• Author / Uploaded
• Usuário Número 1647

• Categories
• Campeonato mundial de ajedrez
• Ajedrez
• Juegos de mesa tradicionales
• Juegos de estrategia abstractos
• Competiciones de ajedrez

Citation preview

Dynamic Decision Making in Chess by

Boris Gelfand with invaluable help from Jacob Aagaard

Quality Chess q ualitychess.co. uk

www.

First edition 20 1 6 by Quality Chess UK Ltd Copyright © 20 1 6 Boris Gelfand

DYNAMIC DECISION MAKING IN CHESS 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. Hardcover ISBN 978- 1 -78483-0 1 3-7 All sales or enquiries should be directed to Quality Chess UK Ltd, Suite 247, Central Chambers, 1 1 Bothwell Street, Glasgow G2 6LY, United Kingdom Phone +44 1 4 1 204 2073 e-mail: [email protected]. uk website: www. qualitychess.co.uk Distributed in North America by National Book Network Distributed in Rest of the World by Quality Chess U K Ltd through Sunrise Handicrafts, ul. Poligonowa 3 5A, 20-8 1 7 Lublin, Poland Typeset by Jacob Aagaard Proofreading by Colin McNab & Andrew Greet Edited by John Shaw Cartoon on page 1 3 by kind permission of Jose Diaz Photo page 47 courtesy of Biel Chess Festival Official FIDE photos by Anastasia Karlovich pages 79, 273 Photos pages 1 5 1 and 24 1 by Ilya Odessky Photo page 1 97 by Maya Gelfand Photos and illustrations on pages 274-277 by Max Meltser Cover design by adamsondesign.com Cover photo by Fred Lucas; effects by Max Meltser Printed in Estonia by Tallinna Raamatutriikikoja LLC

Contents Key to Symbols used Publisher's Foreword Introduction

1

2 3

4

5 6

7

8

Minsk 1979 Petrosian

Tactics at the Top Level The Nature of Tactical Mistakes at the Top Level Compensation T ime Dynamic Masterpieces Dynamic Defence

Appendix - Borenka will Remember! Name Index Game Index Opening Index

4

5

7

19

33

47

79

121

151 197

241

273

278 281

283

Key to symbols used ± +

+-+

Cii

� m

?? !! !? ?! #

White is slightly better Black is slightly better White is better Black is better White has a decisive advantage Black has a decisive advantage equality with compensation with counterplay unclear a weak move a blunder a good move an excellent move a move worth considering a move of doubtful value mate

Publisher's Foreword The last few years have been everything I could ever dream of as a chess writer. My books are selling enough for me to buy new soles for my second-hand shoes, readers from all over the world are telling me that they enjoyed the books, and strong players are even pretending that they were useful for them in their tournament preparation. But the biggest joy has been working with Boris Gelfand on this project. Boris loves chess immensely and it is impossible not to fall in love with the game all over again when discussing it with him. Our analysis sessions have been spirited and enjoyable, and I have been able to learn a lot about the game from them, all of which is hopefully included in this book! Writing a book is a difficult job, even when it is co-writing. You still have to choose the right words, structure, restructure and then restructure some more. A point made in August might be easier to understand if added to a game analysed in February. You get the idea. When the ideas are not in your head, but in someone else's, this does not become an easier process. What has made writing these two books amazing is the time spent with Boris. His warmth and wit dominate our conversations. I laugh more in our sessions than at any other time during a normal week. I will leave you with one extract from one of our conversations in 20 1 4 :

Boris: Jacob: Boris: Jacob: Boris: Jacob: Boris: Jacob: Boris: Jacob: Boris: Jacob: Boris: Jacob:

Hi Jacob, how are you? Good thank you, and you? Is it raining in Scotland? What do you mean ? Water falling from the sky. Well, it's Scotland. It is always raining a little bit. Here we had no rain for two weeks. How warm is it? Heat wave. What does this mean, heat wave? Eighteen degrees. Ha! Here it is thirty-five degrees. You should come and visit!

Then our conversation was interrupted. A siren rang out weakly somewhere outside Boris's house. He stood up immediately.

6

Boris:

Boris Gelfand - Dynamic Decision Making i n Chess Sorry, I will be back in ten minutes.

Boris returned ten minutes later.

Boris:

So, Jacob. Is it still raining in Scotland?

Being a part of this project is an ongoing joy. I hope some of this joy has spilled over into the pag es and makes this not only an instructional book, but also a pleasure to read. Jacob Aag aard Glasgow, May 20 1 6

Introduction At the beginning of this year, I played in a rapid tournament in Estonia. I won one game where my opponent, an old room-mate from training camps with Petrosian, failed to show up. He came up to me later and apologized. He had been very excited about playing against me, and got so caught up in the preparation for the game that he lost all sense of time. When he finally looked at the clock, the game was already over, and he was still in his room. .. Chess is all about time. Each player makes a move, choosing to move only one piece, hoping that all his pieces will be in time to reach the necessary squares. If you are late, the opponent will checkmate you, queen his pawn , take your knight, skewer your rook or maybe j ust run away with half the kingdom and all three princesses. Chess is also about handling the clock. Before each game you are entrusted with a certain amount of time, and you have to do your best to spend it wisely. This book is about dynamics. The things that easily fall victim to time, have an unstable foundation, and erode quickly. This is not an academic textbook about dynamics; it is a deeply personal book, with dynamics as the central theme. Other recurring themes I shall discuss include the influence of engines on modern chess, and the clash between the millennials and my generation, both in chess style and in the public imagination. But if I think one of my experiences or opinions will interest the reader, then I shall mention it, even if it is 'off-topic'. If you can survive these digressions, you will be less shocked later in the book when a recipe for strawberry j am appears...

Early investigations I worked with Eduard Zelkind until I was 1 1 years old. At this time, he moved to the US and asked his friend Albert Kapengut to look after his best student. Ilya Smirin arrived later on, when he moved to Minsk to study at the university in the second part of the 1 980s. I was maybe six or seven when Zelkind gave me tactical exercises to solve at home, sacrifices on h7 and other such basic stuff. I had not learned to read and write yet, so my father would write down the positions and the solutions for me, so I could bring them to the next training session. This continued until I learned to read and write myself. Zelkind had his own system for writing down positions in the most effective way, which others might look at as an unbreakable code. At this time there was no access to such luxuries as a photocopying machine. All exercises would have to be written down by hand at the end of the lesson.

8

Boris Gelfand - Dynamic Decision Makin g i n Chess

Lacer, when I went co Moscow and visited the Central Chess Club, they did have a photocopying machine. le required a good deal of pushing, but Kapenguc always managed co gee chem co copy the games from the latest cop tournaments for me. To explain this co the young er generation: he was downloading TWIC by hand . . . OK, it required quite a bit more work than chat, which we will talk more about in the volume on the opening . Incidentally, one of the explanations for the collapse of the Soviet Union is the arrival of the g enerally available photocopier and other forms of technical prog ress in the 1980s . . . I n che first volume we discussed m y favourite book when I was young, Rubinstein with game annotations by Razuvaev. From this I learned about space and other positional factors. The big hero in my study of dynamics was Efim Geller. More about him in Chapter 1 on the Soviet Championship in Minsk 1 979. However, I should immediately nullify the dichotomy I have just sec up: obviously I was learning dynamics from Rubinstein and Pecrosian as well, and positional thing s from Geller, Tai and Spassky. Chess is a complex game and the cop players have always been able co do a bit of everything. This is natural, as positional play and tactical opportunities are so closely interwoven chat it can at times be hard co see the difference.

Sort of defining dynamics To me, dynamics is the potential of the pieces. Many players are good at calculation, but this does not mean chat they have any feeling for the potential of the pieces. AI; we are not computers, calculation is not enough, we need direction in our calculation and a strong sense of dynamics, so a feeling for the potential of the pieces is important. Technically, a lot of moves are possible and you cannot calculate

everything , so you need a system of some sort co decide which candidate moves you want co focus on. There are some strong grandmasters who attempt co calculate everything, and do so with moderate success, with the key word being moderate. I am chinking of one player in particular. He certainly has his own philosophy about chess, which he chinks serves him best, but as he is not able co defend his point of view here (should he even wane co!) I do not want to mention any names. In pose-mortem analysis, he will sugg est all the moves. Almost like a computer does. Bue as I see it, the human brain does not work like a computer. Our thought processes are not linear; although we should cry co structure chem, we cannot be completely systematic in our chinking. This is why we use techniques such as candidate moves, why we recheck variations, why we look for intermediate moves, why we try to insert some moves. Calculation has co be done smarcly, or it will not be effective. le is easy co waste a lot of energy. AI; I talked about in Positional Decision Making in Chess, in some positions there is not really anything co calculate, and we need to make a decision based on other criteria. This book deals with some situations where we have co make intuitive decisions and some where we have co calculate. In dynamics, both skills are essential. The initiative is a somewhat different topic. You have better piece development or something like this. You need co use the momentum to transform it into something valuable, an attack, a positional advantage, or whatever, before it goes away. Essentially it is Steinicz's idea chat you have co use your advantag e before it disappears. Obviously, this is all about dynamic advantag es. With static advantages, the urg ency is often much less.

Introduction Every player tries to find their own balance between intuition and calculation. The method that works well for one player might not work well for another. In a way I am quite intuitive, but this does not mean that I think this is the 'right' way to be. All I can say is that it works for me. But ignoring the influence of intuition in our thought process would be a mistake.

9

Hort and Jansa. I worked on it when I was young, but luckily my memory is not perfect and the book is still very useful for me now. For those who know my games, it will be no surprise that I was heavily influenced by Polugaevsky's books. Especially the long analysis of the Najdorf variations and of various endgames, most memorably against Geller and Gligoric.

Not exactly calculation and tactics Dynamics is not exactly the same as calculation and tactics, just like a chair is not just wood and bolts. But imagine a chair without the raw materials and you will most likely imagine an empty space on your living room floor. This book is about dynamic thinking, but will constantly deal indirectly with calculation and tactics. I hope this distinction will not confuse anyone. Although many references are made to calculation, this book does not deal directly with the technique of calculation, so I think I should briefly recommend a few other sources for this. I like Jacob Aagaard's Grandmaster Preparation series of books, where the first volume, Calculation, even has a foreword by a clever Israeli guy. . . I had to mention this book, not only because it is very useful, but also since Jacob is my assistant in writing these books, and getting him to type this will embarrass him no end. I also greatly appreciate Perfect Your Chess, a book by V ladimir Grabinsky, a highly successful Ukrainian trainer, written together with his student, the strong Grandmaster Andrei Volokitin. Grabinsky is a world-class trainer who has worked with close to ten players from their early years until 2650 level. The exercises in this book are very difficult, but they will benefit the diligent student immensely. While writing this book, I am going through an old book, The Best Move by

To me, tactics is mainly about patterns and about calculation. I have dealt to some extent with tactics and calculation in this book, but I do not want to offer a big theory of calculation. You can find this in the writings of Mark Dvoretsky and Jacob Aagaard. Aagaard follows Dvoretsky's way of thinki ng, so you will not find any noticeable difference between the advice given by them. I want to re-emphasize that each player has their own approach. What is good for one player might not work for another. There is even an argument to be made for the idea that some approaches work well against some players, while they are ineffective against others. We all have some opponents we score well against, who again have their favourite customers, who again are delighted whenever they have to play us. Alexander Beliavsky told me that he once witnessed Tal, Korchnoi and Stein playing 'winner stays on' blitz . Tal would lose to Korchnoi, who would lose to Stein, who would lose to Tal. There were almost no exceptions to this, he said. Korchnoi had a great score against Tai in tournament games as well. Tal was very intuitive and saw long variations as inspirations more than as knuckle-down-hard calculation. Korchnoi, on the other hand, did not have as great intuition as Tal , but he worked systematically through the variations, finding flaws that brought down Tal's great ideas. This approach also worked very well for

10

Boris Gelfand - Dynamic Decision Makin g i n Chess

Polug aevsky in his Candidates match against Tal. As Kapengut was Polugaevsky's second at that time, I heard a lot of inside stories about this match as a child.

Calculate without a board When I was a young player, a lot of people paid a lot of attention to the fact that I often do not look at the board while I am playing, while now they have gotten used to it. From an early age I was able to keep the board easily in my head. It is an important quality that should be developed. When I play in tournaments, I will discuss the day's games with teammates, seconds or friends over dinner or while going for an evening walk. We hardly ever use a board, but analyse the games blindfold. I will say more about this in Chapter 2, page 39. I have noticed that some young players are not great at this, which I am sure will damage their results.

What does a Grandmaster see during the game? Honestly, I have no idea if amateurs watching the games live overestimate or underestimate how much a grandmaster actually sees during the game. One of the problems is chat most people watch live games with the engine running, or on a website that has the engine turned on by default. I never get tired of pointing out chat looking at the engine makes you blind. It is a rare person who sees past the engine, and can see the board as it is. I cannot do it well, but in my experience, I do it better than most. A big difference between engines and grandmasters is chat an engine looks at every move, while a grandmaster can only look at some of the moves. Obviously he will have a good intuitive understanding of what is

going on, but he will still be limited, and miss things from time to time. There was a survey on ChessBase.com about the top players and their correlation with the engines . On the low end, you found Magnus Carlsen and myself, for different reasons. Clearly this has nothing to do with the actual strength of the players. One of the reasons I have only a vague idea how much my colleagues see during the game is that few books (if any) are written from the perspective of decision-making, as the books in this series are. There are some players who like to pretend they saw everything when they annotate the games afterwards, while some of my more intellig ent and devious colleagues deliberately cry co obscure what they saw and did not see, to avoid g iving anything away. In recent years we have been given a window into the workings of a grandmaster's mind with online transmissions and commentary. We see strong grandmasters commenting on their colleague's games, and some of them seem to see a lot, while others seem to see very little. Certainly this cannot be used as a valid way co predict anyone's playing strength! We also have videos of the players' pose-mortems, or it's done live at press conferences. It is a different format, but still it gives us an insight into how difficult chess is, which is one of the reasons people tend to enjoy watching this form of commentary more than those who simply read aloud from the first line of the computer. I love watching the post-mortems of big tournaments, although I would still prefer to play in chem (please send me more invitations!) . It is of course worth using computers for many things, but not during the game. Chess is a beautiful game where commentators see one thing, computers something else, and the players something different again.

I ntroductio n

Human versus Computer Matches I saw a match recently where a computer played with a material handicap against Nakamura. I find it rather insulting. Actually, I was once offered a similar match, but I suggested that instead I would give the computer pawn odds, but in return the computer had to play with a 286 processor. For some reason these computer people disappeared immediately. My chinking is that of course the programming has improved, but these engine people are mainly riding on the hardware achievements of faster processors. I don't have a problem with this, but there is no reason to seek to humiliate some of the best chess players in the world, and make chess a lesser game in the process. Of course this is not the intention; they did not see the problem. I had another offer at some point, where I also asked a few questions: " Would you allow the computer to play without an opening book?" " No." "Why not?" " Because we want to beat you and increase our sales. And in the end we will meet someone who will play on our terms." Kasparov played a few matches with computers, which I did not enjoy. They were briefly damaging for chess. These matches were interesting for the media and one player rated over 2800, probably because of the remuneration. It seemed to me that people who follow chess did not show great interest in these matches and would much rather have seen Kasparov play in a top event.

A quick detour into ageism I want to quickly sidestep the question of what is the best way to learn to play chess, and talk for a moment about my generation and the one that comes after me. I once read a short

11

piece of statistical exploitation. It showed that the average age of the Top 1 0 was around 3 1 years old, the average age of the Top 2 5 was around 3 1 years old and the average age of the Top 1 00 was, you guessed it, around 3 1 years old. Although this was just one rating list, it was still a strikingly consistent picture. I would certainly have expected to find bigger swings in the average age. But there is this idea at the moment that "chess is getting younger", promoted by a number of retired or semi­ retired top players. But I do not see this as being the case. When I played the match against Anand, we had the oldest World Champion since Botvinnik, defending against the oldest challenger since Korchnoi. Combined, you would have to go back to pre-war records to find a higher average age. The average age of the top players would of course be much higher if these great players from the past had not retired or resorted to globetrotting to holiday destinations, caking nine-move draws and so on. But after many years of hard work, I can understand the desire to relax in the sun. There are those who would argue that Carlsen was already stronger than Anand and me in 20 1 2, and we somehow have an extraordinary case because one player did not play in the qualification. I share their view. In the same year, my favourite football club, the brilliant Barcelona, at the height of their powers, lost in the semi-final in the Champions League to Chelsea, a team old players that was off form in general, finishing only 5 th in the Premier League. Still, they played best in the Champions League and deservedly won it. Also, I do not see the absence of one player from the cycle as that extraordinary. Petrosian was not the best player in the world all the time he was World Champion, and Fischer exited

12

Boris Gelfand - Dynamic Decision Makin g i n Chess

the cycle that ended with Spassky's victory in 1 969. Botvinnik once did not make the USSR Olympiad team, despite being World Champion. And so on. . . But this will of course not change anybody's mind. In the 2 l st century the importance of public opinion has increased tremendously, and nothing seems to be able to compete with it. It makes a greater impression to have moral support from the media than to win qualification events . . . I can understand the attraction people had from roug hly 20 1 31 4 to a match between Aronian and Carlsen, who was rated Number 2 in the world pretty consistently around that time. But chess is a sport and the rating system reflects current form and all results evenly; it does not win championships, and neither does public opinion. But it does not win championships and neither does public opinion. A lot of people who follow both chess and football, as I do, will care tremendously about the FIDE rating system, but know nothing about the current FIFA rankings (or maybe even that such a list exists) . I gave a friend ten guesses to come up with the Number 1 on that FIFA list in January 20 1 6. He tried the usual suspects, Germany, Argentina (who not long after were indeed Number 1 ) , Brazil, Spain, Italy, Netherlands and so on. The correct answer is Belgium. Based on their average results and current form, this was the best team in the world. Belgium has never won any titles, but I am reluctant to downg rade them too much, as it is really a strong team and Euro 20 1 6 might be their moment!

In the same way in chess, the highest-ranked player can at times win the most important tournaments and at times not. It is a sport; the result is decided over the board, not by a public vote. At the time of writing , Vladimir Kramnik is Number 2 in the world rankings. He is a classically-trained player who aimed to be the "future of chess" , when he won Gold with Russia in the 1 992 Manila Olympiad, as well as Gold on Board 6. Since then he has changed his style many times and continues to impress. This demands the highest possible praise. Sadly, he did not manage to qualify for the Candidates in 20 1 6, but as he is 40, the media has not made a meal of this . . . I n 20 1 4 a lot o f chess fans were unhappy that Nakamura was not in the Candidates in Khanty-Mansiysk. He had failed to qualify, and also had not had a great 20 1 3 . I, on the other hand, had won and tied for first in many big events in 20 1 3, but only my wife said I should have gone! All credit to Hikaru: he did not complain, but went to work, played excellently in the following Grand Prix and convincing ly qualified for the 20 1 6 Candidates. At the end of the day, the question is if chess is a sport or showbiz. For me it is clearly a sport.

