Budapest 1992. a2) 12 it.xc6 bxc6 13 h3 it.hs 14 it.gs (otherwise, there is 14 g4? tll x g4 and 14 'l!Ve2 h6 1s 'iii>h2 gs 16 'lli'c4 �b8 17 b3 as 18 tlld2 g4 19 f3 gxh3 20 gxh3 .U.b4 was unclear in P.Van der Sterren-

l.Smirin, Tilburg 1992) 14 ... .U.b8 lS .U.bl it.xf3 16 'lli'xf3 tlld7 (with the idea of ...tlld 7-e6) 17 b4 as 18 bS (18 a3 axb4 19 axb4 tllf8) 18 ...tllx cs 19 bxc6 tlle6 gave Black counterplay in L.Oll-1.Smirin, Rostov on Don 1993. b) 10 h3 prevents ...it.g4, but costs White time. After 10 ... it.e6 11 cs Black has:

bl) 11 ...tllh s 12 tllg s (12 .tbs ! ?) 12 ...tllf4 13 tllxe6 tllxe6 14 .tbs .U.f8!? (14...'lli'xdl lS �fxdl .U.ed8 16 it.xc6 bxc6 is level) lS it.xc6 bxc6 16 'lli'a4 'lli'e 8 17 .U.adl fS 18 exfs (after 18 f3 f4 19 it.f2 gs Black has counterplay on the kingside as well) 18 ...gxfs 19 f3 'iii>h 8 20 'iii>h l .U.g8 21 .U.d2 it.f6 and Black had an active position in P.Van der Sterren­ B.Gelfand, Biel lnterzonal 1993. b2) 11...a6 kind of mimics White's 10th move by preventing any .tbs ideas: 12 'lli'a4 'lli'e7 13 .U.fdl .U.ad8 14 it.c4 tlld4! lS tllxd4 (Black has no prob­ lems after lS it.xe6 tllxf3+ 16 gxf3 'l!Vxe6 17 'iii>g 2 c6) 1s ... exd4 16 �xd4 was Slovenian Team 1.Jelen-M.Tratar, Championship 1993. Here Black should

159

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 play 16 ... ll'ld7! 17 ll'lds ii.xds 18 .l:f.xds c6 19 l:tdd1 ll'lxcs 20 'ti'b4 ii.f8 with a good position. b3) 11 ... 'ti'e7!? 12 ii.bs (this seems logical as Black has omitted ... a6, but Black still obtains satisfactory play; in­ stead 12 'ti'c2 .l:f.ad8 13 .l:f.ad1 was J.Speelman-J.Nunn, Hastings 1987/88, and now Black should play 13 ... a6 ac­ cording to Speelman) 12 ....l:f.ed8 13 'ti'a4 ll'ld4! 14 ll'lxd4 (14 ll'lxes a6! 1s ii.xd4 axbs 16 't\Vb4 ll'lxe4! with the idea of 17 ll'lxe4 .l:f.a4! is good for Black) 14... exd4 1S ii.xd4 c6 16 ii.e2 ll'lxe4! 17 ii.xg7 ll'lxcs 18 't\Va3 ..t>xg7 (Black has won a pawn, but he will have to untangle his pieces) 19 ll'le4 b6 20 .l:f.fc1 ii.xh 3 ! ? (tempting, but 20....l:f.dS is simpler and preparing for the opening of the b-file with 20....l:f.ab8!? is also possible) 21 'ti'c3+ f6 22 ll'lxcs ii.xg2! 23 .l:f.e1! ii.ds and Black had good compensation for the piece in D.Berczes-A.Jankovic, Sara­ jevo 2010. 10...ll'lxdS

Instead 10 ... .l:f.xd8 11 .l:f.fd1 (after 11 ii.gs .l:f.f8! 12 .l:f.fd1 ii.g4 we're back in

160

Line A) 11 ... ii.g4 12 ll'lds ll'lxe4 13 ll'lxe7 .l:f.xd1+ 14 l:txd1 .l:f.c8 1s ll'lds ll'lf6 16 ll'lxf6+ ii.xf6 17 h3 ii.e6 18 b3 was M.Dlugy-A.Fishbein, New York 1991, and here 18 ...l:td8 19 .l:f.xd8+ ii.xd8 looks fairly level. 11 ll'lbs After 11 ll'lds ll'le6 12 ll'lgs ll'lf4! ? (12 ... ll'lxds 13 cxds ll'ld4 is also okay, but not 13 ... ll'lxgs?! 14 ii.xgs with an edge for White) 13 ii.xf4?! ll'lxds 14 ii.d2 ll'lf6 1s f3 ll'lhs 16 .l:f.fd1 ll'lf4 11 ii.fl ii.f6 18 ll'lh3 ii.xh3 19 gxh3 ll'le6 Black was winning in Z.Mijailovic­ V.Kotronias, Vmjacka Banja 2006. 11 ... ll'le6 12 ll'lgs l:te7 13 .l:f.fd1 Alternatively: a) 13 ll'lxe6 ii.xe6 14 f3 c6 (14 ...b6!?) 1S ll'lc3 (1S ll'lxa7 looks risky but could be more testing) 1s ... .l:f.d7 16 .l:f.fd1 ii.f8 was fine for Black in S.Reshevsky­ R.Fischer, Santa Monica 1966. b) 13 ll'lxa7 ll'lf4 (or 13 ...ll'ld4 14 ii.xd4 exd4 1S ll'lxc8 .l:f.xc8 16 f3 ll'ld7 with compensation) 14 ii.xf4 (14 ii.f3 ii.d7 1S ll'lbs h6 is pleasant for Black) 14 ...exf4 1S ll'lxc8 .l:f.xc8 16 f3 ll'ld7 gives Black reasonable compensation for the pawn. 13 ... b6 13 ... c6 is also possible. 14 a4 Instead 14 ll'lxe6 ii.xe6 1S f3 c6 16 ll'lc3 ii.f8 is fine for Black, while 14 cs ll'lxcs 1 S .l:f.d8+ ii.f8 16 ll'lxa7 nxa7 17 .l:f.xc8 is probably best met by 17 ....l:f.e8 (17 ... ..t>g7 18 f3 ll'le8 19 a3 ll'ld6 20 .l:f.d8 h6 2 1 ll'lh3 ll'le6 was also okay for Black

The M a r de/ Pla ta Varia t i o n : White 's E ig h t h M o ve Deviatio ns in B.Larsen-R. Fischer, Monaco 1967, but White could have tried 18 ii.xcs!?) 18 .l:f.xe8 ll'lxe8 19 ll'lf3 f6 20 ii.c4+ was agreed drawn in this even position in J.Piket-J.Nunn, Wijk aan Zee 1991. 14. ..c6 1s ll'ld6 After 1S ll'lxe6 ii.xe6 16 ll'lc3 a clever idea is 16 ... :tb7!. Black frees the f8bishop and discourages White from playing a4-aS because then Black would get play along the b-file. After 17 b4 ii.f8 18 .l:f.ab1 ll'ld7 (with the idea of ... as!) 19 bs .l:f.c8 20 ll'lds ll'lcs 21 bxc6 :txc6 22 as bxas 23 ll'lf6+ ..t>h8 24 J:txb7 lbxb7 2S ll'ld7 ii.d6 White faced an up­ hill battle to draw in L.Portisch-J.Nunn, Am sterdam 1990.

1s ...ll'ld4 This works tactically and secures Black equal chances. Also possible is 1s ... ll'lf4!? 16 ii.fl h6 17 ll'lgxf7 l:txf7 18 ll'lxf7 ..t>xf7 19 as bxas with unclear play in L.Gonda-V.Kotronias, Hungarian League 2007. 16 ii.xd4 exd4 17 as Or 17 .l:f.xd4 ll'le8 18 ll'lxc8 .l:f.xc8 19 .U.d2 h6 20 ll'lf3 .l:f.xe4 21 l:tad1 ..t>f8 V2-V2,

D.Polajzer-M.Tratar, Slovenian Team Championship 199S. 11 .. l:tbS 18 axb6 axb6 19 l:txd4 ll'le8 20 .l:f.ddl 20 ll'lxc8 is more prudent. 20 ... ll'lxd6 21 l:txd6 ii.b7 22 .l:f.a2 h6 23 ll'lf3 l:txe4 Here Black's bishops gave him an edge in L.Gonda-J.Gallagher, Olbia 2008. 82) 8 ...ll'lg4

This is a more enterprising move. 9 ii.gs f6 10 ii.cl In stead 10 ii.h4 is sometimes played, but this combines the negative aspects of the Gligoric Variation - the bishop could be stranded on the king­ side - while giving up the positive as­ pect - White's flexibility with his king. Play will transpose to less critical lines in Ch apters 9 and 10 after 10... gs (10...'it'e8 and 10...ll'lh6 are also possible and could transpose too to the Gligoric) 11 ii.g 3 ltJ h6. Now Black threatens ... g4 and ...ll'lxd4, so White must react:

161

A ttacking C h e s s : Th e King 's I n dian, Vol u m e 1

a) 12 dS Ci:Je7 13 Ci:Jd2 and now both 13 ... Ci:Jg6 and 13 ...fs are covered in note 'c' to White's 12th move in Line B11 of Chapter 9. b) 12 dxes dxes also transposes to the Gligoric after 13 cs, which is con­ sidered in note 'a' to White's 12th move in Line A of Chapter 10, or 13 h3, which is note 'b' to White's 12th move in Line B of Chapter 10. 10 fs This is consistent. Black plays the­ matically and hopes to show that White has been wasting time with his bishop moves. I prefer this to the open 10...exd4 and the unprovoked 10...Ci:Jh6.

11 ii.g s The bishop continues to bounce around. Other moves also do not lead to much: a) 11 exfs exd4!? (11 ...ii.xfs 12 dS Ci:Jd4 13 Ci:Jxd4 exd4 14 ii.xg4 dxc3 1S ii.xfs :xfs 16 bxc3 ii.xc3 17 .l:f.b1 b6 is fine too for Black) 12 Ci:Jbs (after 12 ii.gs 'it'd7 13 Ci:Jbs I like 13 ....l:f.xfs! when White has little better than 14 ii.cl .l:f.f8, which is certainly fine for Black) 12 ... .l:f.xfs 13 h3 Ci:Jges 14 Ci:Jfxd4 l:tf7 (or 14 ... Ci:Jxd4 1S Ci:Jxd4 .l:f.f8) 1S ii.e3 a6 16 Ci:Jxc6 bxc6 17 Ci:Jd4 cs 18 Ci:Jc2 ii.b7 with good counterplay in Bu Xiangzhi­ T.L.Petrosian, Tiayuan 20os. b) 11 dS Ci:Je7 (Black can also play 11...Ci:Jd4!? 12 Ci:Jxd4 exd4 13 Ci:Jbs Ci:Jes 14 exfs d3 transposing to the Gligoric!) 12 Ci:Jgs Ci:Jf6 looks really stupid for White.

...

The position is the same as the main lines of the Bayonet, but White has not played b4 or .l:f.e1. Still, when playing the white pieces one can some­ times get away with a few things and maintain equality. Indeed, after 13

162

Th e M a r de/ Plata Va r i a t io n : Wh ite 's E ig h t h M o ve De viations exfs! ? ll'lxfs! (better than 13 ... gxfs 14 f4 or 14 'ifb3 !?) the position is fairly level. c) 11 dxes ll'lgxes 12 exfs ii.xfs 13 ii.e3 (after the hasty 13 ll'lds?! ll'lxf3+ 14 ii.xf3 ll'ld4 Black has some initiative) 13 ...'it'f6 (or 13 ... 'it'd7) 14 .U.c1 (14 ll'lds ll'lxf3+ 1S ii.xf3 'it'f7 16 'it'd2?! would give Black the initiative after 16 ...ll'les 17 ii.e2 c6) 14....l:f.ae8 1s b3 h6 16 'it'd2 gs 17 ll'ld4 ll'lxd4 18 ii.xd4 and now 18 ...g4 19 ..t>h1 'it'g6 gave Black good play in J.Granda Zuniga-J.Polgar, Aruba 1992. A promising alternative is 18 ... ll'lf3+ 19 ii.xf3 'it'xd4 20 'it'xd4 ii.xd4 21 .l:f.fd1 (21 ii.xb7? ii.d3) 21 ...ii.cs! when 22 ii.xb7 ii.a3 win s the exchange. Returning to 11 ii.gs and here Black has a pleasant choice:

'ife8 822: u. ..J:.f6

BU: 11

...

.

There is even a third option in 11 ...'it'd7!?, which looks funny but is not bad: a) 12 exfs exd4 13 ll'lb s l:txfs ! is note 'a' to White's 11th move, above.

b) 12 dxes ll'lgxes (or 12 ...fxe4 13 ll'lxe4 ll'lgxes 14 ll'lxes ll'lxes with the idea of ...'it'f7) 13 exfs 'it'xfs 14 'it'd2 ii.e6 1s ll'lds 'iff7 16 ll'lxes ll'lxes 17 .l:f.ac1 .l:f.ae8 was about equal in R.Vera­ R.Slobodjan, Havana 1999. c) 12 dS ll'ld4!? 13 ll'lxd4 exd4 14 ll'lbs fxe4 1s ll'lxd4 .l:f.xf2 16 ll'le6 .l:f.xf1+ 17 ii.xf1 (17 'it'xf1 'it'xe6 18 dxe6 ii.d4+ 19 ..t>h1 ll'lf2+ is a perpetual) 17 ...ll'lf6 18 ll'lxg7 'it'xg7 19 'it'd4 ll'lg4 20 'it'xg7+ ..t>xg7 21 ii.d8 c6 reaches an unclear ending. 821) 11...'it'eS

Black keeps pieces on the board. 12 dxes Others: a) 12 exfs gxfs (12 ... e4!? is possible, while after 12 ...ii.xfs 13 h3 ll'lf6 14 dS, as in B.Lalic-V.Zivkovic, Zagreb 2007, 14... ll'ld4!? could be investigated) 13 dxes dxes 14 h3 ll'lf6 1s ll'lds 'iff7 16 ii.e3 h6 gave Black an active position in B.Kantsler-V.Bologan, Moscow 1991. b) 12 ll'lds 'it'f7 13 dxes (the creative 13 ii.e7 ll'lxe7 14 ll'lgs 'it'e8 1s ll'lxc7

1 63

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 'it'd8 16 ll'lce6 - 16 ll'lxa8 exd4 is less clear - 16 ... ii.xe6 17 ll'lxe6 'it'd7 18 exfs gxfs 19 ll'lxf8 .l:f.xf8 20 h3 was S.Lputian­ V.Bologan, Azov 1991, and here 20 ... ll'lf6 looks much better for Black) 13 ... ll'lgxes (or 13 ... dxes, leading to the note to White's 13th move, below) 14 exfs ii.xfs 1s 'it'd2 l:tae8 was level in P. S an Segundo Carrillo-V.Topalov, Ma­ drid 1997. c) 12 dS ll'ld4? 13 ll'lxd4 exd4 14 ll'lbs 'it'es 1s f4! 'it'xe4 16 ii.xg4 fxg4 17 l:te1 'iffs 18 ll'lxd4 ii.xd4+ 19 'it'xd4 was better for White in C.Toth-G.Hemandez, Merida 1993, but 12 ... ll'le7 and 12 ... ll'ld8!? both look very playable for Black. 12 ...dxes Preferring the dynamic pawn struc­ ture that arises to 12 ... ll'lgxes 13 exfs ii.xfs 14 'it'd2 'iff7 with equality in R.Vera-J.Hebert, Montreal 1997. 13 h3 White can also play 13 ll'lds 'iff7 14 ii.d2 (instead 14 ll'ld2 h6 1s ii.h4 gs 16 ii.xg4 fxg4 17 ii.g3 ll'ld4 gave Black a good position in V.Burmakin-V.Bologan, Sochi 2004) 14...ll'lxf2! ? (more enterpris­ ing than 14...ll'lf6 1s ll'lgs 'it'd7 16 exfs gxfs) 1s l:txf2 fxe4 16 ii.e3 exf3 17 l:txf3 'it'd7 18 .l:f.xf8+ ..t>xf8 19 'iff1+ ..t>g8 20 .l:f.d1 (20 ll'lf6+ hf6 21 'it'xf6 looks like a better try, although Black should be fine here) 20 ...ll'ld4 21 ii.xd4 exd4 22 ll'lf6+ hf6 23 'it'xf6 'it'e8 24 ii.f3 'if e3+ 2S ..t>f1 ii.e6 26 'it'xd4 'it'xd4 27 .l:f.xd4 cs 28 .l:f.e4 'itf7 gave Black slightly the more pleas­ ant endgame in O.Csoli-G.Pirisi, Hungar­ ian League 2001.

1 64

13 ...ll'lf6

14 ii.d3 This move has been criticized, but White does not have any advantage: a) 14 cs ll'lxe4 1s ll'lxe4 fxe4 16 ll'ld2 (after 16 'it'ds+ 'iff7 both 17 'it'xe4 ii.ts and 17 'it'xf7+ .l:f.xf7 18 ll'ld2 ii.e6 give Black good play) 16 ... ii.e6 17 ll'lxe4 h6 gives Black a very active position. b) 14 ii.xf6 is the safest. After 14...ii.xf6 1s ll'lds Black can try 1s...fxe4, 1s ...'it'd8 or the ambitious 1s ...ii.d8!?. 14.. .ii.e6 15 �el 'iff7 16 cs Instead 16 exfs gxfs 17 ll'lxes fails to 17 ... ll'lxes 18 .l:f.xes ll'le4. 16...ll'ld7 1

The M a r de/ Plata Var i a t i o n : Wh ite 's Eig hth M o ve D e v i a t i o n s Black already has some initiative. 11 ii.bs After 17 exfs gxfs 18 ii.e3 h6 Black also stands very well. 11...ll'ld4! White is in trouble, since 18 ii.e3 !Lixf3+ 19 'it'xf3 c6 20 ii.e2 f4, with the idea of ...ll'lxcs, is much better for Black. Instead 18 ii.xd7 ii.xd7 19 ii.h4 ii.c6 20 !Ligs 'it'f6! 21 exfs gxfs 22 ll'le2 .l:f.ad8 23 �xd4 .l:f.xd4 24 �5? h6 and 0-1 was l.Sokolov-A.Shirov, FIDE World Champi­ onship, Las Vegas 1999. Here 25 ll'lf3 loses to 2s ...ii.xf3.

17 ll'ld2 White had some edge follow­ ing both 17 ...ll'lhf7, E.Scarella­ D.Valerga, Villa Martelli 2002, and 11 ... ii.d7 18 .l:f.c1, P.Ricardi-G.Kasparov, Buenos Aires (simul) 1992 (a game which Kasparov lost). I think Black can improve, however, with 14...ll'ld4!.

822) 11 ii.f6 ...

This move was Kasparov's choice and it was also played with success by Anand... against Kasparov! 12 ii.xf6 ll'lxf6 Also possible is 12 ...'it'xf6 !? 13 h3 (the alternatives 13 dxes ll'lgxes, 13 exfs ii.xfs and 13 dS ll'ld4!? do not look dangerous) 13 ...ll'lh6 14 dS (14 dxes dxes is relatively unexplored, but Black has done well in practice), but now af­ ter 14 ...ll'ld8 15 cs ll'ldf7 16 cxd6 ll'lxd6

13 exfs Instead 13 dS ll'le7 leaves Black with the active ideas of .. .fxe4 and ...ll'lfs or .. .f4, and 14 exfs ll'lxfs is certainly okay for him. Another possibility is 13 dxes dxes 14 'it'xd8 .l:f.xd8 15 ll'lds (15 exfs can be met by 15 ... ii.xfs with equality or Kas­ parov's suggestion 1S ...e4) 1s ...ll'lxe4! 16 ll'lxc7 .l:f.b8 17 .l:f.fd1 (better is 17 .l:f.ad1 ii.d7 18 .l:f.fe1) 11 ... ii.d7 18 ii.d3 l:tbc8 (Kasparov suggested 18 ... ll'lf6 and 18 ... ll'lcs! looks good too) 19 ll'lds ll'lcs 20 ii.fl ii.e6 , which is already some­ what better for Black. Indeed, after 21 b4? ii.xds 22 cxds ll'lxb4 23 ll'lxes ll'le4 24 ii.c4 ll'lc3 25 .l:f.d2 ll'lbxds 26 g3 ..t>g7 27 ..t>g 2 ll'le3+ White resigned in L.Portisch-G.Kasparov, Linares 1990. 13 ...ii.xts

1 65

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 This is more solid than 13 ... gxfs 14 dxes dxes 1 5 'it'xd8 .U.xd8 16 ll'lds ll'le8 17 l:tad1 when White had some pres­ sure in A.Mikhalchishin-K.Hulak, Por­ toroz/Rogaska Slatina 1993. 14 d5 Instead 14 dxes dxes is fine for Black. His active pieces and dark-square control compensate for the isolated es­ pawn. 14...ll'le7

If White cannot generate any initia­ tive, the exchange of dark-squared bishops will favour Black positionally. 1s ll'lgs After 15 'it'd2 ..t>g7 16 ll'lh4 ii.d7 17 f4 exf4 18 'it'xf4 ll'lfs 19 ll'lxfs+ ii.xfs 20 g4 ii.d7 21 gs ll'lg4 22 'it'xf8+ 'it'xf8 23 .l:f.xf8 .U.xf8 24 ii.xg4 ii.xg4 25 ll'lbs l:tf4 Black was taking over the initiative in G.Kasparov-V.Anand, Geneva (rapid) 1996. 1s...h6 Instead 1 s ... c6 16 ii.d3 ii.g4 (proba­ bly better is 16 ...'it'b6 17 ii.xfs ll'lxfs 18 ll'le6 .l:f.f7 with the idea ...ll'ld4) 17 'it'd2

1 66

'it'b6 18 h 3 ii.d7 19 dxc6 bxc6 20 .l:f.ae1 .l:f.ad8?! (20 ... l:tae8) 21 ll'la4 'ifC7 22 f4 gave White the initiative in D.Rajkovic­ J.Nunn, German League 1990. Black has also tried 1s ... 'it'c8?!, but this is a strange choice of square. After 16 ii.d3 ii.xd3 (16... c6 17 l:tc1!) 17 'it'xd3 White has a slight edge. However, a better version of this idea is 1 s ... 'it'd7!? 16 ii.d3 c6 17 ii.xfs ll'lxfs 18 ll'le6 .l:f.fe8, with the idea of ... ll'lg7 or perhaps even capturing twice on dS to undermine the e6-knight. 16 ll'le6 ii.xe6 17 dxe6 ll'lfs Alternatives are 17 ... c6 and 17 .....t>g7. 18 ii.d3 ll'ld4 19 f4 19 ii.xg6 ll'lxe6 is unclear. 19 .....t>g1 20 'it'e1 ll'lxe6 21 'ifg3 gs

The position is unclear. Now 22 fxes?! dxes 23 'it'xes?? 'it'xd3 24 'it'xe6 .U.ae8 25 'iffs ll'le4! 0-1 was the sudden end of P.Van der Sterren-G.Kamsky, Wijk aan Zee 1994. Instead 22 fxgs hxg s with the idea of . . .ll'lhs is unclear. Black's kingside is rather loose, but his knights are active.

Part II The Classical· Variation d4 �6 2 c:4 g6 3 lbc3 .ig7 4 e4 d6 s ttJf3 o o 6 .ie2 es without 7 o-o

1

..

1 d4 ltJf6 2 C4 g6 3 ltJc3 Ji.g7 4 e4 d6 5 0if3 o-o 6 ii.e2 es

Here we look at those lines where White does not play 7 o-o. There are three different systems to consider and they are vastly different. With the Gli­ goric Variation (7 ii.e3) White main­ tains the tension, which makes it more difficult for Black to formulate a con­ crete plan. The Petrosian Variation (7 dS) does quite the opposite - White fixes the centre immediately. Lastly there is the Exchange Variation (7 dxes), where White resolves the ten­ sion in a rather extreme fashion by trading queens.

167

Chapter 9 Gligoric Variation 7 i.. e 3 ttJg4 8 ..tgs f6 9 ..tc1 ttJc6 and 9 i.. h4 ltJc6

1 d4 ll'lf6 2 c4 g6 3 ll'lc3 ii.g7 4 e4 d6 5 ll'lf3 0-0 6 ii.e2 es 7 ii.e3

The Gligoric Variation is a pretty flexible continuation for White and this makes it very dangerous. White keeps the tension in the centre and does not disclose what he intends to do with his king. It is not easy for Black because White can adjust his plans depending on how Black chooses to continue. The upside to all of this is that the Gligoric is not easy to play for White because Black has eight (!) main continuations. Because of this, White

168

has to be ready to handle the theory of several lines, whereas Black only really needs to know one. White also has to be prepared to play with several differ­ ent pawn structures. This makes it dif­ ficult to play compared to, for example, the Petrosian Variation, where White really only has to understand one structure. Because of this demand on the white player, the Gligoric is more popular amongst titled players than it is at other levels. Before we move on to my principal recommendation, 7 ...ll'lg4, here is a summary of Black's other options: a) 7 ...ll'lc6?! is played far too often. After 8 dS ll'le7 9 ll'ld2 White has the best of all worlds: an active bishop on e3 that is typical of Chapter 1; a knight on d2 that is ready to jump to c4 after c4-c5 as we saw in Chapter 4; and the flexibility to castle on either side of the board (or not at all), as we shall see in the Petrosian Variation of Chapters 11 and 12.

Gligo ric Va ria t i o n : 7 ii.. e 3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c 1 tlJ c 6 a n d 9 ii.. h 4 tlJ c 6 b) 7...tlJa6 i s perfectly respectable. After 8 o-o (not so dangerous is 8 dS .'Llg4) we are in the main lines of 7 0-0 .'Lla6 8 ii..e 3 which is outside of our rep­ ertoire. This approach has its positive side, however, because Black only has to know the 8 ii..e 3 line (8 .l:f.e1 is an­ other major variation after 7 o-o tlJa6). c) 1 ...tlJbd7 similarly is likely to tran spose to the line 7 o-o tlJbd7 8 ii.. e 3 when 8....l:f.e8!? 9 dS tLJhs 10 g3 ii..f8 has recently become fashionable. This is Bologan's DVD recommendation, but he does not even consider 8 dS which leads to what I consider to be a good variation of the Petrosian variation for White after 8...tlJg4 (otherwise 9 tlJd2) 9 �d2 ! (better than the more common 9 �g s) 9 .. .fs 10 tLJgs tLJdf6 (10 ... tLJcs 11 b4) 11 exfs gxfs 12 h 3 tlJh6 13 g4 when White has the initiative.

There is a great trap concealed here too, as after 13 ... ..t>h 8? 14 tlJe6 ii..xe6 1S gs! White is down a whole knight, but Black cannot avoid losing two pieces! d) 1 ...'ike7 was played by Kasparov against Karpov in their 1990 match,

but nowadays Black's exchange sacri­ fice 8 dxes (8 dS is also possible) 8 ... dxes 9 tLJds 'ikd8 (9 ... tLJxds 10 cxds c6 11 d6 is also considered to favour White) 10 ii.cs tLJxe4 11 ii.. e 7! 'ikd7 12 ii..xf8 ..t>xf8 is considered insufficient after 13 'ikd3 !. e) 7 ...exd4 is one of Black's main continuations. Aft er 8 tlJxd4 l:te8 9 f3 c6 the fi nesse 10 ii..f2 ! is considered to favour White (but 10 o-o dS 11 cxds tLJxds! equalizes instantly and 10 'ikd2 dS 11 exds cxds 12 cs tlJc6 13 o-o .l:f.xe3! is another Kasparov exchange sacrifice). f) 7...c6! ? is a fashionable line, al­ though here Black has to know a couple of different structures. One point of his move order is that 8 o-o exd4 9 tlJxd4 can be met with 9 ... .l:f.e8 10 f3 dS when Black has snuck back into the equaliz­ ing line of 7...exd4. Instead White can play 9 ii..xd4 with some chances of an edge, but he often plays 8 dS instead. The blocked position arising after 8...tlJg4 9 ii.g s f6 10 ii..h4 cs remains topical. g) 7 ...h6!? was a favourite of John Nunn's. The point is that after 8 o-o Black can play 8 ...tlJg4 9 ii..c1 tlJc6 10 dS tlJe7 11 tLJe1 fS 12 ii..xg4 fxg4 with an unusual and double-edged pawn struc­ ture. Instead 8 h3 is well met by 8 ... exd4! 9 tlJxd4 .l:f.e8 when 10 f3 would combine poorly with 8 h3 due to White's dark­ square weaknesses on the kingside. 8 dS is also considered harmless due to 8... tlJg4 9 ii..d2 fS (equalizing instantly

1 69

A ttacking C h e s s : Th e King 's I n d i a n , V o l u m e 1 according to Nunn), but after 10 h 3 ! (not, though, 10 exfs ii.xfs! 11 h3? ll'lxf2! 12 ..t>xf2 e4 when Black wins back the piece with a strong attack) 10 ... ll'lf6 11 exfs gxfs (11 .. .ii.xfs 12 ll'lh4) 12 g3! ll'la6 13 'ifcl ..t>h7 14 'it'c2 White's clever manoeuvres gave him a pleasant ad­ vantage in E.Perelshteyn-B.Lopez, US League 2009. Here Black's ...h6 move has created weaknesses on the kingside and hurt his position. We will discuss this idea again in Chapter 11. 7 ...ll'lg4

This is the main line and it is has been Radjabov's frequent choice. It is both principled and forcing. White cannot very well allow his bishop to be captured, so he invariably plays: 8 �gS f6 Now White has a choice of retreats. k t .A.cl. · B: 9 JUi4

Instead 9 ii.d2 cuts off White's con­ trol of d4 and makes little sense. After 9 ... ll'lc6 10 dS ll'ld4 Black is doing well.

1 70

A) 9 ii.cl

This move makes a funny impres­ sion, undeveloping White's pieces. Black has essentially played the moves ...ll'lg4 and ...f6 'for free' and now he has the move as well. However, these free moves do not automatically help Black's position. 9...ll'lc6! This is the most popular move and fits in well with our repertoire. Even so, some players are sceptical about its value and in New in Chess the strong Georgian Grandmaster Mikheil Mchedlishvili wrote that '9 ... ll'lc6 just won't do for Black'. While I agree with the ideas presented by Mchedlishvili, I will dare to challenge his assessment of 9 ... ll'lc6 ... Also possible is 9 ...ll'ld7 when 10 o-o ll'lh6 transposes to the line 7 o-o ll'lbd7 8 ii.e3 ll'lg4 9 ii.gs f6 10 ii.cl ll'lh6. Black has two other independent continua­ tions that are considered respectable in 9 ... exd4 and 9 ... fs, but we will not con­ sider them here. 1o d s

G lig o ric Va r i a t i o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJc6 a n d 9 ii.. h 4 tlJc6 Instead 10 o-o transposes t o the line 7 o-o tlJc6 8 ii..e 3 tlJg4 9 ii.gs f6 10 ii.. c 1 which was covered in Line B2 of Chap­ ter 8, while after 10 h3 Black should avoid 10 ...tlJh6 11 dS! tlJe7 12 h4!, which transposes to the note to his 10th move, below, and play 10...exd4 11 tlJxd4 tlJges 12 tLJxc6 (12 ii..e3 fS) 12 ...tLJxc6 intending 13 .. .fs. 10 tlJd4! Instead 10...tlJe7 seems natural. It looks as though Black is doing well compared to lines in the Mar del Plata, Black is up a few tempi and ready to play .. .fs. However, White has not cas­ tled and because of that he has extra options. Another problem is the e7knight. If this knight cannot contribute to Black's kingside play he often suf­ fers. This is a difficult variation for Black and it is this line for which Mchedlishvili condemned 9 ...tlJc6. The specific reason is that 11 h 3 ! tlJh6 12 h4! causes Black major problems. Some examples: ...

a) 12 ...fs 13 tLJgs! tlJg4 (13 .. .fxe4 14 g4! controls fS and gives White a big

advantage as shown in several games, while 13 ...f4 14 g4!, with the idea of 14_.fxg3 1S fxg3, is also much better for White) 14 h S (or 14 ii..xg4 fxg4 1 S ii.. e 3) 14...tlJf6 1 S hxg6 hxg6 1 6 ii..f3 and White has an advantage across the whole board. b) 12 ...tlJf7 13 h s cs (after 13 ... gs 14 'it'c2 h6 1s g4! c6 - 1s ... ii..xg4 16 tLJxgs! ii..x e2 17 tlJe6 wins for White - 16 tLJh2 the battle can only be decided on the queenside, where White has a space advantage) 13 ... cs 14 hxg6 hxg6 1 S ii.. e 3 ii..d7 16 g 3 ! ? (with the idea of ..t>f1g2 and 'it'g1-h2) 16 ...a6 17 ..t>f1 fS 18 ..t>g2 fxe4 19 tlJxe4 tLJfs 20 ii.gs tLJxgs 21 tlJfxgs 'ife7 22 .l:f.h7 tlJd4 23 ii..g4 .l:f.fs 24 �1 .l:f.xgs 2 s ii..xd7 .l:f.fs 26 ii..xfs gxfs 27 tlJd2 and White was better in M.Mchedlishvili-G.Kacheishvili, Tbilisi 2007. c) 13 .. .fs 14 hxg6 tlJxg6 1s 'it'c2 (White pressures fS and prepares to castle queenside) 1S .. .f4 16 ii..d2 cs 17 o-o-o a6 18 g3 ii..d7 19 l:LJM tlJxh4 20 l:txh4 fxg3 (a better try is 20...bs!?) 21 fxg3 tLJgs 22 .l:f.dh1 bs 23 'it'd3 (23 .l:f.h s also brings Black to the brink of defeat) 23 ...bxc4 24 'it'e3 h6 2S .l:f.xh6 ii..xh6 26 .l:f.xh6 tlJf7 27 .l:f.g6+ ..t>h7 28 .l:f.xd6! tlJxd6 29 �6+ ..t>g8 30 'it'g6+ ..t>h8 was A.Maric-1.Gaponenko, Moscow 2001. Here the simplest is 31 'it'xd6 when White has tremendous compensation for the two exchanges. Now we return to 10...tlJd4!. This move is eminently logical. White has not been developing his

1 71

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 pieces and, although this has granted him some flexibility, it could clearly become a problem should the position open up. This move was not considered by Mchedlishvili, but I believe it com­ pletely solves Black's problems after 9 ii.cl.

11 ll'lxd4 exd4 12 ll'lbs After 12 ii.xg4 dxc3 13 ii.xc8 'it'xc8 14 bxc3 Black is already better following 14 ...'if e7 or 14...fs, while 12 'it'xd4 fS is very risky. One high-level example went 13 'it'd1 fxe4! 14 ii.xg4 'ifh4 1S ii.e2 ii.xc3+ 16 bxc3 'ifxf2+ 17 ..t>d2 ii.g4! 18 cs .l:f.fs! 19 'it'e1 .l:f.xds+ 20 ..t>c2 'it'xg2 21 l:tg1 'it'xe2+ 22 'it'xe2 ii.xe2 23 cxd6 ii.d3+ 0-1, E.L'Ami-L.Nisipeanu, Euro­ pean Championship, Budva 2009. 10...ll'ld4 must have been a shock to L'Ami, who is usually very well prepared. 12 fs 13 0-0 Instead 13 h3 is crushed by 13 ... ll'lxf2 14 ..t>xf2 fxe4+, while 13 exfs also runs into 14 ... ll'lxf2! 14 ..t>xf2 and now 14 ... a6! 1s ll'la3 ii.xfs 16 ii.f3 d3 gives Black a strong attack. The only other option is 13 ii.xg4 ...

1 72

fxg4 14 o-o (14 ll'lxd4 'ife7! 1s 'it'd3 ii.ts gives Black the initiative), which re­ mains untried. 14...'ife7 looks good enough after 1S 'it'd3 (1S .l:f.e1 'it'f6!) 1 s...l:te8 (or 1s...cs!? 16 dxc6 bxc6 17 ll'lxd4 ii.b7, with ideas like ...cs and . ...l:f.ae8) when the e4-pawn gives White problems. White will be forced to play f3, either immediately or after 16 .l:f.e1 ii.ts, when ... gxf3 will give Black good compensation for the falling d4-pawn.

13 ll'les Black should not get carried away with 13 ... ll'lxh2? 14 ..t>xh2 fxe4 1 S ll'lxd4 �4+ 16 ..t>g1 ii.es 17 f4 exf3 18 ll'lxf3 .l:f.xf3 19 l:txf3 �2+ (or 19 ... ii.g4 20 ii.e3) 20 �1 �1+ 21 �2 'ifh4+ 22 ..t>e3 when the white king was safe enough and the material advantage counted in J.Ulko-D.Sokolov, Moscow 200S. 14 exfs White takes the pawn. Instead 14 ll'lxd4 causes no problems at all for Black. After 14_.fxe4 1 S ii.e3?! (safer is 1 S ll'le6 ii.xe6 16 dxe6, but Black has several possibilities here, the surest ...

G lig o ric Va ria t i o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c6 a n d 9 ii.. h4 tlJ c6 being 16 ... c6) 1 S ... �4 16 .l:f.c1 (it is al­ ready too late for 16 tlJe6 ii..xe6 17 dxe6 because of 17 ...tlJf3+! 18 gxf3 ii.es 19 f4 .l:f.xf4, as pointed out by Golubev) 16 ... tlJg4 17 ii..xg4 ii..xg4 18 'ifd2 ii..e S Black already has a stron g attack: 19 f4 (19 h 3 ii..x h3 and 19 g3 � 3 are no bet­ ter) 19 ... exf3 20 tlJxf3 ii..xf3 21 gxf3 .l:f.ae8 and Black was clearly better in A.Rychagov-M.Vachier Lagrave, Russian Team Championship 2009.

14 d3! Black cannot save the d4-pawn, so he uses to it to grab the bishop-pair and gain time for development. 15 ii..x d3 tlJxd3 16 'it'xd3 ii..xfs Black's powerful bishops give him excellent compensation for the pawn. White will also have some difficulties developing his queenside, and the c4pawn may become vulnerable. 17 'it'dl White has several alternatives, but Black always has enough play: a) 17 'it'e2 a6 18 tlJc3 'ifh4 19 ii..e 3 .l:f.ae8 gave Black good compensation in R.Vidonyak-D.Stets, Lvov 2009. ...

b) 17 'it'b3 a6 18 tlJc3 ii..d3 19 l:td1 �4 20 ii..e 3 ii..xc4 was good for Black in E.otero-R. Leitao, Cuba 1993. If 21 'it'xb7 �ab8 22 'ifxe7 .l:f.xb2 and Black has the initiative. c) 17 'ifg3 prevents ...�4. but is an odd square for the queen.

Black has: c1) 17...a6 18 tlJc3 ii.es ! ? 19 f4 ii..g 7 20 ii..d2 'it'f6 21 b3 bS! 22 cxbs 'it'd4+ 23 :tf2 axbs 24 .l:f.c1 was F.El Debs-D.Flores, Campinas 2010, and here 24...'it'cs! ? gives Black good play. c2) 17 ... 'it'd7 18 tlJc3 .l:f.ae8 19 ii..d2 ii.es 20 f4 ii..d4+ 21 ..t>h1 a6 22 .l:f.ae1 bS and Black had counterplay in A.David­ D.Stets, Fourmies 2010. c3) 17...ii..e s immediately is also possible: 18 ii.g s!? 'it'd7 19 'it'b3 a6 20 tlJa3 (after 20 tlJc3 ii..d 3 21 .l:f.fd1? 'iffs threatening ...ii..c 2 is good for Black, while after 21 .l:f.fe1 Black should play 21 ...'it'fs with compensation, rather than allow 21 ...'it'g4 22 .l:f.xes! when White gets counterplay) 20 ... 'ifg7 21 .l:f.ae1 l:tae8 22 tLJc2 (not 22 'it'xb7 ii..d 3) 22 ...ii..xb2 23 tlJe3 ii..c 8 (23 ...ii..d 4!?) 24

1 73

A ttacking C h e s s : The Kin g 's I n dian, Vo l u m e 1 ii.h6 'it'xh6 2 5 'it'xb2 .l:f.e4 was level in D.Yevseev-A.Khruschiov, Peterhof 2009. 11 'it'f6 This active move leaves the Cl-pawn to its fate, but Black will gain sufficient play. After 17...'it'M White should go for 18 ll'lxe7! as in E.Kanter-A.Savina, Dimi­ trov 2007. However, 17...a6 is plausible: 18 ll'ld4 �4 19 ii.e3 .l:f.ae8 20 'it'd2 ii.xd4 (Black basically plays for a draw against his much higher-rated opponent with opposite-coloured bishops, rather than depending on the bishop-pair for long­ term compensation) 21 ii.xd4 .l:f.e4 22 ii.c3 .l:f.fe8 23 f3 .l:f.e2 24 'it'd4 'it'xd4+ 2 5 ii.xd4 .l:f.c2 2 6 .l:f.fc1 .l:f.ee2 27 .l:f.xc2 .l:f.xc2 28 b3 .l:f.d2 29 ii.c3 .l:f.c2 30 ii.f6 rt;f7 31 ii.d8 .l:f.b2 32 .l:f.e1 .l:f.b1 33 rtlf2 l:txe1 34 ..t>xe1 c6 and Black held in L.Fressinet­ D.Sharavdorj, Las Vegas 2009. •••

active and White's queenside is weak. 21 ii.e3 This is sensible. White prioritizes developing his pieces. 21 b6 22 ii.d4 ii.xd4 23 'it'xd4 .l:f.xc4 •••

The endgame is equal. A couple of examples are 24 'it'd3 l:te4 25 b3 .l:f.e2 26 'it'd4 d5 27 a4 'ife4 28 .l:f.ad1 'it'xd4 29 .l:f.xd4 .l:f.d8 30 .l:f.fd1 .l:f.b2 with equality in L.Andonovski-M.Jazadzievski, Struga 2008, and 24 'it'd2 .l:f.e4 25 .l:f.fe1 .l:f.xe1+ 26 .l:f.xe1 'it'xa2 27 f3 d5 28 .l:f.e7 .l:f.f7 29 .l:f.e8+ .l:f.f8 30 .l:f.e7 with a draw in A.Gavrilov-A.Zontakh, Lipetsk 2009. B) 9 ii.h4

18 ll'lxc7 .l:f.ac8 19 ll'le6 After 19 ll'lb5 .l:f.xc4 Black has excel­ lent compensation for the pawn. 19 ii.xe6 20 dxe6 'it'xe6 White has managed to hold on to the pawn and eliminate Black's bishop­ pair, but Black's major pieces are all •••

1 74

G l ig o ric Va riati o n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c 6 a n d 9 ii.. h4 tlJc6 This is much more challenging. White keeps his bishop active and hin­ ders the advance of Black's f-pawn. Be­ cause this line is critical for 7...tlJg4, we will consider two main lines. The move 9 ...tlJc6 is 'consistent', but most players prefer the modem 9 ... g s, which we will consider in the next chapter. 9 tlJc6 Black sometimes plays 9 ... tlJd7 10 o-o tlJh6, which is a solid line generally reached from the move order 7 o-o tlJbd7 8 ii..e3 tlJg4 (8 ... c6 is more com­ mon here, while 8 ... l:te8 is fashionable) 9 ii.g s f6 10 ii..h4 (and here 10 ii..d 2 and 10 ii.cl are seen more frequently) 10 ... tlJh6. The older 9 ... 'it'e8 is not so bad either, but I have complete confidence in my two recommendations ! ...

1o d s Only this move is critical. Instead 10 o-o leads us back to a position consid­ ered in the notes to White's 10th move in Line B2 of Chapter 8. others: a) 10 h 3 tlJh6 11 dS (11 dxes dxes makes little sense because Black plays ... tlJh6 anyway, but this could still

transpose to variation 'b', below) 11 ... tLJe7 and here: al) 12 'if c2 cs 13 g4 tlJf7 is fine for Black. The plan is to play ...ii..d7, ...tlJc8 (to protect the f6-pawn), and then ... ii..h 6, activating the bishop. If White castles queenside, then ... a6 and ... bs will be enticing. One example was 14 o-o-o a6 1s ..t>b1 ii..d 7 16 tlJd2 'it'as 17 l:tdf1 bS 18 ii..g 3 ii.. h6 with a good posi­ tion in V.Gefenas-J.Van Oosterom, cor­ respondence 1996. a2) 12 tlJd2 gs 13 ii.. g 3 fs 14 exfs (or 14 f3 tlJg6 with counterplay) 14 ...tlJhxfs 1 S tlJde4 tlJg6 16 'it'd2 tLJf4 17 ii.fl tlJd4 18 o-o-o a6 19 ..t>b1 ii..fs 20 f3 cs 21 ii..d 3 bS was another good example of Black's possibilities in F.Gheorghiu1 .Sokolov, Kava la 1990. b) 10 dxes dxes (even simpler is 10 ... tLJgxeS; then 11 tLJxes dxes is equal and 11 o-o can be met with 11 ... ii..e 6, intending ...'it'd7 and ...fs, or by 11...tLJxf3+ 12 ii..xf3 ii..e 6, again intend­ ing ...'it'd7 and ...fs), and here 11 o-o tlJh6 is again comfortable for Black. Others: bl) 11 'it'xd8 tlJxd8 is equal (but not 11 ... l:txd8? 12 h3 tlJh6 13 tLJds). b2) 11 tLJds tlJh6 (intending to con­ quer d4 with ... ii..g4) 12 h3 and here 12 ...tlJd4?! is met by 13 tlJxd4 exd4 14 es!, while 12 ... ii..e 6 13 'ifb 3 ! 'it'c8 14 0-0-0 gave White some advantage in V.Korchnoi-V.Nevednichy, Paks 2004. However, Black can instead play 12 ... g S 13 ii..g 3 f S with counterplay, V.Shish­ kin-C.Carmaciu, Bucharest 2006.

1 75

A t tacking Ch ess: Th e K i n g 's I n d i a n, Vo l u m e 1 10 ll'le7 This retreat is more viable than it was in Line A, because it is not possible for White to launch his h-pawn forward with his bishop on h4. The main concern for Black i s still that the e7knight has trouble finding a useful role. It would be nice if the untried 10 ...ll'ld4 worked here too, but it looks a bit too speculative after 11 ll'lxd4 exd4 in view of both 12 'it'xd4 g s 13 ii.g3 fS 14 'it'd1! ll'lf6 (14 .. .f4? 1 s ii.xg4) 1s exfs (1S f3 fxe4 16 fxe4 ll'ld7 gives Black some play) 1 s...ii.xfs 16 o-o ll'le4 17 ll'lxe4 ii.xe4 18 'it'd2 'it'f6 19 l:tae1 'it'xb2 20 'it'xg s and 12 ll'lbs .l:f.e8 13 h3 ll'lh6 (13 ... ll'les 14 o-o cs 1s dxc6 bxc6 16 ll'lxd4 'ifb6 17 ll'lb3 also look insuffi­ cient) 14 f3 cs 1S dxc6 bxc6 16 ll'lxd4 'it'b6 17 'it'd2. Well, I tried. 11 ll'ld2 This natural move unleashes an at­ tack on the g4-knight. Instead 11 h 3 tran sposes back to 1 0 h 3 , while 1 1 o-o will transpose to lines with 12 o-o after 11 ... ltJh6 12 ltJd2 OT 11 ...h S 12 ltJd2. ...

After 11 ll'ld2 Black can retreat his knight or maintain it on g4 for a move. 81: 11 lbh6 82: 11 h s •.•

...

Instead 11 .. .fs 12 ii.xg4 fxg4 13 ii.gs :tf4! ? (perhaps Black could try 13 ...h6 14 ii.e3 gs 1s cs ll'lg6) is possible. Th is looks a bit too specul ative, though, af­ ter 14 o-o (there is no point in playing 14 ii.xf4 exf4 because Black will spend a tempo to force the capture anyway) 14...h 6 1s ii.xf4 exf4 16 ll'lb3 gs 17 ll'ld4 ii.xd4 18 'it'xd4 ll'lg6, as in Y.Shulman­ A.Poluljahov, Gausdal 1994, and here 19 cs would test Black's concept. Bl) 11 ll'lh6 •••

This is Black's main continuation. 12 f3 White stays flexible. other moves show his hand too early: a) 12 g4 ll'lf7 13 'it'c2 cs 14 f3 ii.d7 is an easy-enough-to-handle set-up for Black, especially with White committed to g4.

1 76

G ligo ric Va ria tio n : 7 ii.. e 3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJ c 6 a n d 9 ii.. h 4 tlJ c 6 b ) 12 b4 does really not do that much, other than rule out castling queen side. Black can just begin his play on the other side of the board with 12 ... g s 13 ii.. g 3 fs (or 13 ... tlJg6) 14 f3 tlJg6. c) 12 o-o gs 13 ii..g 3 is committal too: cl) 13 ... tlJg6 14 l:tel (after 14 l:tcl fS 1S exfs ii..xfs 16 tlJde4 tLJf4 17 cs ii.. g6 18 .l:f.e1 tLJxe2+ 19 'it'xe2 tLJfs 20 'it'bs g4! 21 'it'xb7 h S ! Black had a kingside ini­ tiative in L.Van Wely-A.Shirov, Tunja 1989) 14...tlJf4 1 S ii.fl (or 1S tlJf1 fS) 1S ... fs 16 exfs tLJxfs (16 ...ii..x fs! ?) 17 tlJde4 tlJd4 and Black has counterplay. c2) 13 .. .fs 14 exfs (14 f3 f4 1S ii..f2 cs also gives Black sufficient play) 14 ... tlJhxfs 1S ii..h s (to prevent 1S ... tlJg6) 1S ...tlJd4 16 f3 bS!? (a creative idea; instead both 16 ... cs 17 tlJde4 h6 and 16 ... ii..fs 17 tlJde4 tlJg6 look good enough) 17 cxbs l:tb8 18 tlJde4 tLJxbs 19 'i'd2 (19 tLJxgs .l:f.fs) 19 ...h6 was unclear in A.Onischuk-V.Bologan, Poikovsky 2004.

12 gs ...

This is the sharpest continuation. Black initiates play on the kingside immediately. There is a serious alterna­ tive in 12 ... cs. Black tries to block the queenside before beginning his king­ side campaign. This has been popular amongst the top players, but I think it actually gives White quite a pleasant choice. Here 13 a3 prepares 14 b4, but surely 13 :tb1 is a better way of doing this. Instead 13 ii..f 2 can be met with 13 .. .fs, while after 13 g4 tlJf7 it is hard for White to come up with a good plan. If the bishop moves from h4, then ... ii..h6 is possible, but if the bishop maintains its post, then White cannot play h4 and Black can play ...ii..d7, ...l:tb8, ...a6, ...tlJc8, (now f6 is protected), and ...ii..h6. So White should either ini­ tiate his queenside play immediately or else change nature of the position: a) 13 .l:f.b1 has been played fre­ quently by Gelfand: al) 13 ...tlJf7 14 o-o fS 1S b4 b6 16 a4! ? ii..h6 17 ii..f 2 ..t>h8 18 as tlJg8 19 'it'c2 tlJf6 20 .l:f.b2 tlJhs 21 .l:f.a1 ii..d7 22 axb6 axb6 23 l:txa8 'it'xa8 24 .l:f.a2 'it'd8 2s bxcs bxcs 26 .l:f.a7 and White had maintained the initiative in B.Gelfand1.Nataf, Cannes (rapid) 2002. a2) 13 ... gs 14 ii..f2 fs 1S b4 b6 16 bxcs bxcs 17 h3 (17 o-o tlJg6) 17 ...tlJg6 18 g4 fxg4 19 hxg4 l:LJM and Black had counterplay in B.Gelfand-L.Dominguez, FIDE World Championship, Moscow 2001. b) 13 dxc6 is also critical, changing

1 77

A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1 the contours of the position. Black's centre may look imposing, but White is not really threatened by it and he will have good squares for his pieces after 13 ...bxc6 14 b4 ii.e6 and then :

bl) 1S ll'lb3 dS 16 ll'lcs (instead 16 cxds cxds 17 exds ll'lxds 18 ll'lxds ii.xds 19 o-o ll'lfs 20 ii.f2 ii.h6! was to­ tally fine for Black in B.Gelfand­ T.Radjabov, Astrakhan 2010) 16 ... ii.f7 17 0-0 d4 18 ll'l3a4 d4 18 ll'l3a4 ll'lc8 19 ltJ b3 ll'ld6 20 ll'las 'it'e8 21 ll'lcs was pleasant for White in G.Giorgadze­ T.Shaked, Linares 1997. b2) 1S ii.f2 !? dS 16 ll'lb3 ! (16 cxds cxds 17 exds ll'lxds 18 ll'lxds ii.xds 19 o-o ll'lfs 20 ll'lb3 is Gelfand-Radjabov above) 16 ...d4 17 ll'la4 was V.lvanchuk­ T.Radjabov, Monaco (rapid) 2007, when White's position is easier to play. 14 C5 Instead 14 o-o ll'lg6 1S cs will trans­ pose to 1S o-o below, but White does have another option in 14 h3 ll'lg6 1S g4 fxg4 (after 1S ... f4 White will have a free hand on the queenside, while 1S .. .fxe4 1S ll'ldxe4 looks very good for

1 78

him) and now: a) 16 hxg4 ll'lh4 (or 16 ... ll'lf4 intend­ ing ... cs) looks unclear. Black can play ... a6 and _J:tb8 and maybe ... l::tf4, even as an exchange sacrifice. b) 16 fxg4 ll'lf4 17 l::th2 cs 18 ll'lfl a6 19 ll'le3 l::tb8! 20 ii.fl ii.d7 was unclear in J.Gustafsson-V.Nevednichy, Sarajevo 2010. White has a space advantage, but his king does not have an ideal home.

14...ll'lg6 This is more flexible than 14... g4 1S fxg4 ll'lxg4 16 ii.xg4 fxg4 17 o-o ll'lg6 when 18 cxd6 �ess clear is 18 ll'lc4 dxcs!? 19 ii.xcs l::txf1+ 20 'it'xf1 b6 21 ii.e3 ii.a6) 18 ... cxd6 19 ll'lc4 looks a little better for White. 15 g3 White prevents the knight from coming to f4, but he can also allow it by playing 1S 0-0 ll'lf4 16 cxd6 cxd6 17 ll'lc4 .U.f6 !? (17 ... g4 18 fxg4 ll'lxe2+ 19 'it'xe2 ll'lxg4 is also possible) 18 a4 l::t g 6 19 exfs ll'lxfs (not 19 ...ii.xfs 20 g4!, crippling the h6-knight, S.Shipov­ D.Anagnostopoulos, Cappelle la Grande

G ligo ric Va ria tio n : 7 ii.. e3 tlJ g 4 8 ii.. g 5 f6 9 ii.. c1 tlJc6 a n d 9 ii.. h 4 tlJ c 6 1997) 20 ii..d3 .l:f.h6 21 ii..xfs ii..xfs 2 2 �4 .l:f.g6 23 g4 and here 23 ...ii..xe4 24 fxe4 was a little better for White in S.Shipov-M. Pavlovic, Athens 1997. but Black could consider 23 ...ii..c8! ? with the idea of ...h s. 1s ...g4 Black has an interesting alternative in 1S ...fxe4 16 fxe4 (this has always been played; after 16 tlJdxe4! ? Black can play 16 ... tLJfs, which should be somewhat better for White, or else sac­ rifice a pawn with 16 ... ii..h 3 17 ii..e 3! 0ifs 18 ii..xg s 'it'd7 when he has at least activated his pieces, but it is not clear if it is enough for a pawn) 16 ... ii..h 3 17 �e3 b6 and here:

a) 18 cxd6 cxd6 19 tlJc4 'it'd7 20 'ti'b3 (after 20 ii..xg s bS 21 tlJe3 b4 22 ii..x h6 ii..x h6 23 ii..g4 't\Vb7! Black has serious counterplay, and here 21 tlJd2 b4 also gives Black the initiative) 20 ...tlJg4 21 ii.. x gs tlJf2 22 .l:f.g1 ii.. g 4 23 ii.. e 3 ii..xe2 24 ii..xf2 was Y.Shulman-B.Socko, Saint Petersburg 1997. Now 24...ii..d3 ! ? 2 5 .l:f.d1 ii..xc4 26 'it'xc4 ii..h 6 would give Black good play for the pawn.

b) 18 c6! ? and now 18...a6 is cer­ tainly possible, but more enterprising is 18 ... 'it'c8 when 19 ii..xg s ii.. g 2 20 .l:f.g1 �3 21 'it'c2 'it'xh2 22 o-o-o 'it'xg3 is very murky and 19 tlJc4 tlJg4 20 'it'd2 tLJxe3 (20 ...h 6 20 ... ..t>h8) 21 tLJxe3 (21 'it'xe3 'it'd8 regroups) 21 ... a6! ? is un­ clear. 16 fxg4 After 16 exfs gxf3 17 ii..xf3 tlJxfs Black activates his pieces and has no problems. 16...tLJxg4 Black has also tried sacrificing a pawn with 16 .. .f4, but I have trouble believing that this can be sound after 17 tlJc4. 17 ii..xg4 fxg4 18 o-o

White can also delay this move: 18 tlJc4 (18 cxd6 cxd6 19 tlJc4 tlJh8 trans­ poses, but here too Black could con­ sider 19 ... b6) 18...tlJh8 (possibly not best; again there is 18...b6!?, while another possibility is 18...ii..d7! ? with the idea of 19 cxd6 cxd6 20 tlJxd6?! .l:f.xf2 !) 19 cxd6 cxd6 20 o-o tlJf7 21 ii..e 3 ii..d7 22 a4 when White's space advantage

1 79

A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1 brings him a slight but pleasant advan­ tage. Indeed, he maintained an edge after 22 ... ii.f6 23 '1i'e2 ii.gs 24 'lth1 'ii' e7 2S tllb s ii.xe3 26 tllx e3 in J.Ehlvest­ L.Van Wely, Groningen 1993. 18 ... b6!? This move forces White to resolve the queenside tension while preparing a possible ... ii.a6. Instead 18 ... tllh 8 19 cxd6 cxd6 20 tllc4 leads back to the previous note. 19 cxd6 White could also try 19 c6, although this limits his queenside possibilities. Black could continue with 19 ... ii.a6, but this might tempt White into sacrificing the exchange with 20 '1i'xg4!?. Instead after 19 ... hs White could play 20 .lii . e3, so perhaps the best move is 19 ...ii.h6 ! ?. 19 ... cxd6 20 ii.e3 ii.d7 21 '1i'e2 '1i'e7 22 tllc4 l:txfl+ 23 l:txfl l:tf8 24 l:txfB+ '1i'xf8 25 ii.d2

2s ... tt:Jhs! Now that he has developed his pieces and consolidated his position a bit, Black begins a standard manoeu­ vre; the knight is bound for gs.

180

26 a4 ii.h6 27 tllb s ii.xd2 28 tllx d2 After 28 '1i'xd2 ii.xbs 29 axbs '1i'f6, with ideas like ... '1i'g6 and ...tllf7-g S, Black has no problems. 28... a6 29 tllc 3 bS 30 axbs axbs 31 tllf1 Or 31 tll x bs '1i'b8. 31 ... 'ii'cs 32 'ltg2 'ii'cs 33 tll e 3 hs Black has gradually taken over the initiative and soon won in C.Csiszar­ P. S inkovics, Hungarian League 1996. e2) 11 hs .••

Fedorov's favourite move is ambi­ tious but quite risky. Black hopes to gain some space on the kingside before retreating his knight, but now White can open the h-file, which can lead to trouble for Black if he is not alert. On the plus side, White's bishop blocks the h-file for the time being and some of White's dark squares on the kingside will be weakened if he prises open the kingside. Whether or not Black can make use of this is a different matter. 12 h3 tllh 6 13 g4! This is the only real test of Black's ambitious play.