Introduction

A bit about engines I suggest chat chose who scare out young should study chess without computer assistance for years, in order to understand the game before you use this powerful cool. As Kasparov said: the main thing co understand about engines is when co turn it on and when to turn it off. A lack of understanding of working with engines is persistent with some young players (not all of course!) . There is one example I saw recently, where a young player was facing an experienced veteran. The young player had a promising attack. Ac some point he had two interesting possibilities. One of them looked tempting, but did not work, while the other won in a very nice way. The young player quickly went the wrong way and lost even quicker. Not many minutes passed from the end of the game until the young player was complaining

13

about his lack of fortune on Twitter. He did not understand how he could lose a position with +5 and found his opponent co be terribly lucky. He did not mention the face chat he was personally unable co accurately calculate a few moves ahead. Instead of analysing the game and understanding what he did wrong, he succumbed co the attraction of being the centre of public attention. I chink it is fair co say chat such an approach is unconducive to one's chess progress. And this is not only about one player, this is a general trend I see. I want to emphasize the main point: there is a big difference between misplaying an absolutely winning position and missing a brilliant combination. Boch are +5 on the engine, but only one is an 'eternal' +5. To describe chem as interchangeable shows no understanding of chess {or engines) whatsoever.

14

Boris Gelfand - Dynamic Decisio n Making i n Chess

Don't forget Rubinstein Positional Decision Making in Chess focused

heavily on the inspiration I drew from the games of Akiba. He is the foundation of my positional understanding of chess, as explained in that book. Although Rubinstein was not known for possessing great tactical skills (indeed, he did not) this does not mean that I did not pick up some hints of dynamics from looking through his games. The following game was one of the games that influenced me most when I was a young boy.

This looks rather strange. It would be more normal to play something like 1 O.cxd5 exd5 l I ..ie2 or I O .Y;Vc2. Many games have been played from this position; you can study them if you are interested.

10...�e?!?

George Rotlewi -Akiba Rubinstein Lodz 1907

I .d4 d5 2.c!lif3 e6 3.e3 c5 4.c4 c!lic6 5.�c3 �f6 A very interesting pawn sacrifice, which even with a 2 1 st century sensibility looks very powerful . When we looked at it, my first reaction was that it looked absolutely devastating should White accept it. But computer analysis shows this is not the case, and that it was sensible to accept the challenge.

11.J.d3? This j ust loses a tempo. a

b

c

d

e

f

g

h

One of the important things we have learned from this game is that in symmetrical positions, the value of an extra tempo is increasingly important. In this g ame Rubinstein plays very energetically, trying to make useful moves at every turn, aiming to win half a tempo here and there.

6.dxc5 .ixc5 7.a3 a6 8.b4 J.d6 9.J.b 2 10.�d.2

0-0

Critical was: l l .cxd5 exd5 1 2.lDxd5! lDxd5 l 3 .Y;Vxd5

15

I ntroduction We looked at a few different options: 1 3 .. . lll xb4? l 4.axb4 ixb4 t l 5 ..... ,....;,� ��-, �� �

-�!�li�i �� ,� �� �

1 2 ...b 5 1 3bb 5 13.lLixb5 has scored a solid 2-0 for White, but Black can improve with 13...lLib4! 14.llJxd6 .ixd6 15.�xd6 E:fc8 with enough compensation for the pawns. 7

6 5

4 3

....J�f% �� �� ��-0 a

b

c

d

e

f

g

h

22.lLig5! A nice intermediate move. White is planning E:xd3-h3. Black is busted. c) 15 ... a6 is also possible and what the engine suggests. Black is threatening to take on b 5 , which would open the a-file, s o White has to play 16.lLibd4, when after 16...lLixd4 17.lLixd4 White is still much better. The engine says that Black is more or less OK, but I have analysed a lot of positions like this, and believe that it will take between half an hour and an hour with a computer to prove that White has a winning attack. 16..ixf6 .ixf6 16... gxf6 17.lLibd4 gives White a solid edge. The black knight is exchanged before it gets a chance to get into battle, and Black is left without an attack, but facing one on the kingside all the same. 17.lLixd6 .ic3 18.�e2 b3 19.cxb3 llJb4 20.bxc3 E:xc3 2 1.lLid4 Black's attack is not happening. 2 1...e5 22.lLic4 E:xc4 23.ixc4 ixg4 24.�g2 ixd l 2 5.E:xd l exd4 26.E:g l g5 27.fxg5 hxg5 28.E:fl E:b7 29.E:f5 �b6 30.E:xg5t @IB 3 1.E:g8t @e7 32.�g5 t @d7 33.�f5t 1-0 lvanchuk - Smirin, Klaipeda 198 5.

a

b

c

d

e

f

g

h

The engine wants us to play l 6.E:d4 E:c6 17.� e5 E:c5 18. � d6 E:c6 with a repetition. Solve the draw death: take a sledgehammer to your laptop...

1 3.. J:UcS 1 4..ic4?! White should probably vary earlier, on move 12, but if you do reach this position, then a move to investigate is 14.e5.

1 4...�b4 1 5.ti°e2 White might have had better saving chances in the following line: 15 ..ib3 E:xc3! 16.�xc3 E:c8 8

7

6 5

4 3

2 a

b

c

d

e

f

g

h

White is struggling, but not dead yet.

24

Boris Gelfand - Dynamic Decision M aking i n Chess

White should avoid 1 7.ixf6 ixf6 1 8.We3 ia4 1 9.l'l:d2 ixb3 20.Wxb3 lll xa2t 2 1 .dI ll'lc3t 22.bxc3 Wa l t 23.e2 Wxh l 24.f2 ixc3 when his situation is grim. Bue 1 7.We l ia4 1 8.a3 ixb3 1 9.Wxb4 Wxb4 20.axb4 ixc2 offers at lease some hope. Black is certainly better, with his threatened discovered checks, and the weakness of e4 and b4, but the game is not over yet.

1 5 J�xc4! •.

Black's attack flows; if the bishop had reached b3, it might have been a useful defender.

16.flYxc4 gc8 17.flYb 3

8

7

6 5

4 3

2

1

a

b

c

d

e

f

g

h

A very energetic sacrifice, continuing Black's c-file rampag e.

19.flYxc2 White is running out of options. 1 9.tll xe4 tll d 4t wins the queen, while keeping a vicious attack. l 9.l'l:d3 ixh4 20.ll'lxh4 d5 also gives Black a winning attack.

I 7 tllxe4!! •••

Black refuses co lee any defender block the c-file.

18.a3 This does not help, but neither would any other move. 1 8.ll'lxe4 ia4 is devastating. And after 1 8.ixe7 ll'lxc3 1 9.bxc3 l'l:xc3 20.ixd6 ll'lxa2t 2 1 .b2 l'l:xb3t 22.cxb3 ic6 Black wins comfortably.

18... lll xc2!!

19 tixc3 20 ..lel .if6 21 .tieS •.•

There are many wins. Kupreichik finds a simple one.

Chapter 1 - Minsk 1979

21...Aa4 22.i.xc3 .ixc2 23.J.xa5 .ixdl t 24.©xdl dxe5 25.fxe5 .ixe5 26.b4 l:k3 27.a4 ga3 28.© c2 Ad4 29.gdl e5 0-1 When you are watching this live in the tournament hall as a child, you cannot help but be overwhelmed. This is one of the reasons why I think young players should be encouraged to watch top tournament games. If they cannot make it to the tournament hall, then at least follow the games online. Be entirely focused on the games, without online commentary or a mind-numbing engine, trying to find ideas of your own, calculating the various possibilities as the players think about them.

Efim Geller One player in this tournament fascinated me more than the others... Efim Geller's golden years were 1 949 to 1 980. Born in 1 92 5 , he won the USSR Championship qualifier in Tbilisi in 1 949, and subsequently took j oint 3-4th place at the main championship. He became a grandmaster in 1952 and also played for the first time in the Soviet team that year. He was a strong force for these three decades, scoring +6 in almost 200 games against the six World Champions he faced, suffering a majority of his defeats against Spassky, but achieving a plus score against Bocvinnik, Smyslov, Petrosian and Fischer. He was one of the most respected players in the Soviet Union, and acted as second for Karpov for many years. Before Kasparov's 1 993 match against Nigel Short, he asked Geller what he should do against the Marshall Attack, which Short employed at the time. Geller suggested a system with h3, d3 & llibd2 and slow play, which worked well for Kasparov and stayed popular for the next 1 5 years.

25

There were some young players in the tournament. For example, 1 9-year-old Artur Yusupov took second place in his first-ever championship. Artur kindly shared his memories of Geller from this tournament: When I first arrived at the tournament, my impression of Geller was that Grandad had decided to play. I liked that, but at first his results did not impress. In the first seven rounds he made all draws, before winning a fine strategic game against Romanishin. But it was in Rounds 1 0 and 1 1 that everything changed. First, Razuvaev made a horrific blunder in the opening, and lost to Geller in 2 1 moves. Then the next day, Tseshkovsky, in an equal position, blundered his queen right after exiting his adjournment analysis. Geller was an experienced card player, so he immediately realized his luck was in. He transformed completely and played with such energy - beautiful attacking chess. It was truly fascinating. I was able to use this experience much later in my career, in the German Championship, when Alexander Graf blundered his queen against me in an equal position. I remembered Geller, and knew I j ust had to show up and play, and luck would be on my side. In the last round, the next day, I misplayed my position a little, and my opponent offered a draw. I rejected it, because of Geller. I had to play! And immediately my opponent made mistakes. You have to use the luck! This is what I learned from Geller at the 1 979 championship. Actually, Geller had already made an impression on me in the first round against Dolmatov. Sergey had played quickly and confidently in making a draw with Black,

26

Boris G elfand - Dynamic Decision Making in Chess

which was a decent result. Geller, on the other hand, had been thinking a lot, and was low on time when the game ended, as was his habit. When they analysed the game afterwards, Dolmatov would say: "Here I can play this, this or this, but I chose that." Geller said: " Really, you think you can play this?" and then he showed the most beautiful refutations of Sergey's suggestions, one after the other. None of this happened in the game, which was not so interesting. But it seemed that Sergey had been lucky, and by intuition had chosen the only move again and again. For those not blinded by age, but still interested in numbers, let me offer you something solid: by 1 979 Geller was 54 years old and had not played in the Olympiad team for a decade. He was certainly not over the hill though - his highest Elo rating was as recent as 1 976, when 2620 made him Number 8 in the world. By 1 979 he was still Number 28 in the world, whereas Yusupov was Number 1 28 in the world and Dolmatov Number 45. This recollection is quite pleasant for me: absolutely would like it to be the attitude of young players that they have to beat me, no matter what. I have been declared 'finished' for decades. In 1 998 there were a lot of articles where they announced the end of me. After I lost the candidates match to Short in 1 99 1 , I defied the expectations of some people. I went on to win in Belgrade, tying for second with Kasparov in Reggio Emilia 1 99 1 /92 and, shortly thereafter, won in Wij k aan Zee 1 992. After the last of these events, a journalist confided in me, saying that after the match with Short, "I thought you were finished." I was 24 years old... I want to point out that this was a very pleasant person, who spoke from the heart.

I think he meant that I might have needed a few years to recover, but as no one had told me about this earlier, I simply played well and won the tournament.

Boris the Attacker

-

1979

To understand how exciting the following Geller game was to me at the time, it is necessary first to see a game I played j ust before the tournament.

Boris Gdfand Eduard Raisky -

Minsk 1979

1.e4 c5 2.tlif3 d6 3.d4 cxd4 4.�xd4 tli f6 5.�c3 e6 6..ie2 .te7 7.0-0 0-0 8.f4 tli c6 9..ie3 a6 10.a4 ffc7 11.ff el �a5 12.ff g3 tlic4 13.i.xc4 ffxc4 14.e5

8 7 6 5 4 3 2 1

a

14...tli eS!?

b

c

d

e

f

g

h

1 4... lll d5 is more natural, but the move in the game is not bad if Black plays accurately afterwards.

1 5.�e4 b6? Luckily he does not. l 5 ... d5 would allow White to keep a stable advantage with his massive lead in development. A quick shift to the queenside

Chapter 1

-

27

Min sk 1979

would be very effective: 1 6.tll d2 Vfic7 1 7.c4! dxc4 1 8.:gfc l Black is weak on b6, d6 and all the way down the c-file. 1 5 ... f6! was the best move. After 1 6.b3 Wid5 1 7.exf6 tllx f6 1 8.tll x f6t .ixf6 Black managed to equalize in Klovans - Polugaevsky, Yerevan 1 975. I chink White could find an advantage somewhere in this line, but it would not be overwhelming. a

b

c

d

e

f

g

h

2 1 .Vfie5! with the threat of :gxg6t and Wig?#. Black can try 2 1 ...:ga7, but after 22.:gafl mate is close anyway. For example: 22... :ge8 23.:gxf7!

20.�xe5 �c5 The last try. 20... gxf6 2 1 .Vfixf6t c;t>g8 22. .ih6 and wins.

16.£5! White should not waste any time.

16... dxe5 This is the critical test, but as it does not work, Black had to accept chat his position is a disaster. 1 6... exf5 would be poor on account of I 7.exd6, and White has a big advantage with this impressive passer.

17.f6! lt:lxf6 18.�xf6t hf6 19J�xf6 ®h8!?

Black realizes chat l 9... exd4 is hopeless. White not only wins the exchange after 20..ih6, but can choose co go for mate instead: 20...g6

21 ...gxf6 22.�xaS e5 23.�f3! 1-0 After the first nine moves of Geller's I 4ch round game, I was understandably excited:

28

Boris Gelfand - Dynamic Decision Making in Chess

Efim Geller Yuri Anikaev

Today the computer tells us that White was even winning at this point with an amazing sequence.

-

Minsk 1 979

1.e4 c5 2.�a e6 3.d4 cxd4 4.�xd4 tlif6 5.tli c3 d6 6.J.e2 J.e7 7.0-0 0-0 8.f4 �c6 9.J.e3 a6 As we have j ust seen, I won a great game in the Scheveningen j ust before this tournament, so I was very eager to see what Geller would do with the white pieces.

1 6.fxe6 fxe6 1 7.J.g4! llixe3 1 7 ...Wc5 immediately does not improve things. White wins with 1 8.llif5! and Black is mated or loses the queen. 1 8.Wxe3 Wc5

10.a4 J.d7 1 0...Wc? is the big main move.

ll ..if'3 1 1 .llib3! is known to give White a good game. l 1.

tlia5 1 2.°1We2 '!Wc7 1 3.g4 �Uc8 14.g5 �e8 1 5.6 �c4 ••

8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

Black looks OK, but the knight on e8 is so poorly placed that White can crash through down the f-file. 1 9.Wf2! llic7 1 9...Wxg5 20.Wf7t 'it>h8 2 1 .Wfst J.xf8 22.E:xf8# 20.Wf7t 'it>h8 2 1 .Wxe? %Yxd4t 22.'it>h l J.c6

a

b

c

d

e

f

g

h

At this point I felt that Black was at least OK. Yes, the knight on e8 is awkward, but getting the knight to c4 looked like a big achievement. But in reality Black is j ust lost.

1 6.J.h 5!? An amazing idea from Geller, though not the only way to continue.

29

Chapter 1 - Minsk 1979 With the idea of doubling rooks on the f-file. 23 . . . lll d 5 ! ? Black needs t o try something! 24.Wxd6! l'!d8

1 9 . . . c;!.ihs 20 .Wxe? Wd8! Not an easy move to find in your defensive calculations, bur it is necessary. 20 . . . l'!e8 2 1 .lll xe6! leads to an endgame where White has an extra pawn. 2 I .l'!f7 Wxe7 22.l'!xe? lll xe3 23.l'!xd? e5 24.lll de2 lll xh 5 2 5 .l'! c l

25.exd5 ! White wins.

b

a

It is quite easy to forgive Geller for not seeing all of this. We have no idea if he saw the first seven moves of this line, and felt that he would be able to achieve more with what he did, or if he was focusing more on going deep in the direction he went. One thing we do know is that Geller's idea in the game was at least as beautiful and instructive as this variation. The existence of computers today does not reduce the achievements of the heroes of the past, in my opinion.

c

d

e

f

g

h

White is about to win a pawn and will have decent chances in the endgame, although Black still has some active counterplay. I am sitting there in the front row, following the game, thinking the attack is over. The bishop will retreat, after which Black will play . . . lll g7 and . . . l=!f8 with a better position. White simply played too primitively.

16...g6 This is forced. 1 6 . . . lll e5 1 7.ixf7t! is a good example of the dangers Black is facing.

17.fxg6 fxg6 18.�'fl! ll\ e5? This natural move i s the decisive mistake. After the game it is easy to see that Black had to accept the loss of a pawn and play: I 8 . . . lll g7 I 9 .Wf7t 1 9 .ixg6!? lll xe3 would lead to the white bishop being in the way on f7, and is thus not as attractive.

a

b

c

d

e

f

g

h

Then a new wave of the attack began.

30

Boris Gelfand - Dynamic Decision Making i n Chess

19.llJ O! Eliminating the defence of the f7-sq uare.

19... llJg7

1 9 . . . ttJ f7, intending to return to e5 once the f3-knight moves, is refuted by brute force: 8

7

6 5

4

i:� �;i·,·i �..t�A)) � i la" -'---,�.. •� � � �r !� �!�r- >-�

��- 1 � 2,,-� ��-� � � 2 ��ri � 3

�

�

�

JtJ�j .�:rJfj a

b

c

d

e

f

g

20.llJxe5 gm 21.llJf7! lll xh 5 2 l . . . gxh5 was possible, but it would not save Black: 22.�b6! Wc6 23 .�d4 Wc8 24.tli h6t h8 2 5 .Wd2 gives White a winning attack.

� i: fe11�.;r

!��--,��,, � . :.�, �f:;ii'' : ���" � � -��� '�f:;l'ef� ' ,, �� � �f:;/-� � 2i � - /�� j� ��,,f!J � := 6

3

b

B c

d

e

f

g

8 7 6 5 4 3 2 a

h

20.�xg6! hxg6 2 l . tli h4 tli e 5 22.tlixg6! Black's defence of the f7-square did not last. 22 . . . tli xg6 23 .Wf7t �h8 24.Wxg6 White is winning. The threats are too many to count, but we can mention �f7 in passing.

a

White dominates. It is not a direct knock­ out blow, but enough to win without any uncertainty.

h

The main ideas are �xf8t followed by �fl penetrating on the 7th rank, as well as meeting 25 . . . e5 with 26.tli d 5 ! , leading to further suffering for Black. For example: 26 . . . exd4 27.tlixe7 We8 28.�xf8t Wxf8 29.tlid5 and

b

c

d

e

f

g

h

Here I thought: the knight is trapped on f7 , White i s in danger. But again Geller played a move I had not considered at all.

22.llJ d5!! exd5 23.llJh 6t © g7 This allows a beautiful finish, but Black was not getting out of it anyway: 23 . . . �h8 24.�d4t �f6 2 5 .Wxf6t! Obviously Geller would have taken with the queen. 25 . . . �xf6 26.�xf6 And we have returned to the game.

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

24.'!Wf7t!! l3xf7 2s.l3xf7t ©h s 26 ..id4t .if6 27.l3xf6!

Chapter 1 - Minsk 1 979 Black resigned. He 27 . . . lll g7 28.Ei:f7 Ei:g8

a

b

c

d

is

e

getting

f

g

mated:

h

29.�xg7t Ei:xg7 30.Ei:f8t Ei:g8 3 l .Ei:xg8#

1-0 This game might not seem so complicated to us now, but imagine being a small boy in the tournament hall, trying to guess the moves of the grandmasters. I know I used a lot of diagrams in this game, but I am trying to reproduce the experience I had during the game, which was not just an emotional one, one of surprise and turnarounds, but also a visual one. I was watching the games live, watching the players. Back in that time, smoking was still allowed in the tournament hall. Geller was smoking, and Tai was smoking one cigarette after another. This helped to form the atmosphere of the tournament into a 'grown-up area' , where I as a child had been allowed to visit. And didn't they play grown-up chess there! (One more thing about smoking: Back then many sports stars were smokers, not j ust chess players, also tennis stars, runners and footballers. Recently football lost one of its biggest stars, Johan Cruyff, my biggest sporting hero, who died aged 66 from a smoking-related illness. I am glad that smoking is gone from chess, as well as from other sports. Especially I am glad that my opponents are not smoking in my face!)

31

When I arrived home that evening, I was in a state of shock. I had seen something special , a miracle at the chessboard. It is not impossible that it determined the course of my life. It is for this reason that I want to underscore the importance of following top games live. Preferably at the tournament hall, if you have the chance. But if you do not, then use all this wonderful technology we have, with transmissions from all of the top events. If you want to make progress, I think it is important to focus on no more than three games (or less!) and forget all about multitasking. And of course you should watch the games without the engine on. Anyone who has read the first volume of this series, Positional Decision Making in Chess, should be able to say this in their sleep! It can be difficult to work out where you switch off the engine on the website, and it can be tempting to turn it on when the game becomes complicated. But life has always been full of shortcuts to nowhere and it has always been clear that this is exactly where they lead. To me the 1 979 Soviet Championship in Minsk was a rare chance to see the best players in the country in action. It was won by Efim Geller, of course. He won six games and drew eleven, scoring 1 1 V2 from 1 7. By round 1 4 it was already clear that things were heading his way, and his masterpiece against Anikaev only solidified the deserved nature of his victory. Yusupov was second with 1 OY2 and Kasparov shared third with Balashov on 1 0 points. At the closing banquet of the 1 992 Alekhine Memorial in Moscow, where I had tied for first with Anand, I was sitting next to Geller. At some point he raised a toast for me to become World Champion. I remember Efim Geller and the 1 979 Soviet Championship fondly.

Chapter 2 Petrosian

Photo ofPetrosian from the Soviet Sport article

Diagram Preview

On this page you will find a few diagrams with critical moments from the coming chapter. If you want to compare your thinking with the games, you have the possibility. Take as much time as you need or want. This is not a test, but a chance to 'think along' with the grandmasters in the games.

�

b

c

d

e

f

g

h

Do you know this famous combination? (see page 36)

z

!!��II ��-r0

� : r.��r•r � ')) , · , -j � ������ %� ---%��� � r � r� �� -- -�� :� if� � ;·r 5 4

3

2 a

b

c

d

e

f

g

h

a

b

c

d

e

f

g

a

b

c

d

e

f

g

h

What amazing move did Petrosian play? {see page 44)

How can Black start an attack? {see pages 39-40)

a

8

h

What is the most dynamic way to continue? (see page 40)

Chapter 2 The Soviet School of Chess

Many people think there is a Soviet Chess School, which is not really the case. This concept originated after the Soviet players showed great strength in the post-WW2 period, and was supported by a book by Kotov and Yudovich in the late 1 950s. It is very hard for people to understand that this was a metaphor, and not a real thing. In 1 98 5 a delegation of Indian chess players visited the Soviet Union. Playing against their delegation was my first game against a foreign player, A. Mittei. At one point on their trip they were in Moscow, and one Indian guy kept insisting that they show him the Soviet Chess School. The guides tried to explain that it was a metaphor, but the guy was not biting. Anand told me that eventually they had had it with this guy, and j ust pointed at a random building, assuring him that this was indeed the Soviet Chess School, but sadly it was closed on that day. I don't think it makes sense to talk about a Soviet School of Chess. We can talk about regional tradition, where many players developed in a certain direction, influenced by a star player in their city. A clear case is Riga, where Tai influenced players such as Alvis Vitolins, Alexander Shabalov and Alexei Shirov, who play exactly the same brilliant attacking style. I grew up in Minsk, and in my chess you can clearly see the style of Boleslavsky, with serious opening preparation, Sicilian and King's Indian with Black, and main lines in other openings. I did not work with Boleslavsky, but he influenced Kapengut, who influenced other players such as Ilya Smirin, Yuri Shulman and myself. Razuvaev called me Boleslavsky's chess grandson.

-

Petrosian

35

Players from Lviv were influenced by Stein. Players from Odessa were influenced by Geller. In Armenia and Georgia, Petrosian's influence was strong, with a liking for play with knights against bishops. But this was of course not j ust a Soviet phenomenon. In the US, Fischer had a big influence, which is still strong to this day. In Denmark they had Larsen and Nimzowitsch. The Netherlands had Euwe and later Timman. England did not have a guru in the same way, as their players all appeared at the same time. But still you can see a lot of common traits in their style, their originality and thirst for fighting chess. You will find similarities between the players in many countries. When you think about it, it is very obvious why. We learn from each other and try to emulate our heroes.

Real Soviet chess schools Botvinnik's school is famous all over the world, with students such as Kasparov, Kramnik and so on. Tigran Petrosian also had his own chess school, but with a more limited range - he could only select his pupils from the members of the Spartak club, a sports organization, making the choice narrower. Still, a number of kids who later became famous grandmasters took part, such as Igor Novikov, Yury Dokhoian, Svetlana Matveeva and Ilya Smirin. I was invited after I won the U 1 8 Spartak Championship, half a point ahead of Giorgi Giorgadze, who later became a strong grandmaster. I had to join Spartak in order to train with Kapengut, who worked there. The Petrosian school was for players under 1 8 , and usually met twice a year for training sessions. Petrosian was assisted by Alexander Niki tin, who was the trainer of Garry Kasparov then and for many years of his career.

36

Boris Gelfand - Dynamic Decision Making in Chess

In the morning there was a group session, with a lecture. Maybe an explanation of an opening, or the analysis of some games of the participants. In the second half of the day, the group was split in two. Petrosian would take the children he felt had the greatest potential or whom he simply liked. He would show this group some of his own games. In the evening Petrosian would play blitz with us, giving us all sorts of handicaps, such as choosing the square upon which we would get mated. We were young and naive, and with Black I said a l , after which he simply took all my pawns and checkmated me with king and queen. If I had chosen one of the central squares, it would have been more of a challenge for him to win the game. Petrosian was an incredible blitz player. It was quite common for his generation. Fischer, Korchnoi and Tai were also amazing. Actually, Tai won the World Championship in blitz in 1 988, not long before he died.

8 7 6 5 4 3 2 a

14.f;Yh6

b

c

d

e

f

g

h

The following game in this line is very famous: 1 4.g5 ! ? This idea worked wonders in this game, but it does not stand up to scrutiny. 1 4 . . . b4? 14 . . . fxg5 is necessary, but then Black has an excellent score. l 5 .gxf6 i.xf6 1 6.�g l t �h8 l 7.°1Wh6 \We?

This game was played in the tournament where I qualified to go to Petrosian's school. M. Magomedov - Boris Gelfand Kalining rad 1980

1 .e4 c5 2.c!LJf3 d6 3.d4 cxd4 4.c!Lixd4 c!Lif6 5.c!Lic3 a6 Already at that time I had analysed the Najdorf rather deeply.

6.Ag5 e6 7.£4 j_e7 s.Y;Yf3 Y;Yc7 9.0-0-0 c!Libd7 10.g4 b5 1 1 .Lf6 gxf6 I played this system a lot at the time, inspired by some games by Fischer.

12.f5 c!ll e 5 13.f;Yh3 0-0 I had some improvement based on a Spassky game, but it was not really important.

1 8 .tll c6!! A combination that is almost beyond belief. 1 8 . . . tll xc6 1 9.e5 ig5t Not the most accurate defence, but Black was in big trouble anyway: l 9 . . . ixe5 20.f6 ixf6 2 I .id3 ig5 t 22.�xg5 f6 23.�g3 �f7 24.�dg l Wf8 2 5 .�g8t Wxg8 26.�xg8t �xg8 27.tll e4 and White is going to win all the same.

Chapter 2 - Petrosian 20.E!:xg5 f6 2 l .exd6 '1Mi'f7 22.E!:g3 bxc3 23.ic4 cxb2t 24.�b l lll d8 25 .E!:dg l E!:a7 26.d7 E!:xd7 27.fxe6 lll xe6 28 .ixe6 E!:d l t 29.E!:xd l ixe6 30.�xb2 E!:b8t 3 1 .�al ixa2 32.E!:gd3 '1Mi'e7 33.�xa2 We6t 34.E!:b3 1 -0 Kholmov - Bronstein, Kiev 1 964.

14 ... 'itih8 15.g5 :gg8 16.g6 fxg6 17.fxe6 ih7 18.tlJd5 ixd5 19.exd5 f5 20.'itibl Wfc5! 8 7

37

Three times

I went to Petrosian's school three times; it was organized as a training camp. The first time was in 1 980 in Sochi, the second time 1 98 1 in Moscow, and the third time again in Moscow, but in 1 983. On the third occasion, I was only there for a few days, as I had tournament commitments starting halfway through the sessions. This was the last time I attended. Petrosian died in 1 984. The first time I came to Petrosian's school, I was 1 2 years old. I was quite deluded, thinking that Petrosian was a great manoeuvrer, but that in tactics I was stronger than him. But the moment I saw him play blitz, it was immediately apparent to me how excellent his tactics were. One day we played a simul against Petrosian, which was a great honour.

6 5 4 3 2

1

a

b

c

d

e

f

g

Tigran Petrosian - Boris Gelfand

h

This is the small highlight of the game; the moment I remember fondly. My recollection was that I wanted to avoid ig2 and lll c6, and for this reason I moved my queen out early. Anyway, this is the moment I wanted to show. I shall include the remaining moves without comment.

21.ig2 :gaf"B 22.Wfd2 tlJc4 23.Wfd3 if6 24.d b4 25.tlJb3 Wfc7 26.ifl tlJe5 27.Wfd4 hxc3 28.Wfxd Wig? 29.Wf d2 :gb8 30.ie2 Wfc7 31 .:gcl Wfb6 32.Wfd4 Wfxd4 33.tlJxd4 tlJg4 34.gc4 tlJe3 35.ga4 lDxd5 36.tlJb3 lllc7 37.gcl tlixe6 38.gxa6 ie5 39.h3 tlJc5 40.ga7 tlJxb3 41 .axh3 gxh3 42.gc2 ggb8 43.ga2 gxh3 44.ic4 ghl t 45.gcl gxcl t 46.'itixcl gc8 47.b3 d5 0-1

Sochi (simul) 1980

1 .d4 tlJf6 2.c4 g6 3.tlJc3 ig7 4.e4 d6 5.ie2 e5 My idea was to meet 6.d5 with 6 . . . lt:l bd7 to save a move and avoid the Petrosian system. Not a big idea, but I was j ust a kid. Sadly there was a downside as well. 8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making in Chess

38

6.dxe5 dxe5 7.�xd8t ©xd8 8.f4 exf4 9.J.xf4 .ie6 IO.O-O-Ot tlibd7 1 1 .tlif3 tll g4 12J:thfl ©c8 13.tlig5 tlige5 14.tll xe6 fxe6 1 5 ..ig3 c5 16.tlib5 ©d8 17 ..ih4t ©c8 18 ..ig3 gg8 19.©bl a6 20.tll d6t ©b8

a

21.tll e 8!

b

c

d

e

f

g

h

A nice move by Petrosian. Black's position is falling apart.

21 ... ©a7 22.tlic7 gac8 23.tlixe6 .ih8 24.tll g5 h6 25.tlif7 .ig7 26J�d5 �ke8 27.tll xe5 tlixe5 28,gxc5 tlic6 29 ..if3 ggf'8 30.gd5 tll e 5 31 .he5 .ixe5 32.h3 g17 33J�fdl gc7 34.b3 g17 35.a4 .ig3 36.a5 gc7 37.©b2 gce7 38.b4 gc7 39.©b3 gec8 40.c5 ge7 41 .gd7 gc7 42.gds

1-0

a

b

c

d

e

f

g

h

After this game, I was asked to work on this type of position before the next session, which I duly forgot about until the very last minute. I quickly scribbled down a few notes taken from a Geller book, all very obvious. Nikitin suggested that they should send me back, that there was no purpose in keeping me, when I was unwilling to do my homework. But Petrosian said: "The boy has talent, he enjoys chess, so who cares if he forgot to do his homework." So Petrosian sort of liked me. I was really amazed that such a great player would spend so much time with us. There are so many great memories from these training camps. Once Petrosian organized for a bus to collect us and take us to a top­ level ice hockey game. He followed the sport obsessively, but this was not the point of the trip. He wanted to show us the intensity of top-level athletes in battle. The generosity of spirit was incredible. In 1 980 the sessions coincided with the Chigorin Memorial. In the morning we would have our training sessions with Petrosian and Nikitin, and in the evening we would go and watch the games. I have autographs from, for example, Eugenio Torre and John van der Wiel. At times, Petrosian would ask me: "Boris, what do you think of this position?" Very naively, I said something obvious. He replied: "Even my wife can count the pieces. But what do you think, what will the result be?" I predicted that White was going to win the game, but the player {let's not name names) let me down and even lost the game. I tell this story j ust to illustrate the way Petrosian engaged with us all, and made us feel respected and valuable. It was true mentoring. It was an amazing experience and I am sure it gave me a big push. This is part of the reason why I like spending time with young players. Everyone can analyse a variation deeply and

Chapter 2 - Petrosian find a new idea, but to meet a great player is not always possible, to get a tip or two from him. This is something that can make a big difference. I still remember some of Petrosian's tips:

You should think about each move, even in blitz. Do not make senseless moves. This had a big impact on me, both in a good and a bad way. It definitely cost time as well. As with most things, it can be a plus and a minus; it is all about the application. Maybe it is a plus when you are in good form, but when you are in bad form, it can lead to time trouble. This is also a reason why I believe that the introduction of the increment, especially in rapid, was very good for my career. You should never miss a chan ce to solve an exercise! We come across positions all the time. In magazines, books, on the internet or as something a friend shows us. There are many chances to train your thinking. It does not matter if the position is simple or difficult. There are a few top players who are so confident in their calculation that they feel they do not need to solve exercises. I do not want to be presumptuous and think I know best. I simply follow Petrosian's advice. He knew a thing or two about chess. Random solving in action

I practise what I preach. Recently I was having dinner at home with my coach, Alexander Huzman, and we were discussing recent games we had seen. Alex had played in the European Senior Team Championship. Although the field was weaker than expected (Alex won three of his games in 20 moves or less) , he still had some interesting moments and games. He offered me one of them as a positional exercise. We were still eating and talking, while I was chinking about the position.

39

Galina Strutinskaia -Alexander Huzman Halkidiki 2016

t.d4 �f6 2.�a g6 3.g3 .tg7 4.Ag2 0-0 5.0-0 d6 6.c4 tll bd7 7.9c2 e5 s.gdl ges 9.b3 e4 10.�gS e3

a

c

b

1 1 .�h3

d

e

f

g

h

Let's look at the game and the lesser alternatives before we move on to the critical moments. l l ..ixe3? :!'he3! 1 2.fxe3 lt'i g4 and Black is already better. l l .f4?! h6 1 2. ll'i f3 lli g4 1 3 .ll'ic3 llidf6 1 4 . .ib2 Maybe not the best, but the position is already bad. 1 4 . . . .if5 1 5 .�c l This was played in Sideif Sade - Mamadshoev, Azov 1 99 1 .

a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making in Chess

40

1 5 . . . tll f2!N 1 6.l:!fl ih3 would have given Black a crushing attack.

1 l . fxe3 tll g4 1 2.e4 c5 1 3.tll f3

l l. .. �g4 12.f3 �h6 13.�c3? Blundering away the d4-square. 1 3 .ib2 would give even chances.

a

1 5 ....ixd4 16.l:!bl �c5 17.b4 .if5 18.�e4 �xe4 19.fxe4 .lxh3 20 ..ixh3 c6

8 7 65 4 3 21 b

c

d

e

f

g

h

Alex won the game in another 1 7 moves. During the game, Alex was chinking most about:

c

d

e

f

g

h

This is the position of interest. Alex asked me to chink, and I did so while sharing my considerations out loud, so he could follow my line of chinking. Only one game has been played from this position, and it did not present us with a map to follow. 1 3 . . . cxd4 1 4.tll xd4 tll c5 1 5 .tll c3 a6 was the choice in Zohar - Parwicz, email 2005. This feels unenergecic and I chink White could fight for an advantage. After some time, I found a very dynamic solution. 1 3 . . . b5!!

a

a

b

b

c

d

e

f

g

h

I suggested this move after some chinking, and Alex said chat it was indeed the best move, and we agreed chat it would be interesting to include it in this book. The upsides co this move are easy co describe. Lee's cake a look at how play would continue if White is collaborative:

Chapter 2 1 4.cxb5 ib7 1 5 .tll c3 cxd4 1 6.tll xd4!'k8 with a big advantage for Black. Play could continue with 1 7.Wd2 Wf6 1 8.e3 gxc3! 1 9.gfl gxc l ! and Black has a winning position. The next morning, I woke up early and started thinking about the position again. I realized that White had to continue actively as well. It was only while having breakfast that I found how White should play. 1 4.h3!

- Petrosian

41

l 5 . . . tll f2!! This tactical resource is important and could easily be missed. While if there was a pawn on f2, everyone would find it! 1 5 . . . tll g e5 1 6.tll x b5 leaves Black without real compensation for the pawn. 1 6.�xf2 ixd4t 1 7.gxd4 Wf6t 1 8.'it>e3 Wg5t

�,lifl, � 6 . �-. �. !� � �. . �� �; ,. ' � � : , ,� �

�

..... �

;

3

2 1

1 4 . . . cxd4 l 4 . . . tll gf6? l 5 .e5! would be a disaster for Black. l 5 .tll xd4 1 5 .hxg4? d3 1 6.exd3 ixa l 1 7.d4 bxc4 1 8.bxc4 Wb6 gives Black a serious advantage. White will not be successful in trapping the bishop.

8 ,,.\�£�� '))J! ., �a , .- , · --·� .."•

�� · � � : ��f� - -3� ��� � 7

6 3

2 1

..

fm,f .. � :�!-'l . :. �8 8•t 1 �- - �• • � � � � �; � �t'.{J��� ��-0� ..

a

b

c

d

e

f

�

g

�

h

This position is not easy at all. If Black does not react very actively, he could easily find himself lost in only a few moves.