G l ig o ric Va riatio n : 7 il e3 tlJg4 8 il g 5 f6 9 il c1 tlJ c 6 a n d 9 il h 4 tlJ c 6 13... hxg4 14 hxg4 tlJf7

Black would love to play ... 1'.. h 6, ...'ltg7, and ...�h8, but that is not so easy because White's bishop attacks the f6-pawn. Therefore Black needs to move his clumsy e7-knight, but this too is not that easy. The typical plan for Black is ... cs, ...1'..d7 and ...tlJc8, but while Black is making these manoeu­ vres White can start feeding his pieces towards the kingside. 15 '1i'c2 This move prepares castling queen­ side and is the most popular, but oth­ ers are dangerous too: a) lS '1i'b3 also prepares castling, but from here the white queen may slide along the third rank to threaten the black king: 1s ... cs 16 tiJd1 ! ? 1'..d7 (Black could also consider the immedi­ ate 16 ...l:te8 or an evacuation plan starting with 16 ... tiJh8) 17 '1i'h3 (17 '1i'xb7 l:tb8 18 '1i'a6 tlJc8 19 b3 il h6 20 �b2 'ltg7 21 o-o-o '1i'C7 gave Black some compensation for the pawn due to the clumsy position of the white queen in J.Ehlvest-A.Fedorov, Calcutta

1999) 17 ... l:te8 18 tlJe3 looked quite threatening in J.Ehlvest-V.Spasov, Ko­ caeli 2002, but actually here Black could try to create some counterplay on the queenside, starting with 18...'ii' a s!?. b) lS tiJfl cs 16 tlJe3 (Golubev also mentions 16 '1i'd3 1'..d7 17 '1i'h3 l:te8 18 '1i'f3 tlJc8 19 tlJe3 a6 20 a4, but while White may have some advantage, this does not look fatal for Black; his king is safe enough for now and he can try to prepare ... bs) 16 ... 1'..d7 17 1'..d3 was Z.Azmaiparashvili-A.Fedorov, Elis ta Olympiad 1998, and here Fedorov sug­ gests 17 ...tlJc8 with the idea of ... il h6. 15 ...c5 After 1s ...1'..d7 16 o-o-o tlJc8 17 ilg3 il h6 White has a thematic device available.

Indeed, 18 f4! (much stronger than 18 l:th2 'ltg7 19 z:tdh1 l:th8 20 'ltb1 V2-V2, E.Magerramov-A.Fedorov, Dubai 2001) 18 ... 1'.. xf4 19 ilxf4 exf4 20 es! (the g6pawn is the target) 20... tZJxes 21 l:th6 '1i'e7 22 �dhl '1i'g7 23 CZJf3 tlJe7 24 tZJxes fxes (24... dxes 2S cs!) 2S 1'.. d3

1 81

A ttacking C h ess: Th e King 's I n d i a n , Vo l u m e 1 ii.xg4 26 ii.xg6 gave White a winning attack in F. Peralta-V.Spasov, Lorca 2004. 16 o-o-o

16 ... a6 By playing this immediately, Black is ready to play ...bs at any time. The al­ ternative is to play 16 ... ii.d7 first, fol­ lowed by either ...a6 or ... tllc8. Some examples: a) 17 �dfl and here: al) 17 ... a6 18 �h 2 tllc8 19 ii.g3 ii.h6 20 f4 exf4?! (an exchange of mistakes; Black should play 20 ... ..txf4) 21 ii.xf4?! (21 es!) 21 ... .Lf4 22 �xf4 was B.Avrukh-E.lnarkiev, Moscow 2002. Here 22 ... tlle s looks solid enough. a2) 17 ... tllc8 18 ii.g 3 ii.h 6 19 f4 ..txf4! 20 ii.xf4 exf4 21 l:i.xf4 tlle s (but not 2 L .'ltg7? 22 �hf1 tlles 23 g s ! fs 24 exfs ii.xfs 2s tllce4 tllb6 26 '1i'c3 with a big advantage in A.Huzman-V.Sara­ vanan, Biel 2000) 22 l:i.hfl gs 23 �4f2 'ltg7 and Black is okay. If White protects the g4-pawn with a rook, Black will be able to play ... tllc 8-e7-g6. b) 17 �h2 tllc8?! (this abandons the

1 82

g6-pawn, so Black must consider 17 ... a6, which could transpose to the note to White's next move) 18 ii.g3 ii.h6 19 f4! (instead 19 lldhl 'ltg7 20 g S ! ? fxgs 21 'ltb1 gives White compen­ sation, but this is less challenging), and now Black has problems: bl) With 19 ... exf4 Black wants to keep the bishops on the board, but this fails tactically: 20 es 'ltg7 21 exf6+ '1i'xf6 22 tllce4 '1i'e7 (or 22 ...'ii' es 23 ii.h4 with a winning advantage in F.Peralta­ A.Suarez Real, Leon 2008) 23 ilh4 gs 24 ii.xg s ii.xgs 2s �h7+ 'ltxh7 26 tllxg s+ with a mating attack in T.Enkhbat­ Y.Shulman, Connecticut 2002. An im­ pressive attack especially considering that Shulman is a strong proponent of the Gligoric with White! b2) 19 ...ii.xf4 20 Lf4 exf4 21 es tllx es 22 l:i.h6 '1i'e7 23 �dhl '1i'g7 24 tl:Jf3 gives White a strong attack similar to Peralta-Spasov above.

17 �dfl Instead 17 l:th2 can be met by the immediate 17 ...bs. After 18 �dhl Black might try 18 ...b4 or 18 ... ii.d7, with the

G l ig o ric Va ria tio n : 7 il e3 tlJ g 4 8 il g 5 f6 9 il c 1 tlJ c6 a n d 9 il h 4 tlJ c6 idea of 19 ilg3 '1i'c8 when the pressure on g4 will deter White from playing f4 for a while. If then 20 f3 b4 21 tiJb1 as with counterplay. Instead, long after I had first drafted this section, 19 'ltbl was seen in B.Gelfand-H.Nakamura, Amsterdam 2010 (Nakamura had opted for the move order 17 ...1'.. d7 18 l:i.dhl bS), and after 19 .. �b8 20 1'.. g 3 '1i'c8! (again we see this instructive idea) 21 f3 �b7 22 'ital '1i'e7 23 f4 one way to obtain counterplay was 23 ... exf4!? 24 1'..xf4 b4 2S tiJdl as, but Nakamura preferred 23 .. �fb8 24 fxes dxes!?, maintaining a tough, unbal­ anced and approximately level middle­ game. 11 bs 1s 1'..g3 Fedorov suggests 18 tiJd1 ! ? with un­ clear play. One possibility is 18 ... b4 19 ilg3 1'.. h6 20 f4 ilxf4 21 1'..xf4 exf4 22 l:i.xf4 'ltg7 (but not 22 ...tZJes 23 'llt'b 3 ! with the idea of 'ii'h 3) 23 �hfl tZJgs and ...

Black is very solid. 18 ...1'.. h 6 19 f4 b4 20 tiJd1 exf4 21 1'..xf4 1'..xf4 22 l:txf4 'ltg7

Black has achieved everything he could hope for and stands well. After the further moves 23 �f2 l:th8 24 l:tfh2 �xh2 2 S �xh2 1'..d7 26 '1i'd3 tlJg8 27 CZJf2 tlJgh6 28 '1i'g3 '1i'e7 29 tiJf3 �e8 30 '1i'f4 as 31 �hl a4 32 1'..d 3 a3 33 b3 �h8 Black's initiative had become quite se­ rious in B.Gelfand-A.Fedorov, Batumi 1999.

1 83

Cha pter 10 Gligoric Variation 1 i.e3 lt:Jg4 s i.gs f6 g i.. h 4 gs

1 d4 tllf6 2 c4 g6 3 tll c 3 ii.g7 4 e4 d6 S tllf3 0-0 6 ii.e2 es 7 ii.e3 tllg4 8 ii.gs f6 9 ..th4 gs

A: 11 cs B: 11 h3 This somewhat ugly move really makes the g7-bishop look bad, but its counterpart on g3 will have rather lim­ ited scope as well and it may even be­ come a target. This is currently the main line of the Gligoric Variation. 10 ii.g3 tll h6 It is too late for 10 ... tll c6, because af­ ter 11 dS tlle7 12 tllxg s! tllxf2 13 ii.xf2 fxg s 14 ilg4 White is clearly better. Now White has a broad choice:

184

C: 11 dxe5

D: 11 dS A) 11 cs This aggressive move is quite tricky. 11...tllc 6!? More common is 11 ...g4 12 tllh4 tllc6 (White is clearly better after 12 ... exd4 13 '1i'xd4 and 12 .. .fs 13 tllxfs, but 12 ... dxcs with the idea of 13 dxes ii.e6!? could be considered) 13 cxd6 (13

G lig o ric Va r i a t i o n : 7 il e 3 tiJ g 4 8 ilg5 f6 9 il h 4 g5 dxes dxes 1 4 o-o reaches note 'a' to White's 12th move, below) 13 ... cxd6 14 dxes (of course after 14 ds Black can play 14...tiJd4) and now Black has a choice: a) 14 ... dxes looks fairly solid, but with the c-pawn s exchanged Black cannot control the ds-square. Indeed, White maintains a pull while Black is devoid of counterplay after 15 1'..c4+ ltih8 16 '1i'xd8 l:txd8 17 o-o tiJd4 and: al) 18 f4 gxf3 19 tiJxf3 tiJxf3+ 20 :.xf3 looks great for White. After 20.. .fs 21 exfs tiJxfs 22 �afl tiJd6 23 1'..b3 1'..g 4 24 l:i.f6! he was winning in M.Roiz­ B.Socko, Austrian League 2009. a2) 18 �adl is also good. After 18 ... 1'..d7 19 f4 l:i.ac8 20 1'..d3 tiJf7 21 fxes tiJxes 22 1'..b1 1'..e6 23 1'..f2 tiJec6 24 �fs White maintained the initiative deep into the endgame in A.Onischuk1.Smirin, World Team Championship, Bursa 2010. b) 14.. .fxes is more combative, but it's certainly ugly. Black's justification for his strange pawn structure is the floating knight on h4.

White has tried several moves in this position: bl) 15 tiJfs tiJxfs 16 exfs 1'..xfs 17 ilxg4 tiJd4 with equality in Z.Ksieski­ B.Socko, Glogow 2001. b2) 15 tiJbs a6 16 tiJxd6 'ii'a s+ 17 'lt>f1 as played in 5.Slugin-1. Belov, Rus­ sian Team Championship 2009, could be met with 17 ...tiJd4 with compensa­ tion for the pawn. b3) 15 tiJds 1'..e6 16 o-o tiJd4 17 tiJe3 �c8 18 b3 'ii' d 7 19 ilc4 ilxc4 20 tiJxc4 1'..f6 21 f4 '1i'e7 22 fxes dxes 23 tiJfs tiJhxfs 24 exfs h s left Black with the initiative in A.Poluljahov-R.Anton­ iewski, Koszalin 1999. b4) 15 o-o!? tiJd4 (worse is 1s ...1'..f 6?! 16 tiJfs tiJxfs 17 exfs 1'..xfs 18 1'.. x g4, which was pleasant for White in L.Fressinet-D.Stellwagen, European Team Championship, Novi Sad 2009, but Black should consider 1s ... 1'.. e6) 16 .�c4+ 'lth8 17 tiJe2 1'.. e6 (White also re­ tains an edge after 17 ... tiJxe2+ 18 1'.. x e2 1'..f6 19 tiJfs tiJxfs 20 exfs 1'..xfs 21 1'..xg4) 18 1'.. x e6 tiJxe6 19 '1i'd2 tiJf4 20 l:i.adl l:tc8 21 f3! and White had the initiative in A.Giri-M.Vachier Lagrave, Biel 2010. bS) 15 1'..c4+ is the most obvious move and is probably best: 1s ... 'lth8 16 '1i'd2 tiJd4 17 o-o (White also kept an edge after 17 tiJe2 tiJxe2 18 1'..xe2 1'.. e6 19 o-o '1i'e7 20 b3 �ad8 21 l:i.adl in A.Onischuk-E.lnarkiev, Poikovsky 2009) 17 ... 1'.. d7 (this is Black's latest try; 11 ...'1i'e7 18 b3 'ii'as 19 �fd1 1'..d7 20 tiJe2 '1i'xd2 21 l:i.xd2 also left White with

1 85

A ttacking Ch ess: Th e King 's I n dia n , Vol u m e 1 a slight edge in E.L'Ami-G.Agamaliev, Hoogeveen 2008) 18 tlle2 �c8 19 l:tacl �xc4 (Black tries to solve his problems tactically, but White's positional ad­ vantage persists) 20 �xc4 tllxe2+ 21 '1i'xe2 .tbs and Black will win back the exchange. However, White is still better: 22 �dl (more complicated but still favour­ ing White is 22 '1i'c2 ii.f6 23 tllfs tllxfs 24 exfs h S 2 s f3 gxf3 26 l:txf3 ii.xc4 27 '1i'xc4 dS 28 '1i'e2 e4 29 �fl, S.Slugin­ Lyaskovsky, Russian Team Champion­ ship 2010) 22 ...'1i'd7 23 �dcl '1i'e6 24 a4 Lc4 2S '1i'xc4 '1i'xc4 26 J:1xc4 ii.f6 27 'ltfl and White eventually won the ending in S.Volkov-M.Krylov, Moscow 2010. 12 cxd6 After 12 dxes dxes we reach a posi­ tion that is often reached via the move order 11 dxes dxes 12 cs tllc6, al­ though in that case 12 ... ii.e6 is another viable option for Black. Here Black's knight is already on c6, but he can still hold the balance.

a) 13 o-o g4 14 tllh4 ii.e6 lS '1i'a4 '1i'd4! gave Black good play in V.lvanchuk-T.Radjabov, Odessa (rapid) 2007. b) 13 ilc4+ 'lth8 14 '1i'xd8 tllxd8 lS o-o-o (lS tllb s ii.e6 was also satisfac­ tory for Black in the game Y.Shulman1.Umanskaya, St Petersburg 1994) 1S ... tlle6 16 b4 as 17 a3 axb4 18 axb4 c6 19 'ltb2 tllf4 20 ii.xf4 exf4 21 'ltb3 1/2-1/2, P.San Segundo Carrillo-A.Romero Holmes, Vendrell 1996. c) 13 '1i'xd8 is the most dangerous move according to Bologan, who now gives 13 ... �xd8! (rather than 13 ... tllxd8 14 tllb s �f7 lS o-o-o ii.e6 16 h4 g4 17 tllg S! when White was much better in V.Grabliauskas-C.Krogh, Ringsted 1992; here 1S ... tlle6 16 ilc4 is unpleasant but may be a better try) 14 tlld s g4 l S tlld2 tlld4 16 tllxe7 tllc2+ 17 'ltdl tllxal 18 ilc4+ (no better is 18 tllxa8 ii.e6) 18 ... tllf7 19 tllx a8 ii.h6 20 ii.ds ii.d7 21 tllc7 llc8 22 tlle6 ii.xe6 23 ii.xe6 �xcs when Black has equal chances in a complicated endgame. Note that 24 ii.xg4? fails to 24...�c2 with the initia­ tive. 12 cxd6 13 dxes Instead 13 dS just helps Black for­ mulate a plan: a) 13 ... tlle7 14 o-o tllg 6 lS �el fS 16 exfs tllxfs 17 tll d 2 tllf4 18 ii.fl g4 19 �cl h S ! and Black had already seized the initiative in E.Levin-S.Klimov, St Pe­ tersburg 2010. b) 13 ... tlld4!? is also possible: 14 tllxd4 exd4 l S tllb s (not l S '1i'xd4 fS) ...

White has:

186

Glig o ric Varia tio n : 7 il e 3 tlJ g 4 8 ilg5 f6 9 il h 4 g5 lS.. .fS! 16 .i.xd6 fxe4 (worse is 1 6. . .l:i.e8 17 es!) 17 1'.. xf8 '1i'xf8 with some com­ pensation. Here 18 tlJxd4? fails to 18 ...'llt'b 4+. 13 ...dxes

14 1'..c4+ White should seize the diagonal while he can. Black had no probl ems after 14 o-o 1'.. e 6 in P.H.Nielsen­ E.Mortensen, Festuge 1991. 14...'lthB This is better than 14... CZJf7 lS h3! when Black is in a pin and has commit­ ted his knight too early (but lS o-o ilg4 successfully fights for the d4-square). 15 h3 White prevents ...1'..g4, so Black must find a way to create some coun­ terplay. 1s ... tlJd4 16 o-o g4! 17 tlJxd4 exd4 18 tZJds Instead 18 tlJe2 gxh 3 19 '1i'xd4 was P.San Segundo Carrillo-M.Pavlovic, Ubeda 1996, and now the simplest is 19 ...'1i'xd4 20 tlJxd4 fS with counterplay. 18 ... gxh3 19 '1i'xd4 hxg2 20 �fdl The position is very complicated. In-

stead 20 'ltxg2 looks critical, but Black has enough play after 20 .. .fs. One pos­ sible line is 21 es f4 22 Lf4 1'.. e 6 23 ilg3 CZJfs 24 '1i'e4 tZJxg3 2s fxg3 'ii'g s 26 �xf8+ �xf8 27 �el bS! 28 il b3 (not 28 1'.. x bs '1i'd2+ 29 �e2 .i.xds) 28 ...'1i'd2+ 29 �e2 il fS! 30 '1i'e3 (30 �xd2 1'..xe4+ with the idea of ...1'.. x es is equal) 30 ...'1i'xe3 31 �xe3 �e8 32 tiJC7 �xes 33 �xes 1'.. x es 34 tlJxbs and with such limited material on the board, Black's good bishops should ensure a draw. 20 ...1'..g4 21 'ltxg2 White sacrifices an exchange to fight for the initiative. After 21 �d3 or 21 l:td2 Black has counterplay with 21 ... fs. 21... .i.xdl 22 l:i.xdl fS 23 es

The position is unclear, but fairly balanced. White has a strong central presence, but his king is not completely comfortable and Black is up an ex­ change for a pawn. 2 3 ...f4!? other interesting possibilities are 23 ... �g8, 23 ...�c8 and 23 ...tlJg4. 24 CZJxf4

187

A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1 Instead 24 ii.xf4 '1i'h4 2 S ii.g3 '1i'xd4 26 �xd4 tllfs is similar. 24...'1i'xd4 2 s �xd4 tllfs 26 �e4 tllxg3 Black could also play 26 ... �ac8. 21 'ltxg3 z:tfs 28 tlld3 �gs+ 29 'lth2 �fB 30 f4 l:tg6 31 .tbs �dB 32 ii.c4 l:tfs 33 ii.bs z:tds 34 ii.c4 Now Black declined a repetition of moves against his higher-rated oppo­ nent by playing 34... a6! ? in J.Gustafsson-A.Shimanov, European Championship, Rijeka 2010, and even­ tually went on to win. B) 11 h3

This move looks a little funny be­ cause Black usually voluntarily retreats his knight in any case, but it prevents ... ilg4 and the advance ... g4 is discour­ aged because it would open the h-file for White. 11 ...tllc6 Both 11 ... exd4 and 11 ... tt:Jd7 are pos­ sible as well, but the text move fits in best with our repertoire. 12 dS White has an alternative in 12 dxes

1 88

dxes (I do not like 12 .. .fxes 13 cs!: for example, 13 ... dxcs 14 '1i'xd8 followed by lS tllx gs) and now: a) 13 '1i'xd8 tllxd8 14 tllds tlle6 lS h4 (otherwise, lS tlle 7+ 'ltf7 16 tllxc8 l:taxc8 is level, while lS o-o-o l:i.e8 16 h4, as in G.Timoscenko-E.Hagara, Slo­ vakian League 2008, could be met with 16 ... g4, intending ... c6 and then either ... tlld4 OT... tllc s) 1S ... g4 16 tllh 2 tlld4 (16 .. .fs also gives Black counterplay) 17 tllxo �b8 18 ii.d1 fs 19 exfs tllhxfs 20 tllxg4 tllx g3 21 fxg3 e4! and Black had the initiative in S.Emst-D.Stellwagen, Dutch Championship, Leeuwarden 200S. b) 13 o-o ii.e6 (the solid 13 ...'ii'xdl 14 l:i.fxdl ii.e6 also looks playable) 14 '1i'c1 '1i'e8 lS tllds l:i.c8 16 b4 (Yer­ molinsky instead suggests 16 !tel g4 17 hxg4 ii.xg4 18 tlle 3 ii.xf3 19 ii.xf3 tlld4 20 ii.dl '1i'g6 21 ii.c2 which looks unclear) 16 ... g4! 17 hxg4 ii.xg4 18 bS ii.xf3 19 ii.xf3 tlld 4 20 '1i'a3 'lth8 21 l:i.adl (or 21 '1i'xa7 fS 22 exfs tllhxfs with counterplay) 21 -.fS and Black had kingside counterplay in A.Yermolinsky­ J.Fedorowicz, US Championship, Los Angeles 1991 . 12 ... tlle 7 The speculative 12 ...tlld4 13 tllxd4 exd4 14 '1i'xd4 fS lS '1i'd2 f4 16 ii.h2 tllf7 has even been tried by Kasparov (in a rapid game), but it is probably objectively unsound.

(seefollowing diagram)

Gligoric Varia t i o n : 7 11.. e 3 tlJ g 4 8 11.. g s f6 9 1'.. h4 g5

13 '1i'd2!? An independent try. After 13 tiJd2 we tran spose to 9 ... tlJc6 10 h3 tiJh6 11 dS tlJe7 12 tiJd2 gs 13 1'..g3 which was discussed in note 'a2' to White's 10th move in Line B of the previous chapter. 13 ... tiJf7 After 13 ...tlJg6 ! ? 14 o-o-o White has scored well, but following 14... a6 1S 'iit b 1 1'.. d 7 the position is not clear. 14 tiJh2 Or 14 1'..h 2 cs 1S 1'..g 1 a6 16 tiJh 2 fS 17 CZJf1 which looks rather comical.

White managed to win. Instead the flexible 17 ... 1'..d7 is natural, while 17 ...l:tb8 and 17 ...'ii'a s!? are also possible. 14 ...f5 15 exf5 After 1S f3 Black could force the pace with 1S .. .f4 or maintain the ten­ sion with 1s ... tlJg6. 15 ...tZJxf5 16 o-o tiJh8 Black repositions his knight. Similar is 16 ...tlJd4 17 1'.. g 4 tiJh8 when 18 tlJe2 CZJfs 19 11..h s was S.Shipov-V.Golod, Minsk 1993, and here Black could con­ sider 19 ... tlJg6 20 1'.. xg6 hxg6, intending ... e4! ? with counterplay. 11 1'.. h 5 CZJd4 Black does not really have to rush this move and might prefer 17 ... 1'.. d7 ! ?. 18 tlJe4 White could also consider 18 tZJe 2. Then 18 ...tZJfs would leave Black a tempo down on the note to his 16th move, or he could play 18 ... tZJxe2+ 19 '1i'xe2 11..fs, intending ... tlJg6 and maybe even ... e4. White has a grip on the light squares, but his g3-bishop is not much better than its g7-counterpart. 18 ... h6 19 �fe1 1'..f5 20 �acl tlJg6

Now 17 ...fxe4 18 1'..g 4 CZJfs 19 tZJxe4 bS looked okay for Black in S.Volkov­ A.Shomoev, Ulan Ude 2009, although

189

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1 Now White played 21 cs? dxcs 22 l:i.xcs b6 and Black won a pawn in U.Boensch-B.Gelfand, Dortmund 1990. Instead 21 ii.xg6 ii.xg6 leaves Black better with his bishop-pair, so White should settle for 21 ii.g4 with approxi­ mate equality. C) 11 dxes White exchanges in the centre, which often leads to an exchange of queens. 11 ...dxes

12th in Line A, but here Black can also consider 12 ...ii.e6 ! ? covering the a2-g8 diagonal. A sharp try is 12 h4!?, but Black has enough resources: 12 ... g4 13 tllh 2 tllc6 14 '1i'xd8 �xd8 1s tlld s tlld4! (Black has serious counterplay) 16 l:i.c1 (instead 16 ii.d1 fS was Li Wenliang-F.Nijboer, Groningen 2002, but 16 ...ii.d7 ! ? looks simple and good) 16 .. .fS! ? (or just 16 ... c6 17 tlle7+ 'ltf7 18 tllxc8 tllxe2 19 'ltxe2 .!:i.axc8 with equality) 17 exfs (17 tllxa �b8 18 exfs tllhxfs also gives Black good play) 17 ...tllhxfs 18 ii.xg4 c6 19 tlle3 tllx g3 20 fxg3 ii.h6 and Black had the initiative in l.Gulkov­ S.Dyachkov, Voronezh 1999. Cl) 12 '1i'xd8 �xd8

Now White has two main continua­ tions. He can exchange queens imme­ diately with 12 '1i'xd8 or play a little finesse first by going 12 'ii'd s+. C1: U �8

cz� u trds+ Instead 12 h3 tllc6 transposes to the notes to White's 12th in Line B, while 12 o-o c6 is level and 12 '1i'b3 does not achieve much After 12 ... c6 White can­ not castle long. Note too that 12 cs tllc6 was covered in the notes to White's

190

13 tlld s This allows White to gain the bishop-pair, but that does not trouble Black Instead 13 tlld2 ii.e6 14 o-o-o tlla6 1S f3 c6 16 tllb3 tllf7 17 ii.f2 ii.f8 18 �xd8 tllxd8 19 l:i.d1 b6 is equal and after 20 tll a1 tllb7 21 tll c2 �d8 22 l:i.xd8 tllxd8 the players agreed to a draw in

G l ig o ri c Va ria tion : 7 ile3 tlJ g4 8 ilg5 f6 9 il h 4 g5 V.lvanchuk-V.Bologan, Foros 2006. 13 ...tlJa6 More ambitious is 13 ... �d7, but this is obviously riskier. Black clogs up his own development and the g4-pawn is weakened. Moreover, he should not fear the loss of his light-squared bishop because he will get counterplay on the dark squares. 14 0R7+ Instead 14 o-o-o 1'..e6 is similar to lvanchuk-Bologan, above.

14 '>tf7! Black's king will either have to take a square away from his knight (on f7) or his bishop (on f8). From a positional standpoint, it makes more sense to go to f7 because the h6-knight will not necessarily be in a hurry to go to f7, but the g7-bishop will almost certainly want to go to f8. There is also a tactical point, as we shall see. That said, 14...'lt>f8 is also playable: 15 tZJxc8 �axc8 16 a3 tZJcs 17 tiJd2 �d4 (instead Black should play 17 ... �d7 with equality) 18 f3 l:i.cd8 19 1'..f2 ! (but not 19 �dl? tlJa4!) gave White a slight .••

advantage in G.Timoscenko-E.Pinter, Hlohovec 1998. The point is th at 19 ...l:i.xd2 20 1'..x cs+ is check. 15 tZJxc8 �axes 16 �cl White wants to safeguard his queenside. The point of Black's 14th move i s seen in the line 16 a3?! tZJcs 17 tiJd2 l:i.d4 18 f3 �cd8 19 �dl (after 19 1'..f2 �xd2 20 1'..x cs it's not check and Black can just play 20 ... �xb2) 19 ... tlJa4! when White's position falls apart. White tried 16 �dl tiJcS 17 tiJd2 as 18 b3 1'..f8 19 o-o in R.Akesson-G.Jones, Klaksvik 2008, and now 19 ... �d4 20 f3 �cd8 2 1 tiJbl tlJe6 would be good for Black because of his grip on the dark squares. 16 ... tZJcs The immediate 16 ...1'..f8 also looks very good. The bishop threatens a dis­ ruptive check on b4. Now 17 0-0 would be met by 17 ... g4 18 tiJel �d2, while 17 a3 tZJcs also favours Black. 11 tiJd2 as 18 f3 .i.fB 19 1'..f2 �d7

Black is ready to double rooks on the d-file and to increase his grip on the dark squares with ... b6, ... tlJe6 and

191

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1 ... ii.cs. White cannot even castle, so he tries to stir up some counterplay. 20 h4 g4! 21 ..11i. e3 Black has a strong initiative after 21 fxg4 l:i.cd8 22 �c2 tlld 3+ 23 ..11i.xd3 l:i.xd3 or 22 ..txcs ..txcs. 21 �cdB 22 �c2 gxf3 23 gxf3 b6 24 �gl Now Black erred with 24 ... tlle 6? 2S cs! which unleashed White's light­ squared bishop in M.Roiz-1.Smirin, Tel Aviv 2002. Instead 24 ... tlld 3+ 2S ..11i. x d3 �xd3 26 tllfl and here either 26...tllg 8 or 26 ...'lte6 are comfortable for Black, while another idea is 24... tllg8, aiming to bring the knight to d4. After 2S tllfl tlle7 Black need not fear 26 ii.xcs bxcs followed by ...tllc6-d4 and ... ..11i. h6 with a pleasant position for him. •..

C2) 12 'ii' d S+

This is White's most ambitious try. Now Black does not want to exchange queens himself, because after 12 ... 'ii' xds 13 cxds White will have pressure on the c-file.

192

12 'lthB Another possibility is 12 ... tllf7 13 h4 (or 13 o-o-o 'ii' e7 14 h4 c6 lS 'ii'a s tlla6 16 hxg s fxgs, reaching the same posi­ tion) 13 ... '1i'e7! (13 ... g4 14 tllh 2 hS lS f3 wins a pawn) 14 hxgs fxgs lS o-o-o c6 16 'ii'a s tlla6 17 tlle1 tllcs 18 '1i'a3 �e8 with unclear play. 13 0-0-0 Instead 13 cs?! c6 14 '1i'xd8 �xd8 lS tlld2 ..11i.f8 16 tlla4 was V.Korchnoi­ J. Fedorowicz, World Team Champion­ ship, Lucerne 1989, and now 16 ...�d4! 17 b3 tlld7 18 .!:!.cl bS! 19 cxb6 ..11i.b4! 20 l:tc2 axb6 is much better for Black, as pointed out by Mikhalevski. White can also exchange queens with 13 '1i'xd8 l:txd8, although this should not be dangerous: a) 14 tlld2 cs!? (or 14 ... ..11i. e6) lS f3 tllc6 16 o-o-o tlld4 17 ..11i. d3 ..11i.e6 18 tllb3 �ac8 was equal in V.lvanchuk­ T.Radjabov, Sochi 2007. b) 14 tlld s tlla6 1s o-o-o (White cannot even grab the bishop-pair here because, of course, lS tlle7 is not check) 1s ... ..11i. e 6 16 tlld2 (16 a3 c6 17 tllc3 tllcs 18 b4 tllb 3+ 19 'ltc2 tlld4+ 20 tllxd4 exd4 21 tlla4 was S.Volkov-A.Shomoev, Krasnoyarsk 2007, and here 21...b6 looks best, while Black can vary earlier as well, perhaps with 17 ... tlle7! ? or 18 ... tlld 7) 16 ... �d7 17 f3 c6 18 tllc3 ..11i.f8 19 tllb3 b6 20 �xd7 ii.xd7 21 ..tf2 ..11i. e6 22 �dl tllf7 and the position was bal­ anced in E.Perelshteyn-M.Beelby, Tulsa 2008. c) 14 h4 g4 lS tlld2 and now •••

G lig o ric Va ria ti o n : 7 il e 3 tlJg4 8 il gs f6 9 il h 4 g5 15 ... cs!? is possible, but in practice Black has preferred to develop his queen knight:

tZJbs (22 'Itel 1'.. h 6 23 �dl �xe4 24 1'..fl 1'..d7 looks fine for Black) 22 ...�d8 23 'Itel 1'..h 6 24 l:i.dl a6 25 tlJc3 �d4 with a strange situation. The al-knight is trapped, but White can hardly win it. He may exchange it indirectly with 26 'ltbl, but then 26 ... .i.xd2 27 'ltxal?! ilxc3 28 bxc3 l:i.xe4 29 �d8+ 'ltg 7 cer­ tainly favours Black 13 '1i'e7 ••.

cl) 1S ... tlJa6 16 o-o-o 1'.. e6 17 CZJb3 b6! (the inferior 17 ... c6?! 18 tZJas �xdl+ 19 l:txdl �b8 20 a3 gave White an edge in A.Onischuk-A.Shomoev, Sochi 2007, but Mikhalevski's 17 ...CZJf7 18 tZJas l.h6+ 19 'ltbl tiJd6 looks solid enough) 18 f4 gxf3 19 gxf3 c6 20 'ltbl CZJf7 21 f4 exf4 22 1'..xf4 fs and Black had enough play in V.Bhat- B.Sambuev, Montreal 2009. c2) 1S ... tlJc6!? (Black logically takes aim at the d4-square, but there are some tactical issues) 16 tZJds (16 CZJb3 is possible as well, but the text move is certainly critical) 16 ...tlJd4!? (going for it, but 16 ... l:td7 17 CZJb3 �f7 is safer) 17 !Uxo tZJc2+ (not 17 .. Jlb8 18 �cl with a healthy extra pawn) 18 'ltdl tlJxal 19 tlJxa8 CZJfs! (a clever resource) 20 tZJo (not 20 exfs 1'..h 6!) was played in 5.Feller-M.Al Modiahki, Paris 2010. Now Black could justify his play with 20 ... tZJxg3 21 fxg3 �d4! (but not 21...1'..h6 22 tZJds, blocking the d-file) 22

14 'ii' as Slow is 14 h3 and 14... tlJa6 15 'ii' as c6 was fine for Black in P.Wells-L.Kritz, Biel 2004. Instead 14 h4 g4 15 tiJel tlJc6 16 tZJc2 fs 17 exfs was K.Sundararajan­ R.Shetty, Visakhapatnam 2004, and here the obvious 17 ... CZJxfs gives Black counterplay. 14 ...c6 After the natural 14... tlJc6 White has 15 tZJds ! tZJxas 16 tZJxe7 1'..e6 17 tZJds l:tac8 18 tiJd2 with an edge, E.lturri zaga-5.Feller, Moscow 2010. 15 tiJd2 b6 16 'ii' a4 1'.. d 7 17 f3 cs!? Black gives up the ds-square, but he will get d4 for his knight in return.

193

A ttacking C he s s : Th e King 's In dian, Volu m e 1 18 '1i'a3 tllc6 19 tlld s •t1 20 ii.d3 ii.e6

Black may break immediately or he can develop first in order to better fight for the e4-square. D1: 11...f5

02: 11....�7

D1) 11 ...fs

12

exfs

The position was unclear in A.Motylev-T.Radjabov, Wijk aan Zee 2007. Black's kingside set-up may look funny, but he has possibilities to initi­ ate play with ...f5 or ...b5. D) 11 d 5

Closing the position is White's most popular option. After Black breaks with .. .f5 White will gain the e4-square for his pieces. Black will have to either fight for control of this square or gen­ erate enough counterplay to offset White's grip on the light squares. Here a decision must be made.

194

12...g4!? This invention of Kasimdzhanov commits Black to a pawn sacrifice. The alternative is 12 ... tllxf5 (but not 12 ... ..txf5?! 13 h4 g4 14 tllg 5 ! ) 13 tlld2 tlld4 and here: a) After 14 tlld e4 g4! (intending ...h 5 with counterplay) White could still play 15 ii.xg4 ii.xg4 16 '1i'xg4 tllc2+ 17 'ltd2 tllxal 18 �xal with some compensa­ tion, although Black is probably not worse. b) 14 ii.d3 was suggested by Bolo­ gan. After 14 ... ii.f5 (?!) 15 ii.xf5 l:txf5 16 tllde4 White is better, but Black should prefer 14... tlla6, as in E.Karavade­ T.Abrahamyan, Yerevan 2006. c) 14 o-o tllx e2+ (14...ii.f5 15 tllde4 tlld7 16 ilg4 was a little better for

G l ig o ric Varia tio n : 7 ile3 tb g 4 8 il g5 f6 9 i.. h4 g5 White in S.Gross-V.Golod, Ceske Bude­ jovice 1993) lS '1i'xe2 '1i'e8 16 tl:Jde4 1t'g6 17 f3 ilfs (instead 17 ... tl:Ja6 18 a3 g4 19 b4 keeps the knight sidelined) was Ki.Georgiev-V.Bologan, Gibraltar 2006. Here White could play 18 cs im­ mediately with the initiative. 13 tl:Jd2 After 13 tl:Jh4? Black wins a piece with 13 ... 1'..f6. White does not obtain enough compensation after 14 h3 ilxh4 lS hxg4 1'..x g3 16 fxg3 tl:Jf7 with the idea of ... 0,g s. 13 ... 1'..xfs

14 tl:Jfl This move wins a pawn and is therefore critical. 14 tl:Jde4 is less dan­ gerous, leading to 14 ... tl:Jd7 and then: a) lS 1'..d 3 'lth8 (1s ...'1i' e8? !) 16 o-o as 17 '1i'e1 1'..g 6 18 '1i'e2 '1i'e7 was about equal in L.Gofshtein-R.Kasimdzhanov, Hoogeveen 1999. b) lS 0-0, after which the following play is very thematic: 1s ...'1i'e8 16 1'..d3 '1i'g 6 17 '1i'c2 'lth8 18 �adl a6 19 b4 tl:Jf6 20 tl:Jxf6 '1i'xf6 2 1 tl:Je4 1'..x e4! (a typical idea; Black will target the g3-bishop) 22

ilxe4 tbfs 23 '1i'd3 hS 24 f3 (or 24 cs as! with counterplay) 24 ... tl:Jxg3 2S hxg3 'ii'g s 26 'ltf2 �f6 (simpler is 26...1'..h 6) 27 '1i'e3 '1i'xe3+ 28 'ltxe3 1'..h 6+ 29 'ltf2 'ltg7 30 �h l �h8 31 'lte2 1'..g s?! (Black can equalize with 31 ... gxf3+ and then 32 gxf3 iJ..g S OT 32 iJ..xf3 aS!) 32 fxg4 hxg4 33 �xh8 'ltxh8 34 a4 'ltg7 3S as and White kept an edge in R.Hungaski­ D.Lemos, Riobamba 2007. 14...tl:Jd7 Another idea is 14 ... 'ii'g s lS tl:Je3 i..d7 16 tl:Je4 '1i'g6 17 1'.. d3 tl:Jfs with counterplay. Here Huzman gives 18 tl:Jxg4?! tl:Jxg3 (or 18...hs!?) 19 tl:Jef6+ ilxf6 20 1'..x g6 tl:Jxhl 21 1'..fs ! i..g 7! 22 1'..e6+ (Black is also doing well after 22 ilxd7 tbxd7 23 '1i'e2 hS 24 o-o-o �f4!) 22 ...1'.. xe6 23 dxe6 tl:Jc6 24 'ii'd s hS 2S e7+ �f7 and because White will not be able to round up the hl-knight, Black is much better. 15 tl:Je3

1s ...e4 Th is looks like the best move order. Instead 1s ... tl:Jcs can be played when 16 o-o e4 transposes to the main line, but

195

A ttacking C he s s : The Kin g 's In dian, Vo l u m e 1 White can also try: a) 16 b4 e4 17 �cl tlld 3+ 18 ii.xd3 exd3 19 tllxfs tllxfs gives Black some counterplay: for example, 20 o-o (much worse is 20 '1i'xg4? '1i'f6 with the idea of 21 o-o d2 22 l:i.c2 tlle 3) 20 ... tt:Jd4 (other possibilities are 20 ... h s and 20 ...'1i'f6) 21 '1i'xd3 hS, targeting the g3-bishop. b) 16 tllxfs tllxfs 17 ii.xg4 e4 and now: bl) 18 ii.xfs ii.xc3+ (White is much better after 18 ... tlld 3+ 19 'ltfl l:i.xfs 20 '1i'g4 '1i'f6 21 tllxe4 '1i'xb2 22 �dl '1i'c2 23 '1i'e2 '1i'xe2+ 24 'ltxe2) 19 bxc3 l:txfs (Black's compensation looks insuffi­ cient after 19 ... tlld 3+ 20 'ltfl l:txfs 21 '1i'g4+ 'ii'g s 22 'ii'x gs+ �xgs 23 'lte2) 20 o-o 'ii'g s with some compensation. b2) 18 o-o tllx g3 (Huzman prefers 18 ... ii.xc3 19 bxc3 'ii'g s with some compensation) 19 hxg3 ii.xc3 20 bxc3 'ii'g s 21 ii.e6+ tllx e6 22 dxe6 �fs 23 '1i'd4 was agreed drawn in A.Huzman­ B.Avrukh, Pula 2000. Instead of 23 '1i'd4, Huzman suggested 23 'ii' bl. This has since been seen in practice and al­ though White eventually won after 23 ... �es 24 �ael 'ii'fs 2s e7 �xe7 26 f3 l:tae8 27 fxe4 '1i'g6?! 28 'ii' ds+ l:te6 29 '1i'xb7 in H.Grooten-N.Vanderhallen, Vlissingen 2006, Black could have maintained the balance with the move 27 ...'ii'g s ! . 16 o-o tllc s Also interesting is 16 ...ii.xc3 17 bxc3 'ii'g s with an unclear position in F.Atakisi-N.Gavrilakis, correspondence 2006.

196

17 l:tcl White lends the c3-knight some ex­ tra protection and creates the possibil­ ity of playing b2-b4. Others: a) 17 ii.xg4 is untried, but Black should have enough for the pawn after 17 ... ii.xc3 18 bxc3 tllxg4 (or perhaps 18 ... ii.xg4 19 tllxg4 'ii'g s) 19 tllxfs (19 tllxg4 'ii'g s 20 tlle 3 tlld3 gives Black counterplay) 19 ... l:txfs 20 '1i'xg4+ 'ii'g s. b) 17 tllxg4 ii.xg4 18 ii.xg4 ii.xc3 19 bxc3 '1i'f6 gives Black some compensa­ tion. After 20 ii.e6+ (or 20 .!:!.cl tllxg4 21 '1i'xg4+ '1i'g6) 20 ... tllxe6 21 dxe6 '1i'xe6 22 cs tllfs 23 '1i'g4+ 'lth8 24 cxd6 cxd6 2S l:i.fel the game was agreed drawn in R.Markus-N.Djukic, Herceg Novi 200S; 2s ... l:tae8 would be unbalanced, but level. c) 17 '1i'd2 'ii' gs (or 17 ...ii.d7 !? 18 b4 tlla4 forcing White to make a reason­ able exchange sacrifice) 18 l:i.adl a6 19 b4 tlld7 20 cs tlle s (not 20...dxcs 21 bxcs tllxcs 22 ..txa) 21 ii.xes ii.xes 22 tllxfs 'ii'xfs (and not 22 ...'1i'xd2 23 l:txd2 tllxfs 24 tllxe4 when White's extra pawn means more than it does in the game)

G lig oric Va riati o n : 7 il e 3 tlJ g 4 8 ilg5 f6 9 il h 4 g5 23 '1i'xh6 1'..xc3 24 cxd6 �f6! 2S '1i'h4 (White really had to try 2S '1i'e3 or 2S '1i'c1) 2s...cxd6 26 1'..xg4 'ii'e s. Now Black held without much difficulty against his higher rated opponent: 27 f4 exf3 28 J:1xf3 �xf3 29 ilxf3 �f8 30 'lthl '1i'f4 31 '1i'xf4 l1xf4 32 a3 1/2-1/2, A.Huzman1.Sidorenko, Ramat Aviv 2004. 11 ...as Black secures the cs-knight, but also possible is Mikhalevski's 17 ...'1i'd7!? to cover fS and g4. After 18 b4 (18 �el gives Black time to play 18 ... as and 18 ilf4 CZJd3! 19 CZJxfs! 'ii'xfs 20 1'..xh6 1'.. xh6 21 1'.. xg4 '1i'g6 22 l:tc2 e3! gives him counterplay according to Mik­ halevski) 18 ... CZJd3 19 .i.xd3 exd3 20 CZJxfs 'ii'xfs looks okay for Black. 18 tlJxg4 tlJxg4 19 1'.. xg4 'ii'g s 20 1'.. xfs l:i.xfs 21 '1i'c2 e3 22 �eel exf2+ 23 .i.xf2 .:!.af8

24...tlJxe4 25 '1i'xe4 1'.. es Black threatens ... 1'..xh2+. Instead 2s ... 1'..xb2 26 1'.. e3 l:txfl+ 27 l:txfl �xfl+ 28 'ltxfl is also possible and after 28 ... 'ii'hs Black should hold. 26 1'.. e3 l:i.xfl+? This is a mistake. Instead 26 ...'ii'h S! holds after 27 l:txfs (no better are 27 g 3 l:txfl+ 2 8 l:i.xfl �xfl+ 29 'ltxfl '1i'xh2 and 27 h3 �xfl+ 28 �xfl 1'..h2+ 29 'ltxh2 l:txf1) 27 ... 1'.. x h2+. 27 l:txf1 �xfl+ 28 'ltxf1 'ii' h s 29 h3 1'.. xb2 30 '1i'e6+ 'ltg7 31 '1i'd7+ '1i'f7+ 32 'ii'xf7+ 'ltxf7 3 3 a 4 This was M.Roiz-T.Radjabov, Saint Vin­ cent 200S. White has a clear advantage in the bishop ending because Black's pawns are fixed on dark squares and White can create an outside passed pawn on the kingside . D2) 11 ...tiJd7

24 tlJe4!? White return s the pawn to neutral­ ize Black's activity. 24 .lii.xcs is also pos­ sible, but after 24 ... dxcs 2s �xfs �xfs it will not be easy for White to make pro­ gress.

Black makes this move in prepara­ tion for .. .fs, so that the knight can quickly come to f6 to fight for the e4square. 12 tiJd2 fs 13 exfs

197

A ttacking C h es s : The Kin g 's I ndian, Vo l u m e 1 The alternative is 13 f3 tllf6 (intend­ ing ... g4) 14 h3.

Here Black has: a) 14...tllh S lS ii.f2 tllf4 16 g3 tllxe2 17 '1i'xe2 cs (17 .. .fxe4!? looks better, as after 18 tlldxe4 Black can activate his knight with 18...tllfs) 18 g4! f4 was B.Gelfand-1.Nataf, Cap d'Agde 2002. Here 19 h4 gives White an obvious ad­ vantage. b) 14 ... cs lS ..tf2 (White's play has been too slow to consider lS dxc6?! bxc6 with ...ds coming) 1S .. .f4 16 �bl (White tried the slower 16 a3 in E.Bacrot-T.Radjabov, Morelia/Linares 2006) 16 ...b6 17 b4 cxb4 18 �xb4 tlld7 19 a4 tllcs 20 as ii.d7 was solid for Black in B.Gelfand-T.Radjabov, Turin Olympiad 2006. c) With all of White's pawn moves, it is logical to open the position with 14 .. .fxe4 and Black has scored well with this move: lS tlldxe4 (after lS fxe4 Black can play 1S ... g4 or 1 S ... tlld7, intending ...tll cs and perhaps ...l:tf4!?) 1s ...tllfs (Black can also play 1s ...tllxe4 immediately when 16 tllxe4 tllfs trans-

198

poses and 1 6 fxe4 can be met with 16 ... g4 17 hxg4 tllx g4! 18 ii.xg4 'ii'g s) 16 ii.f2 tllxe4! (this weakens White's con­ trol of dS) 17 tllxe4 (17 fxe4 tlld4) 17 ... cs 18 '1i'd2 (18 dxc6 bxc6 19 '1i'd2 dS 20 cxds cxds 21 tllcs tlld4 22 ii.xd4 exd4 is unclear according to Bologan) 18 ... h6 19 o-o-o (or 19 o-o tll d4 with a good position) 19 ... a6 20 g4 tllh4! (20...tt:Jd4 21 h4!) 21 ii.xh4 gxh4 22 ii.d3 (better is 22 gs with complications) 22 ...�f4! 23 'ltb1 l:i.b8 (or 23 ... bs!? im­ mediately) 24 '1i'e2 '1i'e7 2S a3 ii.d7 26 'lta2 bS and Black had the initiative in A.Huzman-V.Bologan, Moscow 2006. Moreover, Black can even improve on this with Bologan's 19 ... bS!? 20 cxbs a6 21 bxa6 (21 b6 �b8 22 'ltbl '1i'xb6 23 h4 gxh4) 21 .....txa6 22 ii.xa6 �xa6 23 'ltbl '1i'a8 24 tllc3 �b8 with a strong initia­ tive. 13 tllf6 ...

This is the point of Black's play. 14 tllde4 Instead 14 o-o ii.xfs is not danger­ ous, but White can play 14 ii.d3 tllxfs lS tllde4 and here:

Gligoric Va ria tio n : 7 il e 3 tlJ g 4 8 il g 5 f6 9 il h 4 g 5 a) 1S ...1'.. h6 1 6 o-o 'lth8 17 cs g4 18 tiJxf6 '1i'xf6 19 tZJbs '1i'e7 20 '1i'e2 and then: al) 20... tZJxg3 21 fxg3 �xfl+ 22 .!:i.xfl dxcs 23 '1i'e4! threatens 24 d6! and gives White a winning attack a2) 20 ... 1'.. g7 21 cxd6 cxd6 22 '1i'xg4 tlJxg3 23 '1i'xg3 1'.. d7 was V.Kramnik­ T.Radjabov, Wijk aan Zee 2007. After 24 '1i'e3 ! Black would not have enough for the pawn. a3) 20 ... a6 21 1'..xfs �xfs (not 21 ... .i.xfs 22 cxd6 cxd6 23 tlJxd6) 22 tiJc3 �f8 23 tlJe4 'ltg8! 24 '1i'c4 1'..f s is okay for Black according to Bologan. b) 1S ...tZJxe4 16 tZJxe4 tiJh4! 17 o-o (or 17 ilxh4 gxh4 with the idea of ...1'..fs) 17 ... g4 18 0.d2 (if 18 cs then 18 ... tlJg6 intending ... tiJf4 or ...h s) 18 ...'ii' g s 19 tlJe4 '1i'e7 20 tiJd2 h S ! ? 21 f3 CZJfs 22 1'..f2 1'..h6 23 '1i'e2 '1i'g7 24 1'..xfs 1'..xfs 2S 1'..e3 l:tf7 (Bologan suggests 2s ... 1'..f4!?) 26 �ael l:i.af8 27 1'.. xh6 '1i'xh6 28 '1i'e3 '1i'xe3+ 29 �xe3 gxf3 30 l:tfxf3 was Wang Yue-V.Bologan, Ji­ angsu Wuxi 2008. Here 30 ... 'ltg7 would leave Black with a slight advantage in the endgame. 14 tZJxe4 15 tZJxe4 1'.. xfs 16 1'.. d3 A good example of Black's possibili­ ties is 16 f3?! g4 17 1'..f 2? (better is 17 o-o) 17 ...gxf3 18 gxf3 tlJg4 19 1'.. d3 CZJxf2 20 CZJxf2 'iih4 21 o-o e4 22 tZJxe4 1'..h 3 23 �el 1'..e s 24 '1i'e2 'lth8 2 s 'lth1 �xf3 26 '1i'xf3 ilg4 27 '1i'f2 ilf3+ 28 'ltgl 1'..x h2+ 0-1, J.Donner-S.Gligoric, Eersel 1968. 16 ...g4 17 0-0 •••

White can also prepare to castle long, but Black has resources: 17 '1i'e2 'lth8 18 o-o-o '1i'e8 19 'ltbl '1i'g6 20 h3 gxh3 21 gxh3 tlJg8! (a typical regroup­ ing) 22 1'.. h4 1'.. xe4! 23 1'.. xe4 �6 24 ilg3 CZJf6 2S 1'..f 3 e4 26 1'.. g4 tlJxg4 27 hxg4 '1i'g6 28 'ital as and Black had the initiative in J.Dorfman-1.Nataf, Mon­ dariz 2000.

White has control of the e4-square, but Black has gained some space on the kingside. White will be able to initi­ ate queenside play with c4-cS, but Black's main concern is the activation of his h6-knight. 17 'lthB The king is a little safer here and Black can consider regrouping with ... tlJg8 at some point. Nataf has used this move frequently, but Black has some choice here and it is not clear which move is best. There are two other logical tries: a) 17 ... 1'..xe4 really tries to force things when 18 1'..xe4 'ii'g s protects the g4-pawn and intends ... tZJfs and ...hs. White has: •••

199

A t tacking C h e s s : Th e Kin g 's Indian, Vo l u m e 1 al) 19 'itb3 !Iab8 20 cs lLifs 21 'it'a3 was suggested by Mikhalevski, but 19 ...b6 looks like a better try. a2) 19 'it'd3 lLifs 20 !Iae1 !If7 21 cs !Iaf8 22 cxd6 cxd6 23 !Xcel a6 24 b3 hS gives Black counterplay. After 2S f4 (forced) 2s ... lLixg3 26 fxgs !Ixf1+ 27 !Ixf1 !Ixfl+ 28 'it'xf1 lLixfl 29 @xfl Black had to work a little bit to draw in R.Markus-V.Sikula, Hungarian League 2004. Instead Black could have tried 2s ...'i'h6 26 ..ixfs !Ixfs 27 fxes !Ixf1+ 28 !Ixfl !Ixfl+ 29 @xf1 h4! with equality. a3) 19 'it'c1 'it'hs (19 ...'it'xcl followed by 20 ...lLifs is also possible) 20 cs !If7 21 cxd6 cxd6 22 'it'c2 lLifs (after 22 ... ..if6 23 !Iad1 !Iaf8 24 'it'a4 lLifs 2s 'it'xa7 lLid4 26 'itb6?? ..id8 White was facing ...lLie2xg3 and resigned in J.Vanheste-J.Piket, Am­ sterdam 1988, but of course he can im­ prove with 26 'it'a3) 23 ..ixfs 'it'xfs 24 'it'xfs !Ixfs 2s !Iac1 e4! gave Black enough counterplay in D.Kuljasevic­ D.Popovic, Zupanja 2008. b) With 17 ... 'it'e8 Black brin gs his queen to g6 before addressing his side­ lined knight.

2 00

White has: bl) 18 !Iel 'if g6 19 'if d2 (19 cs! transposes to variation 'b2') 19...!If7 20 b4 !Iaf8 21 ..ih4 ..ixe4! 22 ..ixe4 lLifs 23 cs ..ih6 24 'it'd3 ..igs 2s ..ixg s 'it'xg s 26 ..ixfs !Ixfs gives Black good play in a position with all the major pieces. The following game serves as a good illus­ tration of Black's chances in this type of position: 27 !Ie2 h S 28 c6 b6 29 'it'a6? h4 30 'it'xa7 !Ixf2 ! 31 !Ixf2 'it'e3 and 0-1 in P.Smirnov-D. Kryakvin, Belorechen sk 2009. b2) 18 cs looks more to the point: 18 ... 'it'g6 19 !Iel lLif7 20 ..ih4 (White was also better after 20 !Icl lLig s 21 lLixgs ..ixd3 22 lLie6 in Yang Shen­ Xiaowen Zhang, Jiangsu Wuxi 2008) 20 ...!Iae8 (20...'it'h6 21 g3 is not such a help to Black) 21 !Ic1 dxcs 22 !Ixcs lLid6 23 'ifa4 ..ixe4 24 ..ixe4 'it'h6 (perhaps better is 24 ... lLixe4 2S 'it'xe4 - but not 2S !Ixe4? !If4! - 2S ...'it'xe4 26 !Ixe4 !If4 27 !Iec4 !Id4 28 @f1 bS 29 !Ic3 a6 as suggested by Mikhalevski, but Black is still worse) 2S ..ig3 'it'd2 was L.Aronian­ T.Radjabov, Morelia/Lin ares 2006. Now 26 !Id1 'it'e2 27 ..id3 'it'xb2 28 !Ixe7 would leave White much better. 18 .l:i:el Instead 18 'it'e2 ..id7 19 f3 'it'e7 20 fxg4 and Vi-Vi B.Gelfand-T.Radjabov, Russian Team Championship 2006, was not very revealing, while 18 f3 'ife7 19 'it'c2 ..ig6 20 fxg4 lLixg4 21 'it'e2 lLih6 22 ..if2 lLifs 23 !Iael ..ih6 gave Black enough play in A.Kovalyov-1.N ataf, Montreal 2008.

G l ig o ric Varia t i o n : 7 j_ e 3 0i g 4 8 j,_95 f6 9 j_ h 4 gs 18 'ife7 After 18 ... j_d7 19 cs 'if e7 20 l:i:c1 0ifs 21 cxd6 cxd6 22 .l:f.c7 White won material in Y.Shulman-1.Nataf, Mont­ real 2008. The alternative, 18 ... 0ig8, was played by Radjabov, but he did not repeat it when given the chance. After 19 f3 (perhaps 19 J:!.c1 OT 19 CS!? later worried the Azeri) 19 ... gxf3 20 'it'xf3 0ih6 21 'it'e2 j_g6 22 j_f2 0ifs 23 �f1 the game was drawn in A.Onischuk­ T.Radjabov, Mallorca Olympiad 2004. 19 !k1 b6 20 b4 as! Black creates some play for himself on the queenside. 21 a3 axb4 22 axb4 .••

22 j_g6 Instead 22 .. Jfa3 23 cs bxcs 24 bxcs j_g6 (24.. JHa8! ? is also possible) 2S j_b1 .l:f.a1 26 'it'e2 �fa8 27 lUC3 j_xb1 28 �xb1 !Ixb1 29 !Ixb1 'if gs was Wang Yue-T. Radjabov, Elista 2008. White still seems a bit better here, but Black gradually e qualized and drew. 23 cs bxcs 24 bxcs !Ias Black eyes the dS-pawn. 2S 'it'e2 •••

After 2S cxd6 cxd6 26 0ic3 j_xd3 27 'it'xd3 the g4-pawn is not under attack and Black can play 27 ... 0ifs with coun­ terplay. 2s l:i:a4 26 0ic3 j_xd3 27 'it'xd3 l:i:b4 Not 27 ... l:i:d4?? 28 'it'xd4. 28 cxd6 cxd6 29 !Ie4 l:i:xe4 30 0ixe4 £Ufs •••

At last Black activates the knight. Now ... hs is coming and it becomes clear that the g3-bishop is more of a problem for White than Black's 'bad' bishop on g7, which may become ac­ tive with ... j_h6. The game V.Akobian1.Nataf, Montreal 2008, concluded as follows: 31 'it'd1 hs 32 h3 j_h6 33 !Ic6 h4 34 hxg4? White had to give up a piece with 34 j_xh4 'it'xh4 3 S 'it'xg4. 34 0ixg3 3S fxg3 j_e3+ 36 @h2 'ifh7 Black can also win with 36 ...hxg3+ 37 @xg3 j_f4+ 38 @f3 'i'h4. 37 'it'c2 Instead 37 �c4 loses to 37 ... hxg3+ 38 @xg3 j_f4+ 39 @f3 'it'h2. 37 hxg3+ 38 @xg3 j_g1! 0-1 •••

•••

201

Chapter 11 Petrosian Variation 7�d-S aS withOut 8

it.gs

1 d4 lLif6 2 c4 g6 3 lLic3 ..ig7 4 e4 d6 5 lLif3 o-o 6 ..ie2 es 7 ds

sian Variation was once a big favourite of Kramnik and he used this system to defeat Kasparov on more than one oc­ casion, so it certainly has to be taken seriously. Personally I have been happy to play the Petrosian with both colours. 1 as! This prophylactic move is the main line and it has been played more than twice as often as all other moves com­ bined. A couple of Black's obvious moves run into difficulties: for exam­ ple, 7 ... lLie8 8 h4! or 7... lLih s 8 g 3 ! fs 9 exfs ..ixfs (9 ... gxfs 10 lLixes!) 10 lLig s lLif6 11 g4! and White has an unshake­ able grip on the e4-square. The old main line was 7...lLibd7 8 ..ig s (8 ..ie3 may be even better) 8 ... h6 g ..ih4 gs 1 0 ..ig3 lLih s 11 h4!. In this line Black will have to depend com­ pletely on piece play, because a subse­ quent .. .fs will hand the e4-square to White. A respectable line is 7 ... lLia6 (which was Gallagher's recommenda­ tion in his 2004 book) 8 lLid2! and now ...

This variation was pioneered by the ninth World Champion. Although it appears to be a simple line, the Petro­ sian Variation leads to a real heavy­ weight strategic struggle. Both sides often have the chance to play on either side of the board. Although it is not considered critical nowadays, it is still a popular choice. The positions are gen­ erally not as sharp as they are in the Mar Del Plata Variation, but the theory does run deep in some lines. The Petro-

2 02

Petrosian Varia tio n : 7 dS a s with o u t 8 j,_9 5 8 ... c6 9 a 3 will sideline the a6-knight, while Gallagher's 8 ... cs looks too stodgy to me after 9 h4 h S 1o lLif3!. With 7...as Black intends to play ... lLia6, from where it can hop to cs at an appropriate moment to pressure the e4-pawn. From a6 the knight will also help to hold up both White's b2-b4 and c4-c5 advances without interfering with Black's development (compared to 7...lLibd7). After White makes the nec­ essary preparations for the b4-advance (commonly !Ib1, b3 and a3), Black will often play ... lLics, in order to meet b4 with ... lLia4!. Of course the knight will need the proper support to make this leap, but we will see that the moves ...'it'e8 and ...j_d7 are often an integral part of Black's plans. After 7 ... as, White has a diverse choice.