, � �· . . ·"m � 8 ��. �8• :t ��-1����1: . .�tt:J�· � ��� � a

b

c

d

e

f

g

h

At this point White can force a draw, should he wish to. Or he can choose to play for three results. 1 9.�d3!? White has every right to play for an advantage. He is a piece up after all. 1 9.�f2 Wf6t is an immediate draw. 1 9 . . . bxc4t This appears to be the most accurate. 20.�c3! I think this is best. 20.bxc4 tll c5t 2 1 .�c3 Wxg3t 22.if3 f5 gives Black the initiative, although White's position is defensible. 20 . . . Wxg3t 2 1 .if3 tll c5

42

Boris Gelfand - Dynamic Decision Making in Chess

This looks OK for Black; we are playing for all three results. White has a piece, but Black has one pawn (and will most certainly get another one soon) and the white king will be in need of perpetual babysitting. I do not think this is a position where the king will be able to protect itself (as in the game below) . White has more pieces, but they also have more work to do! Both 1 3 . . . b5! and I 5 . . . lDf2! are difficult moves to find, making it imperative that we train our ability to find them. 1981

The 1 98 1 sessions were held not long after the Tilburg tournament. One day Nikitin said that he had received a letter from Kasparov with analysis of the game he had lost to Petrosian in that tournament. Kasparov claimed that he had found a way he could have won. We had all seen the game and were amazed by Petrosian's defensive skills. When Nikitin said he had received the letter, Petrosian simply started laughing. They agreed that on one of the following days, instead of the regular lesson, they would discuss the game in front of everyone, so that we would be able to see how Petrosian would refute the analysis. Sadly, I had to leave the day before this session, to play a tournament. Since then 3 5 years have passed, and it still hurts when I think about missing out on this session. The game is also still fresh in my memory, as can be seen in my game against Alekseev on page 223.

However, the notes in that book felt somewhat below the great standard of the series (I am a big fan of Kasparov's books, and have been for more than 30 years; I cannot recommend them enough) . The best annotations of the game, in terms of accuracy of the analysis, are probably those by Jacob Aagaard in Python Strategy, a recently published collection of Petrosian's annotations and articles. Aagaard was able to combine Petrosian's and Kasparov's annotations, and check them carefully with a strong engine, getting us a bit closer to the truth. My comments below are not meant to give the objective truth of the game, but to share my impressions, then and now.

1.d4 d5 2.c4 dxc4 3.4Jf3 t£!f6 4.e3 i.g4 s.J.xc4 e6 6.h3 .ths 7.tll c3 a6 8.g4 i.g6 9.tll e 5 tll bd7 IO.tll xg6 hxg6 1 1 .i.fl c6 12.i.g2 V!!c7 13.0-0 i.e7 14.f4 t£lb6 1 5.gS tll fd7 16.V!!g4 0-0-0 17J�b1 ©bs 1S.b4 t£ld5 19.t£la4 f5 20.V!!g3 tll xb4 21 ..id2 t£ld5 22JUcl ©a7 23.V!!e l i.a3 24J�c2 V!!d6 25J�b3 V!!e7 26.V!!e2 gbs 27.V!!d3 i.d6 28.t£!b2 ghc8 29.tll c4 i.c7

Garry Kasparov -Tigran Petrosian Tilburg 1981

This game has been annotated quite a number of times in various places, most notably in My Great Predecessors, Part III by Garry Kasparov.

a

b

c

d

e

f

g

h

We can see that White has sacrificed a pawn for clear long-term compensation. All his pieces are active on the open files. So, normally, the

Chapter 2 general course of events would be that Black would defend passively, while White would try to find a way to break through.

-

Petrosian

43

33.J.b4 The engine points out 33.Wfb 1 ! was very strong here.

30.a4 b5! An amazing move that sends us back to Steinitz's maxim, that the king should be able to protect himself. More importantly, this move, followed by move 32, changed the course of the game. Obviously, the young Garry, already then incredibly strong and a great master of attack, was convinced he should checkmate his opponent in not too many moves. Probably this feeling made him play the attack without the necessary patience.

31 .axb5 cxb5 32J�a2 A natural move. 8 1

, wJ-J �itw· ��r r� �,gJ%� - - - �� ,,-�%� �')) , ,,%�

������p £!!�,,�� 4 �t.z.J �C§ �(_§� � 3 �: �v;�-- -� �8 21 r ���'�' ��-0� � q � � 6 5

�

,;,,J �

a

32 ... ©b7!

b

c

d

e

f

g

h

32 . . . id6 may be objectively better but I like the game move, as it keeps alive the possibiliry of Black winning the game. I am sure there was a psychological dimension to Petrosian's play. With such a big attack, it is very easy to believe that it is possible to win immediately with White, and as we have reached move 32, there is little doubt that both players were running low on time.

The idea of sacrificing a rook on b5 or a6 in some lines, combined with the manoeuvre lll c4-b2-d3, gives White a completely winning position. Black is in too many pins, and squeezed together in too confined a corner of the board to be able to regroup quickly enough to deal with all the arising threats.

33 ...�eS Placing the queen on the light squares, which is essential, as the knights and bishop all control dark squares only. Petrosian was a fantastic defender and must have felt that White would not be able to break through without sacrifices. And if White did take such risks on, Petrosian was confident in his incredible calculating abilities. I think in general that Petrosian is underestimated as a calculator by the general public. As I said, even I had this impression before I met him. But he really amazed all of us at the training camp.

34.J.d6 Kasparov plays according to his beliefs; moving only forwards. It is not easy to find a move like 34.lll b2!, with the idea of lll d3 later, in the heat of the moment. It can easily feel as if White is losing

Boris Gelfand - Dynamic Decision Making in Chess

44

momentum. It would cake some time looking at the position to realize chat Black will not be able to regroup his forces in order to adequately defend all of his weaknesses.

34 .. J�aS 35.�bl Quite a natural-looking move, although White does not have any direct threat. 8 7 6

Kasparov does not react well to the new challenge, and simply loses a piece. White should have played 36.J.xc?, when the game would probably have continued: 36 . . . bxc4 37.E:b7! E:xc7

� -0 c-- �1� 168 ,�� •�:·· . ,.,. •• � 5 ��··ef ••• ��-�% ��-ef,�J, - - �� 4 � 3 ".,.,%� - �� 8 ,.���."-�.,. �

5

..

2

4 3

a

2 1

36J�ba3?

a

35 ... ©c6!!

b

c

d

e

f

g

h

This is the move, completely changing the nature of the game, and again following Sceinitz's maxim. The main point of the move is a simple double attack on the bishop on d6 and the knight on c4 {now the b-pawn is no longer pinned) . The king also defends the b5-pawn, but the sacrifices were not working at this moment anyway, so this is not the main point. Although objectively the strongest move, and one chat changes the nature of the position, it is the psychological dimension chat had the greatest impact on the course of the game. Noe only is the evaluation of the position no longer in White's favour, he has to find the accurate way co make a draw. But the sheer audacity of this move, the fearlessness of it, threw the young lion completely off the trail. The sudden change was coo much co handle, even for such a great player as Kasparov.

i� � ' b

c

d

e

f

g

h

38.E:xa6t! E:xa6 39.�b5t @d6 40.�xa6t @e7 4 1 .J.xd5 White wins back his piece, but after 4 1 . . .E:xb? 42.J.xb? �b8 followed by . . . �c7, Black has nothing to fear.

36 ...bxc4 37J�xa6t gxa6 38J�xa6t i.b6 39.i.c5 �d8

Chapter 2 The annotations of this game in Predecessors and Python Strategy were both computer­ assisted, but I cannot help wondering if Garry's annotations from 1 98 1 exist in a box in Nikitin's archive. It would be amazing to see his original analysis. (I should say that Kasparov was already known then as a great analyst, and one of his early books, The Test of nme, was based partly on notes he published in chess magazines on the advice of Botvinnik.) This was the second time in a row where Kasparov did not manage to bring an overwhelming attack to its conclusion against Petrosian (the first being in Moscow 1 98 1 , where he also lost) . Kasparov wrote that Spassky gave him a piece of advice after this game: against Perrosian you have to play patiently, you should not try to mate him. With this advice, Kasparov was able to beat Petrosian very convincingly twice the following year.

-

Petrosian

45

which will most likely be included in Volume 4 of this series. This was the last time I saw him. Influenced

It is difficult to overestimate the value of time spent with a great player. Even if there are not so many things you can use immediately, their approach, attitude and reasoning will influence you for years to come. It's hard to explain in words how you benefit, but to me it was really invaluable. Petrosian was very kind to us. He was the first person who showed me a portable tape recorder. He had bought it on a trip abroad, and lent it to us with some music tapes. This was not the first time I saw a tape recorder. At home we had the old style with big reels . . .

In this game, the position suggested that White had a winning attack and computer-assisted analysis confirmed this. But it also showed that White had to make more than one accurate move, including a not at all intuitive knight manoeuvre. Petrosian showed great defensive skills, and managed to gain one last victory against Kasparov - a game that is more famous than all of Kasparov's subsequent wins against Petrosian.

This book i s about dynamics, s o I should admit that the main things I learned from Petrosian were probably not dynamics. But some people say that they can see a lot of his approach to dynamics in my games. The approach to counterplay, the way we both sacrifice the exchange. Unhurried compensation. What can I say? I was definitely influenced a lot by Petrosian, but I do not see myself from the outside, so I cannot say for sure to what extent this is so. I will allow the reader to decide for himself.

1983

Postscript

In 1 983 the training sessions were very brief. It was at the same time as the Soviet Championship, so Petrosian did not have a lot of time for us, but we were taken each day to watch the games. I do not know if he was already ill at this time - it is hard to say, as he played well in the tournament. For example, he won a nice game against Polugaevsky,

In 2004 I was invited (on the proposal of Kasparov) to represent Team Petrosian in a match against "The Rest of the World". Our team consisted of players with some form of connection to Petrosian. This was mainly Armenians, but also included Peter Leko, who married Petrosian's daughter - although we should make it clear that this is Arshak

Boris Gelfand - Dynamic Decision Making in Chess

46

Petrosian's daughter we are talking about! In the first game, against Peter Svidler, I made an exchange sacrifice on c3 in the Sicilian. Peter played a beautiful game and totally refuted my concept. As any serious chess fan will know, Svidler at his best is a really tough opponent! The following day this position arose: Boris Gelfand - Loek van Wely Moscow 2004

Ltll f3 �f6 2.c4 c5 3.tll c3 d5 4.cxd5 tll xd5 5.e3 �xc3 6.bxc3 g6 7 ..ib5t �d7 8.a4 i.g7 9.a5 0-0 10.0-0 tll f6 1 1 .d4 J.5 12.Wfb3 �k8 13.i.b2 �e4 14 ..ie2 Wfc7 1 5.c4 cxd4 16.exd4 gfd8 17.Wfe3 tll c5 18.i.a3 tll e6 19.gfdl .ig4

At this point I was tempted to sacrifice the exchange with 20.d5 ! ? for long-term compensation. Black would soon have to take on f3 as well, giving me the bishop pair against knight and rook, one of the most favourable types of exchange sacrifices. Potential variations could be as follows: 20 . . . tll f4 is not likely to be a great inclusion. After 2 1 .iW i.xa l 22.�xa l i.xf3 23.Wfxf3, Black is under a bit of pressure. For example, 23 . . . e5 24.Wfe3 b6 2 5 .g3, when the black pieces are not harmoniously placed. Or 23 . . . Wfe5 24.�b l , when White will take control of the long diagonal. More sensible is 20 . . . i.xa l 2 1 .�xa l tLl c5 22.i.b2 f6, but after 23.h3 i.xf3 24.i.xf3 I think White has good prospects. But I didn't play it. It was definitely a mistake not to sacrifice the exchange, as the game quickly petered out to a draw. I especially regret chickening out in a tournament played to honour Petrosian.

20.�g5 .ih6 2 l .�xe6 he3 22.tll xc7 .ixe2 23.�d5 i.g5 24.gel .ixc4 25.�xe7t i.xe7 26.l:�xe7 a

b

c

d

e

f

g

h

1/2-1/2

Chapter 3 Tactics at the Top Level

My first game with Magnus Carlsen - Biel 2005

Diagram Preview

On this page you will find a few diagrams with critical moments from the coming chapter. If you want to compare your thinking with the games, you have the possibility. Take as much time as you need or want. This is not a test, but a chance to 'think along' with the grandmasters in the games.

a

b

c

d

e

f

g

h

Which winning move did the World Champion miss? (see page 52)

a

b

c

d

e

f

g

h

Worse out of the opening, I set a cunning trap. Can you find it? (see page 49)

a

b

c

d

e

f

g

h

The choice was between 27 . . . lll e3 and 27 . . lll f4. Which one would you choose? {see page 50) .

a

b

c

d

e

f

g

h

Sometimes it can be difficult to see even the simplest things. White wins! (see page 54)

a

b

c

d

e

f

g

h

Here I found a funny little combination. Can you see it? {see page 5 5)

a

b

c

d

e

f

g

h

Black has sacrificed two pawns in the hope of getting enough counterplay due to the opposite-coloured bishops. With a nice tactical sequence, I managed to make life very difficult for my opponent and subsequently won the game. Can you find it? (see page 66)

Chapter 3

-

Tactics at the Top Level

In this chapter I will use a few combinations (of sorts) from my career to think aloud about various aspects of tactics, and the thought process surrounding them. This is meant to be inspirational and not prescriptive. As I say in this chapter (in case you miss it!), every player will have to find his own individual balance in the thinking process, and decide for himself how much attention he wants to pay to positional and how much attention he wants to pay to dynamic aspects of the position. There is no right or wrong way. Alekhine, Tai, Spassky and Kasparov all paid more attention to the dynamic aspects of chess, while Capablanca, Petrosian, Karpov and Kramnik all favoured the positional side of things. At the moment, the World Champion pays more attention to the positional/technical aspects of chess, while a lot of top-level players care more about the dynamic side. Who would, even for a moment, suggest that any of these guys are 'getting it wrong' and do not understand chess? In order to make the chapter more interactive, I have started by asking you questions relating to six positions. If you have the chance, please rake 5- 1 0 minutes to think about each of them before you continue reading. Chess is easy when you agree with the first line of the engine, but quite a bit harder when you have to think for yourself. . .

49

Boris Gelfand - Michael Adams London (rapid) 2013

a

b

c

d

e

f

g

h

This was a rapid game where the opening did not go well for me. Adams was very well prepared and, even with Black, he had the better game from the opening. I realized this, and knew I had to play very carefully in order to equalize. Then at one moment I found a trap! If you play with a standard time control, the chances that a player of this level would fall for it is minimal, but in rapid . . .

24.tll e 5!? After this move, Mickey replied instantly with:

24... lll d??

Boris Gelfand - Dynamic Decision Making i n Chess

50

25.0-0-0!! I don't think there is a lot more to say about this game (I won the endgame after winning the exchange) , but I would say that it is a fantastic feeling to be able to play this move, after seeing so many positions with this theme since I was a little boy. To some people it might seem unbelievable that this could happen to a top grandmaster, but it is important to understand that the reduced time control really matters. Of course it is possible to play a great game in rapid chess, but here we have a positional game, drifting into an ending. Not having time to do everything, he forgot that chess also includes tactics! It might be hard to believe, but it happened!

46J�h5t i>f6 47J:ld4 gbit 48.i>d2 gb2t 49.i>c3 gc2t 50.i>b4 .ih7 5 1 .gxf4t i>g6 52.:ah3 e5 53.gxc4 gxf2 54Jk7 1-0 Choosing between two similar-looking moves

I played in Bermuda for two years in a row. They were well-organized and strong double round-robins, and I really enjoyed playing there. I won two good games against Shabalov in the first tournament (one of which is on page 1 68) and a very nice game against Harikrishna the following year. But let's see an interesting position I had against Madeja. Bartlomeij Madeja - Boris Gelfand

25 ... llixe5 26.i>xb2 llixf3 27.exf3 i.xc4 28J�d8t i>h7 29Jkl ±

Berm uda 2004

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

I won this game after another 2 5 moves. Both players made a few inaccuracies later on, but I do not think this ending is so interesting, so I will not dwell on it. I promise we will look at truly interesting endings in the next volume.

29 ....id5 30Jk3 i>g6 31 .h4 i>f6 32.a3 i>e5 33.ghs g6 34.g4 h.xg4 35.fxg4 i>d4 36.ge3 f5 37.gxfS gxf5 38.h5 f4 39.gh3 gb7t 40.i>cl .ie4 41 .gdst i>e5 42.h6 c4 43.i>d2 gb2t 44.i>el gb lt 45.gdl gb7

a

b

c

d

e

f

g

h

In this position I have a simple idea: sacrificing the knight on g2. The real question for me was which way to get there, via e3 or f4?

27 ... c!lif4! I decided on this move after careful calculation, but it is not possible for us to consider every move, so we cannot rely solely on brute force. We also have to use our intuition. And in this position the knight simply feels better placed on f4. It is better supported and, compared to the e3-square, it has additional

Chapter 3

-

51

Tactics at the Top Level

possible destinations on e2 and h3. It is thus logical to check this move carefully first. And as a matter of fact, it turns out that Black does not win with 27 . . . lli e3?, as White has a beautiful defence: 28.cxd4 llixg2 29.%Vd l !! (it is not hard to predict that 29.'it>xg2? %Vf.3t leads to a winning attack)

28.�g3 White has no defence, as the computer confirms. There are many places where we should argue against the metal box, but not with these types of tactics. On 28.llig5 Black has 28 . . . %Vd5 with a winning attack. After: 28.cxd4 lli e2t 29.'it>h l

The queen is defending the f3-square. The position looks like a case of Tai-syndrome: with many hanging pieces, White does not have time to take them all. But sadly the old master's magic has not rubbed off, and all Black has is a draw. 29 . . . ixd4t 30.'it>xg2 id5 3 l .ih6! Another sneaky double attack. 3 1 . . .%Vf2t 32.'it>h l White is planning %Vg4 so Black has to force matters. 32 . . . %Vh4 33.%Vxd4 %Vxh3t 34.'it>gl

8

7

I had planned to win the ending after 29 . . . llixc l 30.dxe5 %Vfl t, which would require some work. Instead 29 . . . ib8! shows the power of Black's position, leaving White without a move. The main threat is . . . %Vc7. And after 28.ixf4 the simplest would be to win the exchange with 28 . . . ixf4 29.i:!d l ie3t 30.'it>h l ib3, and Black wins.

6 4 3 2 1

a

b

c

d

e

f

g

h

Black has nothing more than perpetual check. a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making in Chess

52

28 ... �xg2! The simplest, though Black's position is simply winning in various ways.

29.'!Wxe5 29.©xg2 '!Wf3t and . . . i.d5 with mate.