A:

Lines A and B are seen occasionally but are not very dangerous, while Line C is a kind of transpositional device. Line D was fashionable for a while, but 8 j_g5 is the absolute main line and will be considered in Chapter 12. There are a couple of other options: a) 8 h4 lLia6 9 lLid2 is Line A. b) 8 a3 j_d7! (8 ... lLia6 allows 9 b4 because of the pin on the a-file; this is a common theme in the Petrosian) 9 j_e3 (or 9 !Ib1 a4! intending ...'it'e8 and ... lLia6) 9 ... lLig4 10 j_g5 f6 11 j_h4 lLia6 12 o-o 'it'e8 13 lLid2 lLih6 14 f3 lLif7 1 5 @h1 f S gave Black equal chances in l.lvanov-J.Fedorowicz, USA 1989. c) 8 g3 was another specialty of the late GM Igor Ivanov that I have even tried myself. My friend GM Jesse Kraai called this 'a move from outer space' when he saw it and thought chess must be a great game if such moves were playable. 8 g3 is not very danger­ ous theoretically, but Black should be aware of White's ideas after 8 ...lLia6 9 lLih4!. This looks strange, but it is part of White's plan, as we shall see.

8 lbd2

B: 8 .i.e3 C: 8 0-0 D: 8 h3

203

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1 Now we have: c1) 9 ... lLie8 10 o-o fs 11 exfs gxfs 12 f4 e4 13 ..ie3 intending ..id4 and lLig2e3 was D.Vig orito-A.Matikozian, Las Vegas 200S, when we can see White's idea take shape, although even here Black is not without resources. c2) 9 ... lLics 10 f3 lLih s 11 o-o (it is probably better to play 11 lLig2 ..ih3 12 0-o fs 13 exfs gxfs with an unclear po­ sition) 11 ...lLif4! (taking advantage of the knight's position on h4) 12 lLig2 lLixe2+ (after 12 ... lLih3+ 13 @h1 fS 14 exfs gxfs 1s f4 the h3-knight's position looks a bit precarious) 13 'it'xe2 ..ih3 (or simply 13 .. .fs) 14 ..ie3 b6 1s @h1 fs 16 exfs gxfs 17 !Ig1 'it'd7 18 lLid1 !Iae8 19 lLif2 ..ixg2+ 20 !Ixg2 @h8 21 !Iag1 'it'f7 22 b3 lLid7 23 g4 and now, instead of 23 ... !Ig8?! 24 gxfs 'it'xfs 2 s lLie4 when the g-file and e4-square were under White's control in l.lvanov-B.Gelfand, New York 1989, Black could play 23 ...f4! 24 ..id2 e4! 2s lLixe4 lLics with good counterplay. A) 8 lLid2

2 04

With the centre closed White hopes to stir up trouble on the kingside. 8 lLia6 9 h4 Instead 9 o-o transposes to Line D, while 9 g4 lLics 10 h4 is considered in the notes to White's 10th move, below. 9...lLic s T h e most dynamic. Black h a s no reason to fear the advance of the White h-pawn. 9 ... hs is also possible, but it is a bit stodgy: a) 10 lLif1 lLics 11 ..ig s 'it'e8 12 lLid2 ..id7 13 f3 lLih7 14 ..ie3 fS gave Black counterplay in Y.Teplitsky-A.lstratescu, Yerevan Olympiad 1996; White does not have an ideal home for his king. b) 10 f3 c6 (10 .....ih6 is given as equal by Nunn and Burgess) 11 lLif1 lLics 12 ..ie3 and now rather than 12 ... cxds 13 cxds, which hands the c4square to White, Black should play 12 ... ..id7, intending ... a4, ...'it'as, ... cxds and ... bs. c) 10 lLif3 ! ? heads for the weakened gs-square: 10...lLics 11 lLig s c6 12 ..ie3 ..id7 and a complicated strategic struggle lies ahead. 10 h s Instead 10 g4 is somewhat unnec­ essary, but it has been seen often enough in practice. Following 10 ... a4 (10... c6 is also possible) 11 hS gxh S ! ? 12 gs (after 12 gxhs ..ih6! intending ... @h8 the opening of the kingside is welcomed by Black) 12 ...lLig4 13 lLif1 fS we have: a) 14 exfs ..ixfs 1s f3 a3! gives Black good play. One possibility is 16 fxg4 ...

Petro sian Va riatio n : 7 dS as with o u t 8 j,_95 ..i.xg4 17 j_xg4 hxg4 18 'it'xg4 axb2 19 ..i.xb2 lLid3+ 20 @e2 lLixb2 21 'it'hs l:i:f4 22 'it'xh7+ @f8 with a messy position that looks very playable for Black. b) 14 f3 lLif2 (Black could also play 14.. .fxe4 1 S fxg4 j_xg4 16 j_xg4 lLid3+ 17 @d2 'it'xg s+ 18 @c2 lLib4+ 19 @b1 "ihg4 with an unclear position) 1 S ltixf2 fxe4 1 6 @g2 a3 ! 17 !Ixhs exf3+ 18 ..i.xf3 e4 19 j_e2 j_fS gave Black excel­ lent compensation for the piece in the famous game B.Kouatly-G.Kasparov, Evry (simul) 1989.

10 c6 Black looks for queenside counter­ play. This move was relatively untested when Nunn and Burgess wrote in the 1990s, but now there are many exam­ ples and Black has scored well. 11 g4 Others: a) 11 h6 limits White's possibilities. After 11...j_h8 12 f3 lLihs 13 lLif1 lLif4 14 g3 lLixe2 1 S 'it'xe2 fS Black had the initiative in A.Bykhovsky-1.Caspi, Herzliya 2006. b) 11 'it'c2 (it looks odd to place the ...

queen on the c-file) 11 ... a4 12 lLif1 cxds 13 cxds j_d7 14 hxg6 fxg6 (14...hxg6 is also possible - Black is not about to get mated on the h-file) 1s f3 lLihs 16 g3 bS! gave Black good play in A.Dunnington-Z.Lanka, Paris 1990. If White takes on bS then f3 will be hang­ ing. 11 cxds Often is it better to delay this ex­ change because White gains access to the c4- and bS-squares, but here there is a neat tactical point. Black has also avoided the exchange in practice with: a) 11 ... 'it'b6 is a common idea. The queen eyes the b2- and f2-squares, thereby creating some tactical possi­ bilities: 12 f3 (not 12 gs? lLixhs 13 j_xhs lLid3+ 14 @e2 lLif4+) 12 ...j_d7 13 lLif1 cxds 14 cxds !Ifc8 1s lLie3 a4 16 lLic4 'ife7 17 lLia3 ! was A.Aleksandrov­ R.Leitao, Yerevan Olympiad 1996. White's knight manoeuvring has been very time-consuming, but he has man­ aged to grip the bS-square. Instead of the game's speculative 17 ... bs, I would suggest 17 ... 'it'b6 when Black looks okay. b) 11 ...a4 is logical and then: bl) 12 g s should backfire: 12 ...lLixhs 13 j_xhs lLid3+ 14 @f1 'it'b6 1 s !Ih2 and now, instead of 1s ...gxhs?! 16 'it'xhs h6? 17 gxh6 j_f6 18 lLif3 with a win­ ning attack in A.Aleksandrov­ V.Kotronias, Pula 1997. Black should play Huzman's 1s ... a3 ! 16 b3 'it'd4! 17 lLidb1 lLixc1 18 'it'xc1 gxhs when he has a huge advantage. ...

205

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1 b2) 12 f3 ..id7 13 lLidb1! 'it'b6 14 'it'd2!? (White makes some unusual moves, but they are all designed to shut down Black's counterpl ay) 14 ... !Ifc8 1s lLia3 'it'as 16 @f1 l:!.ab8 17 'it'e1 'it'd8 18 @g2 cxds 19 cxds lLie8 20 lLiabS and White was better in C.Gabriel-E.Grivas, Corfu 1999. 12 cxds Instead 12 exds e4! gives Black ob­ vious counterplay: a) 13 g s lLixh s 14 ..ixh s gxh s 1s lLicxe4 l:i:e8 16 'it'xh s .ifs was clearly better for Black in D.Shchukin­ E.lnarkiev, St Petersburg 2001. b) 13 h6 ..ih8 14 gs lLih s 1s lLidxe4 lLixe4 16 lLixe4 l:i:e8 17 l:i:h4 .ifs 18 ..if3 ..ixe4 19 !Ixe4 l:i:xe4+ 20 ..ixe4 'if e7 21 f3 !Ie8 22 'it'd3 lLig3 won back the pawn and left Black with some initiative in A.Voll-D.Petrovic, correspondence 2004.

12 bs! Black takes advantage of the loose pawns on e4 and g4. Instead 12 .....ixg4 does not work, but the idea is worth keeping in mind, despite here 13 ..ixg4 lLid3+ 14 @fl 'it'b6 1S l:i:h2 lLixb2 16 ...

206

'it'b 3 ! being good for White. 1 2 ... a4 and

12 ... ..id7 are possible, but if Black wants to play more slowly it is better to delay the exchange on dS. 13 gs Altematively: a) 13 f3 b4 14 lLia4 lLifd7 is comfort­ able for Black. Even ... ..if6-gS is a possi­ bility. b) 13 hxg6 fxg6 14 f3 b4 1S lLia4 lLifd7 is similarly pleasant for Black. c) 13 a3 and here: cl) 13 ...'it'b6 14 f3 ..id7 1S lLib3 b4 16 lLixcs 'it'xcs (instead 16...bxc3 17 lLixd7 lLixd7 18 bxc3 fS looked too speculative in l.Tsesarsky-M.Tseitlin, Givataim 2000, although Black won) 17 lLia4 'ifa7 looks fairly level. c2) 13 .....ixg4! now works.

After 14 ..ixg4 b4! both 1S lLie2? lLixg4 and 1S lLia4? lLixa4 hang pieces, which leaves: C21) 1S l2Jcb1 l2Jd3+ 16 @fl jVb6 17 !Ih2 lLixg4 18 jVxg4 lLixc1 wins. c22) 1S axb4 axb4 16 !Ixa8 'it'xa8 17 lLibs lLid3+ 18 @f1 lLixg4 19 jVxg4 'it'b8 and both the knight on bS and bishop

Petrosian Va ria tion : 7 ds a s with o u t 8 j,_95 on cl are hanging. c23) 15 lLia2 lLid3+ 16 ..tf1 bxa3 17 bxa3 jVb6 18 !Ih2 lLixg4 19 'it'xg4 'it'd4 20 lLib3 'ifc4 21 lLid2 'it'c2 22 'it'e2 lLif4 intending ....l:f.ab8 looks like fun for Black. 13...lLixhs There is no reason to play the pas­ sive 13 ...lLie8. Black must exploit his tactical possibilities. 14 j_xhs

14... b4! 1s lLia4 Avoiding the challenge is worse. Af­ ter 15 lLib3 bxc3 16 lLixc5 cxb2 17 j_xb2 dxc5 18 j_e2 'it'xg5 White's position was a mess in B.Kohlweyer-D.Stets, German League 2005. 1s...lLid3+ 15 ...lLixa4 16 'it'xa4 gxh 5 is probably even better. After 17 !Ixh5 j_g4 18 !Ih4 j_d7 Black will play ...'it'xg5 with a clear advantage. 16 @fl lLif4! Now 17 j_f3? j_a6+ 18 @g1 'it'xg5+ 19 @h2 was R.Kozlov-P.Gnusarev, As­ tana 2007, when 19 ...j_e2 wins in­ stantly. Better is 17 j_e2, although after

17 ... 'it'xg 5 18 @e1 'it'g2 19 .l:f.f1 j_h3 20 j_f3 'it'g 5 Black has a strong initiative for the sacrificed material.

As in the Gligoric Variation, White allows this bishop to get chased around a bit. 8 ...lLig4 Worse is 8 ... lLia6 9 lLid2 when White has the ideal piece set-up and total flexibility with his king position. 9 j_g5 f6 10 j_h4 Instead 10 j_d2 is not very danger­ ous after 10...f5 (however, 10...lLia6 11 h3 lLih6 12 h4 is a bit annoying) and now: a) 11 o-o lLia6 (or 11 ... lLif6) 12 lLie1 lLif6 13 exf5 j_xf5!? was V.Gallego Jimenez-J.Gallagher, Benidorm 1991. Taking with the bishop is often a de­ cent choice when Black has good con­ trol of the e4-square, but 13 ...gxf5 is also fine. b) 11 exf5 gxf5 (again 11...j_xf5 is fine, as in A.Huss-J.Gallagher, Samnaun 2008) 12 lLig5 lLif6 is variation 'c' below.

207

A ttacking C h e s s : Th e King 's I n di a n , Vo l u m e 1 Instead 12 ... 'it'e8 13 cs!? dxcs 14 'ifb3 @h8 1S 0-0 gave White compensation in B.Zlotnik-A.Morozevich, Moscow 1991, although he soon went wrong and lost. c) 11 lLigs lLif6 12 exfs gxfs (this time 12 ... ..ixfs 13 g4! ..id7 14 lLige4 is to White's advantage) 13 f4 e4 is fairly level.

After 10 ..ih4, White wants to play lLid2, f3 and ..if2 with a flexible and harmonious position. In most cases Bl ack should maintain his knight on g4 until forced to retreat. White generally cannot play f3 without allowing ... lLig4e3, so Black usually waits for White to play h3, which will cost a tempo and somewhat weaken the white kingside. 10 lLia6 10 ... h s 11 lLid2 lLia6 transposes. Another idea is 10 ...lLih6. The point is to play ...gs and .. .fs before White gets in f3 and ..if2. After 11 lLid2 gs 12 ..ig3 lLia6 (12 .. .fs !?) 13 a3 ! ? (or 13 f3 fS when Black gets counterplay following 14 ..if2 g4 or 14 exfs lLixfs 1s ..if2 lLid4) we have:

a) 1 3 ... ..id7 1 4 o-o fS (Black should consider 14 ... g4 or 14... 'it'e8) 1S exfs lLixfs 16 lLide4 looks somewhat better for White. b) 13 ... 'it'e8 14 f3 (14 b4 fS 1S f3 axb4 16 axb4 ..id7 17 l:i:b1 g4 gives Black counterplay according to Avrukh) 14... ..id7 1S ..if2 and now, instead of 1S ... a4? 16 lLibS! which was very good for White in E.Bacrot-E.Relange, French Team Championship 200S, Avrukh suggests 1s .. .fs 16 exfs lLixfs 17 o-o lLid4 with sufficient play. 11 lLid2 No one has tried 11 h3 lLih6. Pre­ sumably Black gets a favourable ver­ sion of the note to his next move. If Black really does though prefer to play with a pawn on hS, he could avoid this possibility altogether by playing 10 ... hS.

...

208

11 hs Instead 11...lLih6!? allows White to reach his desired set-up with 12 f3, but Black still looks fine following 12 ... lLif7 (or 12 ... ..id7 13 o-o 'it'e8) 13 a3 ..id7 14 l:i:b1 lLics 15 b4 axb4 16 axb4 lLia4. Black is also not bothered by 12 a3 ..id7 ...

Petrosian Varia tion: 7 ds as witho u t 8 ii. g s 1 3 !Ib1 lLics (13 ... 'it'e8 14 b4 sidelines the a6-knight) 14 b4 axb4 1S axb4 l2Ja4 with the idea of ... gs and ... fs. 12 a3 The immediate 12 h3 lLih6 has scored well for Black: 13 g4 (instead 13 a3 ii..d7 transposes to the main line, while after 13 f3 lLif7 Black will play ... ii..h 6) 13 ... hxg4 14 hxg4 lLif7 with the idea ...ii..h 6, ... @g7 and ...!Ih8, which gives Black good play. 12 ...ii..d 7

13 h3 White finally tires of having Black's knight hanging around. This move i s generally criticized, but the alterna­ tives are also unimpressive: a) 13 0-0 and now 13 ...lLih6 is com­ pliant, wh ile 13 ...lLics allows 14 b4, so we have: al) 13 ...'it'e8 14 b3 fs 1S l:i:b1 lLics 16 b4 axb4 17 axb4 lLia4 and here White should avoid 18 'it'c2 lLixc3 19 'it'xc3 fxe4 20 lLixe4, since Black has 20 ... lLixh2! when 21 @xh2 fails to 21 ...J:i.f4 and 21 lLixd6 cxd6 22 @xh2 .l:f.f4 gives Black the initiative, as ...e4 is

in the air. Better is 18 lLixa4 ii..xa4 with equality in J.Speelman-A.Strikovic, Oviedo (rapid) 1992. a2) 13 ...'it'e7 14 .l:f.b1 lLics 1S b4 axb4 16 axb4 l2Ja4 17 lLixa4 !Ixa4 18 h3 lLih6 19 f3 was J.Speelman-J.Piket, Tilburg 1992, when Black should just play the obvious 19 ....rf.fa8 with equality, as in­ dicated by Speelman. b) 13 .l:f.b1 lLics and here: bl) 14 b3 g S ! 1S ii..g 3 fS 16 h3 lLif6 gives Black the initiative after either 17 ii..xh s lLid3+ 18 @f1 lLif4 (or 18...lLixhs 19 'it'xh s lLif4), or 17 f3 'it'e8 18 'it'c2 fxe4 19 fxe4 'it'g6, V.Loginov-A.Fedorov, St Petersburg 1997. b2) 14 b4 axb4 1S axb4 lLia4 16 lLixa4 (after 16 'it'c2 Black again seizes th e initiative with 16... gs 17 ii.. g3 fS) 16...ii.. xa4 17 'it'c1 ii..d7 and Black had a good position in G.Danner-V.Spasov, Leon 2001. 13 ...lLih6 14 .l:f.b1 lLics 15 b4 axb4 16 axb4 lLia4

17 'it'c2 Black also had fine play after 17 lLixa4 ii..xa4 18 'it'c1 gs 19 ii..g3 h4 20 ii..h 2 ii..d7

209

A ttacking C he s s : Th e Kin g 's In dian, Vo l u m e 1 21 f3 f5 22 ..ig1 g4 in l.Morovic Feman­ dez-G.Kamsky, Las Palmas 1994. 17 ...lLixc3 18 'it'xc3 gs 19 ..ig3 h4! 20 ..ih2 fs 21 cs g4 22 c6 ..ic8 Black keeps the tension. Kasparov gives 22 ... bxc6 23 dxc6 ..ie6 24 ..ic4 as unclear. 23 hxg4 fxg4 23 ...lLixg4!? looks good too. 24 cxb7 ..ixb7 2s o-o 'it'gs Black had the initiative in E.Bareev­ G.Kasparov, Tilburg 1991. C) 8 0-0

This move is played rarely nowa­ days, but it is still important, especially for transpositional reasons. 8...lLia6 9 'it'c2 We will take this as the main line, but there are several other moves. 9 ..ig 5 is actually the most common move, and after 9 ... h6 10 ..ih4 we are in Chapter 12. Instead after 9 ..ie3 lLig4 10 ..ig5 f6 White has a less flexible version of Line B. White's two remaining alter­ natives to 9 'it'c2 are the typical knight retreats:

210

a) 9 lLie1 lLic5 and here: al) 10 'it'c2 transposes to note 'a' to White's 10th move in our main line, although this way White has managed to avoid 9 'if c2 lLih5. a2) 10 f3 has scored terribly for White after 10 ...lLih5 ! . Indeed, Black has a pleasant choice after both 11 ..ie3 (between 11 ...b6, 11 ...lLif4 and 11 ... f5), and 11 lLid3 (between 11 ...b6 and 11 ...lLixd3 12 ..ixd3 lLif4), so White usu­ ally prevents the knight hop with 11 g3. Then 11.. ...ih3 12 !If2 f5 (or 12. . ...id7) 13 lLid3 lLixd3 14 'it'xd3 (14 ..ixd3 f4 15 g4 lLif6 is similar) 14.. .f4 15 g4 lLif6 (15 ... lLig3 is possible but more risky af­ ter the cool 16 ..id1) 16 ..if1 ..ixf1 17 'it'xf1 lLid7, with the idea .....if6, gives Black good play. b) 9 lLid2 lLic5

and here: bl) 10 'it'c2 transposes to note 'b' to White's 10th move in our main line, again with White avoiding 9 'it'c2 lLih 5 . b2) 10 f3 lLih5 11 lLib3 b6 has done well for Black. b3) 10 !Ib1 will transpose to varia-

Petrosian Va ria t i o n : 7 ds as with o u t 8 j,_95 tion 'b4' after 10...lLie8 11 b3, while 10 ... j_h 6 is also sensible. b4) 10 b3 lLie8 (10 ... j_h6 has also scored well for Black after 11 'it'c2 lLie8 or 11 ...j_d7) 11 a3 fS 12 l:i:b1 lLif6 and now:

b41) 13 b4 is a strange pawn sacri­ fice. After 13 ... axb4 14 axb4 lLicxe4 1S �dxe4 lLixe4 16 lLixe4 fxe4 17 j_e3 Black is doing well after all of 17 ....rf.a2, 17 ...j_d7 and 17 ... 'it'M. The extra pawn may not mean much itself, but Black certainly has no problems. b42) 13 f3 j_d7 (preparing for White's queenside advance) 14 b4 axb4 1S axb4 lLia4 (we will see this idea countless times) 16 'it'c2 lLixc3 (Black could also play 16 ...j_h6 or 16 ... lLih s) 17 "if'xc3 and now Black has several prom­ ising ideas, such as 17 ... lLihs , 17 ... j_h6 and 17 .. .fxe4 18 lLixe4 lLixe4 19 fxe4 �xf1+ 20 j_xf1 'ifh4. Now we return to 9 'ifc2: 9 lLics After this move we reach a position that has occurred hundreds of times, although it usually arises after 7 o-o ...

lLia6 (or 7... lLibd7) 8 dS lLics 9 'it'c2 as. With the move order we are going by, however, Black has another promising option in 9...lLih s.

White has not been able to show anything here. On the contrary, Black has done very well. Some examples: a) 10 g3 fs 11 exfs j_xfs 12 'it'd1 (or 12 lLie4 lLif6 13 lLifd2 lLib4 14 'it'b1 c6) 12 ... lLif6 13 lLih4 j_h3 14 lLig2 'it'd7 gave Black some initiative in S.Hamann­ R.Hiibner, Skopje Olympiad 1972. b) 10 !Ie1 lLif4 11 j_f1 is a bit trick­ ier. After 11 .. .fs 12 j_xf4 exf4 13 es White has good chances, so Black should play 11 ...j_g4!. c) 10 lLie1 lLif4 11 j_f3 fS 12 j_e3 b6 13 exfs j_xfs 14 j_e4 lLics 1s j_xfs !Ixfs 16 g3 lLih 3+ 17 @g2 'it'd7 18 lLid3?! (better is 1 8 j_xcs bxcs 19 lLie4, al­ though Black is fine after 19 ... l:i:af8 or 19 ... a4) 18 ... lLixd3 19 'it'xd3 lLif4+! 20 gxf4 exf4 gave Black a strong attack in J.Nikolac-G.Ligterink, Wijk aan Zee 1976. Even though 9 ...lLih S looks promis­ ing, we will also examine 9 ...lLics in

211

A ttacking C he s s : The Kin g 's I n d i a n , Vo l u m e 1 detail because the position can arise in various ways and doing so provides Black with a couple of good options.

choice between 10...lLie8 11 lLib3 fS (or 11 ...b6) 12 lLixcs dxcs! 13 ..ie3 b6, in­ tending ...lLid6, and 10.....ih6 11 lLib3 ..ixc1 12 l:i:axc1 lLifd7 (12 ... lLixb3 and 12 ... b6 are satisfactory as well) when Black has no problems. 10. h6 11 ..ie3 b6 This so lid move has gradually evolved to become the main line. Black has also tried 11 ... lLifd7, 11 ...lLihs and 11 ... 'ife7, while after the forcing 11 ... lLig4 12 ..ixcs dxcs 13 h3 lLif6 14 lLixes lLixds 1s cxds ..ixes 16 f4 ..id4+ 17 @h2 White has a big centre, so I pre­ fer the text, which keeps more tension in the position. 12 lLid2 ..

10 ..tgs After 10 b3 lLihs Black is even better off than after 9 ...lLihs above. 10 ..ie3 gives Black a choice between the solid 10...lLifd7 or 10...b6 and the more en­ terprising 10...lLig4 11 ..ixcs dxcs 12 h3 when he has the extra option of 12 ... lLih6! ? compared to 11 ... lLig4 in the note to his next move. Again there are also the knight re­ treats: a) 10 lLie1 lLifd7 (this retreat keeps control of both cs and es, but Black has decent options too in 10...b6, 10.....id7 and 10 ... lLie8) 11 ..ie3 fS 12 f3 (after 12 exfs gxfs 13 f4 exf4! 14 ..ixf4 lLies Black has good piece play) 12 .. .f4 13 ..if2 gs and Black is doing rather well. White still has to prepare b4 and Black should get in ... g4 rather easily White's queen being on c2 (as opposed to d1, from where it controls g4) helps Black. b) 10 lLid2 gives Black a pleasant

212

12.....ig4!? This unusual-looking move has be­ come popular. It looks rather anti­ positional to offer a trade of light­ squared bishops, but Black hopes that White must make some concession to make this exchange. The more natural 12 ... lLig4 is also viable. After 13 ..ix94 ..ixg4 14 a3 (14 f3 ..id7 planning . ts gives Black counterplay, so White tries ..

Petros ian Va riation: 7 ds as with o u t 8 il.. g s to ignore the bishop) 14 ..lLia6 15 !Iab1 fS ! (1s ... i.d7 16 b4 axb4 17 axb4 fS has also been played, but the text is sharper) 16 f3 (16 exfs gxfs 17 f4 e4 is unclear) 16 ... i.h s 17 b4 axb4 18 axb4 f4 19 i.f2 g s Black has counterplay. The break ... g4 is coming and if White plays h3, Black will retreat the hS-bishop and play ...h s and ... g4 with even greater effect. 13 f3 This natural move gains a tempo, but Black has a reason for provoking this advance. The natural 13 i.xg4 lLixg4 14 i.xcs bxcs is fine for Black despite the exchange of light-squared bishops. It is not easy to come up with an active plan for White and Black will play .. .fs or ...h s , activating the bishop on h6. Instead after 13 'it'd1 Black should just retreat with 13 ... i.d7, con­ tent th at White has undeveloped his queen. Others: a) 13 h3 i.d7 (happy to have in­ duced a kingside weakness, Black re­ treats, but 13 ... i.xe2 14 lLixe2 lLih s has al so been tried) 14 b3 lLih7 15 a3 fS 16 exfs (or 16 b4 lLia6 and 16 f3 f4 looks good too) 16 ... gxfs 17 f4 exf4 18 i.xf4 'ifh4! 19 !Iae1 lLigs 20 @h 2 l:i:ae8 21 'it'c1 i.es 22 i.xes l:i:xes 23 b4 axb4 24 axb4 f4! gave Black a winning attack in Y.Zilberman-1.Smirin, European Club Cup, Panormo 2001. White cannot avoid detonation on h3 - a far reaching consequence of his 13th move! b) 13 !Iae1 is more challenging. Then 13 ...i.xe2 (Black should consider

the untried 13 ... i.d7 ! ? when it is not clear what White's rook is doing on el) 14 l:i:xe2 lLig4 15 i.xcs bxcs and we have:

bl) 16 lLibs hs 17 a4 i.h6 18 'ifc3 fs 19 f3 lLif6 20 lLib3 ?! (winning the as­ pawn at great cost; instead 20 exfs gxfs 21 lLib3 h4 22 lLixas lLih s with counterplay is given by Panczyk and llczuk) 20 .. .fxe4 21 fxe4 lLig4 with a strong initiative for Black in A.Korobov­ N.Kuren kov, Moscow 2007. White's knights are too far away from the ac­ tion. b2) 16 g3 looks silly, but the follow­ ing play shows the strategic complexity inherent in the Petrosian Variation. With his rook on e2, White wants to play on the kingside him self: 16 ...h s 17 f3 lLih6 (or 17 ...lLif6 18 f4) 18 f4 fS 19 lLif3 and White had the initiative in A.Korobov-P.Czarnota, Cappelle la Grande 2004. 13 i.d1 14 b3 lLihs! This is precisely why Black provoked f2-f3. 15 .l:f.fel ...

213

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 For the pawn Black has accelerated his kingside play.

1s ... ..tf6! Both 1s ...t2lf4 16 ..if1 and 1s .. .fs 16 a3 look too slow, so Black offers a pawn to activate his dormant bishop. 16 ..ixh6 After 16 a3 ..igs White cannot play 17 ..if2? because of 17 .....ixd2 18 'it'xd2 l2lxb3. With his bishop now activated, Black also has good play after 16 !Iab1 ..tgs 17 ..tf2 fs. 16 ..tgs! 11 ..ixgs White cannot take the rook. He gets mated after 17 ..ixf8 ..ie3+ 18 @h1 (or 18 ..tf1 'ifh4 19 t2ld1 l2lg3+) 18 ... l2lg3+ 19 hxg3 'it'xf8 20 l2lf1 'it'h6+ 21 l2lh2 @g7. 11 'it'xgs 18 tlJfl A quick finish was seen after 18 ..if1 fS 19 a3 !If7 20 !Iab1 t2lf4 21 b4? (bet­ ter is 21 g3 !Ih7, although Black has reasonable compensation for the pawn) 21 ...l2lcd3 and 0-1 in R.Wukits­ G.Timoscenko, Oberwart 2003. White is losing the exchange because 22 !Ied1 walks into 22 ...l2lh3+ 23 @h1 l2ldf2 mate. 1s fs

19 exfs gxfs 20 a3 a4 21 b4 lLib3 Black had excellent play l.Naumkin-1.Smirin, lschia 199S.

in

D) 8 h3

...

...

...

214

This is a tricky move that has been favoured by Zvjaginsev. 8 h3 in the Pet­ rosian Variation is very similar to the Makogonov Variation (S l2lf3 0-0 6 h3) and there are several transpositional possibilities: for example, 1 d4 l2lf6 2 c4 g6 3 t2lc3 ..ig7 4 e4 d6 s l2lf3 0-0 6 h3 es 7 dS as 8 ..igS l2la6 9 lLid2 (or 9 ..ie2) 9 ...'it'e8 10 ..ie2 transposes to the Petro-

Petros ian Varia ti o n : 7 dS as with o u t 8 j,_95 sian. Here we will look at lines where White has played an early j_e2, while in the Makogonov Variation (which will be covered in Volume 2), we will look at positions where White delays or omits th is move. There are a couple of ideas behind the subtle 8 h3. White is ready to play the sequence 9 j_g5 h6 10 j_e3 because now Black cannot play ...l2lg4. White then hopes that ... h6 will weaken the black kingside - the h6-pawn itself may be weak, but more importantly the di­ agonal b1-h7 will be weakened after a subsequent .. .fs. This may seem minor, but in practice this detail can create big problems for Black. Another idea for White is to play g2-g4. Then if Black plays .. .fs, White can pl ay gxfs, opening the g-file, and exchanging twice on fS could grant White the e4-square for his knights. If Black does not play ... fs then White may tum his attention to the other wing. The downside to 8 h3 is that it weakens White's kingside, so castling on that side of the board becomes less appealing for the first player. Also the g4-advance creates weaknesses and Black's .. .fs break can be very danger­ ous for White if he is unable to keep the position under control. Generally Black has two different ways to deploy his pieces. He can play ... tLJcs, ... a4, ... c6, ... j_d7 and ... 'it'as to create play on the queenside. However, Black should be careful concerning the move ... cxds. If it is played too early,

White will seize control of c4 and bS and it will be difficult for Black to cre­ ate counterplay, so Black should gen er­ ally delay this exchange until he can successfully control the bS-square. lfhe can do this, then the advance ...bs will not only control c4, but will give him the possibility of playing ...b4. Black's other set-up is seen in the main lines of both Lines D1 and D2, below. Black will play ... 'it'e8, ...t2ld7 and ... tlJdcs. This will usually allow him to get in ...fs because the diagonal for the c8-bishop remains open. It is in these lines that we see the importance of leaving out the weakening move ... h6. Because the gs-bishop is hanging in the air, Black may drive it away first with .. .f6 before playing .. .fs, or play .. .fs straightaway. Having both knights on the queenside may also give Black some tactical possibilities. The moves ... j_d7 and ... a4 are also thematic (espe­ cially if Black does not play .. .fs), but in general I think th at Black should avoid playing ... c6 in this set-up. 8 l2la6 Instead 8 ...tlJh S is not a bad idea, but White can play either 9 g3 (as he would after 7 ...tlJh s), or the exotic 9 l2lg1 ! ? intending 10 j_f3. An interesting idea is 8 ...l2lfd7!?. In­ deed, Panczyk and llczuk give this as their main line even though it has only been played a handful of times. Now 9 g4 tLJcs 10 j_e3 l2lba6 intending .. .fs looks pretty comfortable for Black. In­ stead g h4! l2lf6 (9 ... fs 10 h S ! ) 10 t2ld2 ! ? ...

215

A ttacking C he s s : Th e King 's In dian, Vo l u m e 1 leaves White a tempo up on 8 lLid2 lLia6 9 h4, although this is not exactly fatal in this kind of position.

16...lLicxe4 17 lLixe4 'it'xc2+ and 18 ... lLixe4 with an extra pawn) 1S ... lLifxe4! 16 lLidxe4 (or 16 lLicxe4 lLixe4 17 lLic4 'i't'b4 18 'it'xe4 !Ifc8 19 ..id2 'it'cs 20 ..ie3 and here, instead of repeating moves with 20 ... 'i't'b4, Black can play 20 ... 'it'bs! with the idea of ... .ifs) 16 ...lLixe4 17 'it'xe4 (17 lLixe4 loses to 17 ...!Ifc8 18 lLic3 a3!) 17 .....tfs 18 'i'h4 !Ifc8 19 @d2 !IxC3! 20 bxc3 'it'xds+ and Black had a winning attack in L.Yurtaev­ Y. Shulman, Vladivostok 199S. 9 'it'e8 Instead 9 ... h6 10 ..ie3 shows one idea behind White's 8th move (there is no ...lLig4), and Black has weakened his kingside in order to break the pin. I pre­ fer the text move. ...

9 ..tgs This pin is a typical weapon in both the Petrosian and Makogonov varia­ tions. White tries to entice Black to weaken his kingside. Other moves are less dangerous: a) 9 ..ie3 lLics (9 ...lLih s is also possi­ ble) 10 lLid2 lLie8 11 g4 fs 12 gxfs gxfs 13 exfs (13 'it'c2 lLia6! threatens ... f4 and then 14 exfs .�xfs 1s .�d3? is not possible because of 1s ...lLib4) 13 ... ..ixfs gives Black excellent play. Note that Black's h-pawn would be much more of a target on h6 here, which is one point of White's 9 ..igs. b) 9 g4 is not well timed: 9...lLics 10 'it'c2 c6 11 ..ie3 a4 (better than 11 ...cxds 12 cxds when bS is weak) 12 o-o-o (White cannot even take a pawn be­ cause 12 ..ixcs dxcs 13 lLixa4 runs into 13 ...lLixe4!) 12 ... cxds 13 cxds 'ifas 14 lLid2 ..id7 1S gs (this runs into a great shot, but after 1S lLic4 'ife7 Black threat­ ens ... bs and 16 lLia3 can be met by

216

The move ... 'it'e8 is very typical in the Petrosian Variation. Black breaks the pin without weakening his kingside and observes the a4-e8 diagonal. Here he hopes to show that h3 cost a tempo and weakened the white position, while the gs-bishop may also be mis­ placed. White has two important con­ tinuations:

Petros i a n Variati o n : 7 dS as with o u t 8 j,_ 9 5

01: 10 g4 02: 10 �d2 Dl) 10 g4

This lunge can be the prelude to an attempted kingside attack, but often White is just hoping to limit Black's possibilities. 10 lLid7! Black can also play 10...lLics, but af­ ter 11 lLid2 but the f6-knight is in the way. By playing the king's knight to cs, Black keeps the possibility of playing ...fs and ... j_xfs. Black has also tried to move the knight to g8, but not surpris­ ingly, this is less dynamic. Indeed, 10 ...@h8 11 lLid2 (or 11 l:!.g1 lLig8) 11 ... lLig8 12 h4 f6 13 j_e3 fs 14 f3 (after 14 gxfs gxfs 15 exfs j_xfs 16 lLide4 lLif6 Black had decent play in T.Polak­ S.Novikov, Pardubice 2007) 14 ...f4 15 j_f2 is better for White, because all the play will shift to the queenside, where White has a significant space advan­ tage. 11 !Igl

Others: a) 11 lLid2 lLidcS 12 lLif1 (in stead 12 h4 could be met by 12 .. .fs or 12 ...f6, 12 a3 should be met by 12 ...j_d7, and 12 !Ig1 transposes to the notes to White's 12th move in our main line, below) is the same as variation 'a1' in the notes to White's 12th move below, but with­ out l:i:g1 and ... @h8. Here the position looks even better for Black after 12 .. .f6 OT 12 .. .fS. b) 11 'it'd2 lLidcS 12 l:i:g 1 lLib4! cre­ ates tactical problems for White. Black intends ... @h8 and .. .fs, and White can­ not play a3 because of the fork on b3. c) 11 a3 and now 11 ...lLidcS? is met by 13 b4 because of the pin along the a-file, but Black has 11 ...lLib6 ! .

•••

This i s an unusual square for the black knight, but White has made a lot of pawn moves and so Black adapts to the situation. He plans .. .fs or ...j_d7 and ... lLia4. After 12 lLid2 j_d7 13 h4 practice has seen: c1) 13 ... lLics 14 hs fs (14 ... lLiba4!?) 15 hxg6?! (better is 15 gxfs gxfs 16 j_e3) 1s ...'it'xg6 16 j_e7 fxe4 17 j_xfg

217

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 !Ixf8 gave Black compensation in B.Thorfinnsson-E.Berg, Reykjavik 2009, but 16 .. .fxg4! 17 ..ixf8 !Ixf8 would have been even stronger. c2) 13 ... lLia4! 14 lLibs lLi4cS 1s hs fs 16 gxfs (not 16 hxg6 'it'xg6, hitting the bishop) was agreed drawn here in Sovata L.Nisipeanu-V.Nevednichy, 1999. After 16 ... gxfs 17 l:i:g1 @h8 Black has good counterplay. 11 lLidcs 11 ... @h8 12 'it'd2 lLidcS (J.Alvarez gives 13 ..ih6 ..ixh6 14 'it'xh6 f6 1S h4 as better for White, but I prefer Black after 1S ... ..id7) 13 o-o-o ..id7 (or 13 ... a4) is a good alternative.

to Black in both cases) 1 8...lLixb2 19 'it'e2 'it'd7 is good for Black. The knight can escape with ...lLia4 and 20 'it'xb2 is met by 20.....ixe4 21 lLixe4 'it'xh3+, win­ ning. a2) 16 lLihs with a further divide:

...

12 h4 Instead 12 'it'd2 lLib4! is note 'b' to White's 11th, above. White's main al­ ternative is 12 lLid2 @h8 and here: a) 13 lLif1?! is slow and a bit clumsy after 13 .. .fs 14 gxfs gxfs 1s lLig3 fxe4! and now: al) 16 lLigxe4 .ifs 17 ..ihs lLid3+! (even better than 17...'it'd7) 18 @f1 (al­ ternatives are worse: 18 'it'xd3 'it'xhs or 18 @d2 lLixb2, with a clear advantage

218

a21) 16 ... 'iff7 is tricky: 17 ..ih4 ..ih6! (Black should avoid the complications of 17 ... l:i:g8 18 !Ixg7 !Ixg7 19 ..if6 'it'xf6 20 lLlxf6 !Ig1+ 21 ..if1 ..ixh3 22 'it'hs) 18 ..if6+ 'it'xf6 19 lLixf6 l:i:xf6 gave Black compensation for the queen in A.Aleksandrov-1.Zakharevich, St Peters­ burg 1994. a22) 16 ...'it'g6! 17 lLixg7 (one point of Black's play is that after 17 .l:f.g3 'iff7 the move 18 ..ih4 does not protect f2, but Black has even better in the shock­ ing 17 ... ..ig4!! 18 ..ie7 'it'xhs 19 ..ixg4 'iff7 20 ..ixf8 l:i:xf8 with a strong attack) 17 ... 'it'xg7 18 'it'd2 lLib4 (Black does not miss his dark-squared bishop because his knights are tremendous) 19 o-o-o lLixa2+! 20 @b1 lLixc3+ 21 bxc3 .ifs was winning for Black in O.Jovanic­ Z.Kozul, Ljubljana 2004. b) 13 a3 and now:

Petro s i a n Va ria tion: 7 ds as with o u t 8 j,_95

bl) 13 ...j_d7 14 h4 f5 15 gxf5 (worse is 15 h5 fxg4) 15 ... gxf5 16 h5 j_f6 17 �xf6+ !Ixf6 18 'it' c2 'it'f8 19 o-o-o fxe4 20 lLicxe4 l:i:f4 21 f3 j_f5 22 !Ig2 'it'h6 with counterplay, M. lvanov-0.Cvitan, Cappelle la Grande 1995. b2) 13 ...a4 is simple enough: 14 'it'c2 f5 15 gxf5 gxf5 16 o-o-o fxe4 17 lLidxe4 �f5 gives Black good play.

12 ...@hB!

A typical reaction to White's rook move. Instead 12 ... c6 is inconsistent and looks panicky to me, but Black may be able to get away with it: 13 h5 cxd5 14 h6 j_h8 15 cxd5 (after 15 lLixd5 f6 16 ..i.e3 Black should avoid 16 ... lLixe4 17 �b6 j_e6 18 lLixa8 'it'xa8 19 lLid2, when White is much better according to Piket, and instead play 16 ... j_e6!) 15 ...j_d7 16 a4 lLic7! (16...bs? 17 axb5 �c7 18 b6 lLib5 19 lLixb5 j_xb5 20 lLid2 gave White a clear advantage in J.Piket­ J.Polgar, 1st matchgame, Aruba 1995) 17 lLid2 b5 18 axb5 lLixb5 gave Black counterplay in S.Giddins-R.Leitao, Ant­ werp 1998. 13 'it'd2 f6

I like this little nudging move, al­ though 13 .. .f5 14 gxf5 gxf5 15 j_h6 comes to the same thing. 14 j_e3 f5 15 gxf5 gxf5 16 j_h6

This may look dangerous, but Black has enough resources to hold the bal­ ance. Note that strategically Black does not mind the exchange of dark­ squared bishops as Jong as his king re­ mains safe. 16 ... j_xh6

Or 16 ... l:i:g8 17 j_xg7+ l:i:xg7 18 !Ixg7 @xg7 19 'it'g 5+ 'it' g6 20 'it'e7+ 'it'f7 with equality according to Zvjaginsev. 17 'it'xh6 'it'e7 18 h5 !If6

Instead 18 ...!If7 19 !Ig6 j_d7 20 lLig 5 !Ig7 is also okay, J.lvanov-J.lbarra Jerez, Mondariz 2000. 19 !Ig6 .l:f.xg6 20 hxg6 'it'g7

Black quickly takes over the initia­ tive in the endgame. 21 '1Wxg7+ @xg7 22 exf5 j_xf5 23 gxh7 lLib4 24 @d2 j_xh7 25 !Igl+ @f6 26 lLig5 j_g6 21 f3 !Ihs

Black had some advantage in the ending in J.Alvarez-E.Pupo, Holguin 1997. 219

A ttacking C h e s s : The King 's I n dian, V o l u m e 1 02) 1o ll:id2

17 b4 axb4 1 8 axb4 �xal 19 '1!r'xa1 ll:ib3 20 '1!r'b2 ll:id4 21 it.d3 f5 22 ll:ie2 fxe4 23 it.xe4 ll:ixe2 24 '1!r'xe2 '1lr'a4 25 '1!r'b2 it.d7 with the initiative in J.Tihonov­ A.Fedorov, Minsk 2010. 13 ... il.h6 14 l':tbl Or 14 o-o ll:ic5 15 �a2!? (15 l:tb1)

A flexible move. 10...ll:ifd7 11 a3

Instead 11 g4 ll:idc5 takes play back into note 'a' to White's 11th move in Line Dl, while 11 o-o ll:idc5 looks fine for Black

15 ...f5 16 f3 il.e3+ 17 il.f2 il.xf2+ 18 .l:!.xf2 ll:ibd7 (other possibilities are 18 ... '1lr'e7 and 18 ...il.d7) 19 b4 axb4 20 axb4 �xa2 21 ll:ixa2 fxe4! 22 fxe4 (22 bxc5 e3) 22 ...l:txf2 23 '1¥o>xf2 '1lr'f8+ 24 '1¥o>g1 ll:ia6 with equality in E.Mirosh­ nichen ko-A.Tukhaev, Kharkov 2007. 14...ll:ics

11...f6! 12 il.h4

After 12 it.e3 Black plays 12 ...f5 im­ mediately, with good play. 12 ...ll:ib6 Again we see this move. Instead 12 ... f5? 13 exf5 gxf5 loses an exchange after 14 iLh5, but 12 ... il.h6!? is possible. Then 13 b4 ll:ib6 14 �bl axb4 (or 14...il.d7 when 15 bxa5?! ll:ia4! and ... ll:ia6-c5 is good for Black) 15 axb4 f5 16 o-o it.d7 17 '1!r'c2 ll:ia4 18 ll:ixa4 Jt.xa4 19 '1lr'd3 il.d7 gave Black a good position in D.Brunsteins-Y.Yarmolyuk, corre­ spondence 2006. 13 b3-

13 b4 is not possible, because Black's last move protected the a8-rook. Instead 13 f3 just invites 13 ... il.h6. A good example of Black's play went 14 Ji.f2 ll:ic5 1 5 ll:ib3 ll:iba4! 16 ll:ixc5 ll:ixc5 220

15 b4

After this move Black takes over the initiative. 15 o-o is more solid and after 1 5 ... il.d7 (instead 1 5 .. .f5 16 b4 axb4 17 axb4 ll:ica4 18 ll:ib5 was awkward for Black in M.Krasenkow-A.Vovk, Vlissin­ gen 2009) White has tried: a) 16 '1!r'c2 f5 17 b4 ll:ica4 (17 ... axb4 18 axb4 ll:ica4) 18 ll:ib5 Jt.xb5 19 cxb5 '1lr'f7 was unclear in A.Jorgensen­ P .Hertel, correspondence 2003. b) 16 b4 axb4 17 axb4 ll:iba4 18

Petrosian Varia t io n : 7 d5 a5 with o u t 8 ii.gs

ll:ixa4 il..xa4! 19 '1!r'e1 il..xd2 20 '1!r'xd2 ll:ixe4 21 '1!r'e3 fS snared an important centre pawn for Black in S.Nenciulescu­ D.Matic, correspondence 20os, al­ though White can claim some compen­ sation with his bishop-pair. 15 ...axb4 16 axb4 ll:ica4 17 ll:ibs

After 17 '1!r'c2 ll:ixc3 18 '1!r'xc3 Bl ack has 18 ...�a2, intending .. .fs with a clear advantage. 17 ...'1!r'd8! 18 '1lr'b3 '1¥o>g7 Black gets off the diagonal of the

white queen. This is solid enough, but he could consider 18 ... cs!? or even the forcing 18 ... il..xd2+!? 19 '1¥o>xd2 c6. After 20 dxc6 bxc6 21 cs+ dS 22 cxb6 cxbs 23 '1lr'xds+ '1lr'xds+ 24 exds ll:ixb6 2s il.. xbs �a2+ 26 '1¥o>e1 ii.ts Black has a strong initiative in the ending. 19 �d1 c6 20 ll:ia3 cs 21 bs ll:id7 22 ll:iab1 '1!r'as

Black intends ...'1lr'b4 with the initia­ tive, N.Legky-A.Shirov, French League 199S.

221

Chapter 12 Petrosia n Variation 1 ds a s a ..tgs

1 d4 lLif6 2 c4 g6 3 lLic3 1lg7 4 e4 d6 S tLif3 o-o 6 1le2 es 7 ds as 8 Ji.gs

when it is difficult for Black to create counterpl ay. Black has sometimes tried 9...�e8 here, as in Line D of the previ­ ous chapter, but White has not wasted time on the weakening move h2-h3 in this line. 9 1lh4 lLia6 10 lLid2

This is White's main move in the Petrosian Variation. By pinning the f6knight he makes it difficult for Black to get in .. .fs. This variation has been fa­ voured with White by GM Igor Naum­ kin for over 25 years. 8 h6 ...

In this case breaking the pin is Black's best because retreating to e3 can be met by ...lLig4. It is important to play this move immediately, else White can play 8 ... lLia6 9 lLJd2 h6 10 1le3 222

This is the most common move, but White can also play 10 o-o �e8 11 lLJd2 tran sposing to the main lines. Instead 11 lLiel is sometimes played. After 11 ... gs (worse is 11 ... tLics 12 f3 lLih s 13 lLibs, but Black can keep the game tense with 11 ... 1i.d7!? 12 lLJd3 lLJh 7 13 f3 fS) 12 1lg3 Black has the well-timed 12 ... tLixe4! (a common motif that is al­ ways worth checking ) 13 tLixe4 fS 14 ..\;lhs �e7 15 f3 fxe4 16 fxe4 �xfl+ 17 '1¥o>xf1 tLics 18 �e2 g4! (without this move Black would just be worse) 19 ..\;lxg4 1lxg4 20 �xg4 �g s! when the e4-pawn falls after 21 �e2 �g6 or 21 �xg s hxgs. 10 �eB ...

Generally 10 ... 1i.d7 will just trans-

Pe tros ian Va ria t i o n : 7 dS a s 8 Ji.gs

pose after 11 o-o (or 11 a3 �e8 12 b3) 11 ...�e8.

White's basic plan is to play o-o, a3, b3 (usually necessary at some point if White wishes to avoid his pawns being fixed by ... a4), l':tbl and only then b4. However, the move order is flexible for both sides. Usually White castles here or on the next couple of moves, but he can also delay or omit castling in order to accelerate his queenside play. We will examine: A: 11 a3 (without an early 0-0) B: 11 0-0

the pin on the a-file: for example, 11 ... lLJh 7?! 12 b4! saves White some tempi.

12 b3 lLih7

This is a multi-purpose move. Black prepares either .. .fs or ... hs and ....ii.h6. Another idea is 12 ... tLixe4!? 13 lLJdxe4 fS 14 f3 fxe4 15 tLixe4 il.fs. If Black can fi ght for the e4-square with pieces, as he can here, this idea is usually good enough for equality. 13 l':tbl Instead 13 o-o takes play into the realm of Line B.

Instead 11 b3 lLih7 12 a3 returns play to Line A, as does 11 l':tbl il.d7 12 b3 lLih7 13 a3, while the lunge 11 g4 is well met by 11...lLih7!. A) 11 a3

Here we only examine lines where White delays castling. 11...il.d7 This develops and protects the a8rook. Black must always be aware of

13 ... h s

Black prepares t o activate his bishop 223

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

on h6. This idea is usually more effec­ tive than pushing the f-pawn, but here 13 .. .fS!? is quite viable for tactical rea­ sons: 14 f3 (14 exfs il.xfs! is one point, because 1s ll:ide4? walks into 1S ... il.xe4 16 ll:ixe4 �f4) 14 ... ll:ics (Black does not want the a6-knight to be sidelined; in­ deed, this looks better than 14...ll:if6 lS b4 axb4 16 axb4 ll:ihs 17 1lf2 ll:if4 18 ..\;lf1 when Black's knights are ineffec­ tive) and now: a) lS b4 axb4 16 axb4 ll:ia4 17 '1!r'c2 (not 17 ll:ibs? ..\;lxbs 18 cxbs ll:ic3 and wins) 17...ll:ixc3 18 '1!r'xc3 ll:if6 gives Black comfortable play: 19 cs?! can be met by 19 ... gs 20 1lf2 fxe4 or 19 ... il.bs ! when White may regret his decision to postpone castling for so long. b) lS '1!r'c2 ll:if6 16 b4 axb4 17 axb4 ll:ia4 18 ll:ibs (after 18 o-o ll:ixc3 19 '1!r'xc3 ll:ihs Black has good counterplay: for example, 20 l':tfcl ll:if4 21 1lf1 gs 22 1lf2 '1!r'g6 23 cs g4).

Here 18 ... il.xbs 19 cxbs '1lr'd7 20 o-o ll:ihs 21 l':tfcl ll:if4 22 1lf1 has scored well for White in practice, but instead Black has 18 ...ll:ixds!, intending 19 exds 224

il.xbs 2 0 cxbs e4! with dangerous ideas like ... ll:ic3 and ... gs. 14 f 3 il.h6 1S il.f2

Again White can transpose to Line B with lS o-o.

1s '1!r'e1 ...

Worse is 1s ...fs 16 b4, but 1s ...ll:ics !? is certainly worth consideration. After 16 '1!r'c2 '1lr'e7 17 b4 axb4 18 axb4 ll:ia4 19 ll:ibs White must goes on an adven­ ture to fight for the initiative. Follow­ ing 19 ... c6 there is: a) 20 ll:ia7 cxds 21 cxds (better is 21 exds, but Black's position is healthy here too and even the pawn sacrifice 21 ...e4 looks promising) 21 ...l':txa7! 22 ..\;lxa7 l:tc8 gives Black a very strong ini­ tiative. b) 20 ll:ie7 l':tac8 21 dxc6 (21 '1!r'xa4 l':txe7) 21 ... bxc6 22 ll:ia6 cs 23 l':tal!? (23 bS ll:ib6 is unclear) 23 ... cxb4! ? with a further divide: bl) 24 l':txa4 il.xa4 2S '1!r'xa4 gives Black a pleasant choice between 2s ... il.xd2+ 26 '1¥o>xd2 '1lr'gs+ 27 .ii.e3 '1!r'xg2 and 2s ... '1!r'gs 26 '1!r'xb4 '1!r'xg2 27 �gl '1!r'xh2.

Petro s i a n Varia tion : 7 d5 a5 8 1l g 5

b2) 24 ll:ixb4 �b8 2s ll:ids (after 2s l:txa4 il.xa4 26 �xa4 Black again has the 26 ...�g s idea) 2s ...�gs 26 ll:ie3 ll:ics is pretty level. 16 b4

Instead 16 h4 is very ambitious. White prevents Black from gaining space on the kingside with ...h4, but makes castling kingside less attractive: 16 ...ll:ics 17 �c2 fS 18 b4 axb4 19 axb4 ll:ia4 20 ll:ibs has been played a few times by Zilberman. After 20 ...c6 21 ll:ic7 �ac8 22 ll:ie6 1lxe6 (22...cxdS !? 2 3 tllxf8 ll:ixf8 24 exds e4 was a bit specu­ lative in Y.Zilberman-J. Becerra Rivero, Havana 1998) 23 dxe6 bS 24 0-0 �xe6 2S .rf.bdl ll:if6 matters were unclear in Y.Zilberman-S.Porat, Tel Aviv 2002, al­ though Black was up a pawn. 16...axb4 17 axb4 h4

a ) 1 9 bxcs!? ll:ixcs 2 0 ll:ic4 ll:ia4 (20 ...h 3 ! ?} 21 d6 �gs is very messy. b} 19 bS ll:ib4 20 ll:ic4 and here rather than 20 ... b6 21 d6 cxd6 22 ll:ixb6 OT 20 ...ll:ia2 21 �C2 ll:ixc3 22 �XC3, in both cases with compensation for White, I suggest 20 ... h 3 or 20 ...�fd8!?. 18 ...l:tfbB!?

This prophylactic move looks stronger than the 18 ... �g S 19 �gl c6 (19 ...b6 is safer) 20 cs! of P.Rossiter­ J.Gallagher, British Championship, Swansea 1987. 19 1ld3 �gs 2o l:tg1 cs!

A useful idea to know. Black seizes the initiative right across the board. 21 dxc6 bxc6 22 bs ll:ics 23 g3 hxg3 24 hxg3 lLif8 2S ll:ib3 ll:ixb3 26 .t:!.xb3 ll:ie6

With active pieces and a safer king, Black was much better in G.Danner­ G.Timoscenko, Budapest 1989. B) 11 0-0

A thematic move. Black gains some space and creates various possibilities an the kingside, such as ... il.f4, ...�g s, -�tf6-hS and ... h3. • Wc2

White has also played the pawn sac­ nfice 18 cs!? dxcs and now:

Castling at some point is White's most logical way of playing. 11...ll:ih 7

Here too Black can play 11 ... il.d7 225

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

which will usually transpose to the main line after 12 b3 (or 12 '1¥o>h1 ll:ih7, while 12 a3 can be met by 12 ... a4 13 ll:ibs il.xbs 14 cxbs ll:ics 15 f3 b6) 12 ...ll:ih7 13 a3. 12 a3

White threatens 13 b4. Instead 12 '1¥o>h1 h S 13 f3 il.h6 14 a3 il.d7 15 l':tbl (15 b3 transposes to the main line) runs into 1S ...a4! (if Black does not play this then White will save a tempo on b2-b3), and then 16 ll:ibs b6 or 16 b4 axb3 17 ll:ixb3 b6, with a good game for Black in both cases. White has also been known to go 12 f3 when 12 ... hs, 12 ...il.d7, and 12 .. .fs are all reasonable and may well trans­ pose to the main line. 12 il.d7 Instead 12 ...hs 13 f3 il.d7 trans­ poses, but here Black has to avoid 13 ... il.h6?! 14 b4!. Premature too is 12 .. .fs 13 exfs il.xfs (13 ... gxfs 14 il.hs is annoying) 14 g4! (not 14 ll:ide4?? il.xe4 15 ll:ixe4 �f4) 14 ... il.d7 15 ll:ide4 with an advantage for White in A.Veingold­ G.Kasparov, Moscow 1979.

this, then White will play b2-b4 in one move. Conversely Black should usually wait for l':tbl before playing this ad­ vance, because after the sequence ... a4; b4 axb3; ll:ixb3 White's rook is usually better posted on al, from where it de­ fends the a3-pawn and supports the advance a3-a4-aS.

...

13 b3

The alternatives are also important: a) 13 ll:ibs hs 14 f3 il.h6 15 '1¥o>h1 (15 b3 is the main line and 15 l':tbl a4 varia­ tion 'b', but 15 '1!r'c2!? is possible) 1S ... il.e3 16 l':tbl and here 16 ...a4! can be played because the d2-knight is loose. b) 13 l':tbl is well met by 13 ... a4. This advance is usually good after White has spent a tempo moving his rook. Moreover, if Black does not play 226

Now, after 14 ll:ibs hS (the most flexible, but 14...il.xbs 15 cxbs ll:ics 16 f3 b6 is fine too) 15 f3 il.h6 (Black should probably avoid 15 ... il.xbs 16 cxbs ll:ics 17 b6!?) 16 il.f2 (or 16 b4 axb3 17 ll:ixb3 il.e3+) Black has: bl) 16 ...b6 17 b4 axb3 18 ll:ixb3 il.xbs 19 cxbs ll:ics 20 ll:ixcs dxcs 21 a4 '1lr'e7 22 '1!r'c2 ll:if6 23 l':tal ll:id7 24 il.d3 l':ta7 25 z:ta2 �fa8 26 l':tfal h4 27 '1¥o>h1 '1¥o>g7 28 '1!r'e2 (White intends il.el and as, so Black sacrifices an exchange to keep the position closed) 28 ... z:tas 29 il.e1 il.f4 30 il.xas l':txas 31 il.c4 ll:if6 32 d6! (White gives back a pawn to open lines for his rooks) 32 ...'1!r'xd6 33 l':tdl '1lr'e7 34 '1!r'd3 l':ta8 35 l':ta3 with a slight advantage for White in D.Vigorito­ A.Sherzer, Philadelphia 1997.