29 '!Wflt 30.©hl �h4! 3 1 ..ie4 �6 32 ..id5t .ixd5 33.'!Wxd5t ©hs .••

But the main point remains the same: different approaches are good in different situations. Working out when to use a particular approach is one of the real challenges of playing chess at the top level. Those looking for easy uncontradictable advice will certainly lose their temper when faced with the classics! Every player has a slightly different approach, and all the poor author can do is use the get-out-of-jail card called 'balance'. We all have to find the way of playing chess that is most practical for us, and make our own considerations about how to balance between the practical and deep approach to decision-making. Recently I saw the following game between Carlsen and Giri, where the World Champion, the master of creating practical problems for his opponents, missed a direct win. Magnus Carlsen -Anish Giri Stavanger 2015

0-1 Different approaches

I feel that, in general, if we see a highly promising option and we have checked it and it works, we should go for it, and not spend a lot of time checking other options. Mistakes are born that way! But you could have an interesting discussion about the topic. Capablanca said that if you see a move that is good enough, you should go for it. But this is certainly not the only method. Lasker and later Fischer were attributed with the saying: " If you find a good move, look for a better one." A quick Google search suggests that a gentleman called William Wayte said this in the 1 880s. To be honest, it does not sound like something Lasker would make up to me anyway. I am also not sure that Mr Wayte was talking about a position where there is a direct win.

a

b

c

d

e

f

g

h

The winning concept was original and surprising, but if you get the idea, the rest should fall into place quite easily:

38 ..ifit!! Carlsen had just five minutes on his clock. He played: 38.We4? :B:f8 39.tll xe5 ? (39.tll h4!

Chapter 3

-

Tactics at the Top Level

with the idea E:g3 was still very dangerous for Black) 39 . . . dxe5 40.E:g3 And here Giri managed to find 40 . . . E:xd5!!, when White's attack may look dangerous, but in fact there is no advantage left . . . Carlsen still managed to create a lot of problems for his opponent, who (it turned out) was only halfway through a long defensive task in this game. He eventually held a draw in an ending with a knight against three connected passed pawns.

38 ... @xf7 Black would not be better off after declining to take the bishop. We could give more variations, but it wouldn't necessarily show anything other than that we know how to tap the spacebar while the engine is running.

39.'ilYe4 Black cannot get his pieces back in time for the defence. White has a number of concepts, including Wf5t xd2 lll e4t!

�;·i �Z ' ' �� : �,%��,%%�Brf� %% •� �3 �-� m t n :� ', , � �� 2 '-�=,��� � �-, ,/.- , %-�

8

1 1••···�· 6 �• m • • �

8

a

b

c

d

e

f

g

h

An important move, making it hard for the knight to escape, as well as blocking the attack from g2 on b7. 1 6. c;t>e 1 The only move. l 6.ixe4 Ei:xe4 wins material for Black, and 1 6.c;t>e3 ixd4t 1 7.'if;>xd4 'll c6t picks up the knight on c8 for free. 1 6 . . . Ei:e8 1 7.Ei:d 1

73

a

b

c

d

e

g

f

h

2 1 .tll xe8 Ei:xe8 22.Ei:xd? tll c5 23.Ei:c? if8 Black has good winning chances. 8 7 6 5

4 3 2 1

a

14... Ei:e4!!

b

c

d

e

f

g

h

This is what he missed. Probably he was expecting 1 4 . . . Ei:eS, when l 5 .a3 followed by ixb7 wins for White. You might ask: ' How did I see this?' The simple answer is chat it popped into my head. I saw it when I played 9 . . . Wc5 . Sadly I do not

74

Boris Gelfand - Dynamic Decision Makin g in Chess

remember all the details of my thoughts. For my classical games, I write down how much time I spent for each move, but not for rapid games. But we should also remember that I felt confident about my position, and was sure that I was doing well, so 14 .. J"�e4! did not come to me as a savin g resource, but as a natural tactical opportunity in a good position. I am sure it will come as a surprise to some that I saw this opportunity so early on, but choosing between the two queen moves at move 9 was an important decision, and this was the only line where it felt as if White was trying to refute Black's play. Add to this that it was a blindfold game. Keeping the position in your head does require mental energy, which is not available to do other things, such as calculate variations. At least, this is how it is for me. The time invested in making the ri ght decision at move 9 was quickly recuperated, as there was little choice for Black on subsequent moves. On the other hand, I had to see 1 4 . . . i:!e4! on move 9, as the variation otherwise would lead by force to a bi g advantage for White. This also makes the move easier to find. Especially, as I said, if you believe that the position is fine, and that there should be something!

15.tll c2! This is the lesser evil. 1 5 .e3 lll c6 is simply good for Black. Even worse is 1 5 ..ixe4 lll xe4 1 6.e3 ( 1 6.lll c2 Wi'c5 wins the knight) 1 6 . . . lll c6 1 7.lll xc6 bxc6 1 8 .i:!c l i:!xc8 1 9.Wi'xd7 Wi'b8 when Black is much better. The computer insists that there are ways for White to resist, but then, this seems to be its primary function!

8 7 6 5

4 3 2

1

a

c

b

1 5 ft°f'S!

d

e

f

g

h

••.

This is the first real choice I had since move 9. I think I made this decision rather quickly, without truly calculating the variations. The main difference if the queen went to c5 seemed to be that the knight would take with a gain of tempo on b7, which I did not want. On the other hand, the queen is more active on c5, which could also be an advantage. With limited time controls, you have to guess some of the time, and reserve your time for critical decisions. The serious alternative was: 1 5 . . . Wi'c5 1 6.lll d6 i:!e6 1 7.lll x b7 Wi'b6 1 8 .lll c5

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

This is the line I considered. 1 8 . . . Wfxc5 Black can keep the game alive with 1 8 . . . i:!d6!? 1 9.Wf cl lll c600•

Chapter 3

- Tactics at the Top Level

l 9.ixa8 At this point I did not see how I could continue my attack. It turns out that after: 1 9 . . . llig4 20.e3 ixa l 2 1 .llixal

75

16.�d6?! White is under pressure, and needed to defend well. Here he makes the wrong choice. I saw that after: 1 6.ixe4 llixe4 l 7.Wd5

7

6

5 4 3 2 a

·i � � . . : . : .... . �. �.r.r� � � � � : �.���.���.�?*���r�� 3 ��� � , . . %.�� 8 2 - � . 1 � ��1� �� -11 8 �7 %'l... :

6

..

.

�g

a

b

c

d

e

f

g

h

23 . . . E:d6!! 24.Wxd6 We2t Black gives perpetual check. I would not have wanted to go for this anyway, had I seen it. Starting from a better position, why would I want to fight for a draw with a long line? You could say that I used the method of comparison, deciding that one variation made more sense than the other. At the end of the day, the reason for calculating lines is not to see lots and lots of moves to impress the audience at the press conference, but to see enough to make good decisions that win games.

b

c

d

e

f

g

h

could play 1 7 . . . lli c6 1 8.Wxe4 ixa l 1 9.llixa l E:xc8 and I am well coordinated. But after 20.llic2 it is not clear that White is really worse. The knight is heading for e3. It turns out that I had a stronger option in 17 ... Wxc8 1 8 .Wxe4 ha l 1 9.llixa l d5!, as shown by the computer. After 20.Wxd5 Wh3t 2 1 .Wg2 Wd7 22.Wf3 lli c6 23.'itig2 Wd2 White can still find equality with an accurate defence, but the position is uncomfortable. Black keeps some pressure.

Boris Gelfand - Dynamic Decision Makin g in Chess

76

l 7.ttJb5

lli c6 is also better for Black. A small point is that Black should not fear l 8.ttJc7?! on account of l 8 . . . l'!d6.

17 ... lll c6

�� •�m lm�£%W§i ' , , J� i�l!l ,.,���

s i, m

7

6

: 3

l!ll!l l!I !ii l!l·l!l·l!l·l!I· ,,

l��l!l��!!10�, !!11!1 *'

2 � �� ��1���� � �V-@� n

'0 ""

a

18.l'!cl

b

c

d

e

f

g

h

I did not really consider l 8.b4!? during the game. After l 8 . . . lli xb4 l 9.ttJxb4 Wxb4 20.l'!b 1 We7 Black has a good game. Still, this is what White should have played.

18 ... l'!b8 Black is winning material, transforming his better coordination into somethin g of lasting value.

19.lll d4 l'!xb7 20.lll xe6 fxe6 21 .i.xc6 2 1 .Wd2 l'!c7 does not look like an improvement for White.

21 ...dxc6 Black is better. IfWhite could manage to take on c6 and get his pieces into the game, things would not be so simple. We know from the games of Tai that very often the rook is quite effective against two minor pieces in positions where there are not so many pieces left on the board. So for this reason it is important for Black to play energetically, not to lose the advantage his lead in development gives him.

Best was 22.Wd3 l'!f7 23.'tti g2 ttJ d5 24.l'!hfl + (not 24.f3 ih6!) , when Black is better, but the game goes on. He will have to find a way to make progress, which is not so simple. At least, this was what I was thinkin g durin g the game.

22 ... l'!f7? The computer gives the following variation as stron ger: 22 . . . ttJdS! 23.We4 l'!f7!

8

,�

�"'� �:e2 tt'ld4t 32.\t>f2 gives White a winning position, on account of 32 . . . b6 33.tt'ld? and it is all over.

26 ... h5 27.f4 A funny line is 27.g4! ?, when Black has to find the following defence: 27 . . . \t>IB! 28.gxh5 ©e8 and realize chat White has 29.E:xb7!?.

a

b

c

d

e

f

g

h

After this, White has nothing. The ending is maybe another way co illustrate poor play, but we will skip through it this time around, as it feels far from our topic.

28.fxe5 tl:!c5 29.gd4 a5 30.g4 hxg4 3 1 .tl:!xg4 @£8 32.�f6?! c.i? e7 33,gd6 gh8 34. c.i? g3 gd8 35.gxd8 c.i?xd8 36.@6 tl:! d3 37.e6 Accurate, but not difficult.

37 ... fxe6 38. c.i? g4 c.i? e7 39. c.i? g5 @f7 40.tl:!d7 b5 4I. c.i? h6 bxc4 42.bxc4 c.i? e7 43.�b6 @f6 44.tl:!d7t c.i? e7 45.tl:!b6 1/2-1/2

a

b

c

d

e

f

g

h

But it doesn't really work. Black is OK after 29 . . . E:xb? 30.h6 E:b8! 3 1 .hxg? \tie? 32.tt'lg4 E:g8 33.tt'lxe5 c5 34.tt'ld3 ©d6 and we have

The above is a good illustration of how a winning position is often squandered. Ac one moment, I missed the most accurate way to win the game, going instead for a superior technical position. Then I missed a finesse, and suddenly Black was back in the game, doing fine. The advantage was lose in two mistakes, not one. Now let's go a bit deeper with the following examples. Not that the mistakes will not remain simple, they will . . .

Chapter 4

-

The Nature of Tactical Mistakes at the Top Level

A tactical mistake in three acts

Every player has a different philosophy to annotating games. Kramnik likes to explain his decisions carefully, Giri gives his emotional experience, and Kasparov overwhelms you with variations. All these styles have something going for chem. In the analysis of the following very short game fragment, I will try to draw on all three approaches. I played this tournament in Budapest in 2003, organized by the Hungarian Federation in order to prepare Peter Leko for his World Championship match with Vladimir Kramnik. I did not play so well, and finished on fifty percent. In the last round, I played my old friend Judie Polgar (more about her on page 1 45). In the game, I made a simple tactical mistake and allowed her off the hook. We join the game where she gives me a chance.

93

26 . .!lib4? Judie is trying to ease the pressure on her position. Apparencly, it did not appeal to her to play 26.b4, when the a3-pawn would have been a continual annoyance. Still this was the better choice. I prefer Black, but the advantage is by no means overwhelming.

26 ... axb2 I played this with a reasonable amount of confidence. I had also seen the option to play 26 . . . i1tf c5 27.lll d3 i1tfxf2!?, but as I already had a clear route to an advantage, going for these complications was not necessary.

27.l3xb2 l3fa8? I played this more or less immediately. I could not see a way for White to get out of the uncomfortable pin down the a-file.

Act 1 - The Game

I have a feeling the simplicity of the oversight I made in the game might surprise some people. But it is what it is. Juclit Polgar - Boris Gelfand Budapest 2003

I had seen that 28.l'fo b 1 i1tf a4! is very unpleasant for White, and was quite looking forward to it. But alas, it was not to be.

28 ....L:a2 29.lll xa2 �e8

Boris Gelfand - Dynamic Decision Making in Chess

94

Black might still be ever so slightly better, but I had had enough, and offered a draw, which was unceremoniously accepted.

3 1 .'1Wxb2t '1We5 32.'1Wxe5t dxe5

�� � � � m� m ·� : �.-.,�. ��m;����r:.�·J�%��� � ��� ��� ��,rf � � �� � 8

l/i-1/i

Act 2 Where I thought -

5

4

I had missed my chance

,,_

When analysing the game in 2003, I believed that my real mistake was not taking on f4. But a slightly deeper look reveals that things were not so simple.

a

_

b

•

c

d

e

f

•

g

h

This endgame is quite dangerous for White. My feeling is that it probably cannot be held. It is easy to imagine that White might play g2-g3, when Black exchanges pawns, and later on his h-pawn for the g3-pawn, giving him two connected passed pawns. One of the pawns will claim a piece, and so on. It is important to point out that 33.lt:ld3 'tti g7 34.lt:lxe5 does not work on account of 34 . . . E:e8 winning a piece.

26 axb2 27Jhb2 exf4 ..•

8 7 6 5 4 3

But looking at this now, I am wondering what the evaluation should be after:

2 a

b

c

d

e

28.gabl! f

g

h

This is stronger than what I played in the game, but it is important to remember that when you are at the board, you do not play one move at a time, but quite often play in sequences. When I took on b2, I knew what my next move would be, as I should. After the game, it is natural to analyse the moves as if they were made without any special connection, and we should certainly do this. Bur it is important to remember that this is not how it happened. For this reason, I started this section by giving the previous move, as this is where I could have made a different decision. Analysing this in 2003, my main line went: 28 .i.xe6 E:xa l 29.'!Wxa l i.xc3 30.'!Wc l i.xb2

Yes, this move again.

28 .L.:a2 29.gxa2 gxa2 30.lt:lxa2 'l&c4 •.•

Chapter 4

-

3 1 .�b4! �xc3 32.�d.3 White has managed to create a fortress-like position. In many lines, the f4-pawn is under attack and in others, Black feels the lack of a hiding place for his king. We moved the pieces around a bit to get a feeling for the position; it did not fill me with great confidence. Black is obviously much better, but White is certainly not without defensive resources. 8 7

White can cake with the queen on f3, leading to an ending with decent drawing chances, or play:

37.gxe4!? gblt 38.gel gxelt 39.�xel f2 40.�e3! fl=�t 41 .�gl White is probably still within the drawing margin. Act 3 As expected, it is in the third ace that all is

revealed, and a way forward is found. At lease this is our aim. It turns out chat I should have gone for the most aggressive solution, trying to put my opponent under as much pressure as possible.

6 5

4 3

26 ...�c5!

2

1

95

The Nature of Tactical Mistakes at the Top Level

a

b

c

d

e

f

g

For example:

32 ... Ae5 33.�f3 �d4 34JUI gbs White has counterplay with:

35.�h3! Black's only winning cry is:

35 ... �e4 36.gel f3!

h

8 7 6 5

4 3 2

�

-�

-

��

��-� � ... 3.,,

"� - � ·� �iur/--- - %�-� �� · -%� � . , , %�f� - - - %� � �,� .. . ... :� ���----%� ----·\3rd----%� �c -- · ��r� �f� - - %� � �� 8 f�

1 � -a

b

c

.:

,.. . , . . Y,-� d

e

f

g

h

The queen is better placed on c5, defending the pawn on d6, while attacking c3 and the rook on f2. There is only one drawback co the move, which is the only reply I would ever consider calculating seriously in a game.

27.�d.3 This is the critical move. When you are calculating during a game, you may notice that other moves are possible, but their importance is often very different. With limited time, you

96

Boris Gelfand - Dynamic Decision Making in Chess

want to get into the habit of calculating only the important moves. I know this is difficult to do and I am not trivializing it, I just want to make the point that a grandmaster's brain does not work faster than an accountant who plays chess on club nights. They both do difficult mental work (and both get corrupted by computers!) . But the grandmaster will have a strong instinct when it comes to chess moves - and hopefully the good sense to hire the accountant to help out with his taxes.

28 . . . e4 There are some tactical options such as 28 . . . Wi'xc3! ? as well, but this seems the most promising to me. However, on 28 . . . E:fa8? 29.E:ab l ! still works. This time it is not the queen that is the target on the b-file, but the 8th rank. 29.E:c l Also 29.E:ab l d5+ is very pleasant for Black. The key tactical idea is that 30.i.xd5 E:d7 3 l .E:d2 i.xc3 simply does not work.

Besides 27.lLid3, there is another possible move, as suggested by the engines: 27.E:d2 It is possible this is the better move, once you analyse for some hours, as we have done - though I want to underline that we have not come to a final conclusion on this. There is an important point to be made on this question. First of all, Black has improved his position, meaning that the inclusion of the queen step to the side is an improvement. So who would actually play 27.E:d2 in a game? I am pretty sure that Judit would have moved the knight rather quickly, as would almost any player. White is already under a lot of pressure; I doubt she would have allowed things to get even worse. 27 . . . axb2 28.E:xb2

29 . . . dS!? In general, it seems preferable to keep more pieces on the board. Especially as the bishop on a2 is a likely target. But this does not mean that Black's position is not very promising after 29 . . . i.xa2 30.lLixa2 d5. It is. 30.E:d2 E:d8

��-�. - -�� ��%il-J� : ,,]� �� ,�-� _ �- �� ��� � : � ;;,,, , . , -����,� , , , %� � �.,3D 2 f�� ,� , ����� ,; . ��,�if �@ 8

:�

,,.

%

.

..

3

/" "'Y.�

a

b

c

d

e

...

f

g

h

Black has a stable positional advantage. All I know is that this would have been a very good shot in the game.

Chapter 4

-

This leads us to a highly critical cry for White, which however has some downsides. 3 1 .ixd5?! gad? 32.llJc6 e3 33,gd3 .ixd5 34.llJxdS Wc4!

�� � ',. .. . :� �Wief� � �� �i �.-� �� ,� r� 4 �� 1� �� �� ef.z · · ·. � g �3 � ��r� � � . . . z . . �� � � 2 ��1� �1� �

8

97

The Nature of Tactical Mistakes at the Top Level

:�

�

5

37.llJh3 A logical but still unfortunate retreat. 37.gd4 .ixd4 38.cxd4 e2 39.Wg l hxg5 and Black wins. The e-pawn is too strong. The best chance is 37.We2, although 37 . . . .ixg2t (not the only try) 38.Wxg2 Wxd3 39.llJf3 �h7 is very close to winning. 37 . . . ixg2t 38.�xg2 gxd3 39.Wf3 e2

��-�

a

b

c

d

e

f

g

h

A strong move chat creates threats against d8 and g2, and at the same time prevents White from playing c3-c4. 3 5 . llJ f7t Again the only move worth analysing. 35.gxe3 gxd8 is a slow grind; Black should win. In a practical game, this is what White should choose, but as we are analysing the viability of Black's ideas, and not just following an electronic overlord's commands, it makes more sense for us to look at the moves chat might cast the final resulc in doubt when a human is pondering over it. 35 . . . �gS 36.llJg5 Black wins quickly after 36.llJe5 ixe5 37.fxe5 ixg2t 38.�xg2 We4t!, when the white king will not find safety. 36 . . . h6

a

b

c

d

e

f

g

h

Slightly unnecessary, but still elegant. 40.Wxe2 Wc6t Black wins the knight and soon the game.