Petro s i a n Va ri a t i o n : 7 d5 a5 8 il.g5

b2) 16 ... '1!r'e7 is more flexible. Black does not need rush to exchange on b5. After 17 b4 (17 '1lr'c2 can be met by 17... h4 or 17 ...b6) 11... axb3 18 ll:ixb3 b6 19 a4 f5 20 ll:id2 (White should proba­ bly change gears and play 20 exf5) 20 ... ll:ic5! Black developed pressure against the e4-pawn in D.Vigorito­ A.Matikozian, Los Angeles 2003. Note that 21 ll:ixc7? loses to 21 ... il.xa4. c) 13 '1¥o>h1 ! ? is a way of reaching the main lines of 'B2' and 'B3', below, with­ out allowing Black the option of 13 b3 f5 (after 13 '1¥o>h1 f5 White isn't vulner­ able down the long diagonal, so can play 14 exf5! with advantage). Black can deviate as well, but it is probably better to transpose to the main line with 13 ... h5 14 f3 il.h6 15 b3 (15 �b1 a4! was briefly discussed in the notes to White's 12th, above).

to cover the a3-pawn and play a4-a5, in which case here White could claim to be two tempi ahead. After 14 ll:ib5 Black has: cl) 14...il.xb5 15 cxb5 ll:ic5 16 f3 and although it seems that Black is not so badly off, White has saved time and scored well from here. c2) 14 ... h5 15 f3 il.h6 and now: c21) 16 ll:ib1!? il.xb5 17 cxb5 ll:ic5 18 il.c4 (with the idea of il.f2 and ll:ic3 to target the a4-pawn) 18 ... il.e3 19 ll:ic3 f5 (19 ... b6) 20 exf5 gxf5 21 '1!r'c2 '1!r'g6 22 l':tael '1!r'h6 (22 ...il.d4 also runs into 23 b6! cxb6 24ll:ib5 and 22 ...il.h 6 23 il.f2 is no help) 23 b6! favoured White in J.Nogueiras-A.Zapata, Matanzas 1994. c22) 16 b4 is the most logical when 16 ... axb3 17 ll:ixb3 has scored tremen­ dously for White, even though Black's position is not so terrible. That said, 11 .. .f5 (11 ... b6 18 a4!) 18 exf5 il.xf5 (18 ... gxf5 19 f4! also looks good for White) 19 il.d3 was still pleasant for White in l.Yanvarjov-0. Loskutov, Mos­ cow 1994. Returning to 13 b3:

Instead 13...a4 has been tried too, but White has made a certain gain in that he has reached this structure without spending a tempo on �bl as explained above. In fact, sometimes (after l':tb1) the white rook return s to al 22 7

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 13 h s ...

Black has a sharp alternative in 13 .. .fS!?.

Then : a) 14 f3 is played rather infre­ quently, but it may be best: al) 14...ll:if6 lS il.f2 (after lS l':tbl ll:ihs 16 �e1 ll:if4 17 il.fl gs 18 il.f2 g4 Black has counterplay) lS ...ll:ihs 16 g3 and it is not s o easy for Black t o create play on the kingside. a2) 14...b6 lS �bl f4 16 b4 axb4 17 axb4 gs 18 il.f2 hS transposes to varia­ tion 'a3'. a3) 14... f4 lS l':tbl (White could also play lS il.f2 or even lS b4!? axb4 16 axb4 ll:ixb4 17 �3 ll:ia6 18 �xb7 �8 19 l':tfb1 with the initiative) lS... gs 16 il.f2 hS 17 b4 axb4 18 axb4 b6 (18...�g6 19 cs g4 20 c6 gives White the advan­ tage) 19 cs! bxcs 20 bxcs ll:ixcs 21 il.xcs dxcs 22 d6 (22 ll:ic4 and 22 ll:ibs are good alternatives) 22 ...cxd6 23 ll:ic4 and White had the initiative in J.Kraai­ D.Gross, German League 1998. b) 14 exfs and here: bl) 14. ..gxfs is very sharp after lS 228

il.hs �c8 (Kasparov's exchange sacrifice) 16 il.e7 l':te8 17 il.xe8 �xe8 18 il.h4 e4. White has tried both 19 l':tcl and 19 �c2 here with no clear conclusions. b2) 14...il.xfs looks quite sound, in­ tending lS g4 e4 (1s .. .il.d7 16 ll:ide4 b6 is also feasible) 16 l':tcl e3!.

White has: b21) 17 fxe3 �xe3+ 18 il.f2 �g s 19 h3 (no better are 19 '1¥o>h1 il.d7 20 ll:ide4 �e7 or 19 h4 �f4 20 gxfs il.es 21 lLif3 �g4+ 22 '1¥o>h1 �h3+ with a draw) 19 ... il.d7 2o ll:if3 �f4 is level. b22) 17 gxfs exd2 18 �xd2 ll:ics 19 �dl (worse is 19 il.dl �xfs 20 �e2? ll:igs 21 �xe8+ �xe8 22 ll:ibs ll:ixb3 with a clear advantage, P.Enders­ V.Babula, Berlin 1996) 19 ...�xfs 20 b4 (after 20 il.g4 �f4 21 �el Black can play 21...�f8 or sacrifice his queen with 21 ...�xel+ 22 �xel �xg4+ 23 il.g3 ll:igs 24 �e2 hS) 20 ... axb4 21 axb4 ll:ie4 22 ll:ibs l':tf7 23 il.g4 lLihg S with ap­ proximate equality in l.Naumkin­ D.lsonzo, Lido Estensi 2003. While 13 ...fs is playable, I prefer 13 ... h s to which we now return. Do also

Petrosian Varia t i o n : 7 d5 a 5 8 1lg5

note that White can sidestep 13 .. .fs with 13 '1¥o>h l !? as discussed above. 14

f3 1lh6

This keeps Black's bishop out of e3, but it allows Black to improve the posi­ tion of his queen. 15 '1lr'e7! From here the queen helps to hold up White's cs-advan ce, while it may become active via the gs-square. Typi­ cal ideas for Black include ...h4,...1lf4, ...ll:if6-hS and ... '1¥o>g 7 when a piece may hop into g3 and ... l':th8 will be a possibility too. ...

16 '1!r'c2

Black has activated his bishop, which may come to e3. White can pre­ vent the intrusion, dodge the check or continue his queen side play. We have: 81: 1s i.f2 82: 15 �h1 83: 1s .:b1

Line Bl is independent, whereas Line B2 can easily transpose into B3. 81) 15 1lf2

By far the most common move, but there are others: a) 16 �bl '1!r'g s looks strong, but af­ ter 17 l':tb2 .lii.h3 18 .lii. g3 h4 (Black can force a draw with 18 ... '1lr'e3+ 19 1lf2 '1lr'g s, but must avoid 19 ... '1lr'xC3? 20 l':tc2) 19 f4 (not 19 gxh 3 hxg3) 19 ... exf4 20 ll:if3 '1lr'f6 21 es dxes 22 il..xh4 g s 23 gxh3 gxh4 24 .lii.d3 matters are not so clear. Black could also just play the move 16 ... h4. b) 16 �a2 protects the d2-knig ht. Black should just play 16 ...h4. c) 16 l':tel h4 17 il..fl 1lf4 18 �bl '1!r'g s 19 �b2 ll:if6 intending ... ll:ih s gives Black a strong attack. d) 16 '1¥o>h 1 seems to be asking for trouble because a knight may come to g 3 with check. Following 16 ...h4 17 �gl?! (even after 17 �bl ll:ics 18 b4 axb4 19 axb4 ll:ia4 20 ll:ixa4 �xa4 21 cs h3 22 g3 z:ta2 23 ll:ib3 fS Black had the initiative in V.Antoshin-B.Kantsler, Ni­ kolaev 1981) 17 ....lii . gs 18 '1!r'e1 '1¥o>g7 19 �bl ll:ics 20 b4 axb4 21 axb4 ll:ia4 22 ll:ib3 ll:ixC3 23 '1lr'xC3 lLif6 24 �al ll:ih s 2s 229

A ttacking C h e s s : Th e K i n g 's I n d i a n , Vo l u m e 1

�a8 �xa8 26 l':tal l':th8 27 '1¥o>g1 il.f4 28 il.e3 '1!r'g s 29 �2 fS Black had enduring pressure in P.Kiss-V.Kotronias, Kavala 200S. 16...h4 17 �fdl

11 ... il.f4

Black can also lead with the queen. One successful example of this idea was 11 ...'1!r'g s 18 il.f1 ll:if6 19 l':tab1 ll:ihs 20 b4 axb4 21 axb4 fS 22 cs h3 23 g3 f4 24 g4 il.xg4 2S fxg4 '1!r'xg4+ 26 '1¥o>h1 ll:ig3+! 27 hxg3 fxg3 28 il.e2 g2+ 29 'it>gl h2+ 0-1, G.Flear-R.Douven, Charl­ ton 1983.

Now 2 1...ll:ihs was given as clearly better for Black by Strikovic in Infor­ mant 45, based on 22 cs dxcs 23 bxcs ll:ixcs 24 il.xcs? il.e3+ 2s il.xe3 '1!r'xe3+ 26 '1¥o>h1 ll:ig3+ 27 hxg3 hxg3 and 0-1 in L.Santa Torres-A.Strikovic, New York 1988. However, White later improved with 24 ll:ic4! in L.Santa Torres-M.Al Modiahki, Moscow Olympiad 1994, when he had good compensation for the pawn. Instead I think that Black should play: 21...il.e3! Black has a good position.

18 �ab1

18 ll:ifl '1!r'g s 19 il.d3 ll:if6 20 �ab1 was drawn here in B.Malich­ W.Schmidt, Decin 1976, but Black could certainly continue with 20 ...ll:ics, 20...'i¥o>g7 OT 20 ...ltJh S. 18 ...'1!r'gs 19 il.f1 ll:if6 Black has a couple of alternatives here that should be investigated. 19 ... il.e3 fights for the dark squares and 19 ...ll:ics avoids Black's knight being potentially sidelined on a6. 20 b4 axb4 21 axb4

230

82) 15 'it>hl

Petro s i a n Va ria tio n : 7 d5 a5 8 il.95

This is a tricky prophylactic move. Often play will transpose to Line B3, but here we will only look at independ­ ent lines where White generally avoids or delays l':t bl. 1s il.e3 Black has a couple of alternatives: a) 1s ...'1!r'b8 intends to bring the queen to a7. Then 16 '1!r'c2 il.e3 (or 16 ...'1!r'a7 17 l':tael!) tran sposes to note 'f to his 16th move, below, but White could also play 16 il.f2 when 16 ...'1!r'd8 17 '1lr'c2 gives him an extra tempo over Line Bl, with Black's queen on d8 in­ stead of the more flexible e1-square. b) 1s ... ll:ics is worth considering. Af­ ter 16 l:tb1 (but not 16 b4? il.xd2) 16 .. .fs 17 '1!r'c2 ll:if6 18 b4 axb4 19 axb4 ll:ia4 White has: bl) 20 exfs ll:ixc3 21 '1!r'xc3 il.xfs 22 ..i..d3 l':ta2 23 ll:ie4 .�xe4! 24 fxe4 ll:ig4 gave Black the initiative in l.Naumkin­ A.Khalifman, USSR 1984. b2) 20 ll:ibs was played in l.Naumkin-J.Trapl, Namestovo 1987. Black's best is 20...il.xbs 21 cxbs il.xd2 22 '1!r'xd2 fxe4 23 il.xf6 �xf6 24 fxe4 which Palliser considers slightly better for White because of the open c-file, but after 24 ... '1lr'e7 Black is probably close to equality. The e4-pawn i s also a potential weakn ess and White's bishop is no better than Black's knight. 16 '1!r'c2 This is critical. White does not bother with l':tabl and instead prepares for war across the board. Instead 16 :b1 transposes to Line B3, while 16 ...

z:ta2 '1¥o>g7 17 '1!r'b1 il.cs 18 ll:id1 c6 19 ll:ib2 should be met by 19 ... il.d4 20 ll:id3 ll:ics with equality, rather than 19...cxds?! 20 cxds ll:ic7 21 ll:id3 il.d4 (M. Narciso Dublan-F.Nijboer, Barbera del Valles 2007) when 22 ll:ic4 gives White a large advantage after either 22 ... gs 23 il.e1 '1!r'e7 24 ll:ixas or 22...ll:ibs 23 �cl ll:ic3 24 l':txc3 il.xc3 2S ll:ixd6 '1!r'b8 26 il.e7.

16 fs •••

This move is very natural, but Black must keep in mind that White can also play on the kingside, especially if Black's dark-squared bishop goes wan­ dering off. The position is very rich and there are several possibilities, of which 'f' is the most interesting: a) 16 ... c6 looks premature: 17 il.f2 (even 17 cs!? is dangerous: for exam­ ple, 17 ...il.xcs 18 ll:ic4 intending ll:ia4 with the initiative) 17 ...Lf2 18 �xf2 '1!r'd8 19 l:tffl ll:if6 20 '1!r'b2 h4 21 f4! (with the dark-squared bishops ex­ changed and Black's knights unable to reach the es-square, this is a strong break) 21 ... h3 22 g3 ll:ig4 23 fS ll:ie3 24 231

A t tacking C h e s s : The King 's I n dia n , Vo l u m e 1

�f3 '1!r'b6 25 lLifl ll:ig4 26 '1!r'd2! '1¥o>g7 (26 ...ll:if2+ 27 l':txf2 '1lr'xf2 28 ll:ie3 wins) 27 '1!r'gs ll:if2+ 28 �xf2 '1lr'xf2 29 ll:ie3 il.e8 30 f6+ 1-0 was M.lllescas Cordoba­ B.Gelfand, Linares 1994; 30 ...'1lr'xf6 31 ll:ifs+ wins. b) 16 ... gs is inflexible, but solid: 17 il.f2 il.xf2 18 �xf2 h4 (worse is 18.. .fS?! 19 exfs .lii.xfs 20 il.d3! with a big ad­ vantage in G.Vescovi-A.Bachmann, Sao Paulo 2006) 19 l':tg1!? '1lr'e7 20 g3 '1¥o>h8 21 ll:if1 fs 22 ll:ie3 f4 23 ll:ifs '1lr'f6 24 gxh4 gxh4 25 l':tfg2 with just an edge for White in Liang Chong-Ye Jiang­ chuan, HeiBei 2001. c) 16 ... ll:ics 17 l':tabl fS 18 b4 axb4 19 axb4 ll:ia4 20 ll:ibs (Palliser suggests 20 ll:id1) 20...il.xbs 21 cxbs f4 22 ll:ic4 il.d4 with a strategically unclear position in S.Savchenko-V.Grigoriev, Alushta 1994. d) 16 ... '1¥o>g7 17 �ael '1!r'b8 18 il.d3 il.cs 19 ll:idb1 gs 20 il.f2 h4 21 ll:id1 bS?! (this just weakens Black even fur­ ther on the light-squares) 22 ll:ie3 '1¥o>h8 23 cxbs il.xbs 24 il.xbs '1lr'xbs 25 ll:ic3 '1lr'b7 26 a4 �g8 27 ll:ibs ll:if6 28 ll:ic4 was much better for White in R.Palliser­ G.Jones, British Championship, Torquay 2009. e) 16 ... il.cs 17 il.f2 il.xf2 18 �xf2 fS (18...'1lr'e7 is safer) 19 exfs gxfs 20 l':tel gave White some initiative in R.Palliser­ L.Trent, British League 2002. f) 16 ... '1!r'b8!? and now: fl) 17 il.f2 seems to justify Black's play: 17 ... '1lr'a7 18 il.xe3 '1!r'xe3 19 f4 (Black also reached a good position af­ ter 19 �ael '1!r'h6 20 ll:id1 h4 21 '1!r'b2 232

ll:ics in D.Cummings-P.Cramling, Pula 1997) 19 ... exf4 20 l':tael '1!r'cs 21 '1!r'c1 '1lr'd4 22 �4 �ae8 23 l':tffl '1lr'g7 with a good position for Black in A.Shirov­ J.Polgar, Madrid 1994. f2) 17 l':tael! is Kramnik's plan. White intends ll:id1, '1!r'c1 and il.d3-c2 when all of his pieces will be aimed at the centre and kingside.

Black must take some care: f21) 17 ... '1lr'a7 18 ll:id1 il.cs 19 '1!r'c1 l':tae8 20 il.d3 c6 21 ll:ib1 '1!r'b6 (better is 21 ...'1¥o>g7 22 ll:ibc3 with just an edge for White according to Kramnik) 22 il.c2 '1¥o>g7 23 ll:ibc3 '1lr'a7 24 il.g3 ! was V.Kramnik-J.Nunn, German League 1994. White intends 25 f4! with the initiative. f22) 11 ... il.cs 18 ll:idb1 '1lr'a7 19 ll:id1 c6 20 il.d3 cxds 21 exds (21 cxds l':tac8 looks okay for Black) 21 .. .fs 22 '1!r'd2 '1!r'b6 23 il.c2 (after 23 '1!r'h6 il.e8 24 il.c2 '1lr'C7 Black is ready to evict the white queen with ...'1lr'g7) 23 ... '1¥o>g7 24 ll:if2 il.d4 25 ll:id3 ll:ics and Black had good play in T.Polak-L.Gofshtein, Vienna 1996.

Petrosian Va riatio n : 7 dS as 8 ii.gs 17

2 0 '1¥o>h8 21 '1!r'b2 tt:\cs

exfs gxfs

•••

Not 17 ... il..xfs? 18 il..d3 when Black's bngside is starting to creak.

.-�������-

18 ii.fl

Or 18 �ael '1!r'g6 19 il..d3 and here: a) 19 ...il..c s 20 tt:ldb1 '1¥o>h8 21 il..f2 :gs 22 �gl �g7 and Black had coun­ terplay in R.Tuominen-M.Hartikainen, Finland 1997. b) 19 ...il..h6 20 g4! ? '1¥o>h8 21 gxfS

-

22 b4 axb4 23 axb4 tt:\a4 24 tt:\xa4 z:txa4 25 il..d1 �as 26 f4

-

according to Komarov. That may be a little bit extreme, but Black has good play in any case. 18 il..xf2 •••

After 18 ...il..h 6!?, as played in Y.Pelle­ tier-G.Mohr, Buekfuerdo 1995, switch­ ing back to the queenside with 19 �abl has been suggested by Palliser. 19 z:txf2 tt:\f6 20 �gl! Here we see another plan from Kramnik. White again plays on the kingside.

26 exf4 •••

Opting for piece play, but 26 ... e4 27 tt:\f1 '1!r'g6 28 tt:\e3 '1¥o>h7 also looks okay for Black. 27 �e2

233

A ttacking Chess: Th e King 's I n di a n , Vo l u m e 1

27 l':txf4 �es ends White's initiative. Black's position looks a little loose, but there are not many pieces on the board and he has enough activity to hold the balance.

1 S....lii.e3+ 16 '1¥o>h1 16 .lii.f2 is also possible, though it is less dangerous.

21 ...�g6

Black could also keep e7 covered by playing 27 ...�f7. After 28 �d4 bS! 29 .lii . b3 l':tae8 30 l':txe8 l':txe8 31 �xf4 �e7 Black has counterplay. Moreover, 32 �h6+ gets nowhere after 32...'1¥o>g8 33 �g6+?! �g7 34 �xg7 'it>xg7 and Black was even better in the endgame in S.Cade-B. Benko, correspondence 2004. 28 l:te7 �f7 29 �xf7 �xf7 30 �d4

Not 30 .lii.xh s �g7 31 .lii.f3 ll:ig4 when Black has the initiative. 30...�g7 31 �xf4 This was V.Kramnik-B.Gelfand, Linares 1994. Instead of 31 ... '1¥o>h 7?! Black should have played 31 ... bS! with counterplay. 83)

15

�bl

White gets on with his queenside play. This does not exclude White from operation s on the kingside, however. 234

After 16 ... .lii.xf2+ (16 ... .lii.cs is sound, e.g. 17 .lii.xcs ll:ixcs and now both 18 �c2 �e7 19 b4 axb4 20 axb4 ll:ia4 and 18 b4 axb4 19 axb4 ll:ia4, as in V.Anto­ shin-M.Chiburdanidze, Baku 1980, look fine for Black) 17 �xf2 Black has: a) 11. . .ll:ics 18 �c2 �e7 19 b4 axb4 20 axb4 ll:ia4 21 cs (after the immedi­ ate 21 ll:ibs, Black can play 21 ... .lii.xbs 22 cxbs ll:ib6) 21 .. .fs?! (a lapse; 21 ...ll:ixc3 22 �xc3 fS was better) 22 ll:ibs ! and Black's a4-knight was in trouble, F.Vallejo Pons-F.Jenni, Aviles 2000.

Petro s i a n Va ria tio n : 7 d5 a 5 8 il.g5

b) 17 ... 'fke7 is solid when the posi­ tion resembles a Bogo-lndian. With some pieces exchanged, Black is less worried about his a6-knight being side­ lined, because White's rook will be tied to the defence of the b4-pawn. After 18 b4 axb4 19 axb4 practice has seen: bl) 19 ... h4 20 'fkcl (the more enter­ prising 20 cs!? dxcs 21 bS lLib4 22 lLic4 with compensation has been played a couple of times with success by Porper) 20... l':tfb8 (more solid is 20 ... cs with equality) 21 cs! dxcs 22 bS lLib4 23 �c4, intending d6 or f4, gave White compensation for the pawn in F.Vallejo Pons-0.Cvitan, European Team Cham­ pionship, Plovdiv 2003. b2) 19 ...cs prevents White's sacrifi­ cial idea: 20 dxc6 (20 bxcs tLixcs 21 -':Jb3 b6 was equal in 1.Khenkin­ D.Bokan, Moscow 1989) 20...bxc6 21 -'!Ja4 (21 bs lLics 22 lLib3 lLixb3 23 .l:txb3 -':Jgs 24 il.fl lLie6 heading to d4 or cs can only favour Black) 21 ...l':tfb8 22 cs occurred in B.Zlotnik-A.Kuzmin, Buda­ pest 1989. Now 22 ... il.e6 is level.

16 ...il.c s

Black uses the bishop t o prevent b3b4 and hits the a3-pawn at the same time. Others: a) 16 ... tLics looks inconsistent with the bishop on e3. After 17 'fkc2 (17 lLJbs 'fkb8!? 18 b4 axb4 19 axb4 lLJa4 20 l':tb3 is also possible) 17 ...fs 18 b4 axb4 19 axb4 lLia4 20 lLibs il.xbs 21 cxbs f4 22 lLib3 'fkd7 23 tLias! gs 24 il.f2 il.xf2 2s l':txf2 lLib6 26 lLixb7 White had won a pawn in A.Yusupov-P.Scheeren, Eind­ hoven 1986. b) 16 ... fS 17 exfs (after 17 b4 axb4 18 axb4 f4 Black has good play on the kingside) 17...gxfs 18 f4! exf4 19 il.xhs 'fies is very unclear. Black has played this several times with mixed results, but I do not like Black's ragged pawn structure and open king position. 17 'fkc1

11 ...fs This is the most direct and looks sat­ isfactory. The position is very rich and both sides may initiate play on either side of the board. While the text move is natural, there are other moves worth considering: 23 5

A tt a c k i n g C h e s s : Th e King 's I n dian, Vo l u m e 1

a) With 17 .. .f6!? Black finds a way to initiate play on the kingside without opening things up too much: 18 ll:ia2 gs 19 il.e1 (19 il.f2) 19 ... g4 20 b4 axb4 21 axb4 il.e3 was unclear in l.Naumkin­ V.Tsarev, Moscow 1988. b) 17 ...'1¥o>h8 is a normal waiting move, but here it is slow and well met by 18 ll:ia2!. Now: b1) 18 ... gs 19 il.e1 (19 il.f2!?) 19 ... fs 20 b4 axb4 21 axb4 il.e3 22 exfs il.xfs 23 z:tb3 il.d4 24 ll:ic3 ll:if6 2s ll:ibs il.b6 26 il.f2 was agreed drawn in a fairly level position, l.Naumkin-T.Paehtz, Bu­ dapest 1991. b2) 17 ...fs 19 b4 axb4 20 axb4 il.e3 21 ll:ic3 cs 22 dxc6 bxc6 23 '1!r'd1!? il.d4 24 ll:ia2 ll:ic7 2 s ll:ic1 ll:ie6 26 exfs gxfs 27 ll:idb3 looked better for White in l.Naumkin-J.Fedorowicz, London 1990. Black's position is rather loose. c) 17...c6 increases the tension. Now both 18 dxc6 bxc6 19 ll:ia4 il.d4 and 18 '1!r'b2 il.d4 19 b4 axb4 20 axb4 cs look fine for Black, while 18 il.f2 il.xf2 19 l::txf2 gives Black a pleasant choice be­ tween 19 ...'1!r'd8, 19 ...'1lr'e7 and 19 ...'1!r'b8!?.

236

That leaves: cl) 18 ll:ia4 il.d4 19 il.f2 il.xf2 20 l':txf2 '1!r'd8! (again the position looks like a Boga-Indian) 21 dxc6 il.xc6 22 ll:ic3 ll:ics 23 b4 axb4 24 axb4 ll:ie6 2s bS il.d7 26 ll:ib3 '1!r'b6 27 l':tfl ll:if6 gave Black an edge in A.Yusupov-B. Daml­ janovic, Saint John 1988. c2) 18 ll:ia2 is the best try: 18 ... cxds 19 b4! (Black gets counterplay after both 19 cxds bS! and 19 exds fS!) 19...axb4 20 axb4 il.e3 21 exdS! (better than 21 cxds when Black gets good play with either 21 ... �c8 or 21 ... '1!r'b8) when Black's position looks quite good.

However, actually Black must be careful: 21 ... il.h6?! weakens cs and af­ ter 22 ll:ic3 fS?! 23 '1!r'c2 (23 cs! looks even stronger) 23 ...l':tc8 24 l':tfel g s 2S il.f2 g4 26 cs! White had seized the ini­ tiative by targeting the es-pawn in l.Naumkin-Ye Jiangchuan, Belgrade 1988. A better try would have been 21 .. .fs!?. 18 exfs gxfs 19 il.f2 After 19 f4 e4 Black looks fine, while 19 ll:ia2 '1!r'g6 20 b4 axb4 21 axb4 il.e3

Petro s i a n Va ria t i o n : 7 d5 a 5 8 il.g5

22 il.f2 il.xf2 23 l':txf2 has been seen a few times. Then the untried 23 ... lLif6! looks flexible and best. 19 lLif6 Also satisfactory is 19 ... �g6 20 il.xc S tLixcs (Black has also tried 20 ... dxcs?!, but I could not do that to the a6knight) 21 b4 axb4 22 axb4 lLia4 with play similar to the main line in Z. Franco-C.Checa, Seville 1994. ...

20 il.xcs tLixcs

Compared to Kramnik-Gelfand in Line B2, Black is even better off because White's pieces are passively placed.

21 b4 axb4 22 axb4 tLJa4

Black certainly had no problems in R.Pogorelov-Z.Plenkovic, Budva 2009.

237

Chapter 1 3 Excha nge Variation 7 dxes dxes 8 1fxd8 !txd8

1 d4 ll:if6 2 c4 g6 3 ll:ic3 1lg7 4 e4 d6 5 ll:if3 0-0 6 Ji.el es 7 dxes dxes 8 '1!r'xd8 l':txd8

The 'dreaded' Exchange Variation. This notorious drawing line has gone so far as to discourage some players from even playing 6 ... es. While the Ex­ change Variation does have obvious drawing tendencies, Black should cer­ tainly not give up hope of playing for a win. As Gallagher has mentioned be­ fore, this line is primarily played by a few different types of players. I will classify these opponents as follows: 238

1. Beggars

These players just want to try to make a draw. Usually the player of the white pieces is lower rated or perhaps the tournament situation dictates that White should try to draw. Unless a quick draw suits the black player as well, I absolutely agree with Gallagher that these players should not be granted a draw until every resource has been exhausted. It is my experience that m any beggars will crack at the first sign of trouble. At the very least we should make the game as miserable an experience for them as possible. 2. Psychologists

Some players will play the Exchange Variation anticipating that Black will feel uncomfortable in the endgame. There is some logic to this; most King's Indian players are striving for a sh arp game. Nevertheless, if Black is well pre­ pared this approach will lose its sting. It is important in general, of course, to

Exch a ng e Va r i a t i o n : 7 dxe5 dxe5 8 'jkxd8 'gxd8

be a well-rounded player capable of continuing to play even in equal end­ games. 3. S pecialists

Some players just love to pl ay dry, technical chess. These players believe in White's chances in the Exchange Varia­ tion and are happy to try to prove that he has a little something. These players are the most dangerous, because unlike beggars and psych ologists, they are likely to be well prepared. Special­ ists will have studied the detail s of the Exchange Variation and will be ready to take advantage of lackadaisical play by Black. 9 1i. gs

It is well known that g ll:ixes can be met by 9 ...ll:ixe4 when Black is at least equal. White can also force things with 9 ll:ids. After 9 ... ll:ixds 10 cxds c6 11 l.c4 cxds 12 il.xds ll:id7 13 Ji.gs 'ge8 we have tran sposed to Line A2, below. If Black wants to avoid this, he can play 11 ... bs!? 12 il.b3 il.b7 13 il.gs and then either 13 ...'gd7 or 13 ... 'gc8.

The text reaches the basic position for the Exchange Variation. The pin gives White a certain pressure against both f6 and c?. Indeed, both 10 ll:ids and 10 hf6 followed by 11 ll:ids are threatened. However, if Black can neu­ tralize White's slight initiative, he can feel confident, because White's 9th move left him with a weak d4-square, whereas Black can still guard the dS­ square with ... c6 at some point. From a theoretical perspective, Black really only needs to study Line A, but I will examine three systems in de­ tail. One reason for this is variety some players may face an opponent that frequently uses the Exchange Variation, so it is nice to have more than one weapon. Another reason is circumstantial - Black may want to either play solidly or take some risks, depending on the opponent and the situation. We will explore: A: 9 .�e8 B: 9. ll:ibd7 C: 9... c6 ..

..

Th e first of these lines, 9 ... 'ge8, is a solid equalizer that was favoured by Fischer. This is the line recommended by Gallagher as well. White has great difficulties obtaining even a shred of an advantage against 9 ... l:te8, but be­ cause a great deal of simplification of­ ten occurs, it can also be difficult for Black to try to win. Line B is more com239

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1

bative, but White also has greater chances of getting an edge. Neverthe­ less, White's advantage is not large and this is a good choice as Black for those that want to try to gradually outplay their opponents. Line C is a gambit con­ tinuation. This option is quite popular and it is in decent theoretical shape. 9 ... c6 is the most combative option, but Bl ack does sacrifice a pawn, so he does risk more than in Lines A and B. There are two other moves. a) 9 ... tlla6 is probably not very good. It is worth examining though, if only to compare it to a line we will see later in the Makogonov Variation (in Volume 2). Play continues 10 tllds (10 tllxes l:te8) 10 ... l:td6 11 �xf6 �xf6 12 tllxf6+ l:txf6 13 tllx es l:te6 14 f4 when Black will have trouble getting his pawn back without making any concessions.

The unsatisfactory choice is: al) 14.. .f6 15 tll g 4 hs 16 tllf2 tllcs 17 o-o (or 17 �f3 tllxe4 18 �xe4 fS 19 �d2 fxe4 20 �e3 with a nice advantage) 17 ...tllx e4 18 tllx e4 l:txe4 19 �d3 and White is clearly better. 240

a2) 14....rf.e8 1 5 o-o-o (15 �f3 f6 16 tll d 3 �fs 17 tllf2 tll cs is level) 1s ...f6 16 tlld 7! �xd7 17 .rtxd7 .rtac8 18 �f3 tllcs 19 l:tddl! with the idea of 19 ...tllx e4 20 l:thel. a3) 14...tllcs 15 o-o-o! f6 16 l:td8+ (16 tllg4 tllxe4 17 �f3 is a little better for White) 16 ... �g 7 and now White has several ways to get an advantage, in­ cluding 17 tll d 3 (also good are 17 b4 tll x e4 18 �g4 and 17 tllf3 l:txe4 18 tlld4) 17 ...l:txe4 18 tllxcs l:txe2 19 tlld3 when he has the upper hand. b) 9 ... l:tf8! ? is an idea of Lanka that has been pioneered by Bologan and Shirov.

White has: bl) 10 tll d s tllxds 11 cxds c6 12 �c4 bS 13 �b3 �b7 14 l:tcl as 15 a3 (after 15 a4 bxa4 16 �xa4 cxdS! the main point of Black's play becomes apparent - if we compare this to the note to Black's 12th move in Line A2, we can appreciate that Black's rook is not pinned) 1S ... a4 16 �a2 l:tc8 17 �e3 b4 18 axb4 a3! gives Black good play as shown in several games.

Exc h a n g e Varia tio n : 7 dxes dxes 8 'ilixd8 1:l.xd8

b2) 10 �xf6 �xf6 11 liJds .�d8 12 liJxes leads nowhere after 12 ...1:l.e8 13 l:.d1 c6 14 liJC3 �as 1s lLlf3 l:txe4. b3) 10 o-o-o liJc6 (10 ... liJbd7 is Line B) 11 h3 �e6 12 �e3 .rf.ad8 13 a3 h6 (13 ... liJas ! ? 14 liJd2 liJc6 could also be tried) 14 b4 as lS bS liJd4! 16 liJxd4 exd4 17 �xd4 l:txd4 18 l:txd4 liJg4 19 �xg4 �xd4 20 �xe6 fxe6 21 liJd1 �cs 22 a4 l:td8 with some compensation for the pawn in Wang Yue-V. Bologan, Moscow 2006. b4) 10 liJxes is critical according to Bologan. After 10 ...liJxe4 11 liJxe4 �xes 12 o-o-o liJc6 13 f4 �d4 (or 13 ... �g7 14 .!l:if6+ �g7 1s h4!) 14 liJf6+ �g7 1s lLlds f6 16 �h4 �fs, instead of the 17 �d3 of R.Ravisekhar-V.Bologan, Calcutta 1992, Bologan gives 17 l:thel when White's good centralization gives him slightly the better prospects. We can conclude that 9 ... liJa6 is just dubious, but 9 ... l:tf8 is playable if Black is unhappy with the three (!) recom­ mendations given here.

This is the classical continuation. Black simply protects his es-pawn while breaking the pin on his king's knight. Now 10 �xf6?! .�xf6 11 liJds can be met with 11 ... �d8! when Black will play ...c6 and have everything he wants. 9 ... l:te8 has always been consid­ ered a solid equalizer, but correspond­ ingly it has a drawish reputation. Even so, I think that with careful study, Bl ack will be well prepared to seize the initia­ tive if White gets careless. There are two main lines, with the second being much more popular. A1: '10 0-0-0

A2: 10 ll'id5 Al) 10 0-0-0

A) 9 l:te8 ...

White just continues his develop­ ment, hoping th at his space advantage will mean more than Black's long-term control of d4. Generally some pieces stay on the board in this variation and there is scope for independent play. 10 liJa6 This is a natural way for Black to de...

241

A ttacking C hess: Th e King 's I n d i a n , Vo l u m e 1

velop. The knight covers the c7-square and may go to cs, from where it pres­ sures the e4-pawn and can continue to e6, with an eye towards the weakened d4-square. Instead 10... h6 is also possi­ ble, but it is not really necessary for Black to chase away the white bishop, so I prefer to keep this move in reserve. 11 tlle1

Instead 11 tllx es should be met by 11 ...tllc s ! (11 .. .l:txes 12 .l:td8+ tlle8 13 f4 l:te6 14 �g4 l:tb8 lS es is probably somewhat better for White, but here 12 ... l:te8 13 �xf6 l:txd8 14 �xd8 �h3 is probably okay) 12 �xf6 �xf6 13 f4 c6 when Black wins back the pawn, while 11 tlld2 c6 12 tllb3 (12 f3 tllcs 13 �e3 tlle6 is also fine for Black) 12 ... tlle7 13 f3 tlle6 14 �e3 �f8 lS a3 b6 16 .l:!.d2 �a6 17 �bl was agreed drawn in V.Akobian-A.Yermolinsky, Agoura Hills 2004.

12 tllc2

Instead 12 f3 tllcs 13 tllc2 leads to the same thing, but White could con­ sider controlling cs with 12 tlld3 ! ? here. For this reason 11 ... tll cs might be pre­ ferred. 12 ... tllcs 13 f3

This is a typical position. White can double rooks on the d-file and try to expand on the queenside. Black will often play moves like ... tlle 6, ...�f8, ...tllfd7, and perhaps ... as and ... b6. Ideally Black would like to get in ...�cs too to activate his dark-squared bishop.

13 ...tlle6

By avoiding ... h6 Black is able to play this move with tempo. Black could also maintain the knight on cs for the time being and play 13 ... as 14 �e3 (or 14 b3 �e6 with the idea of ... tllfd7) 14 ...�f8, which is level. 14 �e3 tllf4 11 ... c6

Black has also tried both 11 ...�e6 and 11 ... h6, while another move order for Black is 11 ... tllcs 12 f3 tlle6 (or the immediate 12 ...c6) 13 �e3 c6. 242

Black heads in a different direction and probes the kingside. He has also tried 14.. .il.f8 with the idea of ... �cs. If lS b4 as 16 a3 axb4 17 axb4 he could try 17 ... cs or 17 ... tllf4. However, I would

Exch a n g e Varia tio n : 7 dxes dxes 8 'ilixd8 '!J.xd8

prefer the prophylactic 14...as! ? when 15 tlla4 tlld7 intending ...�f8 looks fine for Black. 15 �fl hs 16 h4 �f8 17 b4 tlle6 18 a3 b6 19 �b2 �b7

Chances were about equal here in Lesiege- Smirin, Biel lnterzonal 1993.

exds e4 14 tlld2 Black can safely play 14...�xb2, while 13 '!J.xds �e6 13 �bs �d7 14 l:txds �xbs 15 '!J.xbs b6 16 �e2 tll a6 is fine for him too. d) The most interesting try is the gambit continuation 12 o-o!?.

A2) 10 tlld s

This forcing move is more popular. It is not necessarily better, but it forces simplification and generally leads to a symmetrical pawn structure.

10... tll x ds 11 cxd s c6

If Black refrains from this move, his backward c-pawn will become weak. 12 �C4

White wants to bring his relatively passive bishop to the nice dS-square. Other moves are also possible: a) 12 dxc6 tllxc6 just helps Black de­ velop. b) 12 d6?! f6 13 �e3 �e6 leaves the d6-pawn weak. Black will play ... '!J.d8 and ...�f8 to round it up. c) 12 '!J.dl is an inferior way to sup­ port the dS-pawn. After 12 ... cxds 13

Black has: dl) 12 ...cxds 13 exds e4 (after 13 ...�g4 14 h3 �xf3 15 gxf3 the bishop-pair was more important than White's fractured pawns in J.Jansson­ G.Gnichtel, European Club Cup, Fuegen 2006, but 13 ...tlld7 14 l:tfdl, as in J.Jansson-H.Harestad, Oslo 2006, and then 14 ... e4! ? is an idea) 14 tlld2 l:tes (after 14...�xb2 15 '!J.abl �g7 16 '!J.fc1 tlld7 17 d6 tllb6 White has obvious compensation for the pawn) 15 �e3 '!J.xds 16 tllxe4 tllc6 17 �c4 '!J.d8 18 '!J.adl �fS 19 tlld 6 gave White the ini­ tiative in J.Jansson-Thomassen, Norwe­ gian League 2008. d2) 12 ... h6 13 �e3 cxds 14 exds fs 15 '!J.fd1 tlld7 16 .l:!.acl f4 17 �cs tllxcs?! 18 '!J.xcs �g4 19 '!J.a was better for White in J.Jansson-K.Tryg stad, Euro­ pean Club Cup, Fuegen 2006, and 19 243

A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1

d6 ! looks even stronger. Instead Bolo­ ga n gives 17 ... e4 1s tll d4 tllxcs 19 l:txcs e3 l:tce8 19 l:td8 'iii>f8 was V.Korchnoi-G.Kasparov, Til­ burg 1991. White remains up a pawn, although it will be very difficult to con­ vert. b) 14 ..ih4 gs 1s fxgs hxgs 16 ..ixgs ll'lcs 17 ..ixf6 ..ixf6 18 ll'ld3 ..ie7 19 ll'lxcs (after 19 l:td4 ..if6 20 es ll'lxd3+ 21 l:txd3 ..ixes Black has compensation, while after 19 es ..ixd6 20 exd6 it is White who has compensation for the sacrificed material) 19 ... ..ixd6 20 ll'lxb7 is complicated, but it looks a little bet­ ter for White. 14 ..if3

Or 14 ..ixf6 ..ixf6 lS ll'ld3 (after lS J:thdl ..ixeS 16 fxeS aS O T 16.. lte7 Black is fine) 1s ... ll'lxd3+ 16 ..ixd3 (16 l:txd3 ?! ..ixc4) 16.....ie7 17 cs ..ixd6 18 cxd6 f6 19 es fxes 20 fxes l:tf8 when the posi­ tion is unclear, but I do not think Black is any worse. 14 ...ll'lhs 13 ...ll'lcs

This is the most popular move, but 13 ... h6 was Kasparov's choice. This al­ ways means something, but it still does not look like a lot of fun for Black. White has: a) 14 ..ixf6 ..ixf6 lS l:thdl (after l S ll'lf3 ..ixc3 1 6 bxc3 ll'lcs White must avoid 17 l:td4 ..ig4 18 h3 ll'lxe4 19 hxg4 ll'lg3 when Black wins, and play 17 ll'ld2 when 17.....ifS 18 es l:tad8 gives Black 2 60

Here 14...h6 lS ..ih4 gs does not work well because after 16 ..if2 g4? 17 ..ixcs gxf3 18 gxf3 Black is just down a couple of pawns. Instead 14...ll'lfxe4 lS ll'lxe4 ll'lxe4 16 ..ixe4 f6 17 ..ixf6 ..ixf6 18 ..id3 also looks in sufficient 15 ..ixhs gxhs 16 ll'ld3! This is better than 16 ..ih4 when Black can simply play 16 ... ..ixes 17 fxes ..ixc4 or try 16 .. .f6 17 ll'lf3 l:tad8 with counterplay.

Exc h a n g e Va riat io n : 7 dxes dxes 8 -.xd8 :xd8

White's es-knight is now attacked twice and .. .f6 is threatened, so the knight has to move. 13 ll'lg4

16 .....ixc3

Instead 16 ... ll'lxd3+ 17 l:txd3 ..ixc4 18 l:tg3 'iW8 19 es is much better for White according to Golubev. 17 ll'lxcs ..ib4 18 ll'lxb7 ..ic8 19 a3!

White is better. Black should proba­ bly play 19 ...l:txe4, but White has sev­ eral moves to keep some initiative: for example, 20 axb4 (other possibilities are 20 l:txc6 ..ixb7 21 l:te7, 20 l:td8+ 'iii>g 7 21 ll'ld6 ..ixd6 22 :xd6 l:txc4+ 23 'it>d2 and 20 cs l:tc4+ 21 'iii>b 1 ..ixcs 22 ll'lxcs .:xcs 23 l:thdl) 20...l:txc4+ 21 'iii> b 1 ..ixb7 22 l:thdl with a slight advantage. c32) 12 ... ll'lhs

This has been played several times but may not be best. Alternatively: a) 13 ll'lf3 ll'lcs 14 ..ie3 (14 ll'ld2 ..ies lS :d8 .rtxd8 16 ..ixd8 ll'lf4 only gives White trouble, while 14 es ll'le6 with the idea of ...ll'lef4 is given by Golubev) 14.....ixc3 lS ..ixcs ..ig7 ! 16 ..ie3 (16 es ll'lf4 17 ..ifl ..if8) 16.....if8 17 l:tddl is given by Golubev, although after 17 ...ll'lf6 18 ll'ld2 ll'lxe4 19 ll'lxe4 l:txe4 20 .rf.d8 'iii>g 7 or simply 17 ...l:txe4 this is probably not much for White. b) 13 ll'ld3 ! ? contains the a6-knight and may well be best: 13 .....if8 (13 .....ixc3 14 bxc3 l:txe4 lS l:td8+ 'iii>g 7 16 l:tel must favour White) 14 l:td8 (similar is 14 l:td4 ll'lg7 1s :d8 l:txd8 16 ..ixd8 ..ie6 17 ..if6, C.Lakdawala­ J.Peters, Irvine 1998) 14...l:txd8 lS ..ixd8 ..ie6 (1s .....ih6+?! 16 'iii>c2 ..ie6 17 ..ih4 ..ixc4 18 ..ixhs gxhs 19 f3 was better for White in J.Gustafsson-A.Matth aei, Hamburg 1993) 16 ..ig s ..ixc4 17 l:td1 ll'lg7 18 f3 ll'le6 19 ..ie3 looks somewhat better for White because of his strong central position. 13 ... ll'lcs

After 13 .....ixc3 14 bxc3 (14 ll'lh6+!?) 14...ll'lcs (Black does better to play 14... ..ixg4, transposing to the next line) lS ll'lh6+ 'iii>g 7 16 f3! ll'le6 17 ..id2! White keeps an advantage according to Golubev. A better move order for Black is 13 ... ..ixg4 14 ..ixg4 ..ixc3 lS bxc3 261

A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1

l:txe4 with the idea of 16 l:td4 l:tes.

a knight landing on the d3-square) 19 es ..ixes White starts to experience some difficulties because of the knight stuck on h6.

14 ll'lh6+

Instead 14 ..ie3 ..ixc3 lS ..ixcs ..ixg4 16 ..ixg4 l:txe4! is a clever trick given by Golubev. After 17 ..ixh s Black can play 17 .....ies! with equal chances. 14...@fs 15 ..ixhs Now lS f3 runs into 1S ...ll'le6! with the idea of ...ll'lhf4. Probably best is lS ..ie3! ll'lxe4 16 ll'lxe4 l:txe4 17 ..ixh S gx hS 18 l:td8+ �e7 19 l:txc8! .l:i.xc8 (worse is 19 ...l:txc4+ 20 @bl l:txc8 2 1 ll'lfs+ @d7 2 2 ll'lxg7 l:tg8 2 3 ll'lxhs l:txg2 24 ll'lg3) 20 ll'lfs+ @f8 21 ll'ld6 l:tce8 22 ll'lxe4 l:txe4 23 b3 as given by Golubev. White has the better pawn structure, but the position is drawish. 1s ... ..ixc3! 16 ..if3

Not 16 bxc3 ll'lxe4. 16 ... ..ies

This is queried by Golubev, who in­ stead suggests 16 ... ..ib4 17 l:tddl ll'lxe4 18 ..ixe4 l:txe4 19 l:td8+ l:te8 20 .l:i.xe8+ @xe8 or 16 ... ..ias!?. 11 .l:i.d2 After 17 l:tddl ..ic7! 18 l:thel fS ! (Black tries to exploit the possibility of 262

17 .....ic7!

This move is an idea of the Ameri­ can IM Joe Fang. Black intends ... ll'le6, hitting the gs-bishop and eyeing the f4-square. The bishop retreat main­ tains control of f4 and uncovers an at­ tack on e4-pawn. All of White's at­ tempts to defend it make some kind of tactical concession, as his line-up of pieces on the cl-h 6 diagonal will prove to be vulnerable. In stead, both 17 ... ..ig7 18 l:thdl ll'le6 19 ..ie3 cs 20 l:tds b6 21 es! and 17 ... ..ie6 18 b3 (or 18 ..ie3 b6 19 b3) 18... as 19 ..ie3 b6 (A.OnischukM.Golubev, Leningrad 1989) 20 @bl! (Golubev) allow White to fight for an edge. 18 .l:i.e2

In stead, 18 l:tel walks into 18 ... ..ias and 18 ..ie3 ll'lxe4 19 l:tel @g7 20 ..id4+ @xh6 21 .l:i.xe4 l:txe4 22 ..ixe4 ..if4 23 ..ie3 ..ixe3 24 fxe3 ..ie6 is a little better

Exc h a n g e Va ria t i o n : 7 dxes dxes 8 -.xd8 'IJ.xd8

for Black. 18 ...'iii>g 71 19 @bl Black has all sorts of tricks: for ex­ ample, 19 h4 f6 20 ..id2 ..if4! 21 ll'lg4 .ixg4 22 ..ixg4 l:txe4 and 19 g3 f6 20 .i..d2 ll'ld3+ 21 'iii> c 2 ll'les. In both cases Black wins back the sacrificed pawn

and maintains the initiative. 19 . ..ll'le6 20 h4 If the bishop retreats, Black simply plays 20 ...ll'lf4. 20 ...ll'lxgs 21 hxgs ..if4 Black wins back the pawn and stands well.

263

Part Ill The Samisch: variation d4 t'i)f6' z c4 g6 3 liJc3 i.g7 4 e4 d'6 5 f3 o-o 6 ..te3 tiJc6 7 ttJgez a6 8 1fd2 :bs

1

1 d4 ll'lf6 2 c4 g6 3 ll'lc3 .i.g7 4 e4 d6 5 f3

ming-up of the kingside pieces, the d4square is sometimes weak s ...o-o 6 .i.e3 The alternatives 6 .i.gs and 6 ll'lge2 will be discussed in Chapter 20. After 6 .i.e3 Black has a wide choice, as we noted in the Introduction. The most popular move today is definitely 6 .. cs, but I had no hesitation in going for: 6...ll'lc6 .

The Samisch Variation is White's other big main line with 4 e4. White creates a strong centre and maintains a great deal of flexibility. There are a couple of downsides to White's scheme of development, however. The move 5 f3 takes away the most natural square for White's king knight and developing it to the e2-square will block in the flbishop. Because of this slight gum2 64

This is the Panno Variation. It is flexible and ambitious. Black takes aim

Th e S d m is ch Va riation

at the d4-square with his knight, aim­ ing for active piece play rather than the protracted manoeuvring that is com­ mon with the more classical 6 ... es and 6... cs. Black holds ... es in reserve and after White overprotects the d4-square with ll'lge2, Black will look to the queenside with ...ftb8, ...a6 and ... bs. This plan takes into account that White may castle long and from a positional standpoint it highlights the fact that the c4-pawn is lacking defence. The Panno is not very popular nowadays, but I think this is mostly due to the success of 6 ... cs at high levels and consequently the decline of the Samisch in general. 6 ...ll'lc6 has re­ mained popular in correspondence play and has been played frequently by Nunn and Gufeld, while Kaspan;>v has been found on both sides of the board. I believe it is due for a fresh look. 7 ll'lge 2 White sometimes plays 7 ..ifd2 first, although it makes little difference unless Black intends to play the

Westerinen Variation of Chapter 20, in which case 7...l:te8 must be played, be­ cause 7 .. ltb8? 8 ds ll'les 9 i..x a7 wins a pawn as the c4-pawn is still protected. This and other early deviations will be discussed in Chapter 20. 1 a6 8 ..ifd2 ftb8 ...

Black is ready to play ... bs and so White must decide on a plan. He may play the aggressive 9 h4, which is con­ sidered in the next three chapters, pre­ pare to develop his kingside with g ll'lcl (Chapter 17), initiate kingside play without 9 h4 (Chapter 18), or play on the queenside (Chapter 19).

265

Chapter 14 Panno Variation g h4 without 9 hs ..•

1 d4 ll'lf6 2 c 4 g 6 3 ll'lc3 .i.g7 4 e 4 d 6 5 f3 o-o 6 .i.e3 ll'lc6 7 ll'lge2 a6 8 'ilfd2 .l:i.b8 9 h4

with 9 ...e 5 or initiate queenside play with 9 . .bs. Although 9 ...h s is considered the main line nowadays and will be the subject of our next chapter, 9 .. es and 9.. bs have not been refuted and they lead to complex play. I imagine that these two lines could prove to be un­ pleasant to face for some white play­ ers, especially those who intend to play the positional 9 h4 h S 9 ll'lcl lines of Chapter 16. .

.

.

A: 9 es B: 9 bs ...

This is the most aggressive move, al­ though White may still play position­ ally in certain lines. The move's basic intention is obvious - White intends to attack the black king. The play very of­ ten is similar to the Sicilian Dragon. Usually Black halts the advance of the white h-pawn at once with 9 ... h s, al­ though there was a time when it was thought that this move was too weak­ ening. He may also strike in the centre 266

...

A) 9 es ...

Black follows the logic that a flank attack should be met with play in the centre. However, because the d4square is well protected, it turns out that Black actually cannot play in the centre after all. Indeed, if 9 ... es is to prove viable, it will be by combining defence and counterattack.

Pa n n o Va ria tio n : 9 h4 with o u t 9 . . . h 5 12 .....id7 13 ..ih6

It is hard to believe that 13 .l:i.cl can trouble Black after 13 ...b5, while 13 o-o-o b5 14 ..ih6 ..ixh6 15 �xh6 trans­ poses to the main line. 13 .....ixh6

Geller's old book only considered 11 b3, but this move just weakens the queenside for no reason and Black has good play after 11...c5, intending ...b5. The text move activates White's king­ side and threatens 12 b4.

This is a common idea that is also frequently seen in the Dragon. Black invites the white queen to h6 hoping that White's queenside will prove to be more vulnerable from its absence there. It is also possible to play 13 ... b5!? as after 14 ..ixg7 'iii>xg7 if White wants to invade with his queen, he must play 15 hxg6 fxg6 first, which could prove to be premature. After 16 'ilfh6+ 'iii> g 8 Black can defend himself laterally along the second rank.

11... cs 12 hs

14 'ilfxh6 bs 15 o-o-o

10 ds ll'las 1 1 ll'lg3

After 12 .l:i.bl I think Black should play Gallagher's recommendation 12 ...h5!, preventing White from having his way all over the board. Instead 12 ...b6 13 h5 ..id7 14 b4 cxb4 15 l:txb4 "fic7 16 h6 ..ih8 17 'ilfb2 b5 18 cxb5 .:.fc8 19 bxa6! was winning for White in V.Kramnik-V.lsupov, Kuibyshev 1990.

White can also play 15 ..ie2 when 15 ...'ilfe7 (this is the most popular move, but I do not like it; the queen should hang around because White has ll'lf5 ideas in the air, but d8 is a good, flexible square! ) 16 o-o-o ll'lxc4 (Black could also try 16 ...'iii>h 8!? or 16 ... l:tb7) 17 ..ixc4 bxc4 18 ll'lfl allows him a typical 267

Attacking C h ess: Th e King 's Indian, Vo l u m e 1

build-up with ll'le3, :d2 and g4. More critical is the immediate 15 ... ll'lxc4 16 ..ixc4 bxc4 17 0-0-0 l:tb4 when Black's queen can still go to either side of the board.

2 1 ..We3 gxhs 2 2 ll'lg3 i s less clear) when Black's attack is the stronger. For ex­ ample, 20 'ilfxf6? .l:i.xb2 and White will not last long.

1s ...ll'lxc4

B) 9...bS

Instead 1S ... 'ilfe7 16 l:td2 .l:i.b7 was seen in Y.Kraidman-H.Westerinen, Gausdal 1983, but this looks too defen­ sive to me. 16 ..ixc4 bxc4 17 ll'lfl ..Was! Now that there is no ll'lfs to worry about, the queen takes up an aggres­ sive post on the queenside. 18 .lld 2

After 18 g4? l:txb2! (also strong is 18 .....ia4! 19 l:td2 l:txb2! 20 ll'lxa4 l:tb7, winning) White resigned in 5 U. Lenhardt- .Crowdy, correspondence 1985. Following 19 'iii>xb2 l:tb8+ 20 @c2 ..ia4+ 21 ll'lxa4 ..Wxa4+ 22 @d2 ll'lxg4! 23 fxg4 c3+ White's position falls apart. 18 ...l:tb4

Black has good counterplay. For ex­ ample, the natural 19 ..Wgs can be met by 19 ...l:tfb8 (19 ... ll'lxhs 20 l:txhs .l:i.fb8 268

Black continues with his plan. 10 hs es

This seems similar to the previous line, but here Black has played ... bS to attack c4 first, so the move ... cs will not be necessary (although it is still possi­ ble). There are a couple of interesting alternatives too: a) 10 ... bxc4 and then: al) 11 g4 ..ixg4!? (Black could inves­ tigate other moves) 12 fxg4 ll'lxg4 13 o-o-o ll'lxe3 14 �xe3 e6 15 hxg6 hxg6 (in stead 1S ...fxg6 16 'ifh3 ..Wgs+ 17 'iii>b1 l:txb2+ 18 'iii>xb2 .rf.b8+ 19 @al ll'lb4 20 ..Wxe6+ ! 'iii>h 8 21 a4 ll'lc2+ 22 @a2 ll'lb4+ 23 'iii>b l ll'ld3+ 24 'iii>al is winning for White) was G. Kasparov-B. S passky, Nik­ sic 1983, and now 16 ll'lgl! looks good for White. a2) 11 hxg6 fxg6 12 ll'lf4 e6 (after 12 ... ll'las 13 o-o-o c6 14 g4 White has

Pa n n o Varia t i o n : 9 h 4 wit h o u t 9 . . . h 5

the initiative, as shown i n some of Murey's games) 13 ..ixc4 dS 14 ..ib3 l:txb3! lS axb3 dxe4 16 o-o-o exf3 17 gxf3 is unclear after either 17 ...0ie7 or 11 ... l'iJas. a3) 11 ..ih6 has been considered a bit of a refutation of Black's play. I am not so sure about this, though: 11 ...0ib4! 12 0ig3 ..ixh6 13 'ilfxh6 0ic2+ 14 'iii>d l l'iJxal lS hxg6 (after 1s l'iJds, instead of 1 s ... e6 16 hxg6 which was winning for White in M.Ceteras-D.Elliott, Mamaia 1991, be­ cause of 16...exds 17 l'iJhs, Black can play 1s ...gS! abruptly ending the white attack) 1 s.. .fxg6 16 l'iJdS ..ie6 (but not 16....l:i.f7? 17 0ixf6+ exf6 18 ..ixc4, win­ ning) and here:

how he can untangle. b) 10...l'iJas 11 0ig3 (11 0,f4?! pre­ vents White from playing ..ih6 and makes little sense) 11...0ixc4 12 ..ixc4 bxc4 and now: bl) 13 o-o-o c6! 14 ..ih6 ..Was 1s ..ixg7 'iii>x g7 16 hxg6 fxg6 17 'it'h6+ 'iii> g 8 18 es dxes 19 0ige4 was J.Peters­ A.Matikozian, Los Angeles 2004, and here 19 ...:f7! looks good for Black. b2) 13 ..ih6 ..ixh6 14 'ilfxh6 and here: b21) 14...g s 1s ..Wxgs+ (1s es!? may also be good) 1 s ... @h8 16 o-o-o c6 17 :d2 l:tg8 18 'ilff4 ..Was 19 0if1 ..ie6 20 0ie3 gave White some advantage in P.Van der Sterren-M.Bosboom, Brussels 1993. b22) 14...l:txb2 !? is a sharp try: lS es (or lS hxg6 fxg6 16 es dxes 17 0-0-0 l:txa2!?) 1s ... dxes 16 o-o-o (not 16 dxes 'ilfd4!) 16 ...exd4 is unclear. If 17 hxg6 (or 17 0ice4 .ifs) 17 .. .fxg6 18 0ice4 .l:i.f7 19 'iii>xb2, Black could try 19 ... cs, 19 .....ifds or 19 ...0ixe4 20 0ixe4 .ifs, in all cases with a total mess. We now return to 10...es:

a31) 17 0,f4 ..if7 18 l'iJfhS was V.Rajlich-K.Zalkind, Budapest 2000. Here 18 ... gxhs 19 l'iJfs 0ie8 20 l:th3 ..ig6 21 ..ixc4+ e6! 22 ..ixe6+ �U7 23 l:tg3 'ilff6 would successfully defend. a32) 17 0ixf6+ l:txf6 18 'ilfxh7+ 'iii>f8 19 'ilfh8+ ..ig8 20 ..ixc4 e6 2 1 l:th7 �U7 22 ..ixe6 lhh7 23 'ilfxg8+ 'iii>e 7 24 'ilfxg6 'it'h8 2S l'iJfS+ 'iii>f8 is unclear. Black is up a lot of material, but it is hard to see 269

A ttacking C h e s s : Th e King 's In dian, Vo l u m e 1 11 d5

Invariably played, but 11 0-0-0 must also be possible. Indeed, we'll briefly discuss it in the notes to White's 11th move in Line A of Chapter 18. 11 ll'las

J.Tarjan-L.Nezhni, USA 1982, and Galla­ gher's suggestion 16 ll'lfl, with the idea of ll'ld2xc4, favour White. 13 0-0-0

...

13 J:tb4!? ..

12 ll'lg3 bxc4

It can be difficult to decide whether to take with the pawn or the knight. Taking with the pawn keeps the offside knight, but this knight is more likely to participate in an attack than White's light-squared bishop, which does not have much of a role. Taking with the knight can still make some sense, though, especially when White has al­ ready spent a tempo moving his bishop or when the black queen can go to the as-square. Here neither of those factors come into consideration, so it is not surprising that taking with the pawn is correct. Instead after 12 ...ll'lxc4 13 ..ixc4 bxc4 14 o-o-o ll'ld7 lS ..We2 (lS hxg6 fxg6) 1s ...ll'lb6 (some correspondence play has shown that 1s ...fs 16 hxg6 f4 17 ll'lhs! is promising for White), both 16 .l:i.h2, from 2 70

This is a good, active move. More­ over, the alternatives may leave the as­ knight looking vulnerable: a) 13 ...ll'ld7 14 hxg6 fxg6 lS ll'lbl! l:tbs (1s ...ll'lb7 16 ..ixc4 just looks good for White) 16 b4 cxb3 17 ..ixbs cs 18 dxc6 (18 ..ie2 is also strong) 18 ... axbs was J.Timman-G.Kasparov, Bugojno 1982. Here 19 cxd7 ll'lc4 20 ..igS ! would be decisive. b) 13 ... ..id7 14 ..ih6 (White has an extra option in 14 ll'lbl or 14 hxg6 fxg6 lS ll'lbl, although the latter is not so clear after 1s ....rf.bS: for example, 16 b4? �8! and White was in some trouble, V.Rajlich-E.Kahn, Budapest 2000) 14... ..ixh6 lS �xh6 .l:i.b4 16 l:td2 'ilfe7 (but not the 16 ... �b8? of A.Harakis­ E.Kahn, Budapest 2000, because of 17 ll'lfs! ..ixfs 18 exfs with a winning at­ tack) 17 ..ie2 leads back to the main line.

Pa n n o Va r i a t io n : 9 h4 with o u t 9 . . . h 5 14

..ih6 ..ixh6 1 5 �xh6

16 ..ie2 ..id7

Instead 16 ... ll'ld7 17 l:tdfl! is given by Kasparov. White intends to play f3f4 with the initiative. 17 ll'lfl

15 .....We7

The alternative is 1s ...'iii>h 8. Black does not commit his queen and creates the possibility of chasing the white queen away with ... ll'lg8. After 16 hxg6 fxg6 17 ..ie2 (not 17 'ilfxg6? l:tg8) 17 .....id7 18 ll'lfl Black has sometimes played the funny 18 ...'iii> g 8!?. This looks like a silly waste of time, but Black is happy to resolve the tension on the kingside so that he may defend later­ ally along the second rank Thus he moves the king back so that the g6pawn is no longer hanging. After 19 ll'le3 ..ifb8 20 :d2 ..ifb6 the position was unclear in H.Rauch-J.Stephan, corre­ spondence 1996. Here Black could also consider 18 ... l:tg8, although using the rook to protect the g6-pawn looks rather inefficient. After 19 ll'le3 (19 g4 is given as good for White by Djurhuus, but this is still not so clear) 19 ...'ilff8 20 �h2 ll'lhs ! 21 g3 �h6 22 'ilff2 �gs 23 :hgl, as in L.Psakhis-R.Djurhuus, Gausdal 1994, both 23 ... :f8 and 23 ...:gb8 give Black decent play.

This is a typical regrouping, espe­ cially when it is clear that the knight will not be hopping into the fS-square. If White starts instead with 17 :d2, Black can try 17 ...c6!? (instead 17 ...l:tfb8 18 ll'lfl heads back to the main line) 18 ll'lfl when, instead of 18...'iii>h 8? 19 hxg6 fxg6 20 �xg6 cxds 21 ll'le3 ! with a big advantage for White in J.Lautier­ P.Svidler, Internet 2004, Black should have played the logical 18 ... l:tfb8. 11 .. AfbB 18 .lld2

White has solidified his castled po­ sition and is ready to build up on the kingside. Black has: a) 18 ... cs?! (this does not do much for the black cause) 19 ..id1! (after 19 g4 ..ia4, with the idea of ... �7. Black can hope to create some counterplay) 19 ... ll'le8 20 hxg6 fxg6 2 1 g4 'ilf g7 22 gs ! and White has the initiative even i n the ending. Now 22 ...�xh6 23 :xh6 ll'lg7 2 71

A ttacking C he s s : The King 's I n dian, Vo l u m e 1

24 f4! prevented ...lllh s and after 24...exf4 2S l:tdh2 White was much bet­ ter in G.Kasparov-V.Loginov, Manila Olympiad 1992. Thus Black should probably leave the cs-square open for a knight. b) After 18 ... ..ie8 Kasparov gave 19 g4 llld7 20 hxg6 fxg6 21 ll:le3 as much

2 72

better for White. Matters are not so clear, however, following 21 ... lllc s 22 gs ..id7 (to stop lllg 4) when White can­ not regroup so easily because both 23 ..id1? llld 3+ and 23 ..if1? llla b3+ are good for Black. c) Another possibility is 18 ... ..ia4 19 g4 llld7 20 hxg6 fxg6 21 ll:le3 lllc s.

Chapter 15 Panrio Vari:ation 9 h4 hS 10 a.o-o

1 d4 ll'lf6 2 c4 g6 3 ll'lc3 .i.g7 4 e4 d6 5 f3 o-o 6 .i.e3 ll'lc6 7 ll'lge2 a6 8 1lfd2 l:t b8 9 h4 hs 10 o-o-o

This is the main line. White contin­ ues playing for an attack, although he may prefer to completely change gears with 10 ll'lcl, which is covered in the next chapter. Before we examine 10 0-0-0, we should look at a couple of other moves which sometimes transpose to the main lines: a) 10 ll'lds bS 11 cxbs (alternatively, 11 ll'lxf6+ .i.xf6 12 cxbs axbs 13 l:tc1

transposes, while 11 0-0-0 is Line A, below) 11 ...axbs 12 l:tcl .i.d7 makes a funny impression . Instead of develop­ ing his pieces, White tries to play all over the board.

Play is likely to continue 13 ll'lxf6+ .i.xf6 (or 13 ...exf6!? when Black has counterplay after 14 g4 hxg4 lS fxg4 l:te8 or 14 ds ll'le7 1s ll'ld4 fS; in the lat­ ter variation, another option is 14... ll'les 1s ll'ld4 fS !? 16 exfs .i.xfs 17 ll'lxfs gxfs with an unclear position) 14 g4 hxg4 lS fxg4 (after lS hS es 16 dS gxf3 17 dxc6 fxe2 18 'ilfxe2 .i.e6 19 'ilfg2 gs 2 73

A ttacking C he s s : The King 's I n d i a n, Vo l u m e 1

Black was better in S.Nikolic­ M.Cvorovic, Kladovo 1991) 1s ...es 16 dS ll'ld4 17 ll'lxd4 (Black has compensation after 17 ..ixd4 exd4 18 gs ..ies 19 ll'lxd4 l:ta8!?) 17 ... exd4 18 ..ig s (following 18 ..ixd4 both 18 ... .l:i.e8 and 18 ... ..ixg4 look good) 18 ... .l:i.e8 19 �f4.

Perhaps slightly surprisingly con­ sidering White's 11th and 12th moves, we have come to a theoretical divide: al) 19 .....ixgs 20 hxgs �e7 21 ..ig2 and here Black can play either 21 .....Wes 22 �xes (22 o-o 'ilfxf4 23 nxf4 l:tb7 was given by Gallagher) 22 ...l:txes 23 l:txc7 ..ixg4 24 l:tfl l:tf8, with an unclear end­ ing, or 21 ...l:tbc8!? with the idea of 22 'it'h2 ..Wes. a2) 19.....ies 20 �f3 and now 20.. .f6 2 1 ..id2 (R.Gunawan-Ye Jiangchuan, Singapore 1987) 21 ... �c8!? 22 ..ie2 d3 gives Black counterplay. Black can also consider the immediate 20...�c8!? 21 ..ie2 d3! when White must find 22 hS (worse are 22 ..ixd3 ..ixg4, 22 �xd3 ..ixg4 and 22 ..id1 ..ixb2) 22 ... ..ixg4! (but not 22 ...dxe2? 23 hxg6 fxg6 24 'ifh3) 23 'ilfxg4 'ilfxg4 24 ..ixg4 ..ixb2 2S 2 74

..if3 ..ixcl 26 ..ixcl b4 with an unclear ending. b) 10 ..ih6 is fairly common and looks more natural than variation 'a'.

Here Black can to transpose to some other lines or explore independent paths: bl) 10... es 11 ..ixg7 (instead 11 dS ..ixh6 12 �xh6 ll'ld4 transposes to variation 'b3' below, but that is dan­ gerous for Black, so he should consider 12 ... ll'las 13 ll'lcl cs, as pointed out by Golubev) 11 ...@xg7 12 dS! ? (12 o-o-o bS takes play into Line (32, below) 12 ... ll'le7 (or 12 ... ll'las! ? 13 ll'lg3 cs) 13 ll'lg3 c6 14 dxc6 ll'lxc6 lS o-o-o ..ie6 16 @b1 ll'le8 (Black could also consider 16 ... ll'ld4!? 17 ll'lge2 ll'lc6 18 �xd6 'ilfxd6 19 .l:i.xd6 ..ixc4) 17 ll'lds bS was B.Spassky-R.Fischer, 8th matchgame, Sveti Stefan 1992. White may try to claim some an edge here, but it is not much and in fact Black went on to win. b2) 10 ... bs is the simplest. Now 11 o-o-o is Line C and 11 ..ixg7 @xg7 12 o-o-o es is Line C32, which are both satisfactory for Black, as we shall see.

P a n n o Va r iatio n : 9 h4 h5 1 0 o - o - o

Instead 11 g 4 is not really sound after 11 ... hxg4 12 hS gxf3 and then:

b21) 13 lllg 3 ..ixh6 14 'ilfxh6 f2+! lS �dl lllg 4 16 ..ifd2 es is winning for Black (Boleslavsky). b22) 13 hxg6 fxe2 (or even 13 .. .fxg6) 14 ..ixe2 (14 ..ixg7 exfl�+ lS @xfl fxg6 should also win for Black) 14.. .fxg6 lS ..ixg7 @xg7 16 �h6+ 'i;f7 looks un­ sound too. b23) 13 ..ixg7 @xg7 14 lllg 3 f2+! ? (14...es l S hxg6 fxg6 1 6 'it'h6+ @f7 is also good) lS �xf2 es was much better for Black in E.Cooke-Y.Zimmerman, Bu­ dapest 2000. b3) 10 ... ..ixh6 11 �xh6 es (instead 11 ... bs 12 o-o-o es is Line Cl, which is a tittle risky for Black) 12 dS!? (here White can steer things towards the aforementioned Line Cl with 12 o-o-o bS) 12 ...ll:ld4 (this does not work out well, so Black should consider 12 ...llle 7 or 12 ...llla s) 13 o-o-o cs ( 13 ...lllx e2+ 14 .i..xe2 with the idea of g2-g4 looks good for White) 14 dxc6 bxc6 (no better are 14... lllxc6 l S cs or 14 ... lllx e2+ l S ..ixe2 bxc6 16 'ilfd2!, as given by Golubev) lS

lllxd4 exd4 16 l:txd4 l:txb2.

If Black does not play this, he will have nothing to show for the pawn. Moreover, the temporary sacrifice looks very appealing, but it does not work out well against exact play: b31) 17 @xb2 'ilfb6+ 18 lllb s axbs gave Black good play in V.Khomyakov­ M.Golubev, Ostrava 1992. b32) 17 es .ifs 18 ..id3 :xg2 19 ..ixfs �as ! 20 ..ife3 l:te8 21 l:td2 'ilfa3+ 22 @d1 :xes 23 ..ie4 l:txd2+ 24 'ilfxd2 lllxe4 2S fxe4 l:txe4!? 26 lllxe4 �f3+ was quickly drawn in R.Meessen­ M.A hn, European Club Cup, Eupen 2006. b33) 17 cs! (Golubev first pointed out the strength of this move) 17 ...:b8 (no better are 17 ....l:i.b7 18 l:txd6! �as 19 @c2 or 17 ..ltf2 18 ..ife3 ..Was 19 l:ta4! ..Wxcs 20 ..Wxcs dxcs 21 llld 1) 18 l:txd6 �as 19 'iii>c2 ..ie6 (19 ...'ilfa3 20 'ilfc1 should also win) 20 .l:i.xe6 fxe6 (after 20 .....ifa3 21 l:td6! :b2+ 22 'it>d3 ..Wxcs, 23 'ilff4! is probably clearer than 23 :d4 'ilfa3) 21 ..ic4 'ilf7 22 es 'ilfb4 23 exf6 ..Wb2+ 24 @d3 l:tbd8+ 2 s .ids! wins. 2 75

A ttacking C he s s : Th e King 's I n di a n , Vo l u m e 1

We can now return to the main sub­ ject of this chapter: 10 0-0-0

Play becomes very sharp, but Black can hold his own in the complications. 10...bs

Black initiates his counterplay. Note that there are certain similarities with the Soltis Variation of the Sicilian Dragon (1 e4 cs 2 ll'lf3 d6 3 d4 cxd4 4 ll'lxd4 ll'lf6 S ll'lc3 g6 6 ..ie3 ..ig7 7 f3 ll'lc6 8 ..ifd2 o-o 9 ..ic4 ..id7 10 o-o-o lk8 11 ..ib3 ll'les 12 h4 h s). Here White must decide how to proceed: A: ulCds • :u: lhf4, :C: 11.i.h6 A) 11 ll'lds

With this move White removes the f6-knight from the defence of the black king. However, this costs time and Black can create counterplay. 11 ... bxc4 12 ll'lxf6+ Consistent. Beating other moves used to be a specialty of John Nunn: 2 76

a) 12 g4 ll'lxds (worse is 12 ...hxg4?! 13 ll'lxf6+ ..ixf6, which is note 'b' to Black's 13th move in our main line) 13 exds ll'lb4 14 ll'lc3 c6 lS ..ixc4 (instead 1s dxc6 ..ie6 16 gxh s ..Was 17 a3 ll'lds 18 h6 ..if6 looks good for Black, while here 18 ll'lxds 'ilfxds 19 hxg6 c3! is given by Nunn) 1s...cxds 16 ..ib3 (or 16 ll'lxds ..ie6 17 ll'lxb4 ..ixc4 18 ll'lc6 'ilfc7 19 ll'lxb8 ..ixa2+ 20 ..ifc3 'ilfxb8 when Black is much better according to Nunn) 16 .....ifb6 (also good is 16 ...hxg4 17 h S gxf3 with the idea of ....ifs as in A.Ewald-M.Fischer, German League 2oos) 17 :hgl (Nunn gives 17 :dgl fS, with the idea of 18 gxh s f4, and here 17 ... hxg4! ? is also possible) 17 ... ..ie6 18 gxh s .ifs 19 l:tgs ..ih6 20 l:tdg1 ..ixgs 21 l:txgs e6 22 hxg6 fxg6 23 h s l:tb7 24 ..Wg2 l:tg7 2s h6 l:tb7 26 l:txfs exfs 27 'ilfxg6+ 'iii>h 8 28 ..Wg2 f4 29 ..ig1 l:tg8 0-1 was A.Kuligowski-J.Nunn, Wijk aan Zee 1983. b) 12 ..ih6 ll'lxds 13 exds ll'lb4 14 ll'lc3 c6 (14 ... ..ixh6!? lS 'ilfxh6 c6 is a good alternative) looks good for Black as well.

P a n n o Va ria t i o n : 9 h4 h 5 1 0 o - o - o

White has tried: bl) lS ..ixc4 cxds 16 ll'lxds ..ixh6 17 �xh6 ll'lxds 18 ..ixds .ifs 19 ..ie4 ..ixe4 20 fxe4 'ilfc7+ 21 'iii>b l l:txb2+ 0-1 was H.Hurme-J.Nunn, Helsinki 1983. b2) lS dxc6 was suggested by Nunn, but after 1S .....ixh6! (worse is lS ... ds 16 g4) 16 �xh6 ..ifb6 (with the idea of ...ll'la2) 17 �d2 ..Wxc6 Black has the up­ per hand. b3) lS g4 �as ! (1s .....ixh6!?) 16 ..ixc4 (both 16 ..ixg7 ll'lxa2+ and 16 'iii>bl ll'lxds should be winning for Black) 16 ... cxds 17 ..ib3 ..ie6 18 ..ixg7 'iii> x g7 19 'iii>bl (19 gxh S l:tfc8! 20 'iii>bl ll'lxa2! 19 l:tdel l:tfc8 20 'it>d1 ll'lc6 21 gxh s l:tb4 was much better for Black in S.Kitte­ M.Zulfugarli, Szeged 1994) 19 ...l:th8 (19 ... hxg4 and ....ifs is winning as well) 20 l:tdgl ll'lc6! (with the idea of ... l:tb4 and ...�6) 21 ..ia4 l2Ja7 22 l:th 2 l:tb4 23 ..ib3 ll'lc6 24 ll'le2 ..ifb6 2s gxhs .ifs+ 26 'iii>a l l:txb3 27 axb3 ll'lb4 28 ll'lc3 ll'lc2+ 29 'iii>a2 l:tb8 0-1 was M.Petursson­ J .Nunn, Lucerne Olympiad 1982. 12 ..ixf6 ...

Here 12 ... exf6 looks too risky: for ex-

ample, 1 3 g 4 hxg4 (or 13. . .ll'lb4 1 4 ll'lc3 cs lS dxcs ..ie6 16 a3 ll'ld3+ 17 ..ixd3 cxd3 18 gxh s) 14 hS gxf3 lS h6 fxe2 16 ..ixe2 and White has a strong attack. 13 g4 ll'lb4 A thematic move. The knight takes aim at both d3 and a2, while the rear c­ pawn is freed to advance, which will both strike at the centre and open a path for the black queen. other moves are ri sky but are not necessarily bad: a) 13 ... es 14 gxh s ll'lb4 (after 14... exd4 lS ll'lxd4 ll'les 16 hxg6 cs 17 gxf7+ 'iii> h 7 18 ll'lfs? ..ixfs 19 exfs :xb2! Black won in J.Herbert-V.Walsh, corre­ spondence 1989, but 18 f4! looks very strong) lS ll'lc3 exd4 16 ..ixd4 ..ie6 17 hxg6 fxg6 18 'ilfh6 ..ixd4 19 'ilfxg6+ ..ig 7 20 ..Wxe6+ 'iii>h7 21 ..ixc4 gave White the upper hand in J.Sammour Hasbun­ J.Stopa, Pawtucket 2008 . b) 13 ... hxg4 14 hS is risky but may be playable: bl) 14... es lS hxg6 fxg6 16 ll'lc3 ll'lxd4? 17 ..ixd4 exd4 18 ..ixc4+ l:tf7 19 ..ixf7+ 'iii>xf7 20 ll'lds ..ig7 21 fxg4 d3 22 l:th 7 �g8 23 ..Wgs 1-0 was Nguyen Anh Dung-Wang Rui Budapest 2000. b2) 14 ... gs!? lS ..ixg s es 16 ..ixf6 'ilfxf6 with an unclear position has been tried with some success. b3) 14 ... gxf3 ! ? (the 'grab everything' policy is sometimes seen in similar structures in the Dragon) lS ll'lc3 gS! 16 ..ixg s ..ixgs 17 'ilfxg s+ 'iii>h 7 was un­ clear in K.Cedikova-D.Vismara, Estensi 2001, although Black did go on to win. 14 ll'lc3 2 77

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1

14 cs ...

Instead 14.....ie6 ! ? may be an even better move order: lS gx hS cs 16 hxg6 �as 17 a3 (after 17 gxf7+ 'iii>xf7 18 a3 Soloviov gives 18 ... cxd4 19 ..ixd4 ll'ld3+ 20 ..ixd3 ..ixd4 with counterplay, while 18 .. J:tfc8! ? 19 es?! cxd4 20 ..ixd4 dxes 21 ..ie3? ll'ld3+ 22 ..ixd3 cxd3 0-1 was V.Quintana Femandez-D.Roiz Baztan, Asturias 1996) 17 .. .fxg6.

Here White has several options: a) 18 axb4 cxb4 19 ll'lbl c3 gives Black the initiative. b) 18 ..Wg2 'iii>f7 19 .l:i.gl .l:i.g8 20 axb4 cxb4 2 1 ll'lbl c3! 22 dS cxb2+ 23 'iii>d2 ..ic3+ 24 'iii>e 2 it.d7 with good counter278

play according to Soloviov. c) 18 f4 ll'lc6! 19 �g2 (after 19 dS l:txb2! 20 'iii> x b2 :b8+ 21 'iii> c2 'ilfxa3 22 l:tb1 ll'lb4+ Black is winning and 19 es? ll'lxd4 20 ..ixd4 cxd4 2 1 1Wxd4 dxes 22 ..Wg1 .ifs 23 ..ixc4+ 'iii> g 7 24 l:.ds 'ilfc7 2 s J:tcs �b7 gave him a winning attack in J.Tarjan-R.Gunawan, Indonesia 1983), and now Black must avoid 19 ...'iii>f7? 20 fS! and play either 19 .....if7 (unclear according to Soloviov) or 19 ....l:i.f7! ? with good counterplay. d) 18 l:tg1 'iii>f 7! 19 l:tgs?! (19 axb4 cxb4 20 ll'lbl c3 19 ..Wg2 is variation 'b') 19 .....ixg s 20 hxg s cxd4 2 1 ..ixd4 ll'lc6 22 ..ie3 l:tb3 23 'iii>c2 .l:i.fb8 24 l:tb1 ll'les 2S ..ie2 �c7 26 f4 ll'ld3 27 ..ixd3 cxd3+ 28 �xd3 �b7 29 ..icl �6 30 ..ife3 .l:i.xc3+! 31 'it>xc3 �c6+ 32 'it>d2 .l:i.h8! was winning for Black in V.Churkin­ S.Soloviov, St Petersburg 2007. e) 18 h S ! ? (very ambitious) 18 ... cxd4 19 ..ixd4 ..igs! (better than 19 ...it.xd4 20 �xd4 ll'lc6 21 �e3 �b6 22 �xb6 l:txb6 23 ..ie2 with a pull for White in the ending) 20 hxg6! (critical and best; White had to avoid 20 ..ie3? .l:i.xf3 ! 21 ..ixg s .l:i.xc3+! when he loses and 20 f4 ..ixf4 21 ..ie3 ..ies 22 ..id4 ..ixd4 23 'ilfxd4 �g s+ is also in Black's favour) 20 .....ixd2+ 21 l:txd2 .l:i.f6 22 axb4 ..Wal+ 23 ll'lbl c3! and now 24 bxc3?! ..ia2 2S ..id3? es! left White in huge trouble in L.Gerzhoy-S.Ganguly, Chalkis 2010, but 24 l:tdh2! cxb2+ 2S 'it>d2 ! 'ilfxbl 26 .l:i.h8+ 'iii> g 7 27 l:tlh7+ 'iii> x g6 28 .l:i.h6+ would have left Black unable to escape the perpetual.

Pan n o Var i a t i o n : 9 h4 h 5 1 0 o - o - o

l.Krasik-D.Naroditsky, U S League 2008. c2) 16 ...cxd4 is better, although af­ ter 17 ..ixd4 ..ixd4 18 'ilfxd4 'ilfxh S 19 l:tdgl White has some initiative.

15 ..ixc4

This move is almost always played, but it may not be the best: a) 1s dxcs? is bad after 1s .....ixC3! 16 bxc3 ..Was ( 16...ll'ld3+ 17 ..ixd3 ..Was is also good) 17 cxb4 'ilfa3+: for exam­ ple, 18 'iii> b l l:txb4+ 19 'iii>a l c3 20 'ilfc1 .:lb2 21 ..ic4 ..ie6 22 l:tds ..ixds 23 i.xds l:tfb8 with a winning attack. b) lS a3?! ll'lc6 (Black can also play 1s ...ll'ld3+ 16 ..ixd3 cxd4 17 ..ixd4 ..ixd4 18 ..ixc4 ..ig7 with counterplay or sim­ ply 1s ... cxd4 16 ..ixd4 ..ixd4 17 'ilfxd4 -Zlc6 with a good Dragon position) 16 dxcs ll'las 17 'ilf c2 hxg4 looks good for Black. c) 1s gxh s ..Was (Black is probably better off with lS ... ..ie6, transposing to the notes to his 14th move, above) 16 ..ixc4 and here: cl) After 16 ... ..ie6?! 17 ds ll'lxa2+ (or 17 .....id7 when, rather than 18 hxg6 ..ixc3 19 'ilfxc3 ll'lxa2+ 20 ..ixa2 'ilfxa2 21 ..ih6 ..Wal+ 22 'iii> c2 ..ia4+ 23 'iii>d3 'ilfxb2, White should play 18 a3!) 18 ..ixa2 l:txb2 19 'ilfxb2 ..ixc3 20 'ilfb3 White had a winning position in

1s ...cxd4 16 ..ixd4 'ilfc7

Or 16 ... ..ixd4 17 'ilfxd4 ll'lc6, which was agreed drawn in the game Kazan A.Severjuk hin-D.Skorchen ko, 2001. Black could certainly continue here; after the queen moves, 18 ... ll'les with the idea of ... hxg4 gives him coun­ terplay. 17 ..ib3 ..ixd4

Instead 17 ... ..ie6 18 ..ixf6 ..ixb3 19 axb3 exf6 20 gxh s (20 'iii>b 1!?) 20 .....Was 21 'iii> bl 'ilfxhs was fine for Black in L.Oll-B.Gelfand, USSR 1984. 18 �xd4 ..ie6 19 gxhs

Alternatively, 19 ..ixe6 fxe6 (Black threatens ... ll'lxa2+) 20 'iii> b l ll'lc6 21 �c4 'ilfb6 22 'ilfxe6+ 'iii> g 7 23 l:th2 ll'ld4 24 �ds ll'lxf3 2s l:tg2 ll'lxh4 26 l:tc2 l:tf2 27 'ilfg s 1/2-1/2 was M.Tochacek­ R.Maliangkay, correspondence 2002; 27 ...l:txc2 28 'ilfxe7+ results in perpetual check. 19 .....ixb3 20 axb3

2 79

A ttacking C h e s s : Th e King 's I n dia n, Vo l u m e 1

This is a strange-looking move. In order to open the diagonal for his fl­ bishop, White prevents himself from playing ii.h6 and lessens his control over the d4-square. 11 bxc4 ...

20 :ifas ..

A common theme - the black queen is able to recapture on h S by threaten­ ing ...'if al+. Instead 20 ...ttJa2+ is possi­ ble: 21 'iii>c2 l:.b4 22 'if d2 l:.fb8 23 hxg6 l:.xb3 24 gxf7+ 'iii>f8 2S l:.bl 'ifc4 26 'ifh6+ 'iii>xf7 27 'ifhs+ 'iii>f8 28 'iVh8+ (not 28 'ifh6+? 'iii>e 8 and the king escaped in O.Haugseth-H.Theting, correspondence 1990) 28 ...'iii>f7 29 'ifhs+ with a draw in W.Schmidt-A.Sznapik, Prague 198S. 21 'iii>b1 'ifxhs

Black was at least equal in l.Nowak­ A.Sznapik, Polanica Zdroj 198S. B) 11 t[)f4

280

Even though this allows White to immediately activate his bishop, prac­ tice has shown it to be the best move. Instead 11 ... es 12 dxes ttJxes 13 cs (13 cxbs may also be good) 11 ...b4 12 ttJcds ltJh7! ? 13 g4 e6 14 gxhs exds 1s cxds gives White good compensation for the piece. Black's main alternative to 11 ...bxc4 is 11 ...ii.d7 and then: a) 12 cxbs axbs 13 ttJxbs (after 13 ii.xbs both 13 ...es 14 dxes ttJxes and 13 ... ttJxd4 14 ii.xd7 ttJxd7 lS ii.xd4 e s 16 ii.e3 exf4 17 ii.xf4 ttJes, as in correspon­ F.Hanssen-J.Simmelink, dence 1987, give Black compensation for the pawn) 13 ... es 14 dxes ttJxes 1 s ttJc3 'ifc8 and Black had good play in J.Timman-J.Nunn, Tilburg 1982. b) 12 g4 ttJxd4 (12 ... hxg4!?) 13 ii.xd4 es 14 ii.xes (instead both 14 ii.a7 l:.b7 1 s ttJxhs gxhs 16 ii.e3 and 14 ttJxhs gxh s are unclear; however, in the latter variation, 14... exd4?! l S ttJxg7 dxc3 16 'ifxc3 'iii>x g7 17 g s was better for White in M.Petursson-S.Los, Bela Crkva 1983) 14...dxes 1s g s b4 16 ttJcds ttJxds 17 ttJxds ii.e6 18 'if e3 ii.xds 19 l:.xds looks a little better for White, although after 19 ... 'ife7 20 ii.h3 (20 'ifcs !?) 20 ... l:.fd8 21 l:.hd1 l:.xds 22 l:.xds l:.d8 23 'if cs ii.f8 24 'ifxe7 ii.xe7 2 s l:.xes ii.d6 26 l:.ds c6! 27 l:.d3 ii.f4+ 28 'iii>c2 l:.xd3 29 'iii>xd3 cs

Pa n n o Variatio n : 9 h 4 h5 1 0 o - o - o

30 b3 the game was drawn in H.Ree­ J.Nunn, Wijk aan Zee 1983.

well from this position.

12 ii.xc4

The pin of the f7-pawn on the a2-g8 diagonal will sometimes come into play now. White should not get carried away with 12 g4?!, though: 12 ... es 13 dxes ttJxes 14 gxh s ttJxf3 1s 'ifg2 ttJxe4! 16 ii.xc4 ttJxc3 17 bxc3 'ife8 18 'ifxf3 ii.g4 19 'if g3 'ifa4 20 'ifxg4 'ifxc4 0-1 was M.Noble-H.Garcia Rojas, corre­ spondence 2004. 12 ...es

13 dxes

The tempting 13 ttJxg6?! backfires after 13 ... exd4 14 ii.xd4 l:.b4! lS ttJxf8 l:.xc4 16 ii.xf6 'ifxf6 when Black is bet­ ter. 13 ...ttJxes 14 ii.b3 'ife8! This is a key move. Black removes his queen from the d-file and defends the es-knight, thereby making ... cs a possibility because a subsequent 'ifxd6 will not exchange queens. Moreover, from the e8 the queen may emerge on the a4-e8 diagonal and the advance ... aS-a4 is supported. Black has scored

15 'iii>b l

This is a natural move, but it does not work out well. White has also tried 1s ttJfdS ttJxds 16 ttJxds ii.e6 and th en: a) 17 ttJxc7? fails to 17...'ifc6+ 18 'ifc2 (18 'ifc3 ii.xb3 19 'ifxc6 ttJxc6 20 axb3 l:.fc8 is great for Black) 18...ttJc4 with a winning attack. b) 17 'ifas? ii.xds (17 ...'ifc6+ 18 'iii>b 1 'ifb7 is also very strong) 18 'ifxds (18 l:.xds 'ifc6+ has the idea 19 'iii>bl ltJc4, winning) 18 ... cs! shows all of Black's ideas in action: 19 l:.d2 c4 20 ii.c2 l:.bs (20 ...'if c8! is even better) 21 'ifxd6 l:.xb2 22 'ifxa6? (22 'iii>xb2 c3+ 23 'iii>c l cxd2+ 24 ii.xd2 was the only chance) 22 ...'ifb8 and Black was winning in H.Ree­ J.Mestel, European Team Champion­ ship, Plovdiv 1983. c) 17 f4 ii.xds 18 exds ltJg4 is very comfortable for Black. d) 17 ii.h6 ii.xds 18 exds ii.xh6 19 'ifxh6 as 20 l:.d4 l:.b4 21 'ifd2 l:.xd4 22 'ifxd4 'ifbs 23 l:.dl l:.b8 gave Black some initiative in B.Belotti-B.Zueger, Chiasso 1991. 281

A ttacking Chess: The King 's I ndian, Vo l u m e 1

e) 17 ii.d4 ii.xds 18 exds as 19 'ifc2 'ifbs gave Black counterplay in A.Finke­ M.Spaans, correspondence 1987. f) 17 'iii>b 1 ii.xds 18 'ifxds cs 19 ii.cl was J.Merriman-G. Lane, London 1993, and here both 19 ...'ifc8!? (supporting the push ... c4) and 19 ... as deserve con­ si deration. 1s ...as

play ...tt:Ja4 anymore) 2 1...ttJcs 2 2 tt:Jg4? (Gelfand indicates that 22 hxg6 tt:Jxb3 23 axb3 l:.xb3 24 ttJexds cxds 2s 'ifxds is a better try, although 2s ... 'ifb8 26 gxf7+ 'iii>h 7 should still win for Black) 22 ... ii.d3+ 23 ii.c2 d4 24 tt:Jds ii.xc2+ 2s 'ifxc2 cxds 26 'ifxcs .l:lxb2+ 27 'iii>xb2 �8+ with a winning attack in V.Dydyshko-B.Gelfand, Minsk 1986. b) 20...ttJcs! may be even stronger: 21 ii.c4 ii.xc3 (21 ... ds looks good too) 22 'ifxc3 tt:Ja4 23 'iff6 l:.xb2+ 24 'iii>al ii.xc4 2S ttJxc4 l:.c2 26 tt:Jxd6 tt:Jc3 27 'ifxC3! was given by Gelfand, but Black can play 26 ...'ife7! ! with the idea of 27 'ifxe7 (27 'ifd4 cs 28 'ifxa4 'ifes+ is winning for Black as well) 21 ...tt:Jc3, mating. 16 ...ii.d7 17 ii.d4

16 'ifc2

The main alternative is 16 ii.d4 tt:Jfd7 17 tt:Jfds c6 18 ttJe3 (after 18 ttJf4 ii.a6 19 ii.xes ttJxes 20 'ifxd6 ii.c4 Black had the initiative in H.Naundorf­ J.Moucka, correspondence 2000) 18 ...ii.a6! (Black eyes the d3-square) 19 ii.xes (19 f4 tt:Jd3 20 ii.xg 7 'iii>xg7 is much better for Black according to Gel­ fand) 19 ...ii.xes 20 g4 (Gelfand also gives 20 ii.c4 'ifc8 21 g4 l:.b4! 22 ii.xa6 'ifxa6 with a strong attack) and now: a) 20 ... ds!? 21 gxh s (instead 21 exds ttJcs 22 ii.c4? loses to 22 ... ii.xc3 23 'ifxc3 tt:Ja4 24 'iff6 ttJxb2 2S 'iii>al 'ifxe3 because 'ifxg6+ is not possible, but 22 'if c2 could be tried because after 22 ...a4 23 ii.c4 ii.xc3 24 'ifxc3 Black cannot 282

17...'iii> h 7!

Black removes his king from the a2g8 diagonal. Instead 17...tt:Jc6 18 ii.xf6 ii.xf6 19 ttJxg6 .l:lxb3 20 ttJxf8 .l:lxc3 fails to 21 tt:Jxd7 ! - Nunn. 18 ttJd3

Instead 18 tt:Jh 3 has been recom­ mended. However, after 18 ... cs 19

Pa n n o Varia t i o n : 9 h4 h5 1 0 o - o - o

ii.xes 'ifxes 20 f4 'ife7 21 ttJgs+ 'iii> g 8 22 'ifd3 a4 23 ii.xa4 ii.xa4 24 ttJxa4 'ifa7 (24....l:lb4 2S ttJc3 .l:lfb8 with a strong attack may be even better) 2S ttJc3 ltJg4 26 .l:ld2 c4 27 'ifxc4?! .l:lfc8 28 ttJbs l:.xc4 29 ttJxa7 l:.c7 Black won a piece in T.Emst-Y.Zimmerman, Creon 2oos. 18...ttJc6 Black could also play 18 ... ttJxd3 19 l:.xd3 ii.e6, but the text is more ambi­ tious.

C) 1 1 ii.h6

19 ii.xf6

After 19 ii.e3 ltJb4 20 ttJxb4 .l:lxb4 Black clearly has the initiative. 19. �xf6 20 ttJds ii.g7 21 ii.a4 Forcing things with 21 ttJxc7 backfires after 21 ... 'ifd8 22 ttJds ltJd4 23 'ifc4 ii.bs 24 'ifc7 a4 2s 'ifxd8 .l::i.fxd8 26 ii.c2 l:.dc8 and Black wins. ..

White wants to exchange the powerful g7-bishop. Now 11 ... es is the main line, but other moves are also of interest. We have: C1: 11 ii.xh6 C2: 11 bxc4 C3: 11 es •.•

•••

..•

21...ttJes ! 22 ii.xd7 'ikxd7 23 .l:ld2

The c-pawn is clearly untouchable: 23 'if xc7? 'ifxc7 24 ttJxc7 ttJxd3 2s l:.xd3 l:.xb2+ 26 'iii>c l l:.c8. 23 ....l:lb7 24 ttJe3 ttJxd3 25 'ifxd3 l:.fb8 Black had a very strong attack in the game R.Vaganian-J.Nunn, Skelleftea 1989.

Instead 11 ...ttJas! ? is relatively un­ explored, leading to 12 ii.xg7 'iii>x g7 13 ttJf4 and here: a) 13 ... c6 14 g4 .l:lh8 (lS 'ifg2 'ifc7 16 gxhs ttJxhs 17 ttJxhs+ .l:lxhs was fine for Black in H.Hurme-J.Nunn, Helsinki 1981) 1s cxbs axbs 16 gs ltJd7 17 es b4 was unclear in Kliesch-W.Wittmann, correspondence 1984. b) 13 ...ttJxc4!? 14 ii.xc4 bxc4 lS 'ife2 c6 16 es?! (16 'ifxc4 looks equal) 16 ...ttJds 17 ttJfxds cxds 18 g4 (or 18 ttJxds ii.e6) 18.. lth8 19 .l::i.d2 e6 (19.. .ii.e6!) 20 f4 hxg4 21 'ifxg4 dxes (21 ...l:.hs!) 22 fxes .l::i.hs 23 l:.f2 ii.d7 24 l:.hf1 .l:lfs 2s hs gave White compensation in U.Schuster-M.Litovicius, correspondence 1997. 283

A ttacking C h e s s : Th e King 's Indian, V o l u m e 1

Cl) 11 ii.xh6 This is a controversial move. Black invites the white queen into his own king's den, but if the king survives, the white queen could be missed in the centre and on the queen side. ...

We have seen this idea before in the 10 ... bs lines. Here, however, White cannot open the h-file so he must aim for an eventual g4-break.

12 'ifxh6

16 l:.d2

12 ...es 13 d s

Instead 13 dxes ttJxes is comfort­ able for Black, while 13 g4 could be met with 13 ...bxc4 or even 13 ... ii.xg4!?. 13 ttJas 14 ltJ g3 bxc4 I think Black should consider the un­ tried 14...b4! ? here. After 15 ttJb 1 ttJb7 Black can play ... as, ...ttJcs, ... ii.d7 and ... c6 with counterplay. This plan is not so common, but it comes up some­ times and is well worth being familiar with, especially when White's kingside play is also rather slow. ...

15 ii.e2

White intends .l:ld2 and ii.d1 to tidy up the queenside, after which he can proceed with his kingside campaign with ttJf1, intending ttJe3 and g4. 1s .l:.b4 ..

284

After 16 ttJfl ii.d7 17 ttJe3 'ifb8 18 l:.d2 'ifb6, 19 ltJfs? is an instructive mis­ take, since 19 ... ii.xfs 20 exfs .l:lb8 leaves Black's attack much the faster: 21 ttJd1 l:.a4 22 'iii>bl ttJb3 23 axb3 'ifxb3 24 fxg6 'ifa2+ 25 'iii>cl 'ifal+ 26 'iii>c 2 l:.ab4 27 ii.fl l:.xb2+ 28 ttJxb2 l:.xb2+ 0-1 was A.Muir-A.Fedorov, Batumi 1999. 16...ii.d7

Instead after 16 ...cs?! 17 dxc6 ttJxc6 18 a3 l:.b7 19 ii.xc4 ii.e6 White has: a) 20 ii.xe6 fxe6 and White cannot play 21 'ifxg6+? l:.g7. b) 20 ttJds ii.xds? 21 .l:lxds ltJd4 (Black tries to improve on 21 ...ttJxds 22 ttJxhs gxh s 23 l:.h 3, as in A.Schneider­ A.Stummer, Budapest 1993, but this loses too) 22 .l:lxd4! with a winning po­ sition in F.Volkmann-A.Stummer, Bu­ dapest 1993. However, instead of 20 ... ii.xds, Black can improve with 20...ltJd4!.

Pa n n o Va riation: 9 h 4 h5 1 0 0 - 0 - 0

c) 20 ttJfs! gxfs 21 exfs ii.xfs (the point is that 21 ...ii.xc4 fails because of 22 l:.h3 ! ) 22 'ifgs+ 'iii>h8 (22 ...ii.g6? 23 'ifxg6+) 23 'ifxfs and White is clearly better. 17 ttJf1 'ifb8

It seems that staying near the king with the alternative 17...'ife7 ! ? may be a bettertry. 18 g4 'ifb6 19 ii.di l:.b8 20 l:.hh2 1 2 g4

Black's position looks good, but is difficult for him to make progress. The tempting 20 ...ttJb3+?! does not seem to work after 21 axb3 'ifas 22 ii.c2 (22 'iii>b l!? Is an interesting alternative) 22 ...'ifal+ 23 ii.bl l:.xb3 24 l:.c2 as (Black should try 24...hxg4!?) 2S ttJe3 and White took over the attack in P.Neuman-Ji.Houska, Klatovy 2002. Perhaps Black must try something like 20...c6 21 ttJe3 'ifd8 here, but if that is the case then he probably should im­ prove earlier, perhaps on move 14 or move 17. C2) 11...bxc4

This is a sharp try.

White can also exchange bishops immediately with 12 ii.xg7 'iii>x g7 and here 13 ttJf4 es 14 dxes ttJxes is note 'b' to White's 14th move in Line (32 be­ low, but he can also try: a) 13 ds ltJb4 (13 ...ttJas is also possi­ ble, while 13 ...ttJes 14 ltJd4 cs 1s dxc6 'ifas was D.Ruzele-B.Gelfand, USSR 1986; here 16 f4 ttJeg4 17 ii.xc4 is good for White, but earlier 14...ii.d7 looks like a better try for Black) 14 ttJf4 c6 lS ii.xc4 cxds 16 ttJfxds ttJbxds 17 ttJxds ii.e6 18 'iii>b1 ii.xds 19 ii.xds as looks okay for Black, but he should not be too complacent: 20 l:.cl a4 21 l:.hdl 'ifb6 22 'ifd4 'ifas 23 l:.c4 a3 24 b3 l:.fc8 2 s ii.c6 'ifa6 26 l:.dcl l:.b6 27 'ifc3 l:.cb8? ! (27 ... 'iii> g 8) 28 g4! gave White an attack in M.Cebalo-Z.Djukic, Yugoslav Team Championship 1983. b) 13 g4 es 14 ii.xg7 'iii>x g7 trans­ poses to the note to White's 13th move in our main line, which is satisfactory for Black, and here 13 ...hxg4 is also possible. c) 13 ttJg3 es 14 dxes (14 ds ttJas) 2 85

A tt a c k i n g C h e s s : Th e K i n g 's I n d i a n, Vo l u m e 1

14... ttJxes 1s f4 (lS 'ifg s?! can be met with 1s ....l:le8 or even 1s ... 'iii>h 7) 1s ...ltJd3+ (better than 1S ...ttJeg4 16 es) 16 ii.xd3 cxd3 17 'ifxd3 ii.g4 18 .l:ld2 'ifc8 is fairly level.

gxf3 17 exf6 exf6 18 'iff4 fxe2 19 ii.xe2 ds! 20 .l:ldg1 ttJe7 21 'ifxf6 ttJfs 22 'if es l:.e8 23 'iff4 'ife7 gives Black counter­ play and now 24 ii.f3 ? l:.xb2! was sud­ denly winning for Black in C.Van Buskirk-J.Peters, Irvine 1997. 13 ds

Black is well placed after 13 ii.xg 7 'iii>x g7 14 ds (or 14 gs ttJd7 with a good game for Black) 14...ltJb4 lS ltJg3 hxg4 16 ii.xc4 and now 16 ...gxf3! 17 a3 ltJg4 18 axb4 l:.xb4, with the idea of ...'iff6, gives Black excellent play for the piece.

12 ... es

It is time to strike in the centre. Others: a) 12 ...'iii>h 7 wastes a tempo: 13 ii.xg7 'iii>x g7 14 ltJg3 l:.h8 and now both 1s gxh s ttJxh s 16 ttJxh s+ l:.xh s 17 ii.xc4 and lS gs ttJd7 16 ii.xc4 look a little better for White. b) 12 ...ttJb4 13 ltJg3 cs 14 ii.xg7 'iii> x g7 1s gxh s (instead 1s ds l:.h8 16 ii.xc4 was V.Salov-A.Khalifman, USSR Championship, Minsk 1987, and here Black should try 16 ...hxg4! ? with the idea of 17 f4 as 18 a3 ii.a6 or 17 a3 gxf3! 18 axb4 cxb4 intending ...'ifc7) 1S ...cxd4 16 'ifxd4 gave White the ini­ tiative in A.Czerwonski-D.Pedzich, Gdansk 1994. c) 12 ...hxg4! ? 13 hS is scary but playable: 13 ... ii.xh6 (not 13 ... gxf3 14 hxg6 fxe2 lS ii.xe2 with a winning at­ tack) 14 'ifxh 6 g s 1s 'ifxg s+ 'iii>h 7 16 es! 286

13 ..ii.xh6 .

Black has several alternatives to ex­ plore as well: a) 13 ...ltJb4! ? 14 ltJg3 hxg4 1s hs (or 1s ii.xg7 'iii>xg7 16 hs gxf3) 1s ... gxf3 looks risky, but my computer thinks that Black is much better... b) 13 ...ltJd4 14 ttJxd4 (14 ii.xg7 ttJxe2+ lS ii.xe2 'lii>xg 7 16 l:.dgl .l:lh8 looks solid enough for Black) 14... ii.xh6 (worse is 14... exd4 lS ii.xg7 'iii>xg7 16 'ifxd4 hxg4 17 f4 with the initiative for White) lS 'ifxh6 exd4 16 .l:lxd4 cs? (for better or worse, 16 ...hxg4 had to be

Pa n n o Va riatio n : 9 h4 h5 1 0 o - o - o

tried) 17 l:.d2 (good enough, but 17 dxc6! with the idea of 11... l:.xb2 18 ii.xc4! is crushing) 17 ...hxg4 18 hs gs 19 'ifxg s+ 'iii>h8 20 ii.xc4 gave White a big advantage in T.Reilly-5.Rutherford, Sydney 199S. c) 13 ... ttJe7!? is an unusual move in the Panno, but it is not bad here: 14 ltJg 3 hxg4 1s hs gxf3 16 hxg6 fxg6 17 ii.xg7 'iii>x g7 18 'tlt'h6+ �7 19 ii.xc4 ii.g4 20 ii.xa6 l:.h8 21 'ife3 'iii>g 7 was R. Filutowski- H. Lassen, correspondence 1997. The position is complicated, but should probably favour Black. 14

'ifxh6 ii.xg4!? 15 fxg4

Worse is lS dxc6 ii.xf3. 1s...ttJxg4 16 'ifd2 ttJf2 The position is unclear.

White has also tried a couple of other moves: a) 12 ds ttJas (12 ... ii.xh6 13 'ifxh 6 takes play back into Line Cl) 13 ltJg3 b4! ? (13 ... ii.xh6 14 'ifxh6 is again Line Cl) 14 ttJb1 is very similar to note 'a2' to White's 13th in Line C32 and is likely to transpose there. b) 12 ii.gs bxc4 13 ttJds ttJxd4 14 ttJxd4 exd4 1s ttJxf6+ (lS 'ifxd4 ttJxds! 16 ii.xd8 ii.xd4 17 .l:lxd4 l:.xd8 18 exds c3 19 bxc3 .l:le8 is at least equal for Black) 1s ... ii.xf6 16 ii.xf6 'ifxf6 17 'ifxd4 'iff4+ 18 'ifd2 'ifxd2+ 19 l:.xd2 ii.e6 20 .l:lc2 l:.b4 and Black was much better in R. Wunsch-1.Firnhaber, correspondence 1986. c31) 12 ttJds

C3) 11...es

This is the ma in line. Black com­ bines central play with the attacks on the wings. We h ave a further split: C31: 12 liJds

(32: 12 hg7

This is considered to be poorly timed because of Black's response. 12 ...ii.xh6!

Instead 12 ...bxc4 13 ttJxf6+ (13 ii.xg7 'iii>x g7 is note 'b' to White's 13th in Line C32) 13 ...ii.xf6 14 ds (not 14 ii.xf8 exd4! with good play for Black) 14... ltJb4 lS ltJc3 l:.e8 16 ii.xc4 as 17 g4! gave White 287

A tt a c k i n g C h e s s : The King 's I n dian, Vo l u m e 1

a strong attack in Lin Ta-Peng Xiaomin, China 1991.

l:.xf7 with equality in A.Moiseenko­ A.Cabrera, Internet 2004.

13 'ifxh6 bxc4 14 g4 After 14 ttJxf6+ 'ifxf6 15 d5 ttJa5

16 ttJc3 exd4 17 l:.xd4 'iff6 18 'iff4

both 16 ttJc3 ii.d7 and 16 ltJg3 'iff4+ 17 'ifxf4 exf4 18 ttJe2 ii.d7 19 ttJxf4 l:.b4 look good for Black.

White has to bail out into an end­ ing, as 18 l:.f4 runs into 18...ttJxa2+! and 18 l:.d2 'ifxf3 also wins for Black. Instead 18 'ife3 looks reasonable, but after 18 ...hxg4 19 fxg4 ii.xg4! 20 ii.xc4 (or 20 l:.xg4 ttJxa2+ 21 'iii>c2 'iff5+ 22 l:.e4 ttJxc3 23 bxc3 l:.fe8 24 ii.g2 'ifd7!? with good compensation) 20...l:.fe8 Black was doing well in R.Grosso­ R.Berdichesky, correspondence 1994. 18 ...'ifxf4+ 19 l:.xf4 ltJd3+ 20 ii.xd3 cxd3

The endgame was even in E.Higuita­ W.Megier, correspondence 1992. 14 ...ttJxds

Black must avoid 14...ttJxd4? 15 ttJxd4 exd4 16 gxh5 ttJxh 5 17 ii.xc4 c6 18 l:.dg1 cxd5 19 l:.xg6+ fxg6 20 ii.xd5+ .l:lf7 21 'ifxg6+ 'iii>h8 22 ii.xf7 ttJf6 23 l:.gl ltJg4 24 .l:lxg4 'ifc7+ 25 'iii>d 1 1-o, as in P.Hummel- E.Gufeld, Los Angeles 1998. However, 14 ... ii.xg4! ? is a good al­ ternative to the text: 15 ttJxf6+ 'ifxf6 16 fxg4 exd4 17 'iff4 'tlt'h8 18 'ifd2 d3 19 gxh5 (19 ttJc3 hxg4) 19....l:lb7 (or 19 ... 'ifxh 5 2o ttJf4 'ifc5) 20 hxg6 fxg6 21 ttJc3 l:.fb8 and Black had the initiative in K.Wiacek-5.Korosec, correspondence 2002. 15 exds ttJb4

Also possible is 15 ... ttJxd4 16 ttJxd4 exd4 17 ii.xc4 'iff6 18 gxh5 'ifxf3 19 hxg6 'if e3+ 20 'ifxe3 dxe3 21 gxf7+ 288

C32) 12 ii.xg7 'iii>xg7

13 dxes

Instead 13 cxb5 axb 5 14 dxe5 ttJxe5 (or 14... dxe5) transposes to our main line here. Others: a) 13 d5 is not well timed and after 13 ...ttJa5 14 ttJg 3 (14 cxb5 axb5 1 5 ttJg3 ii.d7 looks quite pleasant for Black) there is:

P a n n o Va ri a t i o n : 9 h 4 h 5 1 0 o - o - o

al) 14 ...ttJxc4 lS ii.xc4 bxc4 16 ltJfl as 17 ttJe3 ii.a6 18 g4 (18 ttJe2 could be met with 18 ....l:lb4 or even 18 ... ttJd7!? 19 'ifxas ttJcs) 18 ... l:.h8 19 .l:ldg1 ttJd7! and Black had good prospects in l.Mason­ J.lrvin, correspon dence 1996. a2) 14...b4! lS ttJb1 c6 (Black has other good moves too, such as 1S ... ii.d7, with the idea of 16 ii.e2 ii.a4! 17 l:.del ii.b3!, 1 S ....l:lb6 16 'ifgs ttJb7 17 ttJd2 ttJcs and 1S ...ttJd7 16 ii.e2 ttJcs 17 'ife3 'iff6 18 ltJfl 'iff4 19 g3 'ifxe3+ 20 ttJxe3 fs, which was better for Black in M.Ostreic h-P .Pflichthofer, correspon­ dence 1993) 16 'ifgs cxds 17 exds ttJb7 18 ii.d3 ttJcs and again Black was bet­ ter in A.Gual Pascual-M.lllescas Cor­ doba, Ayamonte 2002. b) 13 ttJds bxc4 14 ttJxf6 (14 ttJe3 is safer) 14...'ifxf6 lS dS and here:

bl) 1S ... c3!? 16 'ifxc3 (16 ttJxc3 ltJd4) was O.Reeh-0.Dannevig, Gausdal 1992. Here 16 ...ltJb4 17 'iii> b l as would assure Black of queenside counterplay. b2) 1S ... ttJa7 16 ttJc3 ttJbs 17 ii.xc4 'ife7 (17 ... ii.d7!?) 18 ttJe2 l:.d8 19 'ife3 cs 20 dxc6 'ifc7 21 .ids ii.d7 22 'iii>bl ii.xc6

23 l:.cl ttJa7 was J.Rowson-J.Nunn, Ox­ ford 1998, and here 24 ttJc3 looks a lit­ tle better for White. b3) 1S ...ltJb4 16 ttJc3 ltJd3+ (Hazai's suggestion of 16...c6 is also interesting) 17 ii.xd3 cxd3 18 'ifgs ii.d7 19 .l:lxd3 l:.b7 20 'iii>b1 'ifxgs 21 hxgs f6 22 gxf6+ 'iii>xf6 gave Black counterplay in the endgame in J.Rowson-J.Gallagher, Brit­ ish Championship, Torquay 2002. 13 ttJxes Black can also play 13 ... dxes and here: ...

a) 14 'ife3 'ife7 lS ttJds ttJxds 16 cxds ttJas 17 ttJc3 (17 'iii>bl ltJc4 18 'ifc3 might be a better try) 17 ... ttJb7 18 ii.d3 'if cs 19 'ifxcs ttJxcs 20 b4 ttJb7 21 'iii>b2 ii.d7 22 .l:lcl .l:lfd8 23 l:.hdl c6 24 dxc6 ii.xc6 was level in M.Wiedenkeller­ N.Stewart, German League 2001. b) 14 'ifg s 'ife7 and now: bl) lS ttJds ttJxds 16 exds (after 16 cxds Black can play either 16 ... 'ifxgs+ 17 hxgs ttJas 18 ttJc3 ttJb7 or 16...f6 17 'ifd2 ttJd8 and ...ttJf7) 16.. .f6 17 'ifd2 l:.d8! 18 g4 bxc4! 19 ttJc3 (or 19 gxhs ltJb4 20 ttJc3 .ifs 21 ii.xc4 'ifcs 22 'ife2 289

A tt a c k i n g C h e s s : Th e King 's I nd i a n , Vo l u m e 1

ii.d3! 23 ii.xd3 ttJxa2+ 24 'iii>c2 .rf.xb2+ 2S 'iii>xb2 ttJxc3 with a winning attack) 19 ...hxg4 20 ii.xc4 ltJd4 21 fxg4 .1'.xg4 22 l:.dfl l:.b4! 23 hs .rf.xc4 24 hxg6 .l:lxc3+ 2S 'ifxc3 ttJe2+ 26 'iii> c2 ttJxc3 27 .l:lh7+ 'iii>x g6 0-1 was J .Mestel-E.Gufeld, Hastings 1986/87. b2) lS ttJg3 ! bxc4 16 ii.xc4 .l:lh8? (better tries are 16 ... 'ifcs 17 .ids .rf.b6 and 16 ... 'iii>h 7!?) 17 ttJfs+ ii.xfs 18 exfs with a clear advantage for White in A.Ostrowski-Y.Zimmerma n, Oberwart 1998. Returning to 13 ...ttJxes:

advance, but this i s not really neces­ sary) 16 'iii>bl (after 16 g4? hxg4 17 hS both 11... gxf3 and 17... gs are good for Black) 16 ... ii.c6 17 'ifd4 'ifc8 18 ttJcds ii.xds 19 ttJxds ttJxds 20 exds 'ifts+ 21 'iii>al 'iff6 and Black was better in B.Popelka-P.Boukal, correspondence 1991. c) 14 ltJd4 b4 1s ttJds cs 16 ttJb3 ttJxds 17 'ifxds and now Black can play 11 ....l:lb6 18 ttJxcs l:.c6 19 ttJb3 ii.e6 20 'ifd2 ii.xc4 21 ltJd4 l:.c8 with equality, or 17 ... ii.e6 18 'ifxd6 'ifxd6 19 l:.xd6 ii.xc4 20 ii.xc4 ttJxc4 21 l:.xa6 .l:lfd8 with compensation for the pawn in Katowice E.Soloz hen kin-W.Sc hmidt, 1991. 14. ..axbs 15 ttJf4

14 cxbs

Again there are alternatives to ex­ plore: a) 14 cs? b4 1s ttJds ttJxds 16 'ifxds l:.bs and Black won a pawn in P.Marttinen-A.Pankratov, correspon­ dence 2002. b) 14 ttJf4 bxc4 lS ii.e2 (even worse are lS 'ifd4?! ttJfd7 16 ii.xc4? cs and lS ttJfds ttJxds 16 ttJxds ii.e6 17 ttJe3?! 'iff6, as in M.Myc-B.Socko, Sopot 1997) 1s ...ii.d7 (1s .. :.rf.h8 has been played fre­ quently as prophylaxis against the g4290

1s ... b4

Black could also play 1s ... ii.d7 16 ii.e2 and then either 16 ... .l:lh8 or 16 ...l:.a8, but there is no need to hesi­ tate and 1S ...b4 has scored well for him. 16 ttJcds ttJxd s

There is little point in giving White an extra option with 16 ... cs 17 ttJe3! ?. 17 ttJxd s cs 18 f4

Pa n n o Var i a t i o n : 9 h4 h5 1 0 o - o - o

Instead 18 ttJe3?! ii.e6 (18 ... l:.a8! ?) 19 'ifxd6 (19 f4 ltJg4) 19 ... 'ifxd6 20 l:.xd6 ii.xa2 gave Black a pleasant ending in O.Sande-V.Zagorovsky, correspondence 1981. However, 18 .�e2 is a sharp alterna­ tive. Black seems to have at least a draw after 18 ....l:la8 (other possibilities are the immediate 18... ii.e6 and 18 ... 'ifa5) 19 f4 (19 'iii>bl can be met by 19 ...ii.e6 or 19 ... 'ifa5) 19...ttJg4 20 f5 .'Of2 21 f6+ 'iii>h7 22 .�xh 5 (or 22 ii.C4 .!Llxhl 23 .l:lxhl ii.a6 !) 22 ...ttJxhl 23 'ifg5 (23 .l:lxhl runs into 23 ... .l:lxa2 ! ) 23 ... ltJg 3 (after 23 ... .l:lxa2 24 ii.xg6+ fxg6 25 h5 :.a1+ 26 'iii>c2 'ife8 27 .l:lxal 'ifxe4+ 28 itd2 'ikd4 it's a draw) 24 ii.xg 6+ fxg6 25 0,e7 'ife8 26 h5 ttJxh 5 27 .l:lh1 'iff7?! 28 !!:lxg6 'ifc4+ 29 'iii>d2 'ifd4+ 30 'iii>cl 'ifc4+ 31 'iii>d2 'ifd4+ 32 'iii>cl was drawn in P.Murdzia-P.Hanko, Krakow 2003. However, going back, 27 ... .l:lxf6 28 'ifxf6 :.xa2 looks to be winning for Black.