28.�xf'l axb2 29J�bl gxa2 30.9d2 I will mention a couple of alternatives; the

98

Boris Gelfand - Dynamic Decision Making in Chess

first is easily dismissed, but the second requires greater attention.

to make a decision, not to out-calculate the computer or any other such nonsense.

30.�xd6 !!a l and Black wins. 30.lll d3 E:b8! In 2003 I had analysed 30 ... e4, which also leads to a very pleasant position for Black. But looking at the position now, it seems more natural to defend the passer. 3 1 .lll b4

This is the critical move. I do understand that the engines will not like it, but as we are talking about decision-making, it is really the only move worth analysing. If Black is allowed to win a third piece for the queen, combining two amazing bishops with a rook, the game is over. Therefore, this is the only move that requires accuracy. 3 1 . . .E:xb4 This is of course the idea. 32.cxb4 exf4 33.b5 Again the only move that makes sense. Turn off the engine and look at the board! 33 . . . E:a l 34.b6 .id5!! A very nice point, which you would have to find at move 30 in order to go into this line. Obviously it is not something I would have ever considered even a move earlier, if I had played like this in the game. The point of calculating variations is after all only to learn enough about the position to be able

Now the bishops are coming to e4 and d4, threatening just about everything. Black would win even if his moves were drawn out of a hat. I like the following illustrative line, as it relies solely on the strength of the passed pawn. 3 5 .�d3 .ie4 36.b7 .ixb7 37.�xf5 E:xb l t 37 . . . .id5 ! is the engine's choice, but not something many GMs would find on move 37 in such a complicated game. 38 .�xb l f3 39.gxf3 .ixf3t 40.'it>gl d5 Black wins.

Chapter 4 - The Nature of Tactical Mistakes at the Top Level

3 1 .�dl From here the knight both defends the first rank and attacks the b2-pawn, making it a very logical move. Alternatives are:

99

32.l!Jd3 32.l::1 x b2 l::1 a l t 33.lt>h2 (33 . l!J d l J.e5 and . . . i.f7-h5 is coming) 33 . . . i.e5 34.l!Jd3

3 1 .l!Jd3 l::1 b 8 is simply a tempo worse than on the last move. There is no reason to think deeper about those consequences. 3 1 .h3

� � �� �� ���4--·

8

, ·

� �� ��,--� �- - - �.� r.�� �� �� � 4 �� �� - - "� 8 ��-J�- - �ilm �-�"���8 � 2 .1 • ���� f"' ' "----� >,,.,.,; �

.....

5 3

a

b

c

d

e

f

g

h

I analysed this natural move in 2003. In reply, I also suggested a natural way of consolidating the advantage, 3 1 . . .l::1 c 8, but today the engine offers a more direct approach. Its first choice, 3 l . . . J.c4, is highly artificial, but its second choice is something you could imagine a human playing in a game: 3 1 . . .l::1 g 8! Again, I want to stress that I am not pretending that a grandmaster would calculate the following sequence all the way until move 39. There is a chance he would find it all the way until move 36, although this too is unlikely. Why? Mainly because it is difficult to choose between all of the opponent's defensive resources. When analysing the game as we are doing it, with two organic brains and an occasional look at the engine, you are of course in a much better position to calculate to the end than you would be in a game, where you also have to keep everything in your head.

32 . . . f3!! 32 ... J.d5 is too slow on account of 33.l!J e l + and White can still offer a lot o f resistance. 33.gxf3 J.d5 34.l!J e l l::1 a l ! 3 5 .Wxb2 l::1 a2 36.Wb6 E1e2!

1 00

Boris Gelfand - Dynamic Decision Making in Chess

Simple prophylaxis. White cannot untangle himself. 36 . . . i.e5 37.E:b2! and White survives. With the rook on e2, the knight would hang. 37.Wfb5 E:d2 Black is threatening . . . i.e5 , which will win the game. 38.E:b2 E:d l 39.Wfe2 Axf3t! Black wins.

32.h4 Af7 33.lll xb2 ie5 34.E:e l f3 3 5 .gxf3 E:b8 and Black wins. It is certainly possible to explain the moves, if you are a strong player and used to working with computers. But the moves are exceptionally artificial and have nothing to do with decision-making in a game of chess.

32 .. J�al 33. ©gl i.f7!

At this point, Black is winning, but the way to seal the deal is rather surprising. The key idea is linked to preparing the manoeuvre . . . Ae6-f7-h5 .

� ��� �. � � �� � � � �-

, � �� �� � : �� �� �,-% �� �� -� � �--8

7

6

3

2 1

�

'""

�

,

� � -� �- rt�

i•"" -� !E

31 ... ges!!

----

�

�

a

-

b

� � r,7,� � c

d

e

f

g

h

We will see below why this move is so fantastic. After the game, in 2003, I had only analysed 3 1 . . .Ae5 32.E:xb2 E:a l , which gives White many problems, but does not offer Black a forced win. My main mistake was in the following line: 33.©gl f3 34.gxf3 E:g8t, where I correctly predicted that 35.©f2 f4! was winning for Black, but did not discover that 3 5 .©h l !?+ was a better defensive try.

32.gxb2 This is the main move a human needs to think about. The engine gives the following line as best play for both players:

34.l:k2 Again the only move we need to calculate. All others lead to lost endings in uncomplicated ways.

34 ... i.h5 35.gcl gxcl 36.YlYxcl gel t 8 7 6 5 4 3 2

1

a

b

c

d

e

Black wins a piece and later the game.

Chapter 4

-

The Nature of Tactical Mistakes at the Top Level

It is important to underline a few points I want you to take away from this game and its analysis: )- Most of what you have seen above was discovered after the game. )- My choice in the game was based on a simple tactical mistake. )- Don't for a second believe that most grandmasters would calculate any significant part of this accurately during the game. )- In many places, the engines suggest moves and lines that are irrelevant to the understanding of the game from a human perspective. The main lines above are the lines that would be important to analyse in the game, not the first line of the computer. It is all on a need-to-know basis - and the main lines are the stuff we need to know to understand the position well. )- Calculation is based heavily on intuition. An engine will calculate several million moves per second, while a human will take seconds to calculate only a few moves. So we need to train ourselves to calculate only the variations that will directly impact on our decision-making. )- There are a number of reasons why I have analysed the game in such detail. The main one being the beauty of the lines! But I also wanted to give real human-style lines to explain what directions the game could take and what variations would need to be found. Missing a single move in a combination

There is a part of calculation that we can improve indefinitely, but never perfect - we will always miss things. At times, it is not the combination itself we miss, but a small but essential part of it. The following example is a good illustration of this theme.

101

Boris Gelfand -Vladimir Kramnik London 2013

I .d4 tlif6 2.c4 e6 3.tlc3 .ib4 4.e3 0-0 5 ..id3 d5 6.tlif3 dxc4 7.hc4 c5 8.0-0 cxd4 9.exd4 b6 10.°i'e2 .ib7 1 1 . .ig5 tlbd7 12.�facl

12 ...°i'b8!? A new idea from Kramnik. This is just one of the many novelties Kramnik introduced in London. He came to the tournament with immense determination and showed great chess. But up to this point he had not been able to translate this into any victories. He was increasingly taking more risks. The previous day he had sacrificed a rook against Ivanchuk for the initiative. It was probably too much, but still the game ended in a draw.

13JUdl �k8 14 ..id3 .id6 15.g3 a6 16 ..ixf6 tlxf6 17.tlie4 gxcl 18.gxcl It is important to look at Kramnik's next move in the context of the tournament situation. After six draws, most of them from better positions, he was taking more and more risks in the search for much-needed victories. One win arrived in the next round, when he produced an absolute masterpiece against Peter Svidler.

Boris Gelfand - Dynamic Decision Making in Chess

1 02

Kramnik is trying to keep some of the dynamics of his position going, while I am j ust making moves in accordance with the position, which I was not happy with at all.

19.�ed2? After a long think, I found nothing. During the game I spent quite an amount of time looking at: 1 9.lDeg5! a

b

c

d

e

f

g

h

Lacer it was suggested by some, mainly behind closed doors but also in some cases in public, chat Kramnik had received a bit too much help from his compatriots in the second half of the tournament. Those accusations are simply ridiculous. For chose wanting the full inside story, I strongly recommend Carlsen's Assault on the Throne by Kotronias and Logothetis. Even for me as a participant, there are many new insights in chat book. In the second half of the tournament there were rumours, while we were still playing, that I had 'sold' my game to Kramnik. Those rumours died when I held the game. Please do not listen to these people.

With all the black pieces in the other corner, it is natural to look at an active reaction to a passive-looking move. But we should be careful not to extend such comments too far. After the game, when the engines were switched on, it was obvious to everyone chat White was winning here. But during the game, all that was available to me was the feeling chat an attack on the kingside might work. In order to determine if it did work, I started calculating the most challenging option: a) 1 9 . . . g6

18 ... tlieS? 8 7 6

a

5 4 3 2

1

a

b

c

d

e

f

g

h

b

c

d

e

f

g

h

Black strengthens the defence of the light squares. But it is only the pawns that defend the light squares, so it is natural to go for a direct attack. No, I do not chink like this during the game. I am guided by an inner feeling, not a walking dictionary of platitudes. However,

Chapter 4

-

The Nature of Tactical Mistakes at the Top Level

this is the best way I can explain why I felt that White might be successful with the following move: 20.lll x f7! 'it>xf7 2 1 .lll g5t 'it>f6 22.Wxe6t! The most natural move and the one I looked at as a try to make the combination work. Those with an engine can see that White has other promising continuations here, but this was not at all apparent to me during the game. 22 . . . 'it>xg5

b) 1 9 . . . h6

a

a

b

c

d

e

f

g

h

While looking at this position, I did not find a way to continue the attack. It seemed to me that Black was just winning. However, it turned out that I was wrong. 23.Wh3!! White has this surprising move. The threat is 24.Wh4 mate. Black can defend either with 23 . . . h5 24.We6, when White wins because of Black's inability to defend g6. Or with 23 . . . 'it>f6, when 24.Wxh? leads to similar problems on g6. It should be mentioned that White has other moves that offer him strong attacking chances instead of this directly winning move. But during the game, none of them looked convincing to me, and there is not really a lot more to say about that. Having failed to make the first variation work, I did not seriously look at the alternative:

1 03

b

c

d

e

f

g

h

Why would I ? It made no sense to investigate this, when I was struggling to find anything convincing against 1 9 . . . g6. But for completeness, I should point out that White has two ways to win the game here, all based on the absence of defence on the light squares. It is most likely I would have gone for: 20 ..ih?t Instead 20 . .ig6!! is a computer move, but not an uncommon tactical theme. 20 . . . hxg5 (Black can fight the tide, but after 20 . . . ixf3 2 1 .ixf7t 'it>f8 22.Wxf3 hxg5 23.i.xe6t lll f6 24.l'l:c8t Wxc8 2 5 .ixc8 l'l:xc8 26.Wb? we have an endgame with great winning chances for White.) 2 1 .lll x g5

��� � ��f%· : T���TA! , . . , %� �� ��� ',., ., %� � w 0 � � 4 �� � �� - - 3� �� � �'ft""z="'�, 2 !�z�� 8 .I � � , ., .,;g8 24.Wh?t 'it>f8 2 5 .Whst 'it>e7

1 04

Boris Gelfand - Dynamic Decision Making in Chess

26.%Yg8!!

7

6 5

4 3

2

a a

b

c

d

e

f

g

h

Black is lost. 26 . . . %Yc8 ! ? 27.%Yf7t 'it>d8 28.lLie6t And so on .. . 20 . . . 'it>h8

7

b

c

d

e

f

g

h

23.J.xf5! exf5 23 . . . J.xg3 is better according co the computer, as it gets a pawn for the bishop. White is still winning, but there is another point as well: who would actually find this move? 24.%Yxf5 lL! f6 2 5 .%Yh3t 'it>g8 26.%Ye6t 'it>h8 27.lLif7t 'it>h7 28.lL!xd6

6 5

4 3

2

a

b

c

d

e

f

g

h

2 1 .%Yd3!! hxg5 22.lLixg5 It is quite remarkable chat Black is unable co defend f7. 22 . . . f5 The only attempt. If f7 cannot be defended, then at least the defence of the h7-bishop can be blocked. 22 . . . g6 23.J.xg6! clearly leads co a winning attack. Variations are unnecessary.

a

b

c

d

e

f

g

h

White has a piece less, but is totally winning. Black does not have a defence against g3-g4 with the ideas g4-g5 and l:!cl -c3-h3 mate. The main line to include is: 28 . . . id5 Black can give up the bishop with 28 . . . %Yd8, but even with 23 ... J.xg3 24.hxg3 included, the position would be lose for Black (although then White would have co show some technical skills) . 29.%Yh3t 'it>g8 30.l:!c8t Winning.

Chapter 4

-

1 05

The Nature of Tactical Mistakes at the Top Level

I will add that l 9.tt:lfg5 h6 20.\Wh5!± also leads to an advantage. But as I did not consider it, it is not so relevant.

always good to go back to Moscow, where I have been coming since my teenage years, and win this nice event in the final days of the year.

It is possible that I would have failed to find anything after l 9 . . . h6 as well. I played really poorly that day.

Grigoriy Oparin - Boris Gelfand

19 ...\Wds The rest of the game became rather placid.

20.ie4 gc8 21 .Wffl gxcl 22.'IWxcl Wfc8 23.'1Wc3 c!LJf6 24.hb7 Wfxb7 25.llJe5 tlJd5 26.Wf c6 Wfxc6 27.tlJxc6 'itif8 28.llJc4 ic7 29.c!LJe3 lLixe3 30.fxe3 id6 3 1 .tlJe5 ixe5 32.dxe5 'itie7 33. 'itif2 'itid7 34.e4 'itic6 35. 'itie3 'itic5 36. 'itid3 1/z-1/2

Missing the opponent's defensive resources

The next game was played in a very interesting tournament, which was held during the Western holiday of Christmas. In Russia the seasonal celebrations are at the start of January, while Middle Eastern countries have their important celebrations at other times of the year. For example, at the same time as this tournament was going on, there was an important open tournament in Qatar, with participation from Magnus Carlsen and Vladimir Kramnik, the two top-rated players in the world at that moment in time. The Nutcracker tournament was a mix of games with a classical time control and rapid games, played in the Scheveningen format, where all the players of one team play against all the players of another team. As so often is the case with these things, one team consisted of promising young players and one team consisted of experienced players. I liked this format when I played in Amsterdam in 20 I 0, so I looked forward to this event. Also, it is

Moscow (rapid) 2015

l .e4 c5 2.c!LJf3 d6 3.d4 cxd4 4.lL!xd4 c!LJf6 5.c!LJc3 a6 6.h3 e5 7.llJde2 h5 8 ..ig5 .ie6 9 ..ixf6 Wfxf6 10.tlJd5 Wfd8 l l .Wfd3 tlJc6 12.0-0-0 g6 13.'itibl ig7 14.h4 gc8 15.g3 b5 16.ih3 0-0 17.c3 tlJa5 18.f4 ixh3 19.gxh3 exf4 20.gxf4 tlJc4 8 7 6 5 4 3 2

1

a

b

c

d

e

f

g

h

The opening has not gone well for Black. If White now played 2 I .tt:ld4, then he would have a big advantage after something like 2 1 . . .Ei:eS 22.Ei:g3 with f4-f5 coming. Instead Oparin went for a speculative and highly committal knight sacrifice.

21 .liJg3? The downside to this move is not so much that it sacrifices a piece, but more that it costs quite some time and that it allows Black to act freely on the dark squares along the g7-b2 diagonal.

21. b4! •.

An easy move to make. There is no time to waste.

Boris Gelfand - Dynamic Decision Makin g in Chess

1 06

22.�xh5 The engines suggest that 22.tll e2 is a better move, but no human would ever make it. He went to g3 with the idea of taking on h5, and that is what he did. Chess is not played move by move, as already discussed.

First of all, we should remember chat this is a rapid game, so the players had limited time. Secondly, we should not forget chat while we can say afterwards chat this is the critical moment of the game, this is not so easy to determine during the game. While playing, you can realize chat you have a choice, but you have no possibility of estimating how important this choice will be for the rest of the game.

8 7 6 5 4 3 2

1

This is the critical moment of the game, and there are a few things I would like to say about it.

a

b

c

22 ... tll xb2!

d

e

f

g

h

This move is also quite easy to make. Black has obvious compensation, but also, I would never play something like: 22 . . . bxc3 23.bxc3 gxh 5 , when after 24.i'!g l I would have to find 24 . . . f6, which appears to be necessary, and then hope chat White will not be able to find more than a perpetual. . .

23.'i!?xb2

Thirdly, it is quite easy to come up with an explanation of why taking with one piece rather than another is the right choice, once you have the support of engine analysis after the game. During the game, this information is not available, so don't let it fool you into chinking it should have been easy to choose the right move. In the game, I thought for a few minutes, maybe five, and made a move I thought was quite convincing. It worked wonders in the game, being evidently strong enough to pose Oparin more problems than he was able to deal with in the limited time he had left. But after the game, the computer suggested some surprising variations and conclusions.

23 ... bxc3t?! It turns out that the correct continuation was to upgrade the knight sacrifice to a rook sacrifice: 23 . . . i'!xc3! 24.tll xc3 bxc3t The critical line involves White putting his king in the corner. a

b

c

d

e

f

g

h

Chapter 4 - The Nature of Tactical Mistakes at the Top Level

1 07

to the game continuation (if White had responded with 24. g7 25 .'\Wd5 h6 26.E!:fl

Chapter 5 - Compensation

131

Lee's have a quick sidestep here. Alex likes co talk about "Vaisser's bishops", referring co a game he played three decades ago. le had a lasting impact on his chess understanding, and as I have worked with him for twenty-five years, some of it has rubbed off on me.

Anatoly Vaisser - Alexander Huzman Sevastopol 1 986 b

a

d

c

e

g

f

h

With the threat of We4. 26 . . . ih3 27.lll e8t 1 -0 Cebalo - Murshed, Subocica 1 984. I only found out that this game existed lacer on. I had made several moves on my own ac this point.

8

7

6

.i �.i �J,

. ,,Y,_� �· , . ,�%� �-

!I•'•

"!I '!I !l !I '!I ' !I

: !l!I�%'!1,,,, 1 !l ��p!l!I "/�'1:�·0 !I 3

r�� �"ilm . %�, . , , .� ��� '0

_

. ...