(both 19 ...ltJd4 and 1 9....l:la8 are inter­ esting alternatives) 20 g4 ttJxg4 21 ii.e2 ttJf2 (21 ...ttJf6! ?) 22 f6+ 'iii>h 7 23 ii.xh5 ttJxhl 24 .l:lxhl l:.e8? 25 'ifg 5 l:.e5 26 ttJe7! was winning for White in M.Rivas Pastor-J.Mestel, Marbella 1982. How­ ever, Black has a clear improvement in 24 ... b3! 25 a3 'ife8! 26 ttJe7 'ifa4! when he will be able to exchange queens with ... 'ifc4-c2+. 19 ts

Both 19 ii.d3 and 19 .�c4 can be met with 19 ...ii.e6. 19 ... gxfs!?

This weakens Black's kingside, but it prevents the advance of White's pawn to f6 and does win a pawn. Instead 19 ... .l:le8!? may also be possible. 20 ii.d3 White continues to play for the ini­ tiative. Instead 20 ttJf4 .l:lh8 21 'ifxd6?! 'ifxd6 22 l:.xd6 ttJf2 23 l:.g l fxe4 gave Black the better endgame in A.Czerwonski-K.Chorfi, correspondence 1998. 20 .fxe4 21 ii.xe4 .l:lh8 22 .l:ldfl ii.e6 ••

18 ...ttJg4

This is considered safer than 18 ... ttJc6, but that move is probably okay too if Black is careful: 19 f5 ltJe5

23 ttJf6!?

291

A ttacking C h e s s : Th e King 's Indian, Vo l u m e 1

This simplifies to an equal ending, but Black's pieces are well placed and it is difficult for White to find any real way to attack 23 ...ttJxf6 24 'if gs+ '>W8 25 l:.xf6 ii.xa2 26 'iff4 'ife7 27 l:.xd6 l:.e8 28 .ids

30

Instead 28 l:.h6 'iii>g7 29 l:.xh8 .l:lxh8 'ifg 3+ was agreed drawn in

292

H.Simon-K.Harald, 2000.

correspondence

28 ... ii.xds 29 .l:lxd s 'ife3+ 30 'ifxe3 l:.xe3 31 �xcs l:.e4

Black had no problems at all in J .Griffin-K.Harma n, correspondence 1992, and he even managed to eventu­ ally win.

Cha pter 16 Pan no Variation 9 h4 h5 10 �C1 1 d4 ttJf6 2 c4 g6 3 ltJc3 ii.g7 4 e4 d6 5 f3 o-o 6 ii.e3 ltJc6 7 ttJge2 a6 8 'ifd2 .l:lb8 9 h4 hs 10 ttJc1

White seems to 'change his mind' and switches to positional play. This method of play has obvious similarities with the 9 ttJc1 of our next chapter, but the advance of both h-pawns will often favour one side or the other. 10 es ...

This is the thematic move - Black strikes at the newly weakened d4square. Still, Black seems to have some difficulties in this line and he may wish

to consider avoiding what follows. Smirin has experimented with 10 ... ttJe8 11 ttJb3 fs, but this looks very loosening to me. I think Black could try waiting with 10...ii.d7! ?.

Here White has: a) 11 ii.e2 es looks good, as 12 ds ltJd4 now hits the e2-bishop and should be okay for Black, while 12 ttJb3 exd4 13 ttJxd4 bs! gives him counterplay. White's h4-advance looks out of place. b) 11 ttJb3 is more testing: 11 ... l:.a8 ! ? (instead 11. . .as! ? 12 d s ! ttJes 13 ttJxas cs 14 ttJb3 �6 lS 'ifc2! does not look 293

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

quite sufficient) looks silly but makes some sense. After 12 ii.e2 as 13 o-o a4 14 ttJc1 es Black has counterplay. Per­ haps 12 o-o-o is more critical when Black should still play 12 ... as. This idea could use some tests. 11 ds ttJd4 Black has experimented with 11 ... ltJe7, although not with much suc­ cess. 12 ttJb3

Instead 12 ttJ1e2 does not make much sense when compared to the 9 ttJc1 es 10 ttJb3 lines because the addi­ tion of the moves h4 and ... h s does not work well for White here. Black should play 12 ... cs 13 dxc6 ttJxc6 (instead 13 ... bxc6 is Line B below). After 14 .l:ldl ii.e6 1s b3 (or 1s ttJds bS) 1s ...bs (also possible is 1s ...'ifas 16 ttJds and now, instead of 16 ... 'ifxd2+ which should leave White a bit better in the end­ game, Black could try 16 ... 'ifa3 ! ?) 16 cxbs (16 'ifxd6 'ifxd6 17 .l:lxd6 ltJb4 gives Black reasonable compensation) 16 ... axbs 17 'ifxd6 'ifc8! Black had good play for the pawn in E.Knol­ R.Wojtkowiak, correspondence 2002. After 12 ttJb3 Black has an impor­ tant choice. He can simplify the posi­ tion or he can play a thematic pawn sacrifice. Unfortunately, neither of these seems completely satisfactory for the second player, although he is not without his opportunities after the sharp pawn sacrifice, 12 ... cs. (seefollowing diagram)

294

We will explore: A: 12 ttJxb3 •..

B: 12 cs •••

A) 12 ...ttJxb3 13 axb3 cs

Black tries to keep the queenside at least somewhat closed. No better is 13 ...c6 14 ii.e2 cxds 1s cxds with an edge for White.

Black has trouble in this structure after ... cs because the inclusion of the moves 9 h4 and 9 ...h s makes it difficult for him to create play on the kingside: ... ttJh s is no longer possible and ... fs is very risky. White, however, can play on

Pa n n o Va riation: 9 h4 h5 1 0 ttJc1

either side of the board. His most dan­ gerous plan is to prepare g2-g4. 14 ii.e 2 This flexible move seems best. White has also tried to initiate play on the queenside immediately with 14 b4 and here: a) 14...b6 1s bs axbs 16 cxbs ttJe8 17 ii.d3 (or 17 b4 fs) 17.. .fs 18 'ifc2 'ife 7 (18 .. .f4 19 ii.f2 ii.f6 looks like a better try) 19 'iii>dl 'iff7 (again, 19 .. .f4!?) 20 'iii>c1 ttJf6?! 21 exfs gxfs 22 ii.xfs ttJxds 23 ii.g6 'ife6 24 ttJxds 'ifxds 2s .l:ld1 'ife6 26 ii.e4 ii.b7 was L.Portisch­ J.Nunn, Szirak lnterzonal 1987, and now 27 .l:la7 ii.xe4 28 'ifxe4 would give White a comfortable advantage. b) 14... cxb4 lS ttJa4 b6 (instead 1s ...bs 16 cxbs axbs 17 'ifxb4 ttJe8 18 .ie2 looks very nice for White and here I do not trust 16 ... as either) 16 'ifxb4 as 17 'ifd2 ttJd7 18 ii.d3 ttJcs 19 ttJxcs bx cs 20 .l::i.xas .l:lxb2 21 'ifxb2 'ifxas+ 22 ii.d2 1l'd8 was level in F.Lucas-J.Gallagher, Bargteheide 1988. 14...ii.d7 There is nothing much better: a) 14 ... b6 1s ttJd1 ! ? (1s o-o-o bs) 1s...as?! 16 ttJf2 gave White a big stra­ tegic advantage in L.Christiansen­ T.Magar, Philadelphia 1990. b) 14 ...ttJe8 lS g4 fs? ! (a misguided plan; better choices are 1s ... hxg4 16 fxg4 ttJf6 11 ii.gs bs and 1s ... ttJf6 16 ii.gs hxg4 17 fxg4 bs) 16 exfs gxfs and here both 17 ii.gs and 17 gxfs ii.xfs 18 ttJe4 look good for White. 15 ttJd1!

This allows White to play b4 if he wishes, but more importantly i t pre­ pares ttJf2 and g4. 1s ...'ikb6

Instead 1s ... ttJh7 16 g4 'ifc8 16 b4 b6 17 ttJf2 ttJe8?! (this is not very good, but it is hard to suggest an improvement) 18 g4 fs?! (after 18 ... ttJf6 both 19 ii.gs and 19 l:.gl prepare the g4-advance) 19 gxfs gxfs 20 exfs ii.xfs 21 .l::i.g 1 gives White a clear advantage. 16 ttJf2

White sacrifices a pawn for the ini­ tiative. Also possible is 16 b4! ? 'ikxb4 17 'ifxb4 cxb4 18 ii.d2 when White will quickly regain the pawn with an edge in the ending. 16 ...'ifxb3 17 ttJd3 l:.fc8 18 o-o ii.es 19 .l:la3 'ifb6 20 b4 White had excellent compensation for the pawn in L.Christiansen-J.Nunn, German League 1988. B) 12 ...cs This pawn sacrifice is Black's main resource, but there are still some prob­ lems to solve. 295

A ttack ing C h e s s : Th e King 's I n dian, Vo l u m e 1

13 dxc6 bxc6

Instead 13 ...tt:Jxb3 is well met by 14 c7! 'ifxc7 lS axb3 ii.e6 16 ii.e2. This is perhaps not as bad for Black as it is without 9 h4 and 9...hs thrown in, but White still holds an edge. 14 tt:Jxd4 exd4 15 ii.xd4

spondence 2003) 2 1... gxfs 2 2 'iff4 ii.xe6 23 ii.xa6 was much better for White in S.Giannetto-J.Van Unen, correspon­ dence 2003. c) lS ...'ifas 16 l:.c1 (White could also try 16 'ifg s !? which forces 16 ...'ifd8) 16 ...l:.e8 (also unsatisfactory is 16 ... ii.e6 17 b3) 17 ii.e2 ds 18 cxds cxds 19 es ttJd7 20 f4 tt:Jf8 21 o-o tt:Je6 22 ii.f3 tt:Jxd4 23 'ifxd4 .l:lb4 24 'ifd2 'ifb6+ 2S l:.f2 ii.e6 26 b3 ii.h6 27 g3 .l:ld8 was l.lbragimov-J.Sammour Hasbun, Phila­ delphia 1992, and here 28 tt:Ja4 intend­ ing tt:Jcs is the most convincing. d) 1s ... ii.e6! ? is perhaps the best al­ ternative and it may be this move which resurrects Black's chances with 12 ...cs.

For the pawn Black has some open lines, a slight lead in development and the possibility of opening the centre with ... ds. At first it seems that the in­ clusion of the moves 9 h4 and 9 ...h s should help Black because the white king does not have an obviously secure shelter. However, analysis shows that this is not necessarily the case. 15...:.es

This is the most common move, but of the alternatives Black should pay close attention to variation 'd': a) lS ...cs 16 ii.e3 ii.e6 17 b3 gives White a clear advantage. b) 1s ... ds 16 cxds cxds 17 es tt:Jd7 (instead 17 ....l:le8 18 ii.e2 transposes to Line Bl below) 18 f4 f6 19 e6 l:.e8 20 o-o-o tt:Jb6 21 fs (or 21 g4, as in M.Massimini Gerbino-K.Sakai, corre296

Following 16 ii.e2 (perhaps White should look into 16 l:.c1! ?) 16 ... ds 17 cxds cxds 18 exds (18 es tt:Jd7 19 f4 f6 20 exf6 tt:Jxf6 gives Black some coun­ terplay) 18 ... tt:Jxds 19 ii.xg7 'iii>xg7 20 l:.dl ttJxc3 21 'ifxc3+ 'iff6 22 'ifxf6+ 'iii>xf6 gives Black good chances to hold the ending. Indeed, after 23 b3 l:.fc8 (or 23 ... as! ? 24 .l:ld2 l:.b4 2s �2 a4 26 bxa4

Pa n no varia t 1. o n 9 h 4 h5 1 0 ttJc1 · ·

.l:lxa4 27 .l:lal an d V2-V2 was J . Adler­ E Buk1c, · Bled 1992 ) 24 'iii>d2 as 2S .l:lcl d8+ 26 'iii>c3 .l:lb4 27 l:.bl l:.c8+ 2 8 'iii>d2 a4 an d V2-Vi, Black had held his gran dmaster opponent in . J.FedorOWlCZ- E . Ronneberg, Chicago 199S. 16 ii.e2 After 16 o-o-o 'ifa S Black at least has counterp] aY but 16 l:.d1 11 is a reason' able altemative. Foll 0wing 16 ...d s 17 cxds cxds 18 es ltJd7 (instead 18 ...ttJ h7 19 f4 f6 20 ii.e2 fxe s 21 fxes ii.xe s 22 o-o was much better for White in K.Chorf-F.Balabaev, correspondence 2001) 19 f4 ii.h6 20 'iff2 ttJf6 21 g3 ii.g4 22 .l:ld2 'ifc8 23 ii.e2 ttJe4 24 ttJxe4 dxe4 2S o-o ii.f8 26 ii.xg4 'ifxg4 27 l:.c2 l:.ed8 Black h ad some counterplay in V.Pere. . vertkin-N.Patnc1, correspondence 2 oos .

�

· ·

16...d s

This is the mo st common and was Karpov's c ho1·ce but may not be best. ' lndee d, Black even has two options:

811) 17 ... cxds

The most natural and obvious move is to s1mp . ly recapture. 18 es ltJd7 1 9 f4 .iL � h61 •

. Black must stnke wh 1"le the white . . kmg lS still in the centre. Here Wh1·te has two options: 81:>fi7.mls ll: t7 e5 . .

Worse is 19 .. .f 6 as Black's kingside would then be very weak . However, after the text Black threatens to capture on es. 297

A ttacking Chess: The King 's In dian, Vo l u m e 1

20 'ife3! Instead 20 ttJxds ttJxes 21 ii.c3 ii.b7 gives Black good play, while 20 ii.e3 runs into 20 ...ttJxes! 21 fxes ii.xe3 22 'if xe3 d4 when Black is doing well.

White a strong attack i n the well­ known game A. Karpov-S.Kindermann, Baden -Baden 1992. 26 ts

'ifxes ii.g7 24 'ife3 (not 24 'ifxds ? ii.d4+) 24... ii.xc3 25 'ifxc3 .l:lxe2 Black has good compensation for the ex­ change.

Instead 26 ttJxds ii.xds 27 ii.xb2 �6+ 28 .l:lf2 ii.f8! 29 ii.e2 ttJxf2 30 'ifxf2 'ifxb2 31 l:.xds was R.Alonso­ J .Butterworth, correspondence 2002. Here 31 ...l:.c2! 32 'iii>f l ii.cs ! would have been very strong, since 33 .l:lxcs 'ifcl+ 34 'ife 1 'ifxf4+ wins for Black.

2 1...ttJf6

26 ... gxfs 21 ii.xfs

Now 21 ...ttJxe s?! does not work be­ cause of 22 ii.xes l:.xes 23 'ifxes ii.g7 24 l:.xds (24 'ife3 is also good) 24...�6 25 'ife8+ 'iii>h 7 26 ltJe4!.

If White plays 27 ttJe2 l:.xa2 28 ttJf4 Black has resources, such as 28 ... 'ifa4! 29 ii.xfs ? ii.xfs 30 ttJxds .l:lc1!.

20 ..Jlxb2 21 .l:ld1 After 21 o-o ttJxes! 22 ii.xes .l:lxes 23

22 0-0

White should avoid 22 ii.f3 ? ltJg4 23 'ifcl ttJxes and Black also has good play after 22 'ifcl .l:lxe2+! 23 ttJxe2 ii.g4 24 l:.d3 ii.xe2 25 'iii>xe2 'ifd7!. 22 ...ttJg4 23 'ifg3 il.e6 24 ii.xa6 'ifas 25

21 ....l:lxc3

Thi s brings Black full equality, but 27 ... l:.b4!? is also possible.

ii.d3

2 s...l:.c8!

Black plays as actively as possible. This move was first pointed out by Knaak. Instead 2s ...ii.f8? 26 fs gave 298

After the text, White has no advan­ tage. For example: a) 28 ii.xc3 ii.e3+ 29 'if xe3 .l:lxg2+! 30 'iii>h l l:.h2+ 31 'iii>g 1 .l:lg2+ is a draw. b) 28 'ifxc3 ii.e3+ 29 'iii>h l ttJf2+ (or 29 ... 'ifd8 when 30 'ife1 is forced and after 30...ttJf2+ 31 l:.xf2 'ifxh4+ 32 ii.h 3 ii.xf2 33 ii.xf2 'ifxf2 34 'ifxf2 l:.xf2 35 .l:lal d4 Black is fine) 30 l:.xf2 'ifxc3 31

Pa n n o Vari a t i o n : 9 h 4 h5 1 0 tLlc1

ii.xc3 l:.xf2 and Black's activity gives him at least enough for the pawn. 812) 17 l:.b4!? ...

This is a very enterprising move. Black is willing to sacrifice more mate­ rial in order to create tactical complica­ tions as quickly as possible. 18 ii.cs This seems to be the best move, but White has tried other things as well: a) 18 es is risky and after 18 ... ttJxds 19 ttJxds .l::i.x d4 20 'ikxd4 ii.xes 21 'if cs kg3+ 22 'iii>fl cxds Black has good compensation for the exchange. Then 23 l:.d1 'ikf6 24 'ifxds 'ifxb2 2s 'ifd2 "i!fes led to a black win in S.Jardorf­ U.Rinkis, correspondence 1993. b) 18 ii.f2 cxds (here 18 ...ttJxe4? fails to 19 fxe4 ii.xc3 20 'ifxc3 l:.bxe4 21 o-o .:.xe2 22 dxc6 with a large advantage) 19 l:.d1 'ifas 20 ttJxds (20 o-o dxe4 21 a3 l:.b7 22 ttJxe4 'ifxd2 23 ttJxf6+ ii.xf6 24 �d2 l:.xb2 is equal) 20 ...ttJxds ttJxds 21 a3 .l:la4 (Black could also play 21 ... ii.h6 22 'ifxds 'ifxds 23 l:.xds .l:lxb2 with compensation, while 21 ....l:lxb2!?

22 'ikxas ii.c3+ 23 'ifxc3 ttJxc3 24 l:.d2 .l:lbl+ 2S ii.d1 ttJxe4 26 fxe4 .l:lxe4+ 27 l:.e2 ii.g4 28 o-o is interesting but probably better for White) 22 'ikxas .l:lxas 23 .l:lxds l:.xds 24 exds ii.xb2 re­ sults in an equal endgame. c) 18 ii.xf6 ii.xf6 19 o-o-o and here: cl) 19...'ifas is risky, but playable: 20 a3 .l:lb8 21 f4 (21 dxc6 .l:le6!) 21 ...ii.xc3 22 'ifxc3 'if xc3+ 23 bxc3 .l:lxe4 24 ii.f3 .l:lc4! 2S dxc6 .ifs 26 'iii>d2 was drawn here in M.Massimini Gerbino-R.Alvarez, correspondence 2003. c2) 19 ...cxds is an easy draw accord­ ing to Alvarez. It is not quite so simple, as 20 ttJxds ii.xb2+ 21 'iii>c2 l:.b8 22 .l::i.b1 ii.e6 23 .l:lxb2 ii.xds 24 exds (not 24 .l:lxb8 'ife7+!) 24 ... .l:lxb2+ 2S 'iii>xb2 'ikb6+ 26 'iii>al (Alvarez gives 26 'iii>a 3! ? .l:le3+ 27 'iii>a4 'ifcs 28 'ifb2 'ifxds 29 ii.xa6 'ifd7+ with a draw) 26 ...'ikf6+ 27 'ikb2 'ifxb2+ 28 'iii> x b2 .l:lxe2+ 29 'iii>c3 �8 30 .l::i.dl 'iii>e 7 31 .l::i.d2 l:.xd2 32 'iii>xd2 'iii>d 6 33 'iii>e 3 'iii>x ds 34 �4 f6 3S g4 looks strong, but then 3S ...hxg4 36 fxg4 'iii>e6 37 h s �7 3 8 h6 'iii>g8 39 g s t s 40 'iii>e s 'iii> h 7 does indeed after all look like a draw.

299

A ttacking C h ess: Th e Kin g 's I n dian, Vo l u m e 1

Black is better after both 22 d6?! "ifxh4 and 22 J:adl? cxds 23 "iff3 °ife7 24 .lld4 .llfs.

This looks best. The advanced white c-pawn is dangerous, but it need not be fatal if Black stays active. Instead 23 ...�xf2 ?! 24 .Jlxf2 °ifg4 2S "iff6 "ife6 26 .Jld4 "ifxf6 27 .Jlxf6 .Jlg4 28 C7 gives White an advantage, while 23 ...°ifgS?! 24 .llafl �xf2 2S .Jlxf2 also looks better for White. Black can probably hold after 23 ..."ifd8: for example, 24 �xe2 �xe2 2S �el �xel+ 26 "ifxel .Jle6 27 "ifes "ifds 28 "if xds .Jlxds 29 C7 .llb7, but it will only be after some suffering. 24 J:af1 fs 25 .Jld4 "ife4 Black stays as active as possible. He also managed to draw after 2S ... �h7 26 b4 °ife4 27 �C2 �d8 28 .Jlf2 �d3 29 "ifc4 "ifxc4 30 �xc4 J:e2 31 a4 �b3 32 J:e1 J:xel+ 33 .Jlxel �bl 34 �2 gs 3S .lld4 f4 36 J:d8 J:b2+ 37 �1 .llfs 38 �h8+ and V2-V2 in K.Pecotic-F.Rubio Dobias, correspondence 20os.

22 ..."ifxh4

26 �c1 f4

Black should certainly take the pawn. Worse are 22 ... "ifds 23 .Jlf2 and 22 ... .llfs 23 .llf2.

Here a draw was agreed in J.Camp­ bell-F.Dutu, correspondence 2003. After 27 C7 gs Black has counterplay.

18 ... lLixe41

Without this resource Black would be busted. 19 fxe4 .Jlxc3 20 "ifxc3 �bxe4

Black wins back the piece and an unusual opposite-coloured bishop situation arises. 21 0-0

White can also play 21 0-0-0, but af­ ter 21 ...�xe2 Black has at least a draw: 22 dxc6 (or 22 d6 .Jlfs 23 �hel "ifxh4 24 J:xe2 �xe2 2S d7 "iff4+ 26 .Jle3 "ifxe3+ with a draw) 22...°ifC7 23 .Jld6 "ifb6 24 �hel .Jlfs 2S �xe2 �xe2 26 c7 was L.Hesselberg-A.Serov, correspondence 2007. Here the simplest is 26 ...°ife3+ 27 "ifxe3 �c2+ 28 �bl J:d2+ with a draw. 21 ... �xe2 22 dxc6

23 �f2 e2) 11 es

23 ...�2e6

300

P a n n o Varia t i o n : 9 h4 h5 1 0 lLic1

This move is very dangerous, but it has not been seen much This is proba­ bly due to the fact that it has become well known that this line causes Black problems and thus the second player has been avoiding it altogether. 11 ...lLid7 Instead 17... cs is creative, but looks insufficient. White has: a) 18 .i.xcs lLid7 19 .i.d6 lLixes!? 20 .i.xb8 lLixc4 21 °ifcl .i.fs was suddenly agreed drawn in J.Lautier-1.Smirin, European Club Cup, Rethymnon 2003. White is up a rook, but Black has defi­ nite counter-chances. b) 18 exf6 was suggested by Hazai. After 18...cxd4 19 fxg7 dxc3 20 bxc3 �xg7 21 cxds °ifb6 22 �fl .i.fs 23 g3 it is hard to believe that Black can prove he has enough for the two pawns. 18 e61

White will exchange the dark­ squared bishops, castle queenside and break with g2-g4. This move also mis­ places Black's pieces and is much stronger than 18 f4 .i.h6! 18 cxds lLixes. 18...J:xe6

Instead 18 ...fxe6 19 .i.xg7 �xg7 20 o-o-o °iff6 21 g4 hxg4 22 fxg4 was also

much better for White in J.Lautier­ M.Golubev, Internet (blitz) 2004. 19 .i.xg7 �xg7 20 o-o-o

20...lLif6 21 g41 .i.b7

This is not a nice move to make, but 21 ... hxg4 is met by 22 h S with a strong attack. 22 gs lLieB 2 3 .i.d3 "ifb6 24 lLia4 Instead 24 �hel looks better for White. 24..."ifc7 Here 24...°ifd4 is a better try. 2s lLics White has a clear advantage and won after 2s ...�e7 26 �hel �g8 27 �bl lLid6 28 �xe7 "ifxe7 29 �el °ifC7 30 lLie6 fxe6 31 �xe6 lLie8 32 �xg6+ and 1-0 in R.Molo-P.Diaz Sjostrom, corre­ spondence 2006.

301

Chapter 17 Pa nno Variation 9 !t:Jc1

1 d 4 l2Jf6 2 C 4 g6 3 l2Jc3 .Jt.g7 4 e4 d6 5 f3 o-o 6 .i.e3 lLic6 7 lLige2 a6 8 'ifd2 .rf.b8 9 lLicl

A: 10 �b3

B: 10d5 A) 10 lLib3

This move is not very popular, partly because White spends a long time ma­ noeuvring with his knight. 10 exd4 11 lLixd4 ...

Instead of playing for a direct at­ tack, White simply prepares to con­ tinue his development. There are some similarities between g lLicl and 9 h4 h S 1 0 lLicl, but very often the advances of the h-pawns radically change the play. 9 es ...

Of course Black strikes at the weak­ ened d4-square. Now White can either maintain the tension or advance in the centre. 302

If we compare this to the Glek Variation of the Classical (1 d4 lLif6 2 c4 g6 3 lLic3 .i.g7 4 e4 d6 5 lLif3 0-0 6 .i.e2 es 7 o-o exd4 8 lLixd4 �e8 9 f3 lLic6) Black is up a couple tempi because

Pa n n o Va ria tion: 9 lLic 1

White has played lLige2-c1-b3xd4 rather than lLif3xd4. True, Black has spent these tempi on ...�b8 and ... a6, but these could be very useful moves if he can initiate play on the queenside. Black has also saved the move ...�e8, which is usually played to force f2-f3. This gives him another tempo to try to create queenside counterplay. At this juncture Black has more than one way to continue. Usually he will try to take advantage of his extra moves on the queenside to aim for ... bs, normally preceded by either ...c6 or even ... cs. We have: Al: 11•.•lbxd4 A2: 11••.lLleS

These two continuations are seen with similar frequency. Black could also consider 11...lLihS!? which is a common idea in the aforementioned Glek. Al) 11...lLixd4

Black chooses to simplify the centre.

Now Black can prepare ...bs with ... c6 or he can try to achieve this ad­ vance tactically, aiming to blow up the position while he is still ahead in de­ velopment. All : 13 .c6 ••

A12: 13 cs!? •••

AU) 13 ...c6

A logical move. Black will follow up with either ... bs or ... ds. 14 a4

The alternative is to allow ...bs with 14 o-o. After 14 ... bs 1 s b3 (1s cxbs axbs 16 �fcl "ife7 17 a3 �fd8 18 b4 �bc8 was fine for Black in A.Ker­ E.Gufeld, Wellington 1988) 1s ...bxc4 16 bxc4 (instead 16 .i.xc4 .i.xc4 17 bxc4 "ifas 18 l:l.ab1 cs 19 .i.e3 lLid7 20 lLids "ifxd2 21 .i.xd2 lLies gave Black good counterplay in P.Korning-L.Fell, corre­ spondence 198S) 16 ... "ifas 17 �acl (17 �abl allows 17 ...cs 18 .i.e3 lLig4) 17 ...�fd8 Black intends ...cs followed by ...lLid7 with active play.

12 .i.xd4 .i.e6 13 .i.e2

Here White has: 3 03

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1

a) 18 �fd1 cs 19 .i.e3 lLid7 20 lLids 'ifxd2 21 �xd2 .i.xds 22 cxds (22 J:xds J:b2) 22 ... as 23 a4 was A.Riazantsev­ E.Gousseinov, Istanbul 1998. Now 23 ...l:l.b4 with the idea of 24 .i.bs lLib6 would have given Black counterplay. b) 18 �h l cs 19 .i.e3 'ifa3 (also good is 19 ... lLid7, because 20 lLidS?! is well met by 20 ... 'ifxd2 21 .i.xd2 �b2) 20 J:c2 lLid7 gave Black good play in J. Hjartarson-J.Nunn, Rotterdam 1989. 14...ds

This forces the pace and while Black should be able to make a draw, any winning chances are rather lacking. Black has also tried 14...'ifas lS o-o 'ifh4!? (other possibilities are lS ... cs and 1s ... J:fd8) 16 as cs 17 .i.e3 (not 17 �a4 lLixe4! 18 fxe4 .i.xd4+ 19 'ifxd4 and 0-1 in A.Suarez-S.Kamberi, Dallas 2001) 17 ... lLid7 18 'ifcl ! .i.xc4 19 .l:.a4 .i.xe2 20 .l:.xb4 .i.xfl 21 l:l.a4 .i.bs which is un­ clear.

Worse is 18 ....i.xf6 19 l:l.d1. 19 l:l.dl 'ifd7 20 0-0 l:l.bd8 21 .i.b6 Instead 21 'if e3 'iff7 (21 ...�fe8 looks okay too) 22 .i.f3 .l:.fe8 23 'iff2 .lt.g4 24 .i.es .i.xf3 2s 'ifxf3 'if e6 26 .i.xf6 'ifxf6 27 lLixds 'ifxb2 28 fs gxfs 29 lLie7+ �h8 30 J:xd8 l:l.xd8 31 lLixfs .llf8 was level in L.Brunner-Xie Jun, Sth matchgame, Bern 199S. 21...d4!?

After 21 ...�c8 22 .i.f3 White has a slight but persistent advantage. 22 'ifxd4 'ifxd4+ 23 .i.xd4 Black should meet 23 �xd4 with 23 ... lLid7 24 �d6 lLixb6 2S �xe6 .i.xc3 26 bxc3 lLixa4 (Be liavsky). 23 ... .i.b3 24 J:d3

Wisely avoiding 24 J:d2? lLie4 2S lLixe4 �xd4. 24. ...lt.c4 25 J:d2

15 cxds cxds 16 es

This was A.Beliavsky-J.Nunn, Reyk­ javik 1988. Black has sacrificed a pawn and must play accurately to maintain compensation.

White has to play this way if he wants to obtain anything.

25 ....i.xe2 26 lLixe2 lLie4 27 �ddl .i.xd4+ 28 lLixd4

16...lLid7 17 f4 f6 18 exf6 lLixf6

3 04

After 28 �xd4 �xd4 29 lLixd4 lLics

Pa n n o Variatio n : 9 4J c1

30 as .l:.d8 31 4Jf3 4Jb3 Black is fine. 28 ...tt:Jcs 29 as �ds 30 b4 �fd8 31 bxcs J:xd4 32 �bl

Instead Beliavsky gives 32 J:xd4 �xd4 33 J:b1 .l:.c4 as equal. 32 ...J:c8 33 �fc1 �ds 34 ltxb7 �dxcs 3S �el .l:.cl 3 6 J:be7 J:xel+ 37 �xel J:cs 38 �al 't;f7

Because of the difference in activity of the rooks, Black should draw without much difficulty. A12) 13 ...cs!?

The only consistent move. Black rips the position open and is prepared to offer more material. 17 .i.xcs

White takes everything on offer. In­ stead 17 exds tt:Jxds 18 .i.xcs is also possible, but after 18....l:.e8 White must avoid 19 0-0?! (a better try is 19 l:l.dl when it is not clear that Black h as enough after 19 ... .i.es 20 g3) 19 ... tt:Jc3 ! 20 4Jd6 tt:Jxe2+ 21 'ifxe2 J:xb2 when Black has the initiative, A.Kemmerer­ T.Marten, correspondence 1986. 11...dxe4 18 .i.xf8 'ifxf8

This is an enterprising but risky way to play. 14 .i.e3 bs!?

Black goes all in, but 14...'ifas 15 o-o �fd8 is also possible. lS cxbs axbs 16 tt:Jxbs

Instead 16 o-o?! b4 17 tt:Jds (better is 17 tt:Jbs 4Je8 18 .l:.fdl 'ife7, although Black is fine after 19 tt:Jxd6 'ifxd6 20 'ifxd6 tt:Jxd6 21 J:xd6 .i.xb2) 17 ... .i.xds (or 17 ... tt:Jxds 18 exds .i.fs) 18 exds �e8 19 .lt.c4 4::l d7 leads to a good Benoni structure for Black. 16... ds

White is up the exchange and a pawn, but with a development lead and dark-square play, Black has certain tactical chances. Some possibilities: a) • 4::ld4!? is natural but has not been seen in practice. After 19 ...tt:Jds 20 fxe4 .i.h6 Black has counterplay. b) 19 'ifd6 hurries to exchange queens: 19 ...'ifxd6 (or 19 ...exf3 20 'ifxf8+ 't;xf8 21 gxf3 tt:Jds 22 't;f2 tt:Jf4 23 a4 .i.xb2 24 �abl .i.f6 with reason­ able compensation for the exchange in J .Brennin kmei jer-M.Bos boom, Hilver305

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1

sum 1988) 20 lLixd6 J:xb2 21 lLixe4 lLixe4 (not 21 ... lLids 22 �dl, as in J.Eising-A.Sznapik, Amsterdam 1984) 22 fxe4 .Jlc3+ 23 �fl J:c2 24 l:l.dl J:xa2 gives Black compensation. c) 19 o-o exf3 (Black should not hurry with 19 ...�xbs 20 .Jlxbs "ifcs+ because after 21 �hl "ifxbs 22 "ifd8+ "ife8 23 "ifxe8+ lLixe8 24 fxe4 .Jlxb2 2S �abl .lies 26 a4 the passed pawns give White the advantage in the ending, N.Andri­ anov-A.Kuzmin, USSR 1984) 20 gxf3 (af­ ter 20 �3 Black should play the imme­ diate 20..."ifcs+, rather than 20...J:xbs? 21 .Jlxbs "ifcs+ 22 "iff2 and 1-0 in N.Le Gore-R.Lifson, correspondence 1988), and here 20.. J::t xbs 21 .Jlxbs "ifcs+ 22 �f2 "ifxbs 23 a4 looks better for White, but 20 ... .Jlh6!? with the idea of ...lLids gives Black active play. A2) 11...lLies

Black elects to keep the knights on the board. 12 J:d1 This deters Black from playing ...cs. Other moves are also seen: 3 06

a) 12 o-o-o is uncommon, but inter­ esting. Then 12 ... c6 (worse is 12 ... cs 13 lLib3 .Jle6 14 "ifxd6 "ifxd6 lS �xd6 .Jlxc4 16 lLixcs when White has won a pawn) 13 h3 (13 lLib3 .Jle6) 13 ...°ifC7 14 lLib3 .Jle6 lS "ifxd6 "ifxd6 16 �xd6 leads to: al) 16 ....Jlxc4 17 f4 (this was consid­ ered to be winning by Szuhanek, but matters are not so clear) 17 ... hfl 18 fxes .Jlxg2 19 .l:.gl (or 19 exf6 .Jlxh l 20 fxg7 �fd8 with an unclear position, as 21 �xd8+ �xd8 22 .Jlh6? �d3 is good for Black) 19 ...lLixe4 20 �xg2 lLixd6 21 exd6 b6 is unclear. a2) 16 ... lLihS ! ? 17 lLid2 lLig3 18 �gl lLixfl 19 �xfl .Jlxc4 (better is 19 ... �fd8! with good compensation for the pawn) 20 .l:.dl .Jle6 21 f4 .Jlh6 was V.Varaciuc­ R.Szuhanek, Romanian Team Champi­ onship 1993. Here 22 �c2 ! ? looks a lit­ tle better for White due to his space advantage. b) 12 .Jle2 cs (bad is 12 ...c6 13 o-o bS? 14 f4! b4 lS fxes bxc3 16 "ifxc3 lLixe4 17 "ifc2 with a clear advantage for White) 13 lLic2 (13 lLib3 .Jle6 is pleasant for Black) 13 ....Jle6 and now:

Pa n n o Va riation: 9 CZ'ic1

bl) 14 b3 scores terribly for White after 14 ... 'ifas: for example, lS o-o bS 16 cxbs axbs 17 �fd1 (17 CZ'ixbs l:l.xbs 18 .i.xbs 'ifxbs 19 'ifxd6 CZ'ifd7 is also good for Black) 17 ...�fe8 18 .i.f2 'Lied 7 with a very comfortable position for Black in A.Beliavsky-E.Gufeld, USSR 1979. b2) 14 CZ'ia3 CZ'ic6 lS o-o 'Lid4! (or 1s ...CZ'id7 16 f4 'Lid4 17 .i.d3 bS!? 18 cxbs axbs 19 .i.xbs 'Lif6 with counter­ play in A.Beliavsky-J.Hjartarson, Reykja­ vik 1991), and now both 16 �fdl 'Lid7 17 CZ'ic2 CZ'ixe2+ 18 'ifxe2 'ife7 19 'Lids? l.xds 20 �xds CZ'ib6 21 J:d2 .i.xb2 22 .:b1 'Lia4 (M.Magomedov-K.Maslak, Sochi 2oos) and 16 �acl 'Lid7 17 b3 it'as 18 CZ'iabl bS (Y.Gerasimov­ D.Yevseev, St Petersburg 200S) are very good for Black.

avoid his suggested 1s...b4 due to 16 'Lia4! and prefer 1S ...CZ'ixc4! 16 .i.xc4 bxc4 17 CZ'ixcs 'ifas which is unclear.

12...c6 13 .i.e2 bs

16 ClJC2

14 cxbs

Instead 14 f4? b4! lS fxes bxc3 16 "lifc2 (or 16 'ifxc3 CZ'ixe4 17 'ifc2 'inl4+) 16 ... 'Lig4! 17 .i.xg4 .i.xg4 with a clear advantage is analysis by Kasparov, but after 14 o-o cs lS CZ'ib3 Black should

14...axbs

Black plans ...b4 and ...ds. 15 b 4 c5

Forcing the pace. Black can also play 1 s ... .i.d7 16 o-o �e8 17 CZ'ic2 .i.f8 18 �hl .i.e6 with equal chances. Taking the pawn with 16 CZ'idxbs cxb4 17 CZ'ib1 'ifas 18 CZ'ixd6 looks too risky for White after 18...�d8. White may also try 16 bxcs b4 17 'Lia4 (17 CZ'ib1! looks more critical, as after 17 ... dxcs 18 CZ'ib3 'ifxd2+ 19 CZJ1xd2 .l:.d8 20 .i.xcs .i.e6 21 o-o Black has to prove that he has enough for the pawn), although after 17 ... 'ifas 18 CZ'ib6 dxcs 19 CZ'ixc8 cxd4 20 'Lie7+ �h8 21 .i.xd4 J:fd8 Black has good play: a) 22 o-o CZ'ixe4! 23 fxe4 �xd4 24 'ifxd4 'Lif3+ 2S J:xf3 .i.xd4+ 26 �xd4 'ifcs with a clear advantage according to Kasparov. b) 22 'ifb2 �xd4 (instead 22 ...b3+ 23 'ifc3 'ifxc3+ 24 .i.xc3 bxa2 2 S J:xd8+ 307

A ttacking C he s s : Th e Kin g 's I n dian, Vo l u m e 1

J:xd8 26 i.xes lLixe4! 27 i.xg7+ �xg7 28 o-o lLic3 29 i.c4 .l:.d7 30 i.xa2 lLixa2 31 lLic6 is equal) 23 'ifxd4 b3+ 24 'ifd2 'ifa7 25 axb3 'ifxe7 and Black's active minor pieces give him slightly the bet­ ter chances.

A.Beliavsky-V.Loginov, Azov 1991, but Kasparov's move looks simpler.

16...cxb4 17 lLixb4 i.e6! 18 o-o

B) 10 dS

After 18 lLixbs 'ifas! (even better than Kasparov's 18 ... i.c4) Black is clearly for choice, while 18 'ifxd6 'ifxd6 19 J:xd6 .l:.fc8 gives him a strong initia­ tive for the pawn. 18 . 'if as 19 'ifxd6 lLic4! This initiates a forcing sequence that will lead to a balanced endgame.

27 lLib6 �c6 28 J:b1 c3 29 �b4 i.xa2

Black had no problems in A.Beliavsky-G.Kasparov, Linares 1990.

More logical than 1o lLib3. 10...lLid4

..

20 i.xc4 bxc4 21 lLic6 'ifxc3 22 i.d4 lLixe4! 23 i.xc3 lLixd6 24 i.xg7 �xg7 25 lLixb8 lLif5

Black is down the exchange, but White's knight is yet to escape. More­ over, the a2-pawn is weak and Black's strong c4-pawn will give him enough counterplay to hold the balance. 26 lLid7

White now has a choice of moves to challenge the black knight.

B1t):l. �e:2 B2: U �J Bl) 11 lLi1e2

26 ....l:.c8

Black also managed to draw after 26 ....l:.a8 27 lLics J:xa2 in the later game 308

Pa n n o Va riation: 9 lLic1

This makes more sense than it does with the moves 9 h4 hS included, be­ cause after 11 ... cs White can castle kingside. We examine two lines here, although Black really has three choices because he can also play a pawn sacri­ fice that will be covered in Line B22.

13. .fs .

8:11: 1�...lbke2· 81%: :1it..+t5:

811) 11...lLixe2

This is a simple way of playing. Black aims for a normal position. 12 .i.xe2

Exchanging a pair of knights has left Black with enough room for his minor pieces. He has a flexible position and may go .. .f4, .. .fxe4 or ...lLif4 on the kingside, while playing on the queen­ side may also be a possibility. 14 �b1

12 ...lLihs

Having this move available is one of the main advantages for Black com­ pared to the previous chapter. Instead 12 ... cs is less good because if we com­ pare the position to Line B21 below, White does not have a weakness on b3 after 13 dxc6 bxc6. 13 0-0-0

Other moves are rare in practice and against them Black will also play 13 .. .fs.

Instead after 14 exfs gxfs or 14 g3 lLif6 (Black could also consider 14.. .fxe4 or 14.. .f4) 14 .i.d3 lLif4 lS .i.c2 bS (or 1s ..."ife7) 16 cs b4 17 lLia4 as 18 c6 fxe4 19 fxe4 .i.a6 Black has counterplay. 14 cs is White's main alternative: a) 14.. .f4 is probably premature. Af­ ter l S .i.f2 .i.f6 16 �hfl .i.h4 17 .i.g1 White had a free hand on the queen­ side in L.Kaufrnan-W.Morris, USA 1979. b) 14.. .fxe4!? lS lLixe4 (or lS fxe4 lLif4) 1s ....i.fs looks okay for Black. c) 14 ... �h8 lS �bl b6 16 cxb6 cxb6 17 .l:.cl bS 18 .i.d3 lLif6 combined play on both flanks in C.Ward-M.Hennigan, British League 1998. d) 14..."ife7 lS cxd6 (White's king position looked a bit loose after lS b4 lLif4 16 .i.f1 fxe4 17 lLixe4 .i.fs in 3 09

A ttacking C he s s : Th e King 's I n dia n , Vo l u m e 1

A.Ward-J.Seia, correspondence 2006) 1s ...cxd6 16 .i.d3 bS 17 �bl .i.d7 gave Black a fine position in l.Lutsko­ A.Fedorov, Minsk 1994.

After 18 exfs gxfs 19 g4?! fxg4 20 fxg4 lLixg4 21 .i.gs .i.f6 White may have some play, but Black has a solid extra pawn.

14....i.d7

18 ... bs! 19 cxbs axbs 20 g4 b4 21 lLie2 fxe4 22 fxe4 cs! 23 dxc6 .i.xc6

Black could also play 14..."if e7. There is no real need to determine the place of the hS-knight yet with 14 ... lLif4 or 14...lLif6, however.

15 .i.d3

Or lS .l:.cl "ife7 16 .i.d3 lLif6 with chances for both sides. Instead lS a3 looks a bit pointless and 1s ... �h8 16 .i.d3 "ife7 17 l:l.c1 lLif4 18 .i.c2 cs 19 dxc6 bxc6 (with the idea of ... lLie6, tak­ ing aim at the d4-square) 20 cs? dS! was good for Black in A.Dreev-Ye Jiang­ chuan, Shanghai 2001. White cannot grab the pawn with 21 exds lLixds 22 lLixds cxds 23 "ifxds, because of 23 ... .i.e6 with the idea of ...e4. 15 ...�h8

Black might consider too 1s ..."ife7 ! ?. 16 .l:.hgl!?

White is thinking about playing g2g4 at some point. 16..."ife7 17 .i.c2 lLif6 18 h 3 310

Black has good counterplay against the e4-pawn and another possibility is to break in the centre with ... ds, to open up the diagonal fort he g7-bishop. 24 .i.gs

White will take the f6-knight to re­ lieve the pressure on e4, but Black still has good chances. It was probably pref­ erable to go in for 24 "ifxd6 °ifh7! 2S lLig3 b3! 26 axb3 lLixe4 27 lLixe4 .i.xe4 28 .i.xe4 "ifxe4+ 29 °ifd3 "ifxd3+ 30 l:l.xd3 e4 31 l:l.d7 J:xb3 32 .i.cl when White should hold the ending, but Black still has some chances to press. 24...°ifa7 25 hf6 J:xf6 26 lLic1

This was l.lbragimov-A.Fedorov, So­ chi 1997. Now 26 ... �f2! 27 "ifxd6 b3 would have been strong: 28 .i.xb3 (Black wins after 28 lLixb3 �xc2 29 �xc2 .i.xe4+ or 28 axb3 l:l.a8 29 "ifa3 "ifb6 30 "ife7 .i.f8! 31 "ifxes+ .i.g7 32

Pa n n o Va ria tio n : 9 lLic1

'ife7 'ifas! 33 'ifa3 .i.xe4!) 28... .i.xe4+ 29 �al .i.c2 with the idea of ... e4 (Fedorov) would have been very good for Black. 812) 11... cs 12 dxc6 lLixc6

Note that the pawn sacrifice 12 ... bxc6 13 lLixd4 exd4 14 .i.xd4 will be considered under Line B22, below. The text move looks better, however, as Black gets good chances with much less risk.

lLid4 and Black already has the initia­ tive. b) 13 lLids bs (13 ... lLixds 14 cxds lLias 15 lLic3 bS also looks okay) 14 lLiec3 in creases White's grip on dS, but gives up the d4-square, and after 14 ... lLixds (14 ...lLid4 15 .i.d3 lLixds is fine too) 15 cxds (or 15 lLixds lLid4) 15 ... lLid4 Black has a good position. c) 13 lLicl .i.e6 (Black can also sacri­ fice a pawn with 13 ...bS!? 14 cxbS axb s 15 .i.xbs lLid4 with the idea of ... ds) 14 lLib3 lLias (14...lLie8! ? 15 .i.e2 fS is an­ other, more risky way to play) 15 lLixas (instead 15 lLids lLixds 16 cxds lLixb3 17 axb3 .i.d7 is also level) 15 ...'ifxas 16 lLids (16 'ifxd6 bS 17 cxbs axbs gives Black good counterplay) 16 ... 'ifxd2+ 17 �xd2 lLixds 18 cxds .i.d7 leads to a sterile ending. 13 ....i.e6

After the safer recapture, the posi­ tion resembles a Maroczy Bind where Black has played the rather horrid­ looking ...es. The justification for this is Black's development lead and possibili­ ties of breaking in Sicilian style with ...bs and even ...ds. Black will have to play very actively to hold the balance, but White's position is also not so easy to pl ay. 13 �d1

White logically increases the pres­ sure on the d-file. Other moves have also been tried, but they do not appear to be too dangerous: a) 13 o-o-o .i.e6 14 lLids bS 15 lLiec3

Black develops and counterattacks against the c4-pawn. 14 lLids

This natural move may not be best, but White has trouble achieving any­ thing: 311

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1

a) 14 'ifxd6 'ifxd6 l S l:txd6 .Jlxc4 is certainly okay for Black. b) 14 b3 bS (Black could also con­ sider 14 ... 'if as lS lLids and here not 1s ... 'ifxd2+ 16 J:xd2 when White is bet­ ter, but rather 1s ... 'ifa3!?) lS 'ifxd6 (or lS cxbs axbs 16 'ifxd6 'ifc8 with coun­ terplay) 1s ...'ifxd6 16 �xd6 lLib4 17 lLids lLic2+ 18 �d2 lLixe3 19 �xe3 bxc4 20 bxc4 J:b2 21 J:xa6 lLid7 and Black's activity combined with White's poor development should give him compen­ sation for t he pawns. c) 14 lLicl lLid4 (Black could also consider 14...'ifc8! ? lS lLib3 bS 16 cxbs axbs and now both 17 lLixbs dS and 17 .Jlxbs .l:.d8 followed by ...ds leave him with good play for the pawn) lS lLib3 bS 16 lLixd4 (after 16 cxbs .Jlxb3 17 axb3 axbs 18 .Jlxd4 exd4 19 lLixbs dS 20 es �e8 Black has the initiative) 16 ... exd4 17 .Jlxd4 bxc4 18 .Jle2 lLihS ! ? 19 .Jlxg7 lLixg7 2 0 .l:.cl (Black also looks to be fine after 20 o-o 'ifb6+ or 20 'ifxd6 'ifg s with counterplay against g2 and b2) 20 ... lLihs (20...'ifb6!?) 21 lLid1?! (bet­ ter is 21 o-o 'ifb6+ 22 �hl 'ifxb2 23 'ifxd6 .l:.fd8, but Black is fine here too) 21 ...'inl4+ 22 lLif2 'iff6 23 �c3 lLif4 and Black had the initiative in N.Mishra­ M.Waqar, India 1999. 14... b5 15 cxb5

After both lS lLixf6+ .Jlxf6 (but not 1s ... 'ifxf6?? 16 .Jigs) and 1s lLiec3 lLid4 Black is also doing well. 15 ...axb5 16 lLixf6+

Or if instead 16 lLiec3 then 16 ... lLid4 with good play. 312

16....Jlxf6

17 g3

White tries to get developed. In­ stead 17 lLiC3 lLid4 is very comfortable for Black, while 17 'ifxd6 'if as+ 18 .Jld2 (even worse is 18 'ifd2 lLib4!) 18 ...'ifb6 gives Black a dangerous initiative for the pawn. 11....Jlxa2 18 'ifxd6 .Jlb3

Black could also consider 18...lLid4 19 'ifxd8 �fxd8 20 lLixd4 exd4 21 .Jlf4 .l:.a8! 22 .Jlxbs .Jlb3 when if the rook moves 23 ...d3 will give him a strong initiative for the pawn. 19 'ifxd8 �fxd8

Gufeld's suggestion 19 ... .Jlxd8 is also fine. After 19.. 1Hxd8 Black's posi­ tion was a little easier to play in Z.Polgar-E.Gufeld, Wellington 1988, although the game was still pretty level. 82) 11 lLib3

This is a little more challenging, as White does not clog up his own devel­ opment. Black can simplify or offer a thematic pawn sacrifice.

. P a n n o Varia t i o n : 9 ltJ c1

pawns, while both 1 3 .i.e2 and 1 3 .i.d3 would be met by 13 ... ltJh s .

821) 11 r.-. '"UXb3 12 axb3 cs ...

. Black blocks th e queens1de. Also . '"l.J worth considen·ng is 12 ...r.-.h S at once.

After the text, White can begin queens1"de play or make a committal . kmgs1"de advance. 8211: �3 1>4 8212: 13 g4

Instead 13 dxc6 bxc 6 gives Black . enoug h play agamst the doubled b 313

A ttacking C h e s s : Th e Kin g 's I n dian, Vo l u m e 1 17 lLib6

After 17 lLic3 Black also has 17 ....i.h6! 18 .i.f2 (18 .i.xh6 "ifh4+) 18 ...°ifg5, which threatens mate and he has counterplay with both 19 lLidl f5 and 19 .l:.dl .i.d7. 11 ....i.h6I

Here we see another resource that is available to Black because of the omission of 9 h4 and 9 ...h 5. 18 .i.f2

Instead 18 .i.xh6 "ifh4+ 19 �dl "ifxh6 is fine for Black. After 20 lLixc8 �xc8 21 "ifd2 he can play 21 ..."ifxd2+ 22 �xd2 f5 (V2-V2 in V.Lukov-Z.Lanka, Ge­ neva 1993) or 21 ...'inl4 22 .i.xb5 lLig7 with compensation for the pawn. 18 ...�xb61

Black falls into the 'trap'.

After 2 2 .i.h4!? .i.a6 2 3 °ifb6 �xb2+ 24 �el lLic7! 25 .i.e2 (not 25 "ifxc7? .rf.c8) 25 ... �c8 26 �f2 Black has: a) 26 ...�a2 27 J:e1 .i.d2 28 �bl .i.a5 29 "if xd6 b4 30 "ifxe5 .i.xe2 should fa­ vour Black, although White went on to win in R.Ristorto-C.Blanco Gramajo, correspondence 2000. b) 26 ...b4! ? 27 �el .i.d2 28 .i.xa6 .i.xel+ 29 �fl (after 29 �xel? the ob­ vious 29 ... lLixa6 gives Black a winning position, although 29 ...lLixd5! is even stronger) 29 ... lLixa6 30 "ifxa6 .i.xh4 31 "ifxc8+ �g7 and only Black can be bet­ ter. 22 ....i::tx b2+ 23 �d3?

It is more prudent to play 23 �el when Black should take the perpetual check.

19 "ifas

19 ....l:.a61

Only this queen sacrifice can justify Black's play. 20 "ifxd8 J:xal+ 21 �e2 J:b1 This is better than 21 ....i.a6, which has also been tried. 22 "ifxc8

314

23 ...J:xf2 24 °ifh3 J:d2+ 25 �C3 b4+ 26 �xb4 .i.e3

Black had a strong attack in l.lbragimov-A.Delchev, Linares 1997. 8212) 13 g4

This is more ambitious. White pre­ vents ... lLih5.

Pa n n o Varia tio n : 9 lLic 1

create counterplay on either side of the board with 16 ... bs or 16 .. .f6. 1s ...°ifh4+

This is the other point. Black avoids ruining his pawn structure. Still, his queen may prove to be misplaced on the edge, so care is needed. 16 "iff2 "ifxhs

13 ... hs

This is much more enterprising than 13 ...lLie8 (th is is better than in the analogous line with 9 h4 hS - note 'b' to Black's 14th in Line A of Chapter 16 but it still looks insufficient) 14 h4 fS 1s gxfs gxfs 16 exfs .i.xfs 17 hs and White is better.

This is better than 16..."ifxf2+ 17 �xf2 gxhs 18 .i.d3 h4 19 �e2 �h8 20 b4 when White has a nice edge in the endgame, A.Karpov-Xie Jun, 3rd matchgame, Guanzhou 2000. 17 .i.e2

14 h3 lLih7

Now Black's queen does not feel too comfortable, but there are sufficient resources. 11...lLif6!?

This is a typical idea in this struc­ ture and it is also seen in certain 6... es lines of the Samisch. 15 gxhs

One point is that after lS o-o-o h4! Black keeps the kingside closed and plans ... .i.f6-g s. After 16 gs Black can

An interesting try. Worse is 17 ... .i.f6 18 h4!, but 17 .. .fs is more common when White has: a) 18 f4 °ifh6 and now: al) 19 h4 exf4 20 'i'xf4 gs 21 "iff1 f4 (or 21 ... .i.es !? as suggested by Golubev, while 21....i.d4!? with the idea of 22 hxg s "ifg7 looks good as well) 22 hxgs "ifxgs 23 .i.d2 °ifg3+ 24 �d1 f3 2s .i.d3 315

A ttacking Ch ess: The King 's I n dian, Vo l u m e 1

lLif6 with sharp play that did not look unfavourable to Black in A.Utnasunov­ V.Kuznetsov, Vladimir 2002. a2) 19 exfs exf4 20 "ifxf4 gs 21 °ifg3 .i.xfs ( 2 1....i.es !? is again proposed by Golubev) 22 h4 �be8! 23 �d2 .i.xc3+! 24 bxc3 lLif6 2s "ifxgs+ "ifxgs 26 hxgs lLie4+ 27 �c2 was T.Roussel Roozmon­ K.Lahno, Montreal 2004, and now 27 ...lLixc3+! would give Black a clear advantage. b) 18 exfs! ? gxfs (Golubev suggests 18 ....i.xfs 19 f 4 °ifh6 20 fxes gs with unclear play) 19 f4 'inl6 20 o-o-o e4 21 �dgl �h8 22 �h2 gave White the ini­ tiative in N.Vitiugov-A.Shomoev, Ulan Ude 2009. 18 f4 °ifh8!

e22) 11 cs •••

A familiar pawn sacrifice. Again the omission of 9 h4 hS will change the assessment of certain positions. 12 dxc6 bxc6 13 lLixd4 exd4 14 .i.xd4 �e8

Black has had little success with moves like 14... cs and 14 ... "ifas. but 14... ds is another move order: l S cxds cxds 16 es (or 16 exds �e8 17 .i.e2) 16 ...lLid7 (instead 16 ... �e8 17 .i.e2 is Line B221 below, but here White has the extra option of 17 o-o-o! ? to reckon with as well) 17 f4 f6 and now both 18 e6 .l:.e8 19 .i.e2 and 18 exf6 �e8+ 19 .i.e2 transpose to the notes to Black's 18th in Line B221, but Black should probably avoid this transposition. 15 .i.e2 dS 16 cxds

Here it is best for White to exchange c-pawns immediately, as after 16 es lLid7 17 f4 Black has 17 ....i.h6! with the idea of meeting 18 "ife3?! With 18 ... cs! (Golubev).

19 f5

After 19 fxes lLid7! Black has good counterplay. 19...lLihs

Instead 19 ...bS!? with counterplay may be even better. 20 .i.xhs "ifxhs 21 f6 .i.h6 22 �d2 �h7

The position is balanced but sharp, O.Sanarov-P.Kotenko, Krasnodar 2001. 316

Black has the same choice as he did in the analogous position in Chapter 16.

Pa n n o Va ria t i o n : 9 lLic 1 18 ....Jl h61? .·· ..

·IUJ.: $6�.ocd.S a:22z:::t.��.:"4

8221) 16...cxds

In stead 18 .. .f6 tends to leave Black with a passive position: a) 19 e6 lLif8 20 fs gxfs may look ugly, but the position is more unclear than anything else. b) 19 o-o fxes 20 fxes lLixes 21 �adl! 'ifd6 22 �hl .llfs 23 b3 was slightly better for White in S.Atalik­ M.Golubev, Bucharest 1996. c) 19 exf6 lLixf6 20 .lies �b4 21 o-o 'ifb6+ 22 �h l �xb2 was J.Borges Mateos-A.Escobar Forero, Barcelona 20os, and now 23 'ifcl! would give White some advantage. 19 'ife3

17 e5

After 17 exds �b4 18 o-o (White should avoid 18 .lies �xb2! 19 'ifxb2 lLixds when Black is better after 20 �cl lLif4! or 20 o-o-o .Jlxc3 21 'ifxc3 lLixc3 22 l:l.xd8 �xd8 23 .Jlc4 .llfs with some initiative) 18 ...lLie4 (capturing on d4 is less effective) 19 fxe4 .Jlxd4+ 20 �h l .lies Black has enough compensation for the two pawns according to Golubev. 17 ... lLid7 18 f4

White can also play 18 o-o lLixes 19 �adl, but Black was not without coun­ terplay after 19 ... .Jle6 20 �hl (or 20 .Jlxa6 'ifas 21 .Jle2 �b4) 20 ...J:b4 in Lvov V.Vdovichenko-P.Kruglyakov, 2007. If 21 lLixds (White could instead repeat moves with 21 .lies �b8 22 .Jld4) 21 ...'ifxds 22 'ifxb4 lLic6 23 .Jlxg7 lLixb4 24 l:l.xds .Jlxds 2s .i.C3 �xe2 26 .Jlxb4 �xb2 the endgame is level.

Golubev stopped his analysis here. Indeed, Black will have to be resource­ ful to avoid a clearly worse position. 19...J:xb2 20 �dl

After 20 lLixds �xe2+! 21 'ifxe2 .llb 7 Black has definite counterplay. 20 ...lLifB! 21 .Jlf3 lLie6 22 lLixds lLixd4 2 3 'ifxd4 'ifas+ 24 �1 'ifbs+ 2 5 �g1 .Jlf8 26 l:l.cl �dB

With some active and precise play Black had obtained reasonable com317

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1

pensation in W.Brandhorst-A.Civitillo, correspondence 2007. The game was eventually drawn. 8222) 16 ...�b4

This is probably asking too much of the position because White's kingside has not been weakened with h2-h4. 11 .i.cs

Other moves are also possible: a) 17 .i.f2 cxds 18 �dl "ifas is simi­ lar to the related position after 9 h4 h S (note 'b' t o White's 18th move i n Line B12 of Chapter 16) and should be okay for Black. b) 17 .i.xf6 .i.xf6 18 o-o-o cxds 19 lLixds .i.xb2+ 20 �c2 l:l.b8 21 l:l.b1 .i.es 22 l:l.xb8 .i.xb8 gave Black good com­ pensation for the pawn in S.Simanjun­ tak-J.lvanov, Thessaloniki 2009. c) 17 es lLixds 18 lLixds .llxd4 19 "ifxd4 and here:

cl) 19 ... �xes 20 °ifb6 "if xds 21 �dl "ife6 22 l:l.d8+ .i.f8 23 o-o was T.Sammalvuo-J.Johansson, Turku 1997, and after 23 ...�g7 24 �xf8 �xf8 25 .i.xa6 .i.xa6 26 "ifxa6 �e2 Black is just 318

hoping for a draw according to Golubev. While this is true, Black is ac­ tive enough that he should be able to hold without too much difficulty. c2) 19 ... .i.xes isn't as good for Black as when 9 h4 hS has been thrown in, because he then has ... .i.g3+ available, but here Black's compensation still ap­ pears sufficient to hold: 20 °ifb6 "ifxds 21 J:d1 "ife6 22 o-o .i.f4 23 °ifh4 "ifes (23 ... °iff6 24 g3 l:l.xe2 25 "ifxf4 "ifxb2 26 °ifh6 °ifb6+ 27 �hl is somewhat better for White according to Golubev) 24 �h l "ifgs 25 g3 �xe2 26 "ifxf4 "ifxf4 27 gxf4 28 l:l.b1 .i.xa2 29 �fel l:l.d2 30 l:l.edl l:l.e2 3 1 .l:.el V2-V2 was 1.Sharpe­ J.Slaby, British League 2010. 17 ...lLixe4

Again Black relies on this trick to justify his play. 18 fxe4 .i.xc3 19 "if xc3 J:bxe4 20 o-o J:xe2 21 dxc6

This position should be better for White than it is in Chapter 16 (Line B12), because there is no pawn hang­ ing on h4. 21. "if gs? ..

P a n n o Va riatio n : 9 CZ'ic 1

This fails spectacularly, but Black is struggling in any case. Instead 21 ..."ifd5 22 .i.f2 �8e3 23 "ifxe3 J:xe3 24 .i.xe3 "ifxc6 25 l:l.acl °ifh7 26 �fdl .i.e6 is in­ sufficient according to Van der Tak. The best try looks to be 21 ...°ife7 ! ? 22 "iff3 (instead 22 .i.f2 .i.f5 with the idea of ... .lt.e4 looks okay, but 22 .i.d4!? would test Black's resources) 22 ... .i.b7! 23 cxb7 "ifxc5+ 24 �hl �2e7 2 5 �ael V2-V2 was A.Barreras Garcia-C.Blanco Gramajo, correspondence 2007. After 25 ... �xel 26 "ifxf7+ �h8 27 °iff6+ it's perpetual check.

clearly worse by now in any case, though.

22 l:l.f2

A fantastic, highly-creative shot. White threatens mate on g7.

This is stronger than 22 .i.f2 .l:.8e3 23 °ifc4 �e4 24 °if d3 J:.4e3 25 "ifc4 l:l.e4 26 °ifh3 (it is probably better to allow a repetition with 26 °ifd3 �4e3) 26 ... .i.e6 27 C7 .i.xb3 28 c8°if+ .l:.e8 29 °ifcl "ifxcl 30 .l:.fxcl .i.e6 and Black had a slight initiative in the ending in J.Ehlvest1.Smirin, Connecticut 2003. 22 ....i. h3? Consistent, but losing. Black is

23 .i.f8!!

23 ...�2es

No better are any of 23 ...J:xf8 24 J:xe2, 23 ...�xf8 24 'inl8+ �e7 25 �xe2+ or 23 .. .f6 24 °ifc4+ J:.2e6 25 c7 when White wins. 24 .i.d6! �e3 25 .i.f4! J:xc3 26 .i.xgs �cs 27 .i.f4 .i.cB 28 c7

White's advantage was decisive in J.Michenka-E.Hagara, Trinec 1998.

319

Cba,pter 18 Panno Variation "

,''

'

'

'

'

1 d4 tt:'lf6 2 c4 g6 3 tt:Jc3 �g7 4 e4 d6 5 f3 o-o 6 �e3 tt:Jc6 7 tt:Jge2 a6 8 'tlt'd2 l:tb8

Bl, while 10 �h6 bxc4 (10 ... �xh6 11 'tlt'xh 6 es 12 h4 is Line (1, below) 11 h4 tt:'lb4 12 tt:'lg 3 and now 12 ... �xh6! (this looks even stronger than 12 ... cs 13 �xg7 @xg7 14 dxcs �e6, which is also satisfactory) 13 'tlt'xh6 cs! 14 es (14 dxcs runs into 14 ... tt:Jxa2+! 1S tt:Jxa2 'tlt'as 16 @b1 c3) 14 ... tt:'lg4!! 1s fxg4 �6! 16 l:td2 cxd4 gives Black a strong attack

In this chapter we will look at White's attacking tries other than 9 h4. A; 9 .0·0•.0

8� 9 14 C: t khri A) 9 0-0-0 This move often transposes to other lines. 9 bs 10 h4 Instead 10 g4 is covered under Line ...

320

10 es ...

Instead 10...hs takes play back into the realm of Chapter 1s, while 10 ...tt:Jas 11 tt:Jg 3 (or 11 tt:Jf4 tt:Jxc4 12 �xc4 bxc4

P a n n o Varia tio n : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

13 hS tt:Jd 7!? with the idea of ...cs) 11...tt:Jxc4 12 Ji.xc4 bxc4 13 hS c6 sud­ denly reaches note 'b1' to Black's 10th in Line B of Chapter 14. Black could also consider 10 ... bxc4 with the idea of 11 hS tt:Jxh S !? (usually this move is bad, but there are a few exceptions) 12 g4 tllb 4!, which is interesting.

Even worse is 16 b3 Ji.xb3!. 16...Ji.b3! 17 hs Ji.xa2 18 'tlt'gs tt:'lb3+ 19 �C2 tt:Jd4+ 20 �Cl h6 21 'tlt'e3 l:th8 22 tt:Jd2 cs

Black had a strong attack in T.Meinert1.Gaponenko, Polanica Zdroj 2000. 8) 9 g4

11 d5

11 hS could also come about from the move order 9 h4 bS 10 h S es 11 o-o-o and might be met by 11... tt:Jxh S!? with the idea of 1 2 g4 exd4. 11 ...tt:Jas 12 tt:Jg3 b4 Black should avoid 12 ... hS?! 13 cxbs, but 12 ...Ji.d7 ! ? is possible. Then 13 h S? would be met with the thematic trick that occurs in the main line following 13 ...b4 14 tt:Jb1 Ji.a4 1s l:te1 Ji.b3!. 13 tt:'lbl Ji.d7

This looks preferable to both 13 ... c6 14 dxc6 and 13 ... hs 14 cs!. 14 ilh6?

Instead 14 cs would also be met by 14... Jla4!, but White should probably settle for 14 b3.

A sharp but committal move. White discourages ... h s and now tt:Je2-g3 will not block the g2-pawn. 9... bs Here White can castle or try to at­ tack with his king still in the centre. Bt: �00-o-Q 82:

14...Ji.a4! 15 Ji.xg7 �xg7 16 l:te1

10 h4

Other moves are dubious: a) 10 tt:Jg3? is a common mistake: 10 ... es 11 dS (the ugly 11 tt:Jce2 fares no better after 11 ... exd4 12 tt:Jxd4 tt:Jxd4 13 Ji.xd4 cs! 14 Ji.e3 bxc4 1S Ji.xc4 tt:Jxg4! 16 fxg4 l:txb2 and White's position is falling apart) 11...tt:Jd4 12 Ji.xd4 (both 12 Ji.g2 and 12 Ji.e2 are simply met by 11...bxc4, while 12 �2 bxc4 13 Ji.xc4 321

A ttacking C h e s s : Th e Kin g 's In dian, Vol u m e 1

�xg4! 14 fxg4 tt:Jxg4+ 1S �g2 l:txb2! is crushing) 12 ... exd4 13 tt:Jce2 d3 14 'tlt'xd3 bxc4 1S 'tlt'xc4 l:txb2 and Black is much better. b) 10 cxbS?! is not much better: 10 ... axbs 11 tt:Jg3 (or 11 h4 hS!, while 11 �h6 �xh6 12 'tlt'xh6 b4 13 tt:Jds es! gave Black a good game in R.Schlindwein-R.Schmaltz, Viemheim 199S) 11 ... es (11 ... b4 12 tt:Jce2 tt:Jas 13 b3 cs is also good) 12 dS tt:'ld4 13 �xd4 exd4 14 tt:'lxbs tt:Jxds! 1 s exds 'tlt'e8+ 16 �2 l:txbs 17 �xbs 'tlt'xbs gives Black excellent compensation for the ex­ change.

(again 14 tt:Jb1? �a4!) 14...bxc3 1S 'tlt'xc3 tt:Jxc6 (1S...tt:Jxg4 might be a bet­ ter chance to complicate matters) 16 dxc6 �e6 17 gs has occurred several times. White looks a little better to me, while 17 �c4!? also looks favourable to him.

Bl) 10 0-0-0 11 h4

This can also arise, of course, from the move order 9 o-o-o bS 10 g4. 10... bxc4

Worse is 10 ... es 11 dS tt:Jas 12 tt:'lg3 when White has reached the set-up he was aiming for: 12 ... �d7 (White is also better after 12 ...bxc4 13 h4 c6 14 g S! or 12 ... tt:Jxc4 13 �xc4 bxc4 14 h4 with the initiative) 13 cs (White must avoid 13 h4 b4 14 tt:'lb1 �a4!) 13 ...b4 14 c6! 322

This is the best move. Instead 11 tt:Jg3?! runs into 11 ...es! again, while other moves also do not impress: a) 11 gs ?! (now White will not be able to exchange of the g7-bishop) 11 ... tt:'ld7 12 h4 tt:'lb4 13 tt:'lg3 cs! is good for Black, as 14 dxcs? loses to 14 ... �xc3 (14 ... tt:Jes and 14 ... 'tlt'as also win) 1s bxc3 tt:'ld3+ 16 �xd3 'tlt'as! with a deci­ sive attack b) 11 �h6 could be met with the untried 11 ... �xh6!. After 12 'tlt'xh6 es 13 dS tt:Jas 14 tt:Jg3 'tlt'e7 1s h4 tt:'ld7 16 h s tt:Jcs, with ideas like ... g s, closing lines, as well as ...�d7 and ....l:tb4 Black has excellent counterplay. 11 ... hs

Instead 11 ...tt:'lb4 looks less trust­ worthy: 12 tt:'lg3 cs 13 h S (better than

Pa n n o Va riation: 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

13 dxcs tt:'ld7! ? 14 Ji.xc4 tt:Jes 1 5 Ji.e2 'tlt'as 16 �bl Ji.e6 when Black's attack is very strong) 13 ...cxd4 14 Ji.xd4 (White looks faster in this Dragon-like posi­ tion) 14...tt:Jc6 15 hf6 ! ? (15 Ji.e3 Ji.e6 is unclear) 1s ...Ji.xf6 16 hxg6 hxg6 and now 17 tt:Jfs! ? Ji.xfs 18 'tlt'h2 l:te8 19 gxfs gives White a strong attack.

placed here, but f3 is weak as well, so White must plough forward. 14 gxhs tt:Jxf3 15 'tlt'g2 tt:'ld4 1S ... Jlg4!? is also possible.

12 tt:'lg3

16...Jlg4 17 h6 Ji.h8 18 Ji.e2

16 Ji.gs?!

The pin is not dangerous. Better was 16 h6 Ji.h8 17 Jlxc4 tt:'lg4 with an un­ clear position.

White is probably better off with 12 Ji.h6, taking play into the variation 9 h4 hS 10 o-o-o es 11 Ji.h6 bxc4 12 g4 (Line C2 of Chapter 15), which is fairly unex­ plored and quite unclear.

18...l:txb2! 19 @xb2 'tlt'b8+ 20 �cl 'tlt'b4

By now Black had a winning attack in J.Philippy-J.Schuller, correspondence 1997. 12 ...es!

82) 10 h4

It is better to strike in the centre here than to play 12 ... hxg4 13 h S when White's attack looks very strong. 13 d5

Instead after 13 dxes tt:Jxes 14 gxhs (Black is clearly better after 14 Ji.e2 hxg4 15 f4 tt:Jf3) 14...tt:Jxf3 15 'tlt'g 2 tt:Jes 16 hxg6 fxg6 17 hS tt:Jfg4 Black was bet­ ter across the whole board in A.Dunne­ B.Hlavica, correspondence 2001. 13 ...tt:Jd4

Not only is the knight excellently

The theory of this line is mostly 323

A ttacking C h e s s : The King 's I n dian, Vol u m e 1

based on a 1992 game between Chris Ward and Joe Gallagher. Interestingly enough, both players subsequently wrote books on the Samisch Variation and annotated this game. 10...hs

White has the initiative after both 10 ... bxc4 11 h S es (or 11 ... tllb4 12 tllg 3) 12 dS tllb4 13 tll g 3 and 10...tll a s 11 tll g 3 tllxc4 12 �xc4 bxc4 13 hs. 11 tllg 3

Instead 11 o-o-o would transpose to Line Bl, while it is harder for White to attack the Black king after 11 gs tlld7. Instead after 11 gxh s tllx hs it is diffi­ cult to eliminate the h S-knight as mov­ ing the knight from e2 weakens d4. Finally, 11 �h6 transposes to note 'b2' to White's 10th move in Chapter 1S, which is not really sound for White.

14...tllf3+

The alternative 14 ... tllxc4 looks risky after 1s �xc4 bxc4 16 h S gxh s 17 tllxhs when he has tried: a) 17 ... tllxhs 18 l:txhs fs 19 o-o-o gives White the initiative. This has been played a couple of times, but it just looks too dangerous to me. b) 17 ...tll x e4 is an interesting try, but it should fall short: 18 tllxe4 l:txb2 19 tllhf6+! �xf6 20 tllxf6+ 'tlt'xf6 21 'tlt'd4 'tlt'xd4 (21 ... 'tlt'g7 22 'tlt'xg7+ �xg7 23 �d4+ �g6 24 �xb2 �b7 may be a slight improvement) 22 �xd4 l:te8+ 23 �fl f6 24 �xb2 l:te4 was C.Ward­ G.Buckley, Guildford 1991. White's ex­ tra rook should outweigh the pawns, especially after 2S l:th6 l:txf4+ 26 �g1. 15 �xf3 gxf3 16 hs tllg4 17 o-o-o

11...esl 12 dxes

Not 12 dS? tlld4. 12 ...tllxes 13 �e2 hxg4

White would have a free attack af­ ter 13 ...tllxc4 14 �xc4 bxc4 1S gxhs, as in A.Gupta-N.Vinuthna, Mumbai 2004. 14 f4

11... b4

This was suggested by both Galla­ gher and Ward in their respective books on the Samisch. There are a cou­ ple of other moves: a) 17 ...tllx e3 was played in the aforementioned game between the opposing experts. After 18 'tlt'xe3 b4 19 324

Pa n n o Va ria t i o n : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

tt:Jds c6 20 h6! Ji.f6? (instead 20...cxds 21 hxg7 �xg7 22 fS gives White a win­ ning attack; best is probably 20... Ji.xb2+ 21 @xb2 cxds, but even here White looks better) 21 tt:'lxf6+ 'tlt'xf6 22 fS aS there is: al) 23 tt:'lhs? is a bit overzealous: 23 ... gxhs 24 l:txd6 Ji.e6 (24 ... 'tlt'es 2s 'tlt'gs+ �h7 26 l:tf6 l:tg8 27 l:txf7+ �h8 28 l:tg7 intending 'tlt'g6 is dangerous for Black according to Gallagher) 2S 'tlt'xf3 (instead 2S fxe6 fxe6 gives Black coun­ terplay with his passed f-pawn) 2S ... �h7 26 'tlt'xhs l:tbd8 27 l:txd8 l:txd8 28 fxe6 fxe6 29 'tlt'h4 'tlt'd4 30 'tlt'e7+ V2-V2 was C.Ward-J.Gallagher, London 1992. a2) 23 l:tdf1 a4 24 l:txf3 a3 2s es! was what worried Gallagher. White will maintain some advantage, even in the endgame. a3) 23 h7+ �h8 24 'tlt'h6 'tlt'g7 2S 'tlt'xg7+ �xg7 26 l:txd6 with advantage for White is given by Ward, who admit­ tedly points out that Black should look to improve earlier. b) However, 17 ...bxc4!? has gone unmentioned.

It appears to be logical to me, as Black opens the b-file instead of chas­ ing White's knight into the centre of the board. Now 18 tt:Jds? l:txb2 is bad for White, and Black defends after both 18 Jld4 hd4 19 'tlt'xd4 'tlt'f6 and 18 hxg6 fxg6 19 fS 'tlt'f6. Another possibility is 18 l:tdf1 tt:Jxe3 19 'tlt'xe3 Jlg4 20 h6 Ji.f6 21 es!? dxes 22 tt:Jge4 exf4 (22 ...�h8!?) 23 'tlt'xf4 Ji.fs and Black appears to at least be holding his own in a sharp position (and he is up three pawns at the mo­ ment).

18 tt:Jds c6 19 h6 tt:Jxh6

Ward points out that after 19 ... Ji.xh6 20 fS tt:Jxe3 21 l:txh6 tt:'lg4 22 l:tdh 1 White has a strong attack. Black could instead try 20 ... Ji.g7, but after 21 f6 tt:Jxf6 22 Ji.h6 tt:'lg4 23 Ji.xg7 @xg7 24 tt:'le3 tt:'lxe3 2 S 'tlt'xe3 l:th8 26 es White has pretty good compensation for the pawns. 20 f5

This is Ward's idea. Instead Galla­ gher discusses some complicated lines starting with 20 tt:Jxb4 'tlt'as 21 tt:Jxc6 325

A ttacking C h e s s : The King 's I n di a n , Vo l u m e 1

'tlt'xa2 22 tt:'le7+ �h7 23 l:txh6+ �xh6 24 �d4, but Ward's suggestion looks like a better try. 20 tt:Jg4 21 f6 ...

(1: 9 .txh6 c2: 9 bs .•.

•••

21 ...�xf6

White has serious compensation af­ ter 21 ...tt:Jxf6 22 �h6 tt:Jg4 (or 22 ... tllxds 23 �xg7 �xg7 24 exds l:th8 25 'tlt'd4+ f6 26 l:txh8 'tlt'xh8 27 dxc6) 23 �xg7 �xg7 24 tt:Je3 because of the open h-file and pressure on the d6-pawn (Ward). 22 tllxf6+ 'tlt'xf6

Of course, not 22 ...tt:Jxf6?? 23 'tlt'h2. 2 3 �d4 tt:'les 24 cs l:tbs!?

Instead Ward points out that 24... dxcs? loses to 25 'tlt'h2 l:te8 26 'tlt'h7+ �8 27 �xcs+. 25 cxd6 l:td8

Black intends ....lii. g4 in this sharp position. C) 9 �h6

Very direct play, as White aims to eliminate the King's Indian bishop. Black can now invite the white queen into h6 or he can simply begin his own play on the queenside: 326

Black can also pl ay in the centre with 9 ... es and then : a) 10 �xg7 �xg7 11 dS tt:Jas 12 tt:Jc1 cs is satisfactory for Black. b) 10 dS is well met by 10 ... �xh6 (or 10 ... tllas) 11 'tlt'xh6, which transposes to the notes to White's 11th in Line C1. c) 10 o-o-o is probably White's best. Now both 10 ... �xh6 and 10 ... bs 11 h4 �xh6 12 'tlt'xh6 lead to Line C1. Cl) 9 .hh6 10 'tlt'xh6 •••

Pa n n o Va ria tion: 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

We have seen this idea before. Black hopes that the white queen's absence from the centre and queenside are more important than its ominous­ looking post near the black king. 10 ... es

worse for White is 13 cxbs axbs fol­ lowed by ... b4) 13 ...b4 (13 ...bxc4 14 h4 ild7 1S hS transposes to note 'b' to Black's 13th in Line B of Chapter 14) 14 tt:Jb1 and now the simplest is 14...tt:'lb7 with the idea of ... tt:Jcs (Watson).

Having exchanged the dark-squared bishops, Black logically places his pawns on dark squares. 11 0-0-0

White can also play 11 dS tt:'ld4 (or 11 ... tt:Jas 12 tt:'lg3 cs 13 h4 Ji.d7 14 hs bs which transposes to Line A of Chapter 14) and here: a) 12 tt:Jxd4 exd4 13 tt:Je2 cs is much already better for Black. b) 12 'tlt'd2 cs (or 12 ... tt:Jxe2+ 13 Ji.xe2 ild7 14 g4 bS with counterplay) 13 dxc6 bxc6 (13 ... tt:Jxc6! ? 14 cs Ji.e6 1 S cxd6 bS is a decent alternative) 14 tt:Jxd4 exd4 1s 'tlt'xd4 l:txb2 16 Ji.e2 tt:'lhs! 17 g3 'tlt'as 18 o-o Ji. h3 19 l:tfd1 .&1.c2 20 l:!.d3 l:!.b8 21 'tlt'xd6 1:!.bb2 gave Black the initiative in A.Lewis-J.Hall, British Championship, Morecambe 1981. c) 12 o-o-o cs 13 dxc6 bxc6 (13 ... tt:Jxc6! ? 14 l:td1 Ji.e6 1s b3 'tlt'as also looks quite comfortable for Black) 14 tt:Jxd4 exd4 1S 'tlt'xd4 l:txb2! and, compared to the position with 9 h4 and 9 ... hs inserted (Chapter 1S, note 'b3' to White's 10th move), Black does not need to fear 16 cs l:tb8 17 l:txd6 'tlt'as 18 @c2 Ji.e6! because the g6-pawn is se­ curely defended and l:txe6 is not effec­ tive here. 11... bs 12 h4

Instead 12 dS tt:Jas 13 tt:'lg3 (even

12 ... bxc4

Black has also often played 12 ... exd4 13 tt:Jxd4 tt:Jxd4 14 l:txd4 and here: a) 14...'tlt'e7 1S hS Ji.e6 16 hxg6 fxg6 17 tt:Jds (not 17 cxbs axbs 18 Ji.xbs Ji.xa2) 17 ... Ji.xds 18 cxds c6 19 dxc6 l:tbc8 20 @b1 l:txc6 21 l:td1 may leave White with a slight edge. b) 14... tt:Jhs 1s g4 tt:'lg3 16 l:tg1 tt:Jxf1 17 l:txf1 Ji.b7 (17 ... Ji.e6 looks better to me) 18 hS 'tlt'f6 19 'tlt'e3 bxc4 20 l:txc4 cs 21 l:td1 Ji.c6 is unclear - White's c4rook is oddly placed. c) 14... cs!? 1s l:td2 b4 16 tt:Jds tt:Jxds 17 l:txds Ji.e6 18 hs l:tb7! 19 l:td2 'tlt'e7 was solid for Black in M.Szymanski­ A.Mista, Trzebinia 2000. 13 hs 'tlt'e7 The position is similar to the varia­ tion 9 h4 bS (Line B of Chapter 14) and could even transpose. 327

A ttacking C h e s s : The King 's I n d i a n , Vo l u m e 1 14 g4

21 'tlt'es

Instead 14 dS tt:Jas 1s tt:'lg3 �d7 16 �e2 l:tb6!? (or 16 ...l:tb4, which is the 9 h4 bS main line) 17 l:td2 cs 18 tt:Jf1 (in­ stead 18 dxc6 l:txc6 19 tt:Jf1 �e6 20 tt:Je3 l::t c s is also unclear) 18...tt:Je8!? 19 tt:Je3 tt:'lc7 20 �d1 tt:Jbs 21 tt:'la4 l::t b 7, with a double-edged position, has occurred in a couple of games.

Black should not be tempted by 21 ... l:tbc8+ 22 �bl l:tc2 because 23 l:th1 l:te2 24 gs wins for White.

•••

14 exd4 15 tt:Jxd4 tt:Jxd4 16 l:txd4 �e6 •••

22 �b1

17 �XC4!

This is a very creative idea. White sacrifices a whole rook. The alternative 17 hxg6 fxg6 18 �xc4 �xc4 19 l:txc4 cs! leaves the rook looking funny on c4. 17 cs 18 hxg6 cxd4 •••

After 18 .. .fxg6 19 �xe6+ 'tlt'xe6 20 l:td2 White's rook was well placed and he had some advantage in H.Lahlum­ F.Elsness, Norwegian League 1998.

White could try 22 gS!. Then Black should avoid 22 ... 'tlt'f4+ 23 �bl 'tlt'xf3 24 a3 ! when White wins and instead play 22 ... 'tlt'd4 23 gxf7+ l:txf7 24 �xf7+ �xf7 2S 'il¥xf6+ 'tlt'xf6 26 gxf6 �xf6, but after 27 l:td2! he still has to work to hold the game. 22 d2 23 l:txd2 hxg6 24 'tlt'xg6+ �h8 25 l:th2+ 'tlt'xh2 26 'tlt'xf6+ •••

Here a draw was agreed in Yaplian­ A.Kuzmin, USSR 1982. C2) 9... bS

This move is a bit riskier than 9 ... �xh6.

19 tt:'lds �xds 20 �xds d 3

10 h4

This move is necessary. Now Black threatens 21 ... 'tlt'C7.

Instead 10 o-o-o bxc4! was consid­ ered in the notes to White's 10th in Line A. A different kind of position is reached after 10 �xg7 �xg7 11 cxbs (11 0-0-0 bxc4 12 g4 es! looks good for

21 l::t h 2

Instead 21 @b1 'tlt'es 22 b3 'tlt'c3 leads to a draw after 23 l:!.c1 d2 24 l::t x c3 d1'tlt'+ or 23 gs 'tlt'c2+ 24 �a1 'tlt'c3+. 328

Pa n n o Va riation : 9 0 - 0 - 0, 9 il h 6 a n d 9 g4

Black) 11...axbs 12 dS and now: a) 12 ... tll e s 13 tlld4 b4 14 tllcbs e6 (after 14 ... Ji.d7 1S 'tlt'xb4 e6 Black has good play, but White could consider 1S .l:tc1 ! ?) 1 S tllc6 tllxc6 16 dxc6 es 17 'tlt'xb4 tllh s and Black has counterplay brewing on the kingside. Then 18 o-o-o Ji.a6 with the idea of ...tllf4-e6 gives Black compensation for the pawn. b) 12 ...tlla s 13 tlld4 b4 14 tllcbs es 1S 'tlt'xb4 (worse was 1S tllb3 tllxb3 16 axb3 Ji.d7 17 'tlt'xb4 tllh s! 18 g3 'tlt'gs with a big advantage for Black in T.Sohal-Zhe Quan, Edmonton 2009) 1s ... cs 16 dxc6 exd4 17 .l:tc1 'tlt'b6 18 c7 was F.Ludvigsen-K.Trygstad, Norwegian League 1993. Now after 18 ... .l:tb7 it is not clear that White has enough for t he piece. 10 es ...

Instead 10...Ji.xh6 11 'tlt'xh6 trans­ poses to Line C1 after both 11 ... es 12 o-o-o and 11 ...bxc4 12 o-o-o es 13 hs. Another move order which Black can use is 10 ... bxc4!?, although he should be careful after 11 Jlxg7 (11 hS tllb4! is note 'a3' to Black's 10th move in Line B of Chapter 14, while 11 o-o-o tllb4 was covered in the notes to White's 10th in Line A of this chapter) 11 ... �xg7 12 hS es 13 tllds and now: a) 13 ... �h8 leads to the main line, below. b) 13 ... tll g 8 has also been played a few times, but this looks too passive. c) 13 ...tllxds 14 exds (14 hxg6 is not sound after 14...tllf6 or 14...tllf4) 14... tllb4 transposes to note 'b' to

Black's 12th, below. d) 13 ...Ji.e6 14 tllxf6 'tlt'xf6 1S dS tllb4 16 tll c3 intending Ji.xc4 looks much better for White. 11 Ji.xg7 �xg7 12 hs

12 �hB!? ...

This leads to some spectacular complications and was played in a well-known game. Alternatively: a) 12 ...tllxd4 13 tllxd4 exd4 and now 14 'tlt'xd4 gives White some advantage (but 14 tllds cs, as in A.Ker-B.Spassky, Wellington 1988, is less clear). b) 12 ... bxc4 13 tlld s! tll x ds 14 exds tllb4 1S tllc3 (not 1S hxg6? tlld 3+) 1S ...exd4 16 hxg6 (or 16 'tlt'xd4+ �g8 17 o-o-o Ji.fs) 16 ... 'tlt'e7+ 17 tlle4 fxg6 18 Jlxc4 and White is better because of the open h-file and Black's unstable b4knight. c) 12 ...tllxh S !? may actually be Black's best: 13 g4 exd4 (not 13 ...tllf6 14 'tlt'h6+ followed by 1s tlld s) 14 tllxd4 (or 14 gxh s dxc3 1s tllxc3 gs with unclear play) 14 ... tllxd4 1S 'tlt'xd4+ 'tlt'f6 16 'tlt'xf6+ tllxf6 17 cxbs axbs 18 Ji.xbs ild7 is level. 329

A ttacking C h e s s : Th e King 's Indian, Vo l u m e 1 13 tt:Jds bxc4 14 hxg6 fxg6 15 'tlt'h6 After 15 o-o-o tt:Jxd5 16 exd5 Black

has the clever 16 ... c3! 17 bxc3 (worse are 17 tt:Jxc3 tt:Jxd4 and 17 'tlt'xc3 tt:'lb4) 17 ... tt:'le7 when he has good chances. 1s ...tt::i h s

This is forced. Instead 15 ... .&!.f7? 16 'tlt'xg6 'tlt'g8 loses to 17 'tlt'xf6+!. 16 g4

After 16 o-o-o l:tf7 17 g4 tt:'lf6 18 'tlt'xg6 'tlt'g8 19 'tlt'xg8+ (now 19 'tlt'xf6+? l:txf6 20 tt:Jxf6 fails to 20 ...'tlt'g5+) 19 ... tt:Jxg8 the ending is equal accord­ ing to Gufeld, but 20 �g2 ! ? may leave White a little bit better. 16 ... l:txb2 17 gxhs

Black next two moves are forced he must keep lines closed.

ning in Bowers-Frankema, correspon­ dence 1988. b2) 20...l:txf3? 21 tt:Jef4 l:ta1+ 22 �b2 l:txdl 23 tt:'lg6+ �g 8 24 �xc4+ wins. b3) 20 ... tt:Jxe5 21 tt:Jef4 �g8 22 tt:Jg6 hxg6 23 hxg6 (or 23 �xc4 tt:Jxc4 24 'tlt'xg6+ and White delivers perpetual) 23 ... 'tlt'd7 ! 24 l:th 1 l:tc2+! with impend­ ing perpetual check 20 ... exf4 21 tt:Jxf4?

White could have draw with 21 �xc4 l:ta1+ 22 �b2 l:txd1 23 l:txd1 l:tg 8! 24 tt:'lf6 l:tg7 25 �g8 ! 'tlt'e7 26 �xh7 l:txh7 27 tt:Jxh7 'tlt'xh 7 28 'tlt'f8+ and again it's perpetual check 21 ...l:txf4 22 'tlt'xf4 C3! 23 �C4 J:ta3! 24 fxg4 tt:'lb4 2 5 �bl

11 ... gs 1s l:tg1 g4 19 o-o-o

Instead 19 dxe5 tt:Jxe5 20 o-o-o l:txa2 21 tt:Jef4 �g8 22 tt:Jg 6 hxg6 23 hxg6 'tlt'd7 ! led to a draw after 24 l:thl l:tal+ 25 @b2 �5+ 26 �xal 'tlt'a4+ 27 �b2 'tlt'b3+ 28 �cl 'tlt'a3+ 29 �bl in Meyer­ G.Antoszkiewicz, correspondence 1984. 19 .. Axa2 20 tt:Jef4

White has to find a way to continue his own attack others: a) 20 �h3 l:txe2 21 �xg4 l:tf7 22 �xc8 'tlt'xc8 23 tt:'lf6 �8 24 l:tg8+ 'tlt'xg8 25 tt:Jxg8 tt:'lb4 26 l:td2 l:te1+ 27 l:td1 (not 27 @b2 l:txf3) 27 ... l:te2 with a repeti­ tion. b) 20 dxe 5!? and here: bl) 20...�d7? 21 tt:Jef4 l:txf4 (Black is also not saved after 21 ... l:tal+ 22 �b2 l:txdl 23 tt:'lg6+ �g8 24 tt:Jde7+ tt:Jxe7 25 �xc4+) 22 tt:Jxf4 and White was win330

2s ...�e6!

A fantastic clearance move. Black's queen will come to b8 with devastating effect. 26 �xe6 tt:Jd3 27 'tlt'f7 'tlt'b8+ 28 �b3 l:txb3+ 29 �c2 tt:'lb4+ 30 �xb3 tt:Jds+ 31 @c2 'tlt'b2+ 32 �d3 'tlt'bs+

Here White resigned in V .Bagirov­ E.Gufeld, USSR 1973, in view of 33 �c2 'tlt'e2+ 34 @b3 �2+ 35 @c4 �5 mate.

Chapter 19 Pa nno Variation 9 a3, 9 I:.b1 and 9 I:.c1

1 d4 tt:'lf6 2 c4 g6 3 tt:'lc3 �g7 4 e4 d6 5 f3 o-o 6 �e3 tt:Jc6 7 tt:Jge2 a6 8 'tlt'd2 1:1.bB

A: 9 a3 8: 9 nb1 C: 9 .:.c1

The first two variations are very similar, but there are some small dif­ ferences. Line C is currently quite fash­ ionable and gives both sides a fair bit of scope for creative play. A) 9 a3

In this chapter we examine plans where White plays to gain space on the queenside. Even though he is slow to get his kingside developed, White has a lot of extra space and the position is not exactly wide open. Still, White should take some care because if the position blows up, he may find his lack of development is indeed an issue. Black must play with intent, as even though White's play is slow, it's not always so easy to create counterplay.

White simply wants to expand on the queenside with b2-b4. The down331

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

side to this move is that it weaken s b3, which in turn can leave c4 vulnerable. 9 �d7 Black makes a useful waiting move before committing to a queenside ad­ vance. There are several other option s to consider as well: a) 9 ...bs is also common and can transpose to the main line, but there are a couple of independent possibili­ ties after 10 cxbs axbs 11 dS (or 11 b4 �d7 12 dS tt:Jes 13 tt:'ld4 which trans­ poses to the main line) and here: al) 11 ... tt:Jes 12 tt:'ld4 �d7 and now in practice White always plays 13 b4, which transposes to the main line. However, he could consider grabbing the bS-pawn at once when I do not see anything special for Black. By playing 9 ...�d7 Black avoids this possibility. a2) 11 ...tt:Jas is an alternative. It does not have a very good reputation, but it is not completely clear. After 12 tt:'ld4 es (Black has also tried 1 2...�d7 1 3 b4 cs 14 dxc6 tt:Jxc6 1S tt:Jdxbs tt:Jes, but this looks too speculative) 13 tt:'lc6 tt:Jxc6 14 dxc6 �e6 1S tt:Jxbs dS 16 �a7 Black has: ...

332

a21) 1 6...l:ta8 1 7 �cs l:te8 1 8 �e2 'tlt'b8 19 �4 did not give Black enough compensation in K.Schmitzer-H.Alber, Doernigheim 1994. a22) 16 ... l:tc8 17 �cs l:te8 (after 17 ... dxe4 18 'tlt'xd8 l:tfxd8 19 �e7 l:te8 20 �xf6 �xf6 21 fxe4 White had the better ending in M.Dlugy-J.Howell, Sharjah 198S) 18 tt:Jc3 d4 19 tt:Jd1 and now, instead of the overzealous 19 ... �dS?!, Black could try 19 ...l:tb8 or 19 ... tt:Jhs with some play for the pawn. b) 9 ... es 10 dS tt:'le7 (it is hard to be­ lieve that the move 9 a3 hurts White enough for this plan to suddenly be good) and now both 11 b4 and 11 g4 look pleasant for White, while practice has also seen: b1) 11 tt:'lg3 tt:'ld7 12 �d3 fs 13 exfs tt:Jxfs 14 tt:Jxfs gxfs 1S o-o-o (1s o-o is probably better) 1S ...tt:Jcs 16 �c2 bS and Black had decent counterplay in N.Situru-E.Gufeld, Honolulu 1996. b2) 11 o-o-o tt:'ld7 12 �bl fS 13 �1 tt:'lf6 14 h3 f4 1S �a7 l:ta8 16 �f2 tt:'lhs 17 �d3 gs 18 cs �d7 19 tt:Jb3 gave White some initiative in L.Zsinka­ S.Cigan, Austrian League 199S, as his queenside play is more dangerous than anything Black has going on the other side of the board. c) 9 ... l:te8 is probably not the most efficient use of a tempo. After 10 b4 bS 11 cxbs axbs 12 ds tt:Jes 13 tt:Jd4 �d7 both 14 �e2 and the immediate 14 tt:Jdxbs have scored well for White. d) 9 ...e6 is a typical flexible move tn these variations. Here the idea is little

Pan n o Va ria tion: 9 a 3, 9 l:! b 1 a n d 9 l:!c 1

explored, although Black was doing fine after 10 b4 �e7 11 l:!dl bS 12 cxbs axbs 13 tt:Jc1 es 14 dS tt:'ld4 1S tt:J1e2 tt:Jxe2 16 �xe2 �d 7 17 o-o tt:'lh S 18 l:!c1 fS in J.Brenninkmeijer-R.Douven, Wijk aan Zee 1988. e) 9 ...tt:'ld7!? is a very interesting idea. This flexible move gives Black sev­ eral possibilities: ...tt:'lb6 will attack the c4-pawn and ...tt:Jas becomes feasible because Black can now follow up with ... cs.

White has tried a few things here: el) 10 �h6?! es! 11 �xg7 �xg7 12 dS tt:'le7 13 g4 c6 14 dxc6? tt:Jxc6 was already better for Black in N. S ituru­ V.Tkachiev, Jakarta 1996. If 1S �xd6 then 1S ... �h4+ 16 tt:Jg3 l:!d8 gives Black excellent compensation for the pawn. e2) 10 h4!? is logical because Black's knight has left his king's defence, but after 10...tt:Jas (10...bs is also possible) 11 tt:Jc1 cs Black has good counterplay according to Golubev. e3) 10 l:!d1 ! ? tt:Jas 11 tt:Jc1 cs?! 12 dxcs dxcs 13 �xcs did not give Black enough for the pawn in D.Vigorito-

R.Bumett, U SA 1994, so the flexible 10 ... e6!? could be considered. e4) 10 b4 is the consistent move, but after 10 ... tt:'lb6 the attack on the c4pawn is annoying: e41) If White protects it with the natural 11 tt:Jc1, then 11 ... es exploits the weakened d4-square because after 12 dS (12 tt:Jb3 exd4 13 tt:Jxd4 tt:Jes is good for Black as 14 �a2? just loses to 14 ... cs) 12 ... tt:'ld4 White's lack of control over the b3-square is telling . e42) 11 �d3 could be met with 11...e6!? intending ...dS!. Instead 11 ...�e6 is well met by 12 tt:Jds, but Golubev has suggested 11 .. .fs. e43) White, however, has another way to cover the c4-pawn with 11 �a2.

This looks awkward, but White just wants to consolidate his space advan­ tage: 11...e6 12 cs (12 l:!d1 dS! gives Black good play) 12 ...tt:'ld7 (Black could consider 12 ... dxcs!? with the idea of 13 bxcs tt:'ld7 intending ...b6, but 13 dxcs is also possible, because the tempting 13 ...tt:'lc4 14 �xc4 tt:'les 1S �a2 tt:'ld3+ 16 �dl tt:Jf2+ 17 @cl tt:Jxhl looks too risky 333

A t tacking C h e s s : Th e King 's Indian, Vo l u m e 1

as the hi-knight cannot escape; here 13 ... ttJd7 is more sensible for Black) 13 .l:td1 dxcs (worse is 13 ...b6 ?! 14 cxd6 cxd6 15 dS!) 14 bxcs (14 dxcs �e7 is unclear) 14 ... �e7 (Black could also con­ sider the immediate 14 ...b6!? because 15 dS exds 16 exds tt:Jces 17 c6 tt:Jcs is very strong, while 15 cxb6 tt:Jxb6 looks okay) 15 es (15 dS tt:Jces 16 d6 cxd6 17 cxd6 �d8 15 �c4 .l:td8 16 'itf2 tt:Jas 17 �b4 tt:'lc6 18 �c4 �f8 was unclear in J.Sriram-V.Saravanan, Sangli 2000) 1s ... b6 16 cxb6 tt:Jxb6 17 tt:'lg3 .l:td8 18 �e2 �f8 ! and Black had a slight initia­ tive in B.Alterman-E.Sutovsky, Haifa 1996. Now we go back to 9 ...�d7.

10 b4 bs

Black can still try to play without this move: a) 10 ... .l:te8 11 J::td1 and now 11 ... �c8 and 11 ... e6 are both possible, while 11 ...bS 12 cxbs axbs 13 ds tt:Jes 14 tt:'ld4 e6 15 dxe6 fxe6 16 tt:Jdxbs dS was a speculative attempt in E.Godin­ C.Chase, Somerville 1996. b) 10...es (Black switches plans now 334

that White has committed his queen­ side) 11 dS tt:'le7 and then: bl) 12 g4 c6 13 tt:Jc1 bS gave Black counterplay in A.Makarov-V.Aleshnya correspondence 2002. b2) 12 tt:Jc1 tt:'lhs 13 tt:'lb3 fs 14 o-o-o has been played a few times. Now 14 ... b6!? discourages White's cs­ advance and looks reasonable for Black. b3) 12 tt:'lg3 tt:Je8 (12 ... h s !?) 13 �d3 c6 14 o-o cxds was L.Zsinka-J.Nunn, Lon­ don 1987, and now 15 cxds should leave White with an edge. 11 cxbs axbs 12 ds tt:Jes 13 tt:'ld4

This is the structure that White is generally aiming for after both 9 a3 and 9 l::t b 1. The bS-pawn is a perma­ nent weakness, so Black must try to get counterplay by breaking in the centre with either his c- or e-pawn. 13 ...�eB

Black protects bS. Instead the im­ mediate 13 ... e6 is sometimes played when White has: a) 14 dxe6 fxe6 15 tt:Jdxbs (White should avoid 15 �xbs �xbs when both 16 tt:Jcxbs tt:Jxe4! 17 fxe4 �h4+ 18 �d1

Pan n o Va riatio n : 9 a3, 9 1:! b 1 a n d 9 1:! c 1

cs and 16 tt:Jdxbs tt:Jxe4! 17 fxe4 'tlt'h4+ 18 @d1 l:txbs! bring him trouble) 1S ... tt:Jxe4 16 fxe4 �xbs (or 16 ... l:txfl+ 17 l:!xf1 �xbs) 17 tt:Jxbs (even worse is 17 �xbs l:!xbs 18 tt:Jxbs 'tlt'h4+ 19 �d1 tt:Jc4 20 'tlt'e2 tt:Jxe3+ 21 'tlt'xe3 �xa1) 17 ...l:txfl+ 18 l:txf1 tt:Jc4 19 'tlt'e2 tt:Jxe3 20 'tlt'xe3 �xa1 and Black has excellent play. b) 14 �e2 ! is similar to the main line and scores very well for White. In­ deed, after 14...exds 1 s exds l:te8 16 o-o tt:Jc4 (16 ...'tlt'c8 17 tt:Jdxbs is also in­ sufficient) 17 �xc4 bxc4 18 bS 'tlt'c8 19 a4 he is certainly better.

16 ... �d7 when White is still probably a little better, but Black is in the game and can try to play ... e6 and ... ds or .. ltb7, ...'tlt'b8 and ...l:!c8. 15 o-o exds

Alternatively, 1s ...tt:Jc4 16 �xc4 bxc4 17 l:tfe1 es 18 tt:Jde2 'tlt'd8 19 'tlt'a2! was also insufficient for Black in M.Dlugy­ J.Fedorowicz, Sth matchgame, New York 1984. 16 exds tt:Jc4 17 �xc4 bxc4 18 l:tfe1

Black is in difficulties - he lacks counterplay and the c4-pawn is weak. White may also continue with bS and a4, with a grip on the queenside.

14 �e2 B) 9 l:tb1

14...e6

Instead 14 ... tt:'lc4 1S �xc4 bxc4 16 o-o es 17 tt:Jde2 tt:'lhs was B.Alterman­

M.Hennigan, Santiago 1990. Here White could just go after the c4-pawn with the typical idea of 18 'tlt'a2! when Black probably does not have enough counterplay. Black's best is likely 14... c6!? 1S dxc6 �xc6 16 o-o (or 16 tt:Jxc6 'tlt'xc6 17 o-o tt:Jc4 when Black looks to be okay)

This little rook move is similar con­ ceptually to 9 a3. White simply intends to play b4. There are a couple of differ­ ences though. By playing 9 l:tb1 White is no longer able to castle queenside. Also, the rook will be tied to the de­ fence of the b4-pawn in most positions, so bringing it to a more active square such as cl or d1 is less likely. Another odd little point is that White's queen 335

A ttacking Ch ess: Th e King 's I n dian, Vo l u m e 1

lacks the a2-square - we saw in Line A that White can often attack a pawn on c4 with 'tlt'd2-a2. On the plus side, White has avoided weakening the b3square, so playing tt:Je2-c1 becomes possible, because after the sequence ... es; dS tt:'ld4 White can challenge Black with tt:'lb3. All in all, it is difficult to say which of 9 a3 and 9 l:tb1 is 'better', but I will try to point out the subtle differences in the play that arises. 9 �d7 Again, there are several alterna­ tives: a) 9 ... es perhaps makes a little more sense here than in Line A, as White cannot castle queen side, but 10 dS tt:'le7 11 tt:'lg3 (or even 11 g4) should still fa­ vour White. b) 9 ... e6!? is rare, but rather interest­ ing. ...

Then 10 b4 tt:'ld7 11 tt:Jc1 (probably not the best) 11 ...'tlt'h4+!? (this looks better than 11 .. .fs, which has been seen a couple of times in practice) 12 �f2 'tlt'f6! shows the downside to 9 .l:tb1 336

the b4-pawn needs to be tended to. After 13 tt:J1e2 (not 13 tt:Jb3 tt:'lxb4) 13 ...tt:'lb6 Black has good play. c) 9 ... tt:'ld7 is possible here as well: 10 h4! ? (after 10 b4 tt:'lb6 11 tt:Jc1 Black could play 11 ...e6!? or 11 ...eS 12 dS tt:'ld4 13 tt:'lb3 tt:Jxb3 14 .l:txb3 fS with decent play) 10 ... hs (worse is 10 ... es 11 dS tt:'le7 12 hS fS. while 10 ...tt:Jas makes less sense here because b3 is not weakened; White might continue, for instance, with 11 tt:Jc1 cs 12 dxcs dxcs 13 h S!) 11 �gs ! ? �f6 12 f4 (this is a rather strange move) 12 ... es (Golubev sug­ gests 12 ... �g7 ! ? 13 fs tt:'lb6) 13 tt:Jds �xg s 14 hxgs tt:Jxd4 1s tt:Jxd4 exd4 16 g4 c6 17 gxhs cxds 18 hxg6 fxg6 19 'tlt'h2 �7 20 'tlt'h7+ @e8 21 'tlt'xg6+ .l:tf7 was E.Ghaem Maghami-A.Volokitin, Lausanne 2001. Now 22 'tlt'xd6 looks best: for example, 22 ... 'tlt'e7 23 .l:th8+ l:tf8 24 .l:txf8+ @xf8 2S 'tlt'xe7+ @xe7 26 cxds tt:Jcs 27 es and here White is bet­ ter. d) 9 ... bs is the most direct. After 10 cxbs axbs White has: d1) 11 dS tt:Jas (or even 11 ...tt:Jes 12 tt:'ld4 b4 13 tt:Jcbs e6!?) 12 tt:'ld4 b4 13 tt:Jcbs es 14 tt:Jc6 (not 14 dxe6 fxe6 1S 'tlt'xb4? cs) 14...tt:Jxc6 1S dxc6 when Black can simply play the straightfor­ ward 1s ... ds or even 1S ...tt:Jxe4!? 16 fxe4 'tlt'h4+ 17 @d1 'tlt'xe4, with sharp play that does not look unfavourable for him. d2) 11 b4 and now 11 ...�d7 trans­ poses to the main line, but Black often prefers 11 ...es!? 12 dS tt:'le7.

Pa n n o Va r i a t i o n : 9 a 3 , 9 l:! b 1 a n d 9 l:!c1

Here White has: d21) 13 tt:'lg3 �d7 and now: d211) 14 �d3 c6 1S dxc6 �xc6 16 o-o dS 17 �cs d4 18 tt:Jce2 tt:'ld7 19 �d6 l:!c8 20 f4 was A.Dreev-Z.Efimenko, Moscow 2002. Here Dreev considers 20... hs to be unclear, while Hazai's suggestion of 20 ...exf4 21 'tlt'xf4 fs 22 exfs tt:Jes! 23 �xes 'tlt'ds is interesting as well. d212) 14 �e2 c6 1S dxc6 �xc6 16 o-o dS 17 �cs (or 17 exds tt:Jexds 18 �cs l:!e8 with unclear play in 1.Khenkin-J.Balcerak, German League 2001) 17 ... d4 (17 ... l:!e8! ?) 18 tt:Jd1 tt:'ld7 (worse is 18 ... tt:Je8, as in S.Savchenko­ A.Fedorov, St Petersburg 1996) 19 �d6 l:!c8 20 tt:Jb2 tt:'lb8! 21 �cs tt:Ja6 and Black's impending possession of the bishop-pair gave him good chances in S.Jenkinson-C.McKenzie, correspondence 2001. d22) 13 tt:Jc1 is a favourite of Haba's: 13 ... �d7 14 tt:Jb3 tt:'lh s (Black could also consider 14... c6 1S dxc6 �xc6 16 tt:Jas �a8 with the idea of ... dS) 1s �d3 (White does well to get himself castled;

instead 1s tt:Jas fs 16 �d3 fxe4 17 fxe4 c6 left him with an uncomfortably­ placed king in P.Haba-P.Bobras, Koszalin 1999) 1s ... tt:Jf4 (Golubev sug­ gests 1s .. .fs 16 o-o f4 17 �a7 l:tb7 18 �f2 gs 19 tt:Jas l:!b8 when Black may lose the bS-pawn, but he has some kingside counterplay) 16 o-o tt:Jxd3 17 'tlt'xd3 fs 18 tt:Jas f4 19 �a7 l:!a8 20 �f2 has been reached a couple of times with by Haba. He won both games, but was considerably higher rated in each.

White has made some progress on the queenside, but Black has the bishop-pair and potential kingside play: d221) 20 ... .rf.b8 21 a4 bxa4 22 tt:Jc6 tt:Jxc6 23 dxc6 �e6 24 tt:Jds l:!f7 2s bS gave White the initiative in P.Haba­ F.Voelzke, Kiel 2001. d222) 20...'tlt'e8 21 l:!fc1 gs 22 l:!b2 �f6 23 tt:Jb1 was P.Haba-M.Raubal, Loosdorf 1993, and now 23 ...\ib8 24 l:!bc2 l:!c8 would be unclear. e) 9 ...l:!e8 is more sly here than it was in Line A because White cannot castle long. After 10 b4 Black has: 337

A t tacking C h e s s : The King 's In dian, Vo l u m e 1

el) 10... �d7 is rather vague. White could play 11 g3, 11 tt:Jc1, 11 a4 or even 11 bS!?. e2) 10 ... es 11 dS tt:'ld4 (11 ... tt:Je7 is better, but then ... .rf.e8 looks silly) does not work after 12 tt:'lxd4 exd4 13 �xd4 tt:Jxe4? 14 fxe4 l::txe4+ 1S tt:Jxe4 'lli'h4+ 16 tt:Jf2 �xd4 17 �d3, winning. e3) 10...bs leads to play similar to what we have already seen. White is bet­ ter after 11 cxbs axbs 12 dS tt:Jes 13 tt:Jd4 �d7 14 tt:Jcxbs e6 1S dxe6 fxe6 16 �e2. e4) 10 ... tt:Jd7 !? sees Black aim for similar play to 9 a3 tt:'ld7 (note 'e' to Black's 9th move in Line A).

White has: e41) 11 cs?! dxcs (11 ...bs!?) 12 bxcs (even worse is 12 dxcs tt:'ldeS) 12 ...b6 13 dS tt:Jas 14 tt:Jf4 was Zhang Zong­ Y.Kruppa, Moscow 2004. Now 14...tt:Jes with the idea of 1S .l:tb4 bS, controlling the c4-square, is pleasant for Black. e42) 11 �gs was suggested by Golubev. Here 11 ... tt:Jb6 12 'tlt'd3 (not 12 tt:Jc1? tt:Jxd4) 12 ... h6!? 13 �h4 (or 13 �e3 e6!) 13 ... as!? 14 a3 (14 bS? tt:'lb4) 14...axb4 1S axb4 .l:ta8 is unclear. 338

e43) 1 1 tt:Jc1 e s 1 2 dS tt:'ld4 1 3 tt:'lb3 tt:'lxb3 (alternatives are 13 .. .fs and 13 ... 'tlt'M+ 14 �f2 'tlt'f6 when Black is supporting d4) 14 l::txb3 fS and al­ though the e8-rook is misplaced, Black has a fairly normal and playable posi­ tion. e44) 11 a3 tt:'lb6 12 tt:Jc1 es (in stead 12 ... e6 13 cs tt:'ld7 14 tt:'lb3 b6 1S cxd6 cxd6 16 bS was a little better for White in V.Malakhatko-J.Radovanovic, Port Erin 2oos) 13 dS tt:'ld4 14 tt:'lb3 tt:'lxb3 (again 14.. .fs and 14 ... 'lli'h 4+! ? 1s �f2 'tlt'f6 are alternatives) 1S .l:txb3 fS is similar to variation 'e43'. e4S) 11 h4!? tt:'lb6 (Black could con­ sider 11 ... hs, but 11 ... es 12 dS tt:Je7 13 hS fS 14 hxg6 hxg6 1s �h6 tt:'lf6 16 exfs was a little shaky in J.Santos-E.Agdes­ tein, Thessaloniki Olympiad 1984) 12 tt:Jc1 (12 tt:'lg3 could be met with 12 ... hs or 12 ... es 13 ds tt:'ld4 14 hs 'tlt'e7) 12 ... es 13 dS tt:'ld4 14 tt:'lb3 probably favours White somewhat, e.g. 14... tt:'lxb3 1S l::t x b3 hS 16 �e2 and the position is similar to some we saw in Chapter 16. Finally, we return to 9...�d7:

Pan n o Varia ti o n : 9 a 3 , 9 :t:! b 1 a n d 9 :t:! c 1 10 b4 b s

After 10...es 11 dS CfJ e 7 both 1 2 CfJg3 and 12 g4 look good for White. 11 cxbs axbs

Black should be very conscious of the type of passive positions that we saw in the main line of 'A' and so try to at least cause White some problems. 12 ds CiJes 13 CiJd4 'tlt'e8 After 13 ... e6 14 �e2 (14 dxe6 fxe6 15 CiJdxbs CiJxe4 is less good than the analogous operation in Line A because here there is no rook hanging on al!) 14...exds 15 exds .rf.e8 16 o-o White has a familiar advantage.

b) In Line A the position after 18 CiJde2 was good for White. The only difference here is that White has played .l:tb1 instead of a3, but now he does not have 'tlt'a2 at least. Here 18 ...CfJh S 19 g4 CfJf6 20 CfJg3 is still bet­ ter for White according to Golubev, so perhaps Black could try 19 ...CfJf4!? 20 CiJxf4 exf4 21 �xf4 .l:ta8 with counter­ play; his pieces are active and White's king position has been loosened. C) 9 l:tc1

14 �e2

14 ...e6

Black should consider too 14 ... cs!? 1 5 dxc6 �xc6 16 o-o �d7. 15 0-0 CfJc4 16 �XC4 bXC4 17 .l:tfel e5

White has a choice: a) 18 CiJc2 CfJhS 19 CiJa3 fS 20 exfs �xfs 21 CfJe4 was 5 .Savchenko­ M.Golubev, Alushta 1995. Now Golubev gives 21 ...CfJf6! 22 �gs (22 CiJxf6+ �xf6 23 .l:tbc1 'tlt'a4 is good for Black) 22 ... C3! with good play for Black.

This line is very fashionable. White does not initiate any action just yet, but he does prevent 9 ...bs? because 10 cxbs axbs 11 CiJxbs .rtxbs 12 .rtxc6 would win a pawn. Moreover, if Black plays too passively, White will slowly develop his kingside and look forward to a nice space advantage in the mid­ dlegame, so Black must play with some purpose. 9 �d 7 ...

Here too this is a good, flexible move. Black also protects the c6-knight which will allow him to play ...bs. In339

A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1

stead both 9 ... tLias and 9 ... tLld7 allow White to happily play his intended 10 t2Jd1!. Other moves also seem insuffi­ cient: a) 9 ...ifd7 (the novel idea is to pro­ tect the c6-knight with the queen, while ...�d8 becomes a possibility)

10 tLlds (after 10 t2Jd1 bS 11 cs dxcs 12 .l:!.xcs .l:!.d8 13 tLif2 e6 14 g3 l2Je8 Black has achieved his desired set-up and stands well according to Kaufman) 10 ....l:!.e8 (not 10...bs? 11 cxbs axbs 12 �xc6!, while 10...e6 11 tLixf6+ il.xf6 12 t2Jc3 is slightly better for White accord­ ing to Kaufman) 11 h4 h S 12 g3 lLih7 13 il.g2 es 14 o-o ifd8 1s .l:!.fd1 lLif8 16 cs il.e6 17 cxd6 cxd6 18 dxes il.xds 19 ifxds dxes 20 ifxd8 �exd8 21 iLh 3 and the bishop-pair gave White a healthy advantage in A.Dreev-A.Cabrera, Albox 200S. b) 9 ... es 10 dS lLie7 11 lLig3 (or 11 cs) 11... l2Je8 12 il.d3 fs 13 exfs tLixfs (13 ... gxfs 14 tLihs) 14 tLixfs gxfs 1s o-o looks like a worse version of the classi­ cal 6... es line - Black's queenside moves are pointless here. 340

c) 9 ... e6!? was Black's choice in a game between two members of the world's elite and could certainly use further testing.