2

I

� a

19.hd4

b

'!I c

,,,,.

e

8 7 6 5 4 3 2 a

Y,

'!l�f«;L d

l .d4 d5 2.c4 c6 3 .lll f3 lll f6 4.lll c3 dxc4 5 .a4 ig4 6.lll e5 ih5 7.g3 e6 8.ig2 ib4 9.lll xc4 lll d5 1 0.Wd3 ie7 1 1 .0-0 0-0 1 2.aS lll a6 1 3.lll e5 lLi ab4 1 4.Wd2 c5? 14 ... f6 with even chances was more prudent. 1 5 .lll x dS exd5 1 6.dxcS ixc5

f

g

h

During the game I also thought chat l 9.!!xd4± was interesting, but I was dissuaded by something. I remember noticing chat l 9 . . . !!b8 was possible, as White does not win a pawn with 20.!!a4 a6 2 l .!!xa6? on account of 21 . . . !!xe3! 22.fXe3 ie6 and Black wins a piece. I was not desperate co win back the pawn, but I did not wane co allow Black co play . . . !!b8 and untangle his position. In shore, it makes no sense co cake with the rook when I can cake with the bishop.

b

c

d

e

f

g

h

1 7.lll d3! lll xd3 1 8.Wxd3 We? 1 9.WxdS! ixe2 20.!!e l 20.ig5 ! ? was also very strong, but Vaisser had been bewitched by a beautiful idea. 20 . . . ic4 2 1 .!!xe? ixd5 22.ixdS! Axe?

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making i n Chess

1 32

23 . .ie3!! This is the idea. After the natural 23.ixb?? Black plays 23 . . . �ab8 24.a6 .ic5 and White does not have sufficient compensation for the exchange. He needs both bishops to be able to dominate his opponent. 23 . . . �ab8 23 . . . �fb8 24.a6 bxa6 2 5 .�xa6 also leads to an endgame where White has an extra pawn. 24.ixa7 �fd8 Vaisser gave the following line back in 1 986: 24 . . . �a8

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

2 5 . .ie3! (25 . .ib6 �ab8 is less clear) 2 5 . . . .ib4 (He does not say, but one can presume that Vaisser was planning to meet 25 . . . �ab8 with 26.if4, when 26 . . . �fd8 27.ixb8 �xb8 does not work - there is no salvation in the opposite-coloured bishop ending. White plays 28.�a4! and pushes the pawns up the board. Black is completely lost.) 26.ib6 The bishop is forced here, but with the black bishop having moved to a worse position. 26 . . . �ab8 27.�a4 id6 28 . .ia7 �a8 29.ie3 �ab8 30.b4 �fd8 3 1 ..if3 White is close to winning. As far as I can see, this all holds up to computer scrutiny. It might seem hard to believe for the youngsters, but we were able to analyse our games decently even before the invention of the engine . . . 25 . .ixb8 �xd5 26.�c l !

White has a technically winning ending, but things can still go wrong of course. The rest of the game is not directly connected to our theme, but it is quite interesting, so I have taken the liberty of including it. This is a chess book after all. . . 2 6 . . .g 6 27.�c7 if6 28.�xb7 �xa5 29.b4 �a l t 30. g2 gb3 47.if2

47 . . . i.xf2 Vaisser gives the following variation: 47 . . . gxb5 48.gd2 gb4 49.'tt> g3 'tt> e4 50.i.xd4! gxd4 5 1 ,ge2t ©d5 52.ge5t 'tt> d6 53.gxf5 'tt> e6 54.gxh5 @f6 where Black reaches a drawn rook ending. Of course he would have to do some defending, but first of all, Alex knows his stuff, and secondly, we played with adjournments back in the day, which would have allowed him to refresh his memory about the finesses of this ending. So Vaisser is right in his analysis again, both objectively and from a practical point of view. But Black is doing better in the game than Vaisser believed. 48.'tt> x f2 'tt> d4 49.ge5 h4 50.'tt> g2 h3t 5 1 .'tt> f2 ga3 52.b6 gb3

1 33

53.b7 Vaisser gives 5 3.gxf5 gxb6 54.©g3 gb3t 5 5 .'tt> g4 gb2 56.gh5! gxh2 57,gxh3 and White wins. But Karsten Muller pointed out chat Black is within the drawing margin after 53 . . . ©e4!, although in practical terms, the defence is by no means easy. 54.gf6 gf3t 5 5 .'tt> e l gb3 56. 'tt> d 2 @f3

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

This is where Muller stops his line, but I found chat this is only the beginning! There are two continuations which I really liked: a) 57.f5 'tt> f4 5 8 .'tt> c2 gb5 59.'tt> d 3 E!b3t 60.'tt> c4 gb2 6 1 .gh6 'tt> g 5! 62.'tt> c3 gb l 63.gd6 gb5! (63 . . . 'tt> x f5 64.'tt> c4) 64.f6 'tt> g6 and Black draws. b) 57.'tt> c2 gb5 5 8 .'tt> d3 and Black has to find a very counter-intuitive move.

Boris Gelfand - Dynamic Decision Making in Chess

1 34

65J'hh2 E!:xb6 66.E!:h6! Instead Black has to find 58 . . . @g4!, which is really surprising. Whenever you start to analyse chess, you always end up finding amazing stuff! 53 . . . E!:xb7 54.@g3 E!:g7t 5 5 .@xh3

Anyway, it is easy to explain why I felt optimistic at this point: I had Vaisser's bishops, squeezing Black hard down the two diagonals, but I had not given up the exchange!

19 ... c!bes

I was planning to meet l 9 . . . E!:e8 with 20.E!:c l lt'i d7 (this is what I expected) 2 1 .E!:fd l (2 l .id5 is premature because of 2 l . . . E!:e7) . And if now 2 1 . . .lt'ie5 White has 22.id5 with excellent compensation. For example: 22 . . . E!:e7 23.ixe5 E!:xe5 24.ixf7t with an overwhelming advantage.

5 5 . . . E!:g8 ? A blunder. Correct was 5 5 . . . E!:gl ! 56.E!:xf5 @e4 57.E!:g5 E!:fl 58.E!:g4 @f5 and Black is ready to cake on f4 and claim a draw. Again pointed out by Vaisser, of course (by the way, 58 . . . E!:xf4?? 59.@g3 should be avoided) . 56.E!:xf5 @e4 57.E!:g5 1-0

But maybe 20.E!:fe 1 is even more precise, as 20 . . . lt'i d7 2 1 .E!:xe8 'Wxe8 22.lt'i b 5 is also very bad for Black. Against l 9 . . . 'Wa5 White has many good moves. I was planning 20.ih3, winning the exchange, as 20 . . . E!:e8 2 l .ixc8 E!:axc8 22.ixf6 gxf6 23.lt'id5 is a disaster for Black.

When I talk about my feeling for a position, it is important to understand that it is always based on 40 years of continuous study of chess games, and the conclusions I drew from these games. Others will see things differently, but not necessarily worse or better! 8

a

7

20.c!bdS!?

6

b

c

d

e

f

g

h

I decided that my position was very strong and that it was time to cash in. I do find it very tempting to calculate a long line to a clear advantage when I have such an overwhelming position as I have here. I guess it is a question of attitude.

5 4 3 2 a

b

c

d

e

f

g

h

Chapter 5

-

The other move I seriously considered during the game was 20.f4, with attacking ideas involving f4-f5 . This was also strong and would be the natural first choice of a player with a different sensibility.

20 ... �c7 21 .:Scl �xd5 2 1 . ..�a6 22.f4! is very strong for White. This could have arisen by transposition had I chosen 20.f4, but against that option it is not necessarily the best line for Black.

22.J.xdS ge7 This is a very instructive position. White is a pawn down, but has total domination. It reminds me of some of Rubinstein's games, which is not so surprising, as I played the Rubinstein system.

a

23.:Sfel

b

c

d

e

f

g

h

I also considered 23.'1Wf3 ih3 24.E:fe l E:xe l t 25 .E:xe l '1Wd7 26.E:c l , taking o n b 7 on the next move. Naturally White is better, but I was looking for more.

23 ... :Sxelt During the game, I thought 23 . . . g 6 was the best defensive chance. But analysing the position now, White looks quite a bit better.

Compensation

1 35

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

It is natural to ask Black what he is actually planning to do, by playing 24.g2, taking away the h3-square from the light-squared bishop. Black has only one move: 24 . . . ig7 2 5 .ixg7 This is the most promising option. (White can also play 25 .E:xe7 '1Wxe7 26.ixf7t '1Wxf7 27.'1Wxf7t xf7 28 .E:c7t e6 29.E:xg7±, but I am not too fond of this possibility. Black has more drawing chances than I would like.) 25 . . . xg7 26.'1Wc3t g8 27.'\Wf6 ie6

a

b

c

d

e

f

g

h

28 .ixe6 (28 .E:xe6? allows Black to survive because of 28 . . . E:xe6!, which is easy to miss) 28 . . . fxe6 29.E:xe6 And now both 29 . . . E:f7 30.'1Wd4 and 29 . . . E:xe6 30.'1Wxe6t g7 3 1 .h4 are awfully close to lost for Black, if not just simply lost. My plan during the game was to play 24.E:e4, forcing 24 . . . ifS, as White otherwise brings the other rook to the e-file. 25.E:xe7 '1Wxe7 26.'1Wxb7 '1Wxb7 27.ixb7 E:e8 28.idS!

Boris Gelfand - Dynamic Decision Making in Chess

1 36

23.Ei:fe l . During the game, my opponent certainly felt the difference, and did not manage to come up with the best defence. It is possible that 20.f4 was objective stronger, but it is a bit hard to criticize a move that was good enough during the game, j ust because Black could have offered a bit more resistance if he had found the most accurate line. a

b

c

d

e

24.Ei:xel Wfc7

Black's position is very unpleasant. We can add a few moves: 28 . . . a5 (28 . . . Ei:e7 29 ..if6 is also nasty) 29.Ei:c7 .ie6 30 . .ixe6 fXe6 3 I .Ei:a7 White seems to be winning, although there will be plenty of technical difficulties. So my chances in this line were probably better than I expected.

I think he missed something around here.

23 . . JWe8 24.Ei:xe7 Wixe7 allows a simple tactic:

4

8 7 6 5 4 3 2

8 7 6 5 3 2 1

25.Ei:e8! The human move, which he clearly hoped was not very strong. a

b

c

d

e

f

g

h

25 . .ixa7! After 25 . . . .ie6 26 . .ie3 White has a clear advantage in the endgame. Black's pieces are not playing, while White's queenside majority is likely to come into play soon. In the end it turns out that 23 . . . .ih3! was the best chance. After 24.Wif3 Ei:xe l t 25 .Ei:xe l Wid7 we have transposed t o the 23.Wif3 line mentioned above. White is better, but Black can resist for a long time. I should point out that this does not mean that 23.Wif3 was j ust as good as the game move

I knew during the game that his intention was to play the line: 25.Wif3 .ie6

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

Chapter 5 - Compensation As the ending after 26 . .ixe6 fxe6 27.!!xe6

Vff fl is not all bad for Black. But the engine comes up with a beautiful and instructive suggestion: 26.!!xe6!! fxe6 27 . .ixe6t ©h8 White has huge compensation. 28.h4! White is rolling his pawns up the board. There is no way for Black to bring his pieces into the game without losing them, making his position untenable. An alternative could be: 28.Vf!h5 The direct threat is .if5 . Black has only one move. 28 . . . Vf!c l t 29.©g2 Vf!c6t Black is in time to bring the queen back to e8 without allowing Vff f7 -g8 on the way. 30 . .id5 Vf!e8 3 I .Vff f5 ± 2 8 . . . !!e8 This is an illustrative line, showing that Black cannot j ust wait. 28 . . . d5 29.Vff h 5! is immediately over. There is no check on c6! 29 . .!d5 Vf!d7 30.©g2 a5 3 1 .h5 a4 32.h6 !!d8 33.g4! !!b8 34.g5 !!c8 3 5 .Wf4 b5 36.g6

�i �� -�-J• 6 �� �!�� -� �� -� �j� 54 �4·1.r ��Uffi,·� ,-�."� � , , , -� �� ""' "� � ��@� �� 2 -----��r% •�-�r% %,i8%, ,Jfl Wh l t and wins. The rest is simple. 44 \Wg4! 45.'it>g2 \Wf3t 46.©h2 \Wxe3 47.6 \Wc5 48Jffl b4 49.f6 b3 50Jkfl c2 5 1 .YlYcl e3 52.f7t ©f'8 53J�f5 b2 54.\Wxb2 cl=YlY 55.\Wxg?t ©xg7 56.gg5t •..

Compensation

141

In this tournament, things were not going well for me, so I was lacking a little in confidence with my play. Especially I did not feel comfortable at the prospect of defending a technical ending against Vladimir Kramnik, who is famous for extracting the utmost from every piece he has in this type of position. Vladimir Kramnik - Boris Gelfand Wijk aan Zee 2008

a

b

c

d

e

f

g

h

Spassky resigned, not having the audacity to believe that Petrosian would play 56 . . . 'it>h6, although this is the move given in my database.

0-1 As I said, this made a big impression on me, and we shall see below that I managed to use this idea with success. But this does not mean that it was new when Perrosian played it. Looking in the database, we find Saint-Amant - Staunton, Paris (2 1 ) 1 843. Not the greatest game in history, but it is possible that Petrosian knew this example and was, consciously or subconsciously, inspired by it. Let's move on to my example. As we shall see, simply copying the idea is a simplification of what is happening. Every position is different and has its own challenges, despite the immense help knowing a lot of thematic options gives us.

a

29Jkl

b

c

d

e

f

g

h

The position might look innocent, but I knew I had to be careful. If I exchanged on c l , it should be fine with the knight on e4, but when you play against a guy who has great technique, you try not to give him even an inch.

29 Jk4 30 .ifl ••

•

Of course I could exchange rooks here, but I did not want to exchange before I knew I could put the knight on c4. When you analyse this position, you will find that I did not play the only move, but as I said, I did not want to relax prematurely.

30 tld6!? ••.

30 . . . lll gS ?! is a curious idea, but after 3 1 .i.xc4 dxc4 32.Wc2! it is hard to make it work.

Boris Gelfand - Dynamic Decision Making in Chess

1 42

safe. This is chess. There are always details and the game goes on (except when it doesn't!) . All we can do is try to make the best possible decisions in the time we have available. 32.i.f2 Wfc2 33.Wfxc2 l:!xc2 34.Ad3 l:!a2 34 . . . l:!c7 3 5 .Ae l;!; looks a bit unpleasant. 3 5 .l:!cl l:!xa4 36.l:!c7

a

c

b

31 .Wfa3

d

e

f

g

h

This is not challenging. 3 U!d l ! This was the only try. One o f the core ideas is co take on c4 and play d4-d5. This would have forced me to play accurately. 3 1 . . .Wfe4 3 1 . . .l:!xa4? is inadvisable. White has a very strong continuation: 32.°Wb3 l:!c4 33.i.b4! and Black loses material. 33 . . . l:!c8 34.Wfa3 l:!d8 3 5 .l:!e l followed by exchanges on d6, l:!e8t and l:!xa8. For this reason Black should probably play 3 1 . . .l:!c8 32.Wfa3 Wfe6 where his position has improved:

a

b

c

d

e

f

g

h

White has abandoned the c-line and Black is ready co put the knight on c4. But White can still try a few things, and the draw is not yet

There is a final dimension to the exchange sacrifice that I will j ust briefly mention: it is aesthetically pleasing.

3 1 . Wfe4 32 .lfl :Sxcl 33.Wfxcl c!Lic4 ••

•

Chapter 5 - Compensation

34.Yfdl .ib7 35 ..id3 Yfe6 36.Yfh5 tll d6 37.Yfe2 Yfxe2 38.i.xe2 ©f7 39 ..iel .icS 40.©fl tll c4 4I ..ib4 g6 42.©el a6 43.bxa6 ha6 44..ic3 .ib7 45.©fl .ic6 46 ..idl tll d6 47.©e3 ©e6 48 . .ib3 l/2-1h

The following game is our last on this theme, and I would not have included it in this book if my editor had not stressed the need for it. I do not consider it much of an achievement, as all that happened in it was that I : a) Used a recent discovery i n the opening by another player, and b) Used a very common exchange sacrifice to eliminate my opponent's defence on the dark squares.

1 43

This wonderful idea was introduced the year before by Shomoev, and immediately made the line untenable for Black. Amonatov was not very active at the time of this game, and he might simply have missed this latest development.

13 ... gxf5 14.gxf5 d5 The stem game went like this: 14 . . . tll bd7 1 5 .!!g l @h8 1 6.%Ye l tll e5 1 7.%Yh4 tll h 7 1 8 .%Yxh5 Yf e7 l 9.!!g3 tll d 7 20.!!agl tll d f6 2 1 .Yfh4 tll xe4

� �£. �!�� -,i ·

8 7 . �. · -....�El�. •••• · � .,�.

6

.

...

%

.

/.

54 �� �� �� �.

� ----7. � � � � �

,, . . ..}. .. -/.� -!�� 3 ���· · ·:·-:.-8 �· · ·�� 2 � 7. � � .

But what this game can show is that simply applying common themes can win easy victories, even on the top board in the Olympiad. Boris Gelfand

-

Farrukh Amonatov

Tromso Olympiad 20 1 4

I .d4 tll f6 2.c4 g6 3.tll c3 .ig7 4.e4 d6 5.tll a 0-0 6 ..ie2 e5 7.0-0 exd4 8.tll xd4 ge8 9.6 c6 10.©hl lll h 5 l l .g4 tll f6 12 ..if4 h5 13.tll f5 !

•••••

a

b

� c

d

-�

� e

f

g

h

22.%Yxh7t! @xh7 23.!!xg7t @h8 24.fxe4 ixf5 2 5 .exf5 !!g8 26.!!7g3 ?! 26.ig5! %Ye5 27.!!xf7, with if6t coming, was lethal . 26 . . . !!xg3 27.!!xg3 !!g8 27 . . . !!e8 offered more resistance, but White is close to winning all the same. 28 .!!h3t @g7 29.ih6t @f6 30.!!e3 Yfc7

Boris Gelfand - Dynamic Decision Making i n Chess

1 44

3 I .id3? A blunder. 3 l .lll e4t @e7 32.if4! was the most convincing way to clinch the full point. 3 1 . . .Wi'b6 32.lll e4t �xf5 ? Returning the favour. After 32 . . . �e7 33.f6t �d8 34.lll xd6 Wi'xb2 it would be White who had to hold the balance. 33.lll g3t �g4 34.if5t �h4 3 5 .ge4t 1 -0 Shomoev - Ganguly, Moscow 20 1 3.

thematic. A quick look in the database shows that there are 5448 games with this typical sacrifice, where White won. I will have seen a lot of them in my time. One of them is this very early example:

Joseph Henry Blackburne - Adolf Anderssen London 1862

15.cxdS cxd5 16.ggl I vaguely remembered reading somewhere that 1 6 . . . dxe4 was the main line. I had not paid attention to the move played in the game. In the book Best Attacking Games of 2012-2015, Csaba Balogh tells a story about talking to another top GM, who apparently had analysed this till move 2 1 , but said that I was fortunate to get there first, referring to computer preparation. I have to admit I got there on my own. If this is a good thing or not, will be up to others to decide.

16 ... @hs 8

a

b

c

d

e

f

g

h

2 1 .gxh7t �g8 22.gxg7t! �xg7 23 .Wi'h6t �g8 24.if6 1 -0 Not very sophisticated for the modern palate, but this is exactly the point. Blackburne found this idea on his own and we, having seen the impact over and over, can copy it with a high degree of certainty. My novelty is an improvement over 1 7.e5 ixf5 ! 1 8.exf6 Wi'xf6 where Black had a decent game and even went on to win in Petrukhina Chigaev, Tomsk 20 1 3 .