White has: cl) 10 g3 can be met by 10...ifd7!? when Kaufman gives both 11 tLlf4 es 12 dxes tLixes and 11 il.g2 bS!. Instead 10 ...t2Je7 11 il.g2 bS 12 b3 c6 13 o-o as 14 es! l2Je8 1S f4 likely gave White an edge in R.Ponomariov-S.Ganguly, Span­ ish Team Championship 2010, al­ though even here Black wasn't without his counter-chances and might have preferred 1S ...bxc4!? 16 bxc4 il.a6 to the stodgy 1s .. .fs of the game. c2) 10 t2Jd1 could be met by 10 ... t2Jd7 (A.Gupta-A.Lahiri, Bhubanes­ war 2009) or 10...t2Je7 (V.Malakhatko­ K.Neumeier, Oberwart 2006). c3) 10 il.g s was suggested by Golubev. Here Black should play 10 ...bS! 11 cxbs axbs 12 tLixbs �xbs 13 il.xf6 (or 13 .l:!.xc6 tLixe4!) 13 ...il.xf6 14 �xc6 il.b7 1s �c2 il.gs 16 ifd3 �b6 when he has compensation for the pawn according to Kaufman.

Pa n n o V aria t i o n : 9 a 3, 9 I:. b 1 a n d 9 I:. c 1

c4) 1 0 b3 tLid7 11 h4! ? (Kaufman suggests 1 1 g3 tLie7 12 il.g2 bS 13 tLid1 cs 14 dxcs dxcs 1s o-o lLic6 16 f4) 1 1...hs 12 g3 tLie7 13 g4! ? hxg4 14 fxg4 lLif6 1S gS lLihS 16 il.g2 es 17 il.f3 bS 18 dxes il.xes 19 t2Jf4 il.xf4 (or even 19 ...tLixf4 20 il.xf4 bxc4) 20 il.xf4 bxc4 21 bxc4 il.e6 was unclear in V.Anand­ A.Grischuk, Linares 2009.

18 exds ifes 19 lLic6 il.xc6 20 dxc6 I:.fe8 21 �f2 was D.Krumpacnik­ M.Muhutdinov, Budapest 1991, and now 21 ... .i::txbs!? 22 tLixbs ifxbs would give Black very good compensation for the exchange.

10 bs ...

1o tLid1

This is another important part of White's plan. Now ... bs can be met with cs and the knight will swing over to f2. Instead 10 dS tLies with the idea of ...cs should be fine for Black, while 10 b3 bS (Black could also try 10 ... es 11 dS tLie7) 11 cxbs axbs 12 dS tLies 13 lLid4 ife8 14 il.e2 (14 a3 b4 looks okay for Black) 14 ... e6 (14...b4! ?) 1S b4 leads to: a) 1s ...exds 16 tLixds tLixds 17 exds l2Jc4 18 il.xc4 bxc4 19 I:.xc4 il.bs was V.Raicevic-M.Golubev, Belgrade 1991. Here White should have played the simple 20 I:.xc7 according to Golubev when Black does not have enough for the pawn. b) 1s ...l2Jc4 16 il.xc4 bxc4 17 bS exds

Thi s is consistent, even though White can 'push by' the black b-pawn. Others: a) 10... es 11 dS l2Je7 12 cs (or 12 tLif2 lLie8 13 g3 cs 14 b4 b6 1s il.g2 fs 16 o-o lLif6 17 I:.b1 with some advantage to White in J.Rowson-K.Arakhamia Grant, British Championship, Scarborough 2004) 12 ... tLle8 (here Black should con­ sider the rather unusual 12 ... dxcs 13 il.xcs I:.e8 with the idea of ... c6, trying to use his development lead) 13 b4 fs 14 tLiec3 l2Jf6 1S il.e2 �h8 16 tLib2 lLic8 17 l2Jc4 was very good for White in A.Graf-J.Nunn, German League 2001. White's position looks like a very good version of a Classical Variation. b) 10 ... e6 11 tLif2 I:.e8 (11 ...tLle8 12 g3 fs 13 il.g 2 lLif6 14 o-o fxe4 1s fxe4 ife7 16 b4 es 17 dS lLid8 18 cs was 341

A ttacking C h e s s : The King 's I n dian, Vo l u m e 1

good for White in E.Tomashevsky­ A.Fier, World Team Championship, Bursa 2010, but Black might retain the tension with Schmaltz's 14...ife7) 12 iLgS ife7 13 g3 es 14 dS l2Jd4 1S t2Jxd4 exd4 16 il.g2 iff8 17 o-o h6 18 il.xf6 il.xf6 19 iLh3 il.xh3 20 tLixh3 c6 was fine for Black in A.Dreev-D.Kokarev, Mumbai 2010. c) 10...�e8 11 tLif2 bS 12 cs dxcs 13 .l:!.xcs es 14 dS lLie7 1S ifc2 c6 (a better try is 1s ... il.f8!? with the idea of 16 .l:!.xe7 tLiexdS! 17 exds tLixds with active play) 16 dxc6 tLixc6 17 �xc6 il.xc6 18 ifxc6 and Black's compensation was insufficient in A.Graf-J.Nunn, German League 2003. d) With 10... as Black switches to more traditional King's Indian methods and tries to hold White up on the queen side.

now, instead of 1 8 tLixc7 ifxe7 19 h S l2Jd4 which gave Black reasonable play in 1.Khenkin-Z.Efimenko, German League 2008, White could try the an­ noying 18 il.gs ! f6 19 il.e3 with a pleasant advantage. d2) 11 g3 b6 12 il.g2 es 13 dS lLie7 14 o-o tLihs 1S tLif2 (Hazai suggests 1S g4 lLif6 16 lLig3 l2Je8 17 t2Jc3 fs 18 gxfs gxfs 19 f4 with a clear advantage) 1s ...fs 16 f4! (Black had sufficient play after 16 b3 .l:!.f7 17 a3 iff8 18 t2Jd3 h6 19 b4 axb4 20 axb4 gs 21 exfs tLixfs 22 il.f2 lLif6 in l.Novikov-Z.Efimenko, Montreal 2004) 16 ... exf4 17 tLixf4 lLixf4 18 il.xf4 and White's space advantage gave him the more pleasant position in E.Gasanov-A.Fedorov, Minsk 2006. 11 C5

11 cxbs axbs has also been played, but this looks strange and should not cause Black any problems. 11 e6 ...

Practice has seen: d1) 11 t2Jdc3!? es 12 dS lLie7 (12 ...t2Jb4!? may be a better try) 13 lLibS ! b6 14 g4! ? l2Je8 1S lLig3 c6 (Black could also lash out with the thematic 1s ... fs) 16 dxc6 tLixc6 17 h4 tLle7 and 342

Alternatively, 11 ...dS 12 es l2Je8 gives up a lot of space, as does 11 ... es 12 dS tLle7 13 c6, while 11 ... dxcs 12 �xcs leaves Black a bit looking a bit

Pa n n o Va r i a t i o n : 9 a 3 , 9 I:. b 1 a n d 9 �c 1

soft on the c-file, but the position is still not so clear after 12 ...e6 13 g3 ike7. 12 l2Jf2 b4

Instead 12 ... ike7 is just met by 13 g3, while 12 ... t2Je7 13 t2Jd3 makes it dif­ ficult for Black to move his queenside pawn s. However, the text move gains some space and will allow Black to ac­ tivate his minor pieces on the queen­ side.

il.xg s tLixes gave Black good play in l.Gromova-E.Akatova, Chelyabinsk 2008. b) 16 I:.fd1 as 17 ikc2 ?! ikc8 18 f4 ika6 19 I:.d2 es (perhaps even better is 19 ... ds!? 20 es tLlfs) 20 fxes dxcs 21 dxcs tLixes and again Black had good play in A.Stanoev-K.Berbatov, Plovdiv 2010. 16 ...cxd6

13 g3

White has a couple of other moves worth considering here. a) 13 t2Jd3 as 14 g3 t2Je7 1s il.g2 il.bs 16 o-o t2Jd7 17 �fd1 �a8 18 ikc2 ikc8 19 a4 il.a6 20 b3 .rf.d8 21 f4 dS and Black held firm in 1.Khairullin­ E.Gorovykh, St Petersburg 2010. b) 13 lLig3 as 14 il.d3 t2Je7 1S 0-0 il.bs 16 tLie2 t2Jd7 17 cxd6 cxd6 18 b3 il.xd3 19 t2Jxd3 dS 20 es ikb6 was also solid for Black in M.Nikolov-K. Berbatov, Kyustendil 2010. 13 ...tLie1 14 il.g2 il.bs

Black has less space, but his position is flexible and not without dynamic potential. 15 0-0 tLld7 Also possible is 1s ...l2Jc6! ?. 16 cxd6

White does not need to hurry with this exchange, but it is also not clear what else he should do. Indeed, in prac­ tice Black has found good counterplay: a) 16 f4 dxcs 17 dxcs tLie s! 18 fxes 'i!t'xd2 19 il.xd2 il.xe2 20 �fe1 il.bs 21 l2Jg4 tLic6 22 l2Jf6+ �h8 23 il.f4 gs 24

The Croatian IM Srjdan Sale and the Belorussian GM Alexei Fedorov have debated this position a couple of times. Black has an interesting Hedgehog­ type position: a) 17 I:.fd1 as 18 il.f1 �c8 19 t2Jd3 �xc1 20 I:.xc1 ika8 21 b3 �c8 22 tLib2 'i!t'a6 23 t2Jf4 il.xf1 24 :i'.xf1 dS left Black doing well in S.Sale-A.Fedorov, Abu Dhabi 200S. b) 17 .rtc2 and now Black tried 17... dS 18 �fc1 es in S.Sale-A.Fedorov, Abu Dhabi 2006, but White looks better prepared for the opening of the centre. Instead Kaufman proposes 17 ...ikas 18 b3 �fc8 19 �fc1 ikd8 with equality.

343

c'hapter 20

·pa:nno .variation Other Lines

1 d4 lLif6 2 c4 g6 3 l2Jc3 il.g7 4 e4 d6 5 f3 o-o 6 il.e3 lLic6

In this chapter we look at some odds and ends, as well as a couple of other systems for Black. A: 7 1i'd2.

a6 8 0-0-0

·

9� 7 �ge2 a6 8 �Cl

C: 7 �e2 a6 8 •d2 :b8 9 lc11 and 9 d5 D: 7 �ge2 llb8 8 •d2 :ea E: 7 �ge..2 a6 8 •dz :ea F: 6 .igs �c6

344

W e begin our coverage of White's early deviations with Line A, where White delays or omits playing tLige2. Line B instead covers lines where White delays or omits playing ii'd2. In Line C we come back to the main line 7 tLige2 a6 8 ii'd2 .l::tb 8 and here we look at what happens when White plays something we have not covered yet on move 9, namely 9 �d1 and 9 dS. Line D is the Westerinen system with 7 ...�b8 and 8 ... �e8. This is a tricky line and if White is inaccurate then Black gets a wonder­ ful version of the ... cs pawn sacrifice. Direct approaches are more dangerous, however. Then in Line E we look at a kind of mixed system where Black omits ... .l::tb8 and instead combines ...a6 with ....l::te 8. Finally, Line F covers 6 il.gS l2Jc6. Note that in the Panno the fairly fashionable move 6 tLige2 does not have any real independence after 6 ... l2Jc6, .as 7 il.e3 just inverts White's usual 6th and 7th moves while 7 il.gs

Pa n n o Va ria tio n : O t h e r L in es

transposes to Line F. Instead White could play 7 dS, but after 7 ... tLies nei­ ther 8 lLig3 c6 nor 8 l2Jd4 cs look dan­ gerous. Black can also meet 6 tLige2 with 6 ... a6!?. Then White has hardly anything better than 7 il.e3 or 7 il.gs when in both cases 7...l2Jc6 just brings us back to normal lines. A) 7 ii'd2 a6 8 0-0-0

There is really no reason to castle so soon, but this move has been played by some strong players. Instead 8 il.d3 also looks a bit naive, but it is not so terrible. After 8 ... es 9 dS l2Jd4 Black may not be better but he should not be worse: 10 tLige2 tLihs!? (10 ... t2Jd7 11 o-o cs 12 dxc6 bxc6 13 b4 was a little more pleasant for White in G.Kamsky-J.Hjartarson, Biel lnterzonal 1993) 11 o-o and now Black could play 11 .. .fs or 11 ...cs.

might try one of: a) 9 g4 e s 10 dS l2Jd4 11 t2Jce2 (not 11 tLige2? tLixf3) allows the interesting 11 ...tLixf3! ? 12 tLixf3 il.xg4 13 il.g2 tLixe4 with good compensation for the piece. b) 9 il.h6 il.xh6 10 ii'xh6 es 11 dS (instead 11 tLige2 bS transposes Line C1 of Chapter 18) 11...t2Jd4 12 tLige2 cs (or even 12 ... tLixe2+ 13 il.xe2 il.d7) 13 dxc6 bxc6 14 t2Jxd4 exd4 1S .i::txd4 .i::tx b2! is good for Black, as we saw in note 'c' to White's 11th move in Line C1 of Chap­ ter 18. c) 9 h4 and here 9 ...es seeks to pun­ ish White's omission of tLige2. After 10 dS l2Jd4 11 tLige2 cs 12 dxc6 bxc6 13 t2Jxd4 exd4 14 il.xd4 il.e6 Black has managed to play the pawn sacrifice with White's king committed to the queenside. Still, I am not sure if I be­ lieve in this version, as White's king is not in terrible danger and he has not spent a lot of time moving his king's knight, which is usually some justifica­ tion for offering the pawn. Instead of all this, Black could simply play 9 ... h s with a likely transposition t o the main lines of Chapter 1S. 9 cxbs

s .. bs!? .

Black seeks to punish White for his committal play. It is also possible to play 8 ...�b8 when White should proba­ bly just play 9 tLige2, which transposes to Line A of Chapter 18, although he

White should probably consider ig­ noring the offer and getting on with it with 9 h4. 9... axbs 10 il.xbs tLias!

Black already has excellent play. With the kings castled on opposite wings, the initiative is worth more than the pawn. 345

A ttacking Chess: Th e King 's Indian, Vo l u m e 1 1 2 il.xa6 .i::txa6 1 3 ifd3 ifa8! 14 tLige2 .i::tb 8

11 �b1

White has also tried 11 il.h6 c6 12 il.xg7 �xg7 13 il.d3 il.e6! (this was queried by Hjartarson in his notes, but it looks good to me; instead 13 ... ifb6 14 es tLlds 1s exd6 ! exd6 16 tLixds cxds 17 tLie2 gives White better chances to weather the storm) 14 b4? (this is the only move according to Hjartarson, but it looks insane; that said, 14 �bl 'i!t'b6 gives Black excellent compensation) 14...l2Jc4 1S il.xc4 il.xc4 16 ifb2 ifb6 17 lLih3 �fb8 18 .l:!.d2 (or 18 a3 ifa6 19 tLib1 cs - Hjartarson), and now 18 ... es 1 9 .l:!.hd1 ifxb4 20 ifxb4 .i::txb4 gave in Black some advantage L.Christiansen-J. Hjartarson, Szirak ln­ terzonal 1987, but with 21 dxes dxes 22 tLif2 .l::ta3 23 �c2 White could have put up a lot of resistance. Instead 18 ...ifxb4 19 ifxb4 .l::txb4 is more flexi­ ble and looks good for Black, but 18 ... ifa6! with the idea of ... cs looks best of all.

Black clearly has an excellent game. It looks as if White has castled queen­ side in a Benko Gambit. 15 il.c1 e6!

Black creates an outpost on c4 for his knight. 16 h4 ds 17 hs l2Jc4 18 hxg6 hxg6 19 b3 cs! 20 dxcs tLld7!

All of Black's pieces are participat­ ing in the attack. 21 exds exds

Even stronger is 21 ... .l:!.xa2 ! !

11...il.a6

Also strong is 11 ... c6 12 il.d3 il.e6 with excellent play. 346

2 2 tLixa2 (instead 2 2 ifxc4 tLies! wins the queen which is tied to the de-

Pa n n o Va ria t i o n : O t h e r L i n e s

fence of the b3-pawn) 22 ... t2Ja3+ (22 ... ifa4!? also looks good, but the text is simpler) 23 il.xa3 ifxa3 24 l2Jd4 tLixcs 2S ifc2 tLixb3 26 'i!t'b2 t2Jd2+ 27 �xd2 �xb2+ 28 .l:!.xb2 il.xd4 and although the material balance is similar, Black has a much better version of the game con­ tinuation. 22 ifxds .i::tx b3+ 23 axb3 .!:!.al+ 24 �c2

8. . ..l::tb8 9 h S bs 1 0 ifd2 transpose to the main lines of Chapter 14. Instead 8...h s 9 ifd2 �b8 takes play t o the start of Chapter 1S, while White can also play 9 tLic1 with an odd mix of set-ups that was tried three times in the 1992 Fischer-Spassky match:

�xcl+ 25 tLixcl lLie3+ 26 �bl tLixds 27 .i::t x ds il.xc3 28 �xd7 ifa3

Black still has the initiative, but White managed to hold out in M.Petur­ sson-E.Gufeld, Hastings 1986/7. B) 7 tLige2 a6

8 tLlcl

Here we look at lines where White delays playing ifd2. Moves like 8 .i::tb1 and 8 a3 are likely to transpose to Chapter 19, although the latter gives Black the interesting option of 8 ... il.d7 9 b4 'ir'b8!? supporting ... bs with the queen. After 8 h4 there are many transposi­ tional possibilities. Here 8 ... es 9 dS tLias 10 lLig3 cs 11 hS :i'.b8 12 if d2 and

a) 9 ... es 10 dS l2Jd4 (Black was left with a passive position after 10 ...tLle7?! 11 il.e2 lLih7 12 t2Jd3 fS 13 a4 lLif6 14 lLif2 as 1S ifc2 cs?! 16 0-0-0 b6 17 .l:!.dg1 in B.Spassky-R.Fischer, 12th matchgame, Sveti Stefan 1992) 11 t2Jb3 tLixb3 12 ifxb3 �h7 13 il.e2 il.h6 14 il.xh6 �xh6 1S o-o-o �g7 was pretty solid for Black in B.Spassky-R.Fischer, 28th matchgame, Belgrade 1992, al­ though with a clear plan of �dg1 and g4, White's play is easier. b) 9 ... tLld7!? 10 t2Jb3 as 11 a4 t2Jb4 12 il.e2 b6 13 g4 hxg4 14 fxg4 cs 1S hS cxd4 16 t2Jxd4 tLics 17 tLlds il.b7 gave Black a dynamic position in B.Spassky­ R.Fischer, 30th matchgame, Belgrade 1992. 8 es 9 ds ...

This time 9 t2Jb3 exd4 10 t2Jxd4 34 7

A ttacking C h e s s : T h e King 's I n dian, Vo l u m e 1

t2Jxd4 11 il.xd4 il.e6 12 il.e2 c6 should favour Black, as he has saved a tempo on the superfluous ... �b8. 9 ...l2Jd4 10 t2Jb3

Again, after 10 tLi1e2 tLixe2 11 il.xe2 tLih s Black has saved time by omitting ....l::tb8. 10...tLixb3

Black can also play 10 ...cs 11 dxc6 bxc6 12 l2Jxd4 exd4 13 il.xd4 .l::tb8 and now 14 ii'd2 transposes to the main line of the pawn sacrifice with 9 tLic1 (Line B22 in Chapter 17). The text move gives the position its own character. 11 ii'xb3

After 11 axb3 Black can play 11...cS to block the queenside or the immedi­ ate 11 ... tLih s. 11...cs

This was Kasparov's choice, but 11 ...tLih s, 11 ...t2Jd7 and 11 ...b6!? are other possibilities.

Black could also try 1 s...�fb8. 16 h4 �fc8 11 hs tLlds?!

This should not work. Instead 11 ... tLld7 18 ii'b4 tLics may leave White with a small edge, but the position is still quite unclear and with ideas like ...l2Je6-d4 1 do not think Black should be too unhappy. 18 exds cxds 19 �xds?!

Instead White could refute Black's play with 19 cxds! il.a8 (not 19 ... il.xds? 20 ii'xb8! ) 20 ii'as e4 21 hxg6 hxg6 22 f4! according to Kasparov. 19 ... il.xds 20 tLixds ii'e6 21 ii'a7

Perhaps Timman missed 21 ii'as �bS!. 21.. AaS 22 ii'b7 �ab8 23 ii'a7 .l::t a 8 24 ii'b7 �ab8

With a draw by repetition in J .Timman-G. Kasparov, Moscow 1981. C) 6 il.e3 l2Jc6 7 l2Jge2 a6 8 ii'd2 .i::t b 8

12 dxc6 bxc6 13 o-o-o ii'e7

Now we come back to the main line to tie up some loose ends. We examine: 14 ii'b6

This is the only way to cause Black any trouble. 14...il.b7 15 g4 �ab8

348

('1:9 Jlcl2. C2:'t'd5

P a n n o Variatio n : Oth e r L in e s

Otherwise, 9 lLig3 looks funny, but after 9 ... es 10 dS l2Jd4 11 tLige2 we have just transposed back to Line Bl of Chapter 17. Instead 9 a4 also looks strange. Black could play 9 ... es (or even 9 ... as!? to fix the queenside) 10 dS and now both 10 ... tLias 11 tLic1 cs and 10 ...l2Jb4, with the idea of 11 as cs, look fine.

11...tLixes 12 il.e2 bsl

Very aggressive. The simple 12 ...il.e6 is another option. 13 cxbs axbs

C1) 9 �d1

14 .il.xbs

This is kind of an odd move. It is cer­ tainly not clear that the rook belongs on the d-file, while White's develop­ ment still lags and he can no longer castle queenside. This move does, how­ ever, serve to illustrate some of Black's dynamic possibilities quite well! 9...�e8

I like this move, but 9 ...bs and 9 ...il.d7, with the idea of 10 tLic1 es, are possible as well. 10 tLic1 es 11 dxes

After 11 dS l2Jd4 12 tLi1e2 tLixe2 13 il.xe2 tLih s Black has a normal position and White has spent a move on the strange �dl.

If White cannot take the pawn at all, then Black has achieved his strategic aims very easily and has an improved version of the variation 9 tLic1 es 10 t2Jb3 exd4 11 t2Jxd4 tLies (Line A2 of Chapter 17). The alternative is 14 tLixbs tLixf3+ ! 1S gxf3 tLixe4 1 6 fxe4 '1lr'h4+ 17 il.f2 ifxe4 18 ifds (18 �g1 �xbs) 18 ... ifb4+ (better than 18 ... �xbs 19 ifxe4 .i::txe4 20 il.d4 il.xd4 21 �xd4 .i::txd4 22 il.xbs �b4 when Black is really playing only for a draw) 19 �fl (19 ifd2 ifxbs) 19 ... il.b7 and Black is much better. 14...tLixe4! 15 fxe4 �xbs!

It is one blow after another. 16 tLixbs t2Jc4 17 iff2

17 il.gs ifd7! does not solve White's problems. 11...�xe4 18 o-o tLixe3 19 ifxf7+ �h8 20 ifxc7 ifxc7 21 tLixc7 tLixdl 22 .i::tx dl il.d4+

349

A tt a c k i ng C h e s s : Th e King 's I n d i a n, Vo l u m e 1

his fair share of the chances. 9...tLies Also possible is 9 ...tLias!? when Black will meet any of 10 lLig3, 10 tLic1 and 10 l2Jd4 with 10 ...cs, aiming for queenside play. 10 l2Jd4

23 �fl?

This loses quickly. Instead 23 �h1 il.g4 24 �fl il.xb2 was a better try, al­ though Black's dominating bishops give him excellent winning chances. 23 ...il.g4 24 .l::t d2

24 Ite1 �f4 is mate. 24 ...il.e3 25 �C2 �d4

White resigned here in F.Gheorghiu­ L.Kavalek, Amsterdam 1969. C2) 9 dS

After 10 lLig3 cs 11 il.e2 e6 the play heads for a Benoni position, but 10 ... c6! ? keeps the tension and gives the game a more original flavour. The same can be said for 10 tLic1 cs (again there is 10 ... c6!?) 11 il.e2 e6 12 o-o (12 a4 exds 13 cxds is a decent enough Modem Benoni for Black) 12 ... exds 13 exds (13 cxds allows 13 ... bs) 13 ... .i::te8 14 a4 h S ! ? with an unusual Benoni. 10...cs

With White's knight centralized, 10 ... c6 is untried, probably because of the simple 11 il.e2, rather than 11 f4!? tLieg4 12 dxc6 tLixe3 13 ifxe3 es! with excellent play for Black .. 11 tLic2

After 11 dxc6 bxc6! (White has a pleasant Maroczy Bind after 11 ... tLixc6 12 il.e2 il.d7 13 0-0) 12 f4? cs! Black is doing well after either 13 t2Jb3 l2Jc6 14 il.e2 (or 14 h3 tLih s) 14 ... l2Jg4 or 13 fxes cxd4 14 il.xd4 dxes when b2 will fall after the impending exchange of queens. 11...e6

This advance looks logical enough, but if it were that good the Panno would never be seen. Play often heads into a Benoni structure that gives Black 350

Also possible is 11 ... il.d7 12 a4 e6 13 t2Ja3 (13 il.e2 exds 14 cxds bS gives Black a good version of the Modem Be­ noni) 13 ... exds 14 cxds tLih s !? 1s il.e2 ifh4+ (or 1s .. .fs!?) 16 il.f2 iff4 when 17

Pa n n o Variatio n : Oth e r L i n es

il.e3 �4+ repeats. 12 a4 exds 13 cxds

c4-pawn is loose. If Black wishes to play this system against 7 ii'd2, however, then he should employ the move order 7 ... .l::te 8 (7 ... .l::tb8? 8 dS does indeed win a pawn because c4 is still protected) 8 l2Jge2 .l:!.b8. 8 ....i::te 8

Black has quite a decent Modem Benoni because White has spent so much time moving his king's knight. 13 ...tLihs Black begins active operations im­ mediately, but 13.. Ji.d7 should also suffice.

This sly system was christened the Westerinen Variation by John Watson in recognition of its development by the Finnish Grandmaster. Black can now avoid the exchange of his dark­ squared bishop following il.h6, but playing 8 ... �e8 does delay Black's queenside counterplay by a tempo. The really clever point of this line is when White plays an early tLic1, as we shall see.

14 .il.e2 fs

Black has good counterplay. D) 7 t2Jge2 .l::tb 8

9 h4

8 ii'd2

Note that 8 dS does not win a pawn after 8 ... tLies 9 il.xa7?! .l::ta8 because the

This direct move is probably best, as it is in the main lines. Of course, though, White has a broad choice: a) 9 o-o-o is similarly direct and dan­ gerous: 9...a6 10 g4 (or 10 h4) 10...bs 11 h4 tLias (instead 11...hs 12 lLidS! ? has scored terribly for Black) 12 lLig3 tLixc4 13 il.xc4 bxc4 14 hS and now, instead of 351

A ttacking C h e s s : Th e King 's I n d i a n, Vo l u m e 1

14...cs 1s dxcs (also very good for White is 1s hxg6 hxg6 16 dxcs ii'as 17 il.h6 il.h8 18 il.f4, as in T.Marinelli-F.Portisch, Budapest 1989) 1s ... ii'as 16 hxg6 fxg6 17 cxd6 il.e6 18 il.d4 when White had a huge advantage in Nguyen Anh Dung­ F.Donguines, Manila 2001, Black should consider 14...c6!?. b) 9 g4 is relatively unexplored. After 9 ... a6 10 h4 bS (or perhaps 10 ... hs) 11 h S Black could try 11 ... bxc4, 11 ...tLias or 11 ... es. c) 9 .i::t d1 a6 is more likely to arise from 8 ... a6 9 �d1 �e8 and this position was considered in Line Cl, above. d) 9 dS will lead to similar play to that of Line C2. Black should not take this plan too lightly, though.

For example, 9 ...tLies 10 l2Jd4 cs 11 tLic2 e6 12 il.e2 exds 13 cxds il.d7 (ob­ vious, but perhaps too slow) 14 a4 a6 1S t2Ja3 and White's long knight tour proved rather purposeful by clamping down on the bS-square in S.Brudno­ D.Vigorito, Providence 2008. Very inter­ esting, however, would have been 13 ...bs!? 14 il.xbs ii'as! 1s il.xe8 tLixe8 352

1 6 t2Jd1 ii'a6 with tremendous activity for the exchange. e) 9 a3 (of course White can pursue queenside plans as well) 9...t2Jd7!? (more interesting than 9...a 6 10 b4 which could transpose to Line A of Chapter 19) 10 b4 tLlb6 11 tLic1?! (this is pleasant for Black, but both 11 ii'a2 e6! and 11 ii'd3 e6! give Black counterplay as well) 11...eS 12 dS l2Jd4 (White's lack of control over the b3-square leaves him struggling for equality) 13 il.d3 c6 14 o-o cxds 1s cxds il.d7 16 il.xd4 exd4 17 tLibs il.xbs 18 il.xbs �e7 19 il.d3 .i::tc7 20 tLie2 �bc8 gave Black some ini­ tiative in D.Vigorito-J.Watson, Phila­ delphia 1993. I may have lost this game, but at least I made a good friend! f) 9 .i::t b 1 is probably a better way of implanting the queenside plan:

9 ... a6 (9 ...tLld7 is possible, but now after 10 b4 lLib6 11 t2Jc1 es 12 dS l2Jd4 13 t2Jb3 Black would rather have his king's rook on f8 to support the .. .fs­ advance) 10 b4 bS (it is probably better to temporize with 10 ... il.d7) 11 cxbs

Pa n n o Va ri a t i o n : O t h e r L i n es

axbs 12 ds tLies 13 l2Jd4 il.d7 14 tLicxbs (the patient 14 il.e2 ! ? is also possible, saving the consumption of the bS­ pawn until White completes his devel­ opment) 14 ... e6 1S dxe6 fxe6 16 il.e2 t2Jxf3+ 17 gxf3 es was L.Polug aevsky­ E.Gufeld, Riga 197S. Now 18 il.c4+ �h8 19 o-o exd4 20 t2Jxd4 would leave Black with very little for the pawn. g) 9 tLic1 is quite common, but Bl ack will now reveal a key point: 9 ... es 10 dS?! (10 t2Jb3 exd4 11 t2Jxd4 is safer, when Black could head for the lines of Chapter 17 with 11 ...t2Jxd4 12 il.xd4 a6 or try 11...ds!? 12 cxds tLixds 13 tLixds t2Jxd4; then 14 o-o-o cs 1s il.gs ii'd6 is unclear and 14 il.xd4 ii'xds 1S il.xg 7 ii'xd2+ 16 �xd2 �xg7 is level) 10...t2Jd4 11 t2Jb3 cs! 12 dxc6 bxc6 13 t2Jxd4 exd4 14 .il.xd4.

Compared with the pawn sacrifice from Chapter 17 (after 8...a6 9 tLic1), Black is much better off having played ... �e8 instead of ... a6. Following 14 ... ds (even 14... cs!? is possible, with the point 1 s il.e3 dS! 16 cxds tLixds and now 17 tLixds .l::txb2 18 ii'd3 il.e6 17

ii'xds il.xc3+ 18 bxc3 ii'xds 19 exds �xe3+ gives Black plenty for the pawn, but 17 �d1 �xb2 18 ii'xb2 il.xc3+ 19 ii'xc3 tLixc3 20 �xd8 .l::txd8 21 il.xcs or 21 �g1 should prove more testing) 1S cxds cxds White has: g1) 16 il.bs runs into 16 ...�xbs! 17 tLixbs tLixe4! 18 fxe4 .i::txe4+ 19 �2 (al­ ternatively, 19 �fl il.a6 or 19 �d1 il.g4+ 20 �cl when after both 20...il.xd4 21 t2Jxd4 ii'c7+ 22 ii'c3 ii'f4+ and 20 ....i::te 2 21 ii'f4 ii'c8+ 22 t2Jc3 �e4 Black wins) 19 ... ii'M+ 20 g3 .i::txd4 21 t2Jxd4 il.xd4+ and White can resign. g2) 16 es tLih s! (16 ...t2Jd7 has also been played, but the text is better) and now:

g21) 17 f4 is well met by 17 ...f6, which leaves: g22) 17 o-o-o is the safest and then 11 ... il.xes 18 il.xes �xes is about equal. g23) 17 il.bs (ambitious, but Black has a wonderful idea) 17 ...�xes+! 18 �f2 (18 il.xes il.xes gives Black a very strong initiative for the exchange) 18 ... a6! 19 il.e2 �xb2 (or 19 ...ii'h4+ 20 g3 �xb2 !) 20 ii'xb2 ii'h4+ 21 �fl ii'xd4 353

A t tacking Ch e s s : The King 's I ndian, Vo l u m e 1

and Black was winning in A.Saidy­ H.Westerinen, Netanya 1969. Now we return to 9 h4:

9 a6 ...

Black has also tried 9 ... hs, but 10 o-o-o (not 10 tLic1?! es with similar play

to variation 'g' above) 10 ... a6 11 il.h6 (White has also fiddled with all of 11 ii'e1, 11 lLif4 and 11 tLlds) 11 ...il.h8 (11...bs 12 g4!) 12 g4 (or 12 ii'e3 ! ? es 13 dS tLias 14 lLig3 cs when both Watson's suggestion of 1S il.d3 bS 16 tLif1 and 1S il.e2 il.d7 16 tLif1 bS 17 .i::td2 tLixc4 18 il.xc4 bxc4 19 ii'e2 l2Jh 7 20 g4, as in J.Tarjan-A.Whiteley, Hastings 1976/77, look promising for White) 12 ... es (Black could try 12 ...b s or even 12 ... hxg4!?, but the whole position looks precarious for him) 13 iLg S! looks promising for White. 10 hs bs

Black could consider too 10 ... es 11 hxg6 fxg6 12 dS tLias 13 lLig3 cs or even 10 ... lLixh S !? 11 il.h6 (11 g4 lLif6 12 il.h6 il.h8 13 il.g s also looks dangerous; Black could try 13 ... bs or 13 ... e6) 11 ... es 12 il.xg7 �xg7 13 g4 and here: 354

a ) 1 3 ... t2Jxd4 1 4 t2Jxd4 exd4 with a further split: al) 1S tLie2 ii'f6 (1S ... tLif6!? 16 ii'xd4 was feared by Watson, but 16 ...il.xg4! 17 fxg4 �xe4 gives Black excellent play) 16 gxh s ii'xf3 17 h6+ �8 18 �h4 .i::txe4 19 .i::txe4 ii'xe4 20 ii'xd4 ii'xd4 21 t2Jxd4 il.g4 22 �2 and White had the better chances in W. Browne-J.Watson, Berke­ ley 1977. a2) 1S lLidS! looks strong, since 1s ... c6 16 gxh s cxds 17 hxg6 hxg6 18 ii'xd4+ f6 19 ii'xds il.e6 20 ii'd2 must favour White. b) 13 ... exd4 is probably best: 14 gxh s (14 tLlds tLies 1s il.g2 lLif6 16 ii'h6+ �g8 gives Black counterplay, as there is no mate) 14... dxc3 1S ii'xc3+ �g8 16 o-o-o gives White some com­ pensation, but this still looks playable for Black; ... tLie s is available, while after hxg6 fxg6 Black can defend along his second rank. However, if White delays the exchange on g6, then ... g s may even be possible to keep the h-file closed. 11 hxg6 fxg6

Pan n o Variation: O th e r L i n e s

Certainly 11...hxg6?! looks too risky.

12 il.h6

Untried is 12 cxbs axbs 13 dS, but this may well be good for White as the bS-pawn is weak. 12 ... il.hS 13 il.gs es Instead 13 ... �f8 has been the most popular choice, which seems strange. Both 14 o-o-o and 14 cxbs axbs 15 lLif4 look promising for White. However, De Firmian and Watson's 1980 suggestion of 13 ... e6!? is still untried! 14 cxbs

Browne and Watson considered 14 tLlds t2Jxd4 15 t2Jxd4 exd4 16 il.xf6 (in­ stead 16 ifxd4 tLixdS! 17 ifxds+ il.e6 18 il.xd8 il.xds 19 il.xe7 .l::tb7 20 il.xd6 il.xc4 is fine for Black) 16 ...il.xf6 17 'i!t'h6 il.g7 18 ifxh7+ 'itf7 19 �h4 to be win­ ning for White (we must keep in mind that this was in 1980), but 19 .. Ji.fS! (worse is 19 ....l::th 8?! 20 .l::tf4+ il.fs 21 .i::txfs+! gxfs 22 'ili'xfs+ �g8 23 0-0-0, which gives White good compensation for the exchange) 20 �f4 ifgs 21 '1!r'h4 (and not 21 o-o-o? .l::th 8) 21 ... il.f6! 22 '1lr'h7 + il.g7 leads to a draw by repetition.

14. ..axbs 1 5 tLlds

Now 1s ...l:!.f8? 16 if c2 (16 �cl also looks good) 16...il.d7 17 'ir'b3 il.e6 18 tLixf6+ 'ili'xf6 19 il.xf6 il.xb3 20 il.xh8 l2Jb4 21 �d2 tLixa2 22 dxes �xh8 23 l2Jd4 il.g8 24 e6 cs 2 5 l2Jc6 �b6 26 tLle7 was winning for White in D.Studzinski­ S.Szczepaniec, correspondence 1991. However, Black can improve with 1S ... t2Jxd4! when the position is un­ clear: for example, 16 t2Jxd4 exd4 17 ifxd4 tLixe4! with complications that are not unfavourable for Black. E) 7 tLige2 a6 8 ifd2 �es

Here Black omits ....l::tb8 rather than 355

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

... a6. This is kind of a half-measure be­ tween the main lines and the Westerinen Variation. 9 h4

This is always critical, but White has the usual array of ideas. The queenside ideas 9 a3 and 9 .l::tb1 will likely trans­ pose to the notes to White's 9th in Line D, if not to Chapter 19, while 9 �cl il.d7 10 b3!? (instead 10 tLid1 could be met by the immediate 10 ... bs when Black has benefited from omitting ... .l::tb8) 10 ... e6!? was J .Piket-M.Golubev, German League 2002. Other attacking moves are less good: 9 o-o-o allows 9 ... bS! immedi­ ately; the unexplored 9 g4 �b8 was mentioned in note 'b' to White's 9th move in Line D; and the direct 9 il.h6 il.h8 10 h4 e s ! allows Black to use the e8-rook with 11 dS l2Jd4 12 tLixd4 exd4 13 tLie2 (13 ii'xd4? tLixe4) 13 ... cs 14 dxc6 bxc6 1S t2Jxd4 tLixe4! 16 fxe4 .i::txe4+ 17 tLie2 .l::tb 8, which gives him a strong initiative for the piece. White can also play 9 tLic1 es

10 dS (after 10 tLlb3 exd4 11 tLixd4 356

Black has the usual choice between 11 ...tLixd4 and 11 ... tLies) 10...t2Jd4 11 tLlb3 (11 tLi1e2 cs 12 dxc6 gives Black the additional option of 12 ... tLixc6 !? 13 lLidS bS 14 il.b6 ii'd7 1S lLie7 .l::tb8 16 tLixe8 ii'xe8 with some compensation for the exchange in the well-known game A.Beliavsky-G.Kasparov, Moscow 1981) 11 ... cs (instead 11 ... tLixb3 12 axb3 may be okay too, but Black has played ... .l::te 8 instead of ... �b8, which hardly favours him) 12 dxc6 bxc6 13 tLixd4 exd4 14 il.xd4 dS!? (14 ... .l::tb8 transposes to the main-line pawn sac­ rifice - Line B22 of Chapter 17) 1S cxds cxds 16 es tLld7, which has been seen a few times. This version of the pawn sacrifice should help Black because ... .l::te8 is often more useful than ....l::tb8. 9 ... hs 10 o-o-o

10 tLic1 again allows the thematic 10 ... es 11 dS l2Jd4 12 t2Jb3 (after 12 tLi1e2 cs 13 dxc6 Black has the extra possibility of 13 ... tLixc6!?) 12 ... cs!. Once more, compared to lines with ... �b8, Black probably has a slightly improved version of the pawn sacrifice with his rook on e8 instead: 13 dxc6 bxc6 14 tLixd4 exd4 1s il.xd4 dS!? (1s ....l::tb8 would transpose to the main lines of the pawn sacrifice covered in Line B22 of Chapter 17) 16 cxds cxds 17 es tLld7 18 f4 f6 19 e6! lLif8!? (instead 19 ... �xe6+ 20 il.e2 has been played a couple of times with little success) 20 fS gxfs (Black has little choice; he must try to use piece play to compensate for his deranged structure) 21 il.e2 tLixe6

Pa n n o Varia t i o n : O t h e r L in es

22 o-o? (White probably felt that this was the safest, but 22 o-o-o! ? looks bet­ ter; his king is pretty safe on the queenside and castling long brings pressure against d5 more quickly) 22 ...t2Jxd4 23 ifxd4 il.h6! and Black had serious counterplay in S.Kasparov­ D.Bocharov, Bhubaneswar 2010. 10...bs

14...gs!?

Black has achieved this advance without the preparatory move ... �b8, but this is not such a great accom­ plishment because as long as White does not capture on b5, Black would benefit from having his rook on the b­ file after all.

Black keeps the h-file closed. Instead 14 ... e5 15 hxg6 fxg6 16 d5 l2Jd4 was the risky continuation of A.Komev1.Kumosov, Tomsk 2001. Here Gallagher gives 17 t2Jxd4 exd4 18 il.xd4 il.xd4 19 ifxd4 ifg5+ 20 �d2 if e5 21 ifxc4 gxf3 22 ifxe7 with the initiative for White and Golubev continues 22 .. lte7 23 ifd8+ �e8 24 ifh4! .l::ta7 25 .i::td h2. This looks scary, although I do not see any­ thing clear for White after 2 5 ....l::t g 7. 15 il.xgs es 16 il.xf6 ifxf6

11 tLlds!

This is considered critical. White cannot easily exchange off Black's g7bishop, so he trades off Black's defen­ sive knight instead. 11..bxc4 12 tLixf6+

Instead 12 g4 hxg4 13 tLixf6+ il.xf6 14 h 5 transposes, whereas 12 ...tLixg4?! 13 fxg4 il.xg4 14 il.h6 did not hold up in J.Lautier-C.Maier, Versailles 2006. 12 ...il.xf6 13 g4 hxg4 14 hs

17 fxg4

Instead 17 d5 gxf3 18 dxc6 fxe2 19 il.xe2 iff4 20 il.xc4 il.e6 21 .i::t dg1+ �h8 357

A ttacking Chess: Th e King 's I n d i a n , Vo l u m e 1

was given as equal by Lautier, but 21 .i::th g l+!? with the idea of �dfl looks promising for White. Black may im­ prove earlier with 17...ifxf3 !?, intend­ ing 18 ...l2Jd4 with counterplay as the e4-pawn is weak. 17...exd4 18 gs iff3 19 g6 il.g4!

A suggestion by Golubev as an im­ provement over 19.. .fxg6?! 20 t2Jxd4 t2Jxd4 21 ifxd4 with a winning position in J.Lautier-J.Piket, Cannes 1990. Later he got to try it against Lautier himself!

il.xc4+ tLixc4 26 ifxc4+ ife6 2 7 ifxc7+! if e7 28 ifc4+ ife6 when Black is some­ how holding on. 20...�xe4?

Tragic. Instead 20 ...tLies 21 gxf7+ �xf7 transposes to variation 'b' above, but Black's best looks to be 20 .. .fxg6! and here: a) 21 �el (to take on d4) 21 ...c3 !? 22 bxc3 il.xhs 23 t2Jxd4 t2Jxd4 24 ifxd4 iff4+ 25 �b2 ifes gives Black good chances. b) 21 hxg6 and now Golubev gives 21 ...�g7 (not 21 ... c3 22 bxc3 .rf.ab8 23 t2Jxd4!) 22 �el d3 23 il.g2 (23 .l::t g 3 ifxe4 24 il.g2 iffs is good for Black) 23 ...iff2! 24 .rf.gfl! ifxe2 25 �xe2 dxe2 with compensation for the queen. 21 gxf7+ �xf7 22 il.g2

And White won significant material in J.Lautier-M.Golubev, Odessa (rapid) 2006. 20 .l::!.gl

F) 6 iLgS

This could well be a mistake. Golubev mentions a couple of alterna­ tives: a) 20 t2Jxd4 ifxdl+ 21 ifxdl il.xdl 22 il.xc4 �g7!? 23 .l:!.g1 il.g4! (not 23 ... .i::tx e4?! 24 h6+! �6 25 g7! when White is magically better) 24 tLixc6 (24 .i::t x g4 tLies! was Black's point) 24 ... .i::txe4 25 il.ds �ae8 26 il.xe4 �xe4 and Black is certainly okay. b) 20 gxf7+ is probably best: 20 ... �xf7 21 .i::tg1 tLies 22 ifxd4 iff6 and now one possibility Golubev gives is 23 �bl! ? .i::tab8 24 t2Jc3 !? il.xdl 25

This move can be annoying to those who like to play 6 il.e3 es, but it en­ courag es Black to play 6 ... cs without

358

Pa n n o Va ria ti o n : O t h e r L i n e s

offering a pawn. While completely satisfactory, 6 ... cs does not exactly fit into our repertoire and there is nothing wrong with continuing with our favourite system. 6 l2Jc6 Just like after 6 tLige2, Black could also consider 6 ... a6 and only then 7 ii'd2 l2Jc6. Against 6 il.e3 this would give White the extra option of 7 il.d3!?, but here the d4-pawn would be weaker, while after 6 tLige2 a6 the move il.d3 is simply not possible. ...

il.h6, we will generally transpose to 6 il.e3 lines with a quick il.h6. Indeed, the immediate 9 il.h6 would take play into Line C of Chapter 18. Independent play will be examined here, chiefly after the options:

F1: 9 l:tc1 F2: 9 d5

7 tLige2

There is nothing to be gained by de­ laying this move: for example, 7 ii'd2 a6 8 tLige2 tran sposes, while 7 dS tLies 8 ii'd2 (8 f4 l2Jed7 9 lLif3 tLics with ideas like ... c6, ... ii'b6/aS and ... l2Jg4 is good for Black) 8 ... c6 is comfortable for Black, as it is not easy for White to develop his kingside naturally without the loosen­ ing f4-advance. 1 a6 8 ii'd2 �b8 So now we have reached our main set-up, except that White's bishop is on gs instead of e3. If White plays a quick ...

Instead 9 tLic1 does not mix well with il.gs because d4 lacks protection: for example, 9 ... es 10 dS l2Jd4 11 t2Jb3 cs 12 dxc6 bxc6 13 t2Jxd4 exd4 14 ii'xd4 .rtxb2 with a great position for Black. White can also play 9 h4 hS (riskier is 9 ... bs 10 hS) when both 10 il.h6 and 10 o-o-o bS 11 il.h6 transpose to the main lines with 6 il.e3 (Chapter 1S). Instead White has tried 10 tLlds, but after 10 ... bs (also possible is 10 ...l2Jh 7!? 11 il.h6 il.xh6 12 ii'xh6 e6 13 t2Je3 bS with unclear play in J .Piket-F.Nijboer, Groningen 1992) 11 o-o-o bxc4 12 g4 tLixds 13 exds l2Jb4 14 t2Jc3 cs! Black enjoyed good play in E.Vegh-L.Vogt, Eger 1984. 359

A ttacking C h e s s : Th e King 's I n dian, Vo l u m e 1

Fl) 9 �cl

have been good for White, Black should play 14... b4! lS llldl (or lS lllce2 es! 16 dxe6 fxe6 17 h3 gs 18 fxg s hxgs 19 Ji.xg s lllf2! - Golubev) 1s ... es! 16 dxe6 fxe6 17 h3 gs 18 fxgs lll h7 19 lllh s hxg s 20 hxg4 gxh4 21 lll x g7 'it>xg7 which looks okay for Black according to Golubev. 10 lll a s!? ...

We have seen this move before, but here White actually intends to head into a Benoni structure rather than expand on the queenside. 9 Ji.d1 lo ds Instead 10 llld1 follows the recipe from Line C of Chapter 19, but it is not clear that White's bishop should be on gs in this kind of position. Black should continue 10 ... bs (or even 10 ... e6!?) 11 cs e6. The subtle 10 b3 is more popular. Then after 10 ... bs 11 ds lll e s (11 ...llla7! ? 12 lllg 3 looks strange and 11 ...llla s?! 12 cxbs axbs 13 llld4 looks good for White, as 13 ...b4 14 lllcbS! is strong) we have: a) 12 cxbs axbs 13 llld4 'ii'e 8 (not 13 ...b4?! 14 lllcbS!) 14 f4 b4! 1s fxes bxc3 16 'ii'e3 (16 'ii'xc3 lllxe4) 16 ... dxes and Black has good counterplay accord­ ing to Golubev. b) 12 f4 llleg4 13 lllg 3 (13 h 3 b4!) 13 ...h6 14 Ji.h4 and now, instead of 14 ... lll h7, as in A.Dreev-M.Golubev, Alushta 1994, when lS Ji.e2 would ...

3 60

Of course 10...lll e s is also possible, but after 11 b3 Black will have some concern over the stability of his es­ knight. 11 lll g3 The e2-knight will have to move at some point anyway, so there is no rea­ son to go 11 b3?! cs 11 llld4 cs 12 lllc2 bS and Black's play is well underway.

11 cs ...

A good alternative is 11 ...bs!? 12 cxbs axbs 13 lllx bs (White could also try 13 b4 lllc4 14 Ji.xc4 bxc4, but this looks like a decent version of this kind of structure for Black - White's pieces are oddly placed) 13 ... Ji.xbs 14 'ii'xas Ji.xfl lS 'it>xfl �xb2 16 'ii'xe7 'ii'a8! with excellent play.

Pa n n o Va riati o n : O t h e r L i nes 12 Ji.d3

No better is 12 Ji.e2 bs 13 cxbs (or 13 b3 bxc4 14 bxc4 �b4 15 llld1 Ji.a4!) 13 ... axbs with good play for Black. 12 ...bs 13 b3 bxc4

Hazai suggests 13 ... es here, but prefer to focus on the queenside.

Another possibility is 9 ...lllas 10 lll c 1 cs. After the text, we have a split: F2a.: 10 .!tla3 F22: 1042Jd4

I

14 bxc4 �b4 1s llld1

This was S.Mamedyarov-1.Chepa­ rinov, Antalya 2004. Now 1s .. .Ji.a4!? 16 o-o lll d 7 with ideas like ...lll b 6 and ...llle s, attacking the c4-pawn, would give Black good counterplay.

F21) 10 lllg3 c6

The alternative 10... cs is fine too, but the text is the more challenging idea.

11 Ji.e2

Black has good play after 11 f4 llle d7 11 a4 'ii'a s. 11. ..bs 12 cxbs cxds!

F2) 9 dS

9 llle s ...

A nice pawn sacrifice. It is not easy for White to keep control. 361

A ttacking Ch e s s : Th e King 's Indian, Vo l u m e 1 13 exds

White has also tried 13 Ji.xf6 Ji.xf6 14 lllxd5 Ji.g7 15 a4 axb5 16 axb5 (16 Ji.xb5 Ji.d7! gives Black good compen­ sation) and here:

15 Ji.xf6

This is a concession, but 15 lllc 3 Ji.a8 with the idea of 16 o-o 'ii'b6+ (Smirin) gives Black counterplay. 1s ...exf6!

a) 16 ...Ji.e6 17 o-o Ji.xd5 was L.Polugaevsky-J.Nunn, European Team Championship, Plovdiv 1983. Now 18 exd5 'ii'b6+ 19 'it>h1 �fc8 20 �acl with the idea of f4 and !k6 would give White the upper hand according to Polugaevsky. b) 16 ... Ji.d7!? 17 f4 (after 17 'ii'a 5 e6 18 'ii'xd8 !Ifxd8 19 lll C 3 both 19 ... d5 and 19 ...�dc8 give Black good compen­ sation for the pawn) 17 ...lllg4 18 'ii'a 5 'ii'e8 19 J:.d1 (not 19 o-o? Ji.d4+) 19 ... h5 (19 ...Ji.e6!?) 20 o-o Ji.xb5 21 Ji.xb5 !Ixb5 22 'ii'a7 !Ixb2 23 h3 lllf6 24 lllxe7+ 'it>h7 25 !Ixd6 'ii' a8! gave Black good coun­ terplay in M.Sadler-L.Vogt, Altensteig 1992. c) 16 ... e6!? 17 lllc 3 'ii'c 7 18 o-o l:.d8 also looks fine for Black. The two bish­ ops promise him good long-term com­ pensation for the pawn. 13 .. axbs 14 lll x bs Ji.b7 .

362

Worse is 15 ...Ji.xf6 16 llle4 Ji.g 7 17 o-o and White has organized his forces. 16 f4 lll d 7 Black intends ...f5 with good play. 17 fs 'ii'b6 18 llle4 Ji.a6! 19 lllbxd6 llle s With White's king stuck in the cen­ tre, Black had good compensation in l.Novikov-1.Smirin, Las Vegas 1999. F22) 10 llld4

10 ...cs

Pa n n o Va ria t i o n : O t h e r L in es

Black kicks the knight with gain of time. Another idea is 10 ... �e8!? 11 a4 (or 11 Ji.e2 cs 12 ct:Jc2 bs!) 11 ... e6. 11 ct:Jc2

11 ... ct:Jhs!?, with the idea of .. .fs, are legitimate alternatives. 12 a4

To prevent ...bs. 12...e6

11 ...'ii'e B!?

Black must play very deliberately. Both 11 ... e6?! 12 f4 and 11 ...'ii'a s 12 a4 miss the mark, while after 11 ... bs 12 cxbs axbs 13 Ji.xbs (13 ct:Jxbs ct:Je8 ! ?) 13 ... !Ixbs 14 ct:Jxbs ct:Jc4 1s 'ii'e 2 �as+ 16 'it>f2 ct:Jxb2 (the main point is that 16 ... 'ii'xbs fails to 17 b3) 17 a4 Black did not have enough in A.Dreev-1.Smirin, Biel lnterzonal 1993. I like the clever text move, but 11 ... ct:Je8!?, aiming to support ... bs, and

13 Ji.e2

Instead 13 ct:Ja3 clamps down on the bs-square, but White's development is lacking and 13 ...exds 14 cxds Ji.d7 1s as ct:Jhs with the idea of .. .fs gives Black counterplay. 13 ... exds 14 cxds

Now 14 exds? runs into 14 ... ct:Jxc4. 14... bs

Black has achieved his goal and has good play.

363

Index of Variations

1 d4 lllf6 2 c4 g6 3 lllc 3 Ji.g7 4 e4 d 6 and now: Part I: s lllf3 0-0 6 Ji.e2 es 7 0-0 Part II: s lllf3 0-0 6 Ji.e2 es without 7 0-0 Part Ill: S f3 Part I: 1 d4 lt:Jf6 2 C4 g6 3 lt:Jc3 Ji.g7 4 e4 d6 S lt:Jf3 0-0 6 Ji.e2 eS 7 0-0 lt:Jc6 8 dS 8 dxes - 1s6 8 Ji.e3 8...!Ie8 - 1S8 8 ...lllg 4 - 161 8 llle 7 9 lllel 9 llld2 as 10 a3 Ji.d7 11 b3 11 ...lll c8 - 80 11...lll e8 - 83 11 ...c6 12 !Ibl - 86 12 !Ia2 - 88 12 Ji.b2 - 89 9 b4 lllh s 1o lll d2 - 93 10 'ii'c2 - 9S 10 cs - 97 10 g3 - 101 10 �e1 fs 11 lllg s lllf6 12 Ji.f3 c6 13 'ii'b3 - 107 13 bs - 110 13 Ji.b2 - 113 ...

364

In dex of Varia t i o n s

13 Ji.e3 - 118 12 f3 12 ... c6 13 'it>hl - 126 13 Ji.e3 - 128 12 ...'it>h8 13 Ji.e3 - 132 13 lll e 6 - 134 13 cs - 137 13 �bl - 139 9 Ji.gs - 14S 9 Ji.d2 - 149 9 a4 - lSl 9 'it>hl - 1S3 9 lll d 7 10 Ji.e3 lo lll d 3 ts 11 Ji.d2 lllf6 12 f3 f4 13 cs lllg 6 14 !kl - 29 14 cxd6 cxd6 lS �Cl - 20; lS lt:Jf2 - 27 13 g4 - 34 lo f3 fs 11 g4 'it>h8 12 Ji.e3 - 39 12 lll d3 - 40 12 lll g2 - 42 12 h4 - 44 •.•

10... fs 11 f3 f4 12 Ji.f2 gs 13 a4

13 b4 - 49 13 llld3 - Sl 13 lll bs - ss 13 �Cl - S8 13 ... as

13 ...lll g 6 - 64 14 lt:Jd3 b6 15 b4

lS Ji.el - 67 1s...axb4

16 lll bs - 69 16 lllxb4 - 73 365

A ttacking C h e s s : The King 's I n d i a n , Vol u m e 1 Part

II:

1 d4 lt:Jf6 2 C4 g6 3 lt:Jc3 Ji.g7 4 e4 d6 S lt:Jf3 0-0 6 Ji.e2 eS 7 Ji.e3

7 ds as 8 llld2 - 204 8 Ji.e3 - 207 8 0-0 - 210 8 h3 - 214 8 Ji.g s h6 9 Ji.h4 llla6 10 llld2 'ii'e8 11 a3 - 223 11 o-o lllh 7 12 a3 Ji.d7 13 b3 hs 14 f3 Ji.h6 lS Ji.f2 - 229 lS 'it>hl - 230 lS l:.bl - 234 7 dxes dxes 8 'ii'xd8 l:.xd8 9 Ji.gs 9 ... !Ie8 - 241 9 ...lllbd7 - 247 9 ... c6 10 lll xes !Ie8 11 o-o-o lll a6 12 lllf3 - 2S3 12 f4 - 2 S S 1 2 !Id6 - 2S9 1 ...lllg4 s Ji.gs f6 9 Ji.h4

9 Ji.cl lll c 6 10 ds 10 ... llle 7 - 171 10 ... llld4 - 171 9...gs

9 ...lll c 6 10 ds llle 7 11 llld2 11 ... lllh 6 - 176 11 ... hs - 180 10 Ji.g3 lll h 6 11 ds

11 cs - 184 11 h3 - 188 11 dxes - 190 11 ...fs - 194 11 ...lll d7 - 197 Part I l l : 1 d4 lllf6 2 c4 g6 3 lll c 3 Ji.g7 4 e4 d6 S f3 0-0 6 Ji.e3

6 lllg e2 - 3S9 6 Ji.gs lllc 6 7 lllg e2 a6 8 'ii'd2 !Ib8 9 !Icl - 360 366

I n dex of Varia t i o n s

9 ds - 361 6 lllc6 7 lllge2 7 'ii'd2 - 34S .••

1 ... a6

7 ...!Ib8 - 3S1 8 'ii'd 2

8 lll c l - 347 8 h4 - 347 8...!I b8 8 ... l:.e8 - 3SS 9 h4

9 lllc1 es 10 lll b 3 exd4 11 lllxd4 11 ...lllxd4 - 303 11 ...lll e s - 306 10 ds llld4 11 lll 1 e2 - 308 11 lllb3 11 ... lllxb3 - 313 11...c s 1 2 dxc6 bxc6 1 3 lllxd4 exd4 1 4 Ji.xd4 !Ie8 1s Ji.e2 ds 16 cxds 16 ... cxds - 317 16 ...l:.b4 - 318 9 0-0-0 - 320 9 g4 - 321 9 Ji.h6 9 ... Ji.xh6 - 326 9 ...bs - 328 9 a3 - 331 9 !Ibl - 33S 9 �c1 - 339 9 �d1 - 349 9 ds - 3so 9... hs

9 ... es - 266 9 ...bs - 268 10 0-0-0

10 lllc1 es 11 ds llld4 12 lllb3 367

A ttacking C h e s s : Th e King 's I n d i a n , Vo l u m e 1

12 ... lllxb3 - 294 12 ... cs 13 dxc6 bxc6 14 lllxd4 exd4 15 Ji.xd4 !Ie8 16 Ji.e2 ds 11 cxds - 297 17 es - 300 10 bs 11 Ji.h6 11 llld s - 216 11 lllf4 - 280 ...

11 es ...

11...Ji.xh6 - 284 11 ...bxc4 - 285 12 Ji.xg7 - 288 12 lll d s - 281

368

2free updates of this book available within a year of publication at www.everymanchess.com