7 6 5 4

17 ... @xg7 18.1;Ygi t @hs 1 8 . . . @f8 1 9.ih6t �e7 20.ig5 and Black is caught in deadly pins.

3 2

19.i.gS! a

17.gxg7!

b

c

d

e

f

g

h

Not a very difficult novelty, even without the help of the engines. Eliminating the defence of the dark squares in this way is incredibly

Exploiting the weakness of the dark squares in all directions.

19 ...i.x6? The final mistake, but Black's position was horrible anyway.

Chapter 5

-

For example, 1 9 . . . � bd7 20.�xd5 �g8 2 1 .h4 does not offer Black an easy way out of the pins. And 1 9 .. .'Wd6 20.Wfd4 � bd7 2 1 .�xd5 �e5 22.�d l also wins material in the long term. For example: 22 . . . �xd5 23.exd5 Wh7 24.Wfh4 b6 25 . .ib5 .ib7 26 ..ixd7 Wfxd7 27.ixf6 Wfxf5 28.Wf e4 Wfxe4 29.fxe4 and White has excellent winning chances. Nonetheless, this was the last chance.

20.tll xd5 l£!bd7 2 U�dl! 8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

Bringing the last piece into the attack ends the fight. White wins back the material investment with dividends. Instead, taking on f5 would allow the white attack to lose its momentum.

21..J�gS 22.tll xf6 gxg5 23.'i!¥xg5 Y!rxf6 24.Y!rxh5t ©g7 25.exf5 l£!b6 26.Y!rg4t ©£8 27.Y!rb4t 'i!¥e7 28.'i!¥c3 f6 29J�gl gcs 30.Y!rd2 1-0

Compensation

145

An obvious exchange sacrifice Boris Gelfand -Judit Polgar Khancy-Mansiysk (3 .1) 2009

write about the World Cup in Khancy­ Mansiysk lacer in this book {page 227) , but I want to talk a little about my opponent in this game. Judit Polgar is a unique personality, who has definitely shown chat women can play chess at the highest level. It was not a fluke either, as she kept her level for decades. She is definitely a hero for other female chess players, but more importantly, in my opinion, is that she played really attractive chess. Judit managed to hold the Number 1 spot in the Women's rating list from January l st 1 989 until she retired from professional chess after the 1 Och round in the 20 1 4 Chess Olympiad in Tromso. As far as I know, no other female athlete has been Number 1 for 25 consecutive years in any sport. Quite an achievement. I played her for the first time in 1 989, when she was only 1 3 years old; quite a nice game by the way, with a lot of instructive moments. Maybe there will be space for it later in this series. Judit had a bone-crunching style when she rose to fame as a j unior, but after she worked for a few years with Lev Psakhis, she realized chat the direct approach did not always work, although she would still strive for an attack whenever possible.

l .d4 tll f6 2.c4 e6 3.tll a b6 Judit played the Queen's Indian for a very long time. I believe it was Lev Psakhis who convinced her to play it first. She also chose this opening against me in our last game in 20 1 3 .

4.g3 .ia6 5.Y!ra4

Boris Gelfand - Dynamic Decision Making in Chess

1 46

general ideas and relied on them, but not well enough to tell them apart.

8 7

10.gdl �e4?!

6

As mentioned above, 10 . . . lLia6 was the right move. Curiously, Judit had already played this correct option against Van Wely and later repeated it against Fridman.

5 4 3

1 1 .�xe4

2 a

b

c

d

e

f

g

h

There are many options here and I have played them all. The one I chose in this game is not critical, but it is certainly playable. It was especially popular in the 1 980s, but with the development of theory, Black managed to solve the main problems. Still, there is some venom in the line, as we shall see in the game. Many strong players, including Kramnik, Mamedyarov, Leko, and lvanchuk have also played like this.

If I had played I l .Wfc2 lLixc3 1 2.Wfxc3, she could have played 1 2 . . . lLi a6 and returned to this plan under good circumstances.

1 1 ...he4 12.J.f4

5 ...J.b7 The critical move.

6.i.g2 c5 7.dxc5 .L:c5 7 . . . bxcS is the alternative. Karjakin likes to play like this. It is neither better nor worse than Judit's move.

8.0-0 0-0 9.lLlc3 J.e7 There is nothing wrong with this move, but it seems to me that Judit later mixed up the two main plans and played an unhea!thy­ looking crossbreed. One plan is to play 9 . . . lLi e4 1 0.lLixe4 i.xe4, while the other is to play 9 . . . i.e7 with the idea . . . lLia6-c5. So when I say that this line is not entirely harmless, we can already see it here. Judit seemed unaware of the finesses, or was maybe seduced by the dangerous substance called 'limited knowledge', where she knew the

a

b

c

d

e

f

g

h

This is the problem with Black's move order. If the bishop was still on cS, Black could play . .. Wfe7 or . . . dS with the idea of playing . . . lLi c6 later on, when the bishop definitely belongs on the active square cS. Instead she has lost a tempo withdrawing the bishop, as the knight no longer has a solid route to cS.

12 ...tfcs This is somewhat awkward. Black is not in time to develop her pieces . 1 2 . . . .if6 This could lead to a curious tactical sequence, although I am sure I would have chosen 1 3.Wi'a3 or 1 3.E:d2.

Chapter 5 - Compensation 1 3.id6!? This was suggested by the engine, and looks mindboggling at first, but the tactical point behind it is really nice. 1 3 . . . ixb2?! 1 3 . . . E:e8;!;

.�

·rP'- - -����.,%���,v,.. . . %�, 6 ..... i:�• . . , �,;, ,� · -� �

8 7

ta' ;?•' � &

•• �

� �

c

d

e

g

f

But she is not in time.

14...i.xd6 1 4 . . . if6 1 5 .°Wa3 does not make things easier for Black.

I think Black's position is very unpleasant here, and Judit made a decent decision by trying to change the nature of the game. It was one of her strong features: she was very inventive, even in inferior positions.

.

b

14.id6!;!;

1 5.gxd6 �c5!?

.

%� � �� : 1�� !�� ,��� -r � �� � � ,.· .,,�� · ·r· . � :� . . .. %=· · · a

1 47

h

1 4.E:ab l !! ixb l 1 5 .E:xb l if6 1 6.ixf8 'Wxf8 1 7.tll e5! ixe5 1 8.ixa8± Black does not have any compensation for the exchange as the a7-pawn is hanging.

13.gacl

The computer might say that White is only slightly better here, but if Black does not manage to get out of her passive position, she would definitely lose, so this computer assessment is extremely misleading. Note that Black is not in time for 1 5 . . . tlia6, as after 1 6.b4 White is simply better.

16.�dl

I could also have played 1 3.id6, which was very strong, but I like to make useful moves first in order not to take premature action.

A natural move; the queen has nothing to do on a4.

13 ... gds

The computer points out that 1 6.E:cd l was also good enough. 1 6 . . . id5 ? does not work here, as 1 7.cxd5 'Wxd6 1 8 .dxe6 and 1 9.tll d4 wins a piece.

Judit is still trying to reroute the knight. 1 3 . . . tli a6 still does not work on account of 1 4.'Wxd7.

a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making i n Chess

1 48

16 ....idS!?

2 l .llJd6t 'it>h8 22.e3± is also pleasant for White. 2 1 . . .'it>h8 22.llJxd? �ad8

Trying to spice things up. The computer wanes to play 16 . . . llJ c6 to give up a pawn. Of course it is difficulc to beat the engine, but White has no weaknesses, so against a human I would be very optimistic after winning the pawn.

17Jixd5 Because Black has changed the course of the game, White now has co make some good decisions in order to exploit the advantage. Unfortunately for Judie, I was in excellent form in this tournament. This exchange sacrifice was however forced upon me.

17 ... exdS

8

61

�rn�

� -illl � l , , , %W •tt:� ,�%.; . -.� �n".,m�, , %� � �i

� � � : �., � .· 3 ��r � 2 !���!�1� ��z..,,z�,?. 'i[ "" ,,

�---�m

,,,,.

m

z

b

a

m

c

d

e

g

f

h

23.llJxf8! �xd5 24.cxd5 llJ d8 This does not hold. White has: 25.llJe6! llJxe6 26.dxe6 Wfxe6 27 . .id5! Wf d7 28.e4 With a technically winning position. All of this is easy to work out with the assistance of the engine, but at the board the whole line did not seem so convincing to me.

18 .. J�fB

a

18.llJgS!?

b

c

d

e

I also considered 1 8 . . . llJ c6 1 9.cxd5 Wfe7, but I realized I could play 20.llJ h3 followed by d5-d6 and .ixa8, with a huge advantage. White might not keep the extra pawn, but he is dominating his opponent. f

g

h

It turns out I had a choice between two good moves, although I was not so sure during the game. l 8.llJe5! Wf e7 l 9.llJxf7 During the game I could not make this work on account of: 1 9 . . . �fg It is only with the help of a computer chat we can see the following: 20.W/xd5 llJ c6 2 I .llJe5t

a

b

c

d

e

f

g

h

Chapter 5 - Compensation

1 49

22.�xd4 l0xd4

19 ..ixd5 This is the most natural move for a grandmaster. It is so tempting to try to calculate it till the very end, and it's possible too. Also, the bishop attacks the weakness on f7, as well as pins the knight down the d5-a8 diagonal. What's not to like? Still, it was not unreasonable to play l 9.cxd5!? with obvious compensation. This is all based on an amazing computer tactic: l 9 . . . Wd6

8 �� u �-·7 -� U i U. i U. i 6 � ·� ..... % u. � � �� � � �: .,-�. ��',,,,,,;� �� 3

�m � ��wJtJ �m·ef · 'm"-' � ��c- - � 1 � �� �� � %

2

a

23J�dl!

a

b

c

d

e

f

g

b

c

d

e

f

g

h

I had to see this in advance, as 23 . .ixa8? would be a blunder due to 23 . . . ll'ixe2t. All of these tactics came easily to me during the game. The only time I spent more than 1 0 minutes o n any move was o n move 1 1 . h

20.ll'ie6!! fxe6 2 1 .dxe6 Wxd l t 22.E:xd l tl'i a6 23.e? E:fe8 24 . .id5t @h8 25 . .ixa8 E:xa8 26.E:xd? E:e8 27.E:xa? tl'i b4 28.a4 @g8 29.E:b? and White has very good winning chances in the ending, with four pawns for the knight. I never looked seriously at this; I was happy with having a strong bishop on d5. It felt right. This is a clear difference between human intelligence and brute-force chess, which you have to be an electron to come up with.

23 ... l0c6 24.t0f6t rlJg7 25.l0xd7 �Uc8 I have to admit that I cannot remember if I had seen the next move at move 1 9, or if l had just relied on something showing up. Both are entirely possible; we are at the natural end of standard human calculation.

19 ... l0c6 20.Wd3 The idea.

20 ... g6 21 .�e4 �d4 This was forced, as far as I could see. 2 1 . . .We? 22.Wc3 is worse, and on 2 1 . . .Wb4 I had planned 22.ll'if6t @g? 23.tl'ixd? E:fc8 24.Wf3 with a triple attack. Black can try 24 . . . Wxb2, but White can play whatever he likes. For example 25 .E:c3, as I had planned.

a

26J�bl!!

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Makin g i n Chess

1 50

After this move, there is little left to say; it is a matter of time before White wins.

38.©f3 �c5 39 ..ic2 gc7 40J�d2 c!ll d7 41 .e4 �£8 42.e5 �d7 43. ©g4 b5 44.h4 gxf4 45.gxf4 a5 46.©hs a4

The idea of the move is that the b2-pawn is supported. White then wins back the exchange, keeps an extra pawn, repels the little counterplay Black has, and then advances. It is all fairly simple and there really is nothing left to analyse after this move. White is much better without it as well, but once you understand the strength of placing the rook on b l , the decision is easy.

26 .. J3c7 Black has to play this, as after 26 . . . f6 27.b4 the game is over immediately.

27.c!ll e5 c!ll xe5 28.ha8 gxc4 29 . .id5 The engine wants to play 29.f4, which is j ust wrong. The right attitude is to keep everything under control, remove all counterplay, and then advance slowly. Which is what I did in the game.

29 ... gc2 30.©fl g5 3 1 ..ib3 gd2 32.©el gd7 33.h3 h6 34.£4 �g6 35.e3 c!ll f8 36J3dl ge7 37.©fl c!ll d7 8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

a

b

c

d

e

f

g

h

47.gg2t @£8 48.©xh6 gc4 49.©g5 �c5 50 ..i.5 gel 5 1 .h5 ©e7 52.h6 ghl 53.h7 c!ll e6t 54.©g4 �£8 55.gc2 �xh7 56.gc7t ©e8 57.gb7 c!ll f8 58.gb8t ©e7 59.gb7t ©e8 60.gxb5 ggl t 6 I . ©f3 gfl t 62. ©e3 1-0 One of my best games in this tournament. The only thing I feel is open to criticism is whether I should have played l 8.c!ll e 5 instead. But sadly you cannot win the same game twice. The following day Judie beat me to take the match into tiebreaks. That game can be found in her book A Game of Queens and was, according to Judie, the reason why she decided to write this wonderful trilogy about her career.

Chapter 6 Time

With Alex Huzman and Maxim Rodshtein Khanty-Mansiysk 2009

Diagram Preview

On this page you will find a few diagrams with critical moments from che coming chapter. If you wane co compare your chinking with the games, you have the possibility. Take as much time as you need or wane. This is not a test, but a chance co 'chink along' with the grandmasters in the games.

8

7

6 5

4 3

2

Time co unleash your tactical brilliance! (see page 1 63)

a

A small intuitive decision co scare with. (see page 1 54)

a

b

c

d

e

f

g

h

A more difficulc question: How co deal with the threat co f2? (see page 1 60)

b

c

d

e

f

g

No time co waste! (see page 1 78)

h

How would you initiate an attack? (see page 1 65)

a

b

c

d

e

f

g

Time for a deep chink. How should White build up the attack? (see page 1 66)

h

a

b

c

d

e

f

g

h

I did not see this idea; can you see the refutation? (see pages 1 8 1 -2)

Chapter 6 It is apparently impossible to win time in a game of chess; no matter what you do, your opponent gets to move as often as you do. But as you get to know chess a bit better, you know that some moves make no progress, and that if you can force your opponent to make this sort of move, you are indeed winning time. This is what this chapter is about: making more useful moves than your opponent in the short term. Momentum has been described many times in chess history, but never as precisely as by the first World Champion, Wilhelm Steinitz: If you have an advantage, you need to attack or you will lose the advantage. He was of course talking about a dynamic advantage, like for example a lead in development {static advantages, like extra material or better structure, are not connected with the same type of urgency) . These days it is rare to win a game against a grandmaster where he takes your pawns while you attack. The following blindfold game was an exception to this. Boris Gelfand -Teimour Radjabov Monte Carlo (blindfold) 2007

-

Time

1 53

Radjabov decides on an ambitious continuation, getting the two-bishop advantage. But on the other hand, White gets a strong centre.

7 ..ig3 lll h 5 8.c3 e6

This was a critical moment.

9.tll g l!? I played this funny-looking move to make him show his intentions. I was not sure if he was considering playing . . . f5 on one of the next few moves.

9 ... tll xg3 Basically, he has to play this.

It was a good year for me in the blindfold part of the Melody Amber tournament, where I think I came second.

9 . . . tll f6 would not make much sense of Black's previous moves.

1 .d4 llif6 2.llif.3 g6 3 ..ig5

l0.hxg3

There are not so many theoretical discussions in this tournament, so I decided to just 'play a game' and get the pieces out in a simple way.

I am planning to play f2-f4 and tll f3, so now it is up to Black to decide which type of pawn structure he wants.

3 ....ig7 4.lli bd2 d6 5.e4

l0 .. e5

5 .e3 would be safer, but this was not the intention.

5 ... h6 6.i.h4 g5

.

Pretty natural. He could also have played 1 0 . . . c5, but I would reply l l .dxc5 dxc5 when the g7 -bishop is not

1 54

Boris Gelfand - Dynamic Decision Making in Chess

very strong, as it is blocked in a typical way by the b2- and c3-pawns. It would result in a complicated game. Maybe I would again play f2-f4 and lll f3. So instead he went for a more standard King's Indian approach.

8

7

6 5

4 3

2

1

1 1 .dxe5! This decision was correct, as it turns out. White gains a few tempos. I was thinking that 1 1 .d5 was a reasonable move as well, but I now see that Black has l 1 . . . g4!, when he is all right. White can try 1 2 .ibSt, but after 1 2 . . . ©f8 1 3 .ie2 h5 the black king is fine. Black can organize his pieces in many ways, all of them perfectly okay for Black. To mention just one, Black can play . . . tll d 7-f6 and . . . ih6, when he does not look worse to me.

l l ....ixe5 If he had played 1 1 . ..dxeS , I would change to a slower game, and play l 2.ie2 with the intention 1 2 . . . 0-0 1 3 .ig4, when White has the better structure; the bishop is bad on g7.

12.lll gf3 J.g7 13.J.b5t!

8 .1�.-; •. 1.t, , , j,· � %� � ·------

�� ·� , . %� ��-d-1- � � 5 � .,t- - f� 4 u �- �------%3 � �� �- t[Jwr 2 8ri %m�wti8%� , %·- - %� � � d ,�

7

6

�

....

....

. .

1

a

b

c

d

e

f

g

h

This intermediate check is really important. White manages to use his slight lead in development to ask Black some awkward questions. It is frequently unpleasant to face such a check. It is fashionable at the moment in some lines of the Queen's Indian and the Gri.infeld.

13 ... c6 This is an awkward move to make, and maybe not the best. 13 ... lll d?!? l 4.lll d4 may look a bit unpleasant for Black, but with a few accurate moves, Black's position is perfectly playable. 1 4 . . . a6! is best here, when 1 5 .lll fS axb5 1 6.lll xg?t ©f8 is fine for Black. But personally I would not gravitate towards this in a blindfold rapid game . To me the most natural move was 1 3 . . . lll c6. You could imagine the game continuing 14.lll c4 a6 1 5 .ixc6t bxc6 1 6.Wfd3 and the game would go on. Black has played rather slowly and has an unkempt structure. But on the other hand, he has two bishops.

14.J.e2 0-0 15.�c4 Suddenly White has some mmanve, and Black has to make a difficult decision.

1 5 �e7? •.•

Chapter 6 This leads to serious problems, but Black is also worse after l 5 . . . d5 1 6.exd5 cxd5 l 7.tt:le3 i.e6 1 8 .tt:ld4, when he has to take on d4 to stay in the game:

-

Time

1 55

develop all of his pieces, and White can create threats against the weak h6-pawn.

18 ...1Wxa2?! This is too risky; now Black gets killed. But it is easy to understand why Radjabov did not fancy playing 1 8 . . . �xd6 1 9.E!xd6 E!e8, when White has a pleasant choice, including 20.tt:lh2!?, and also 20.i.c4 with the point that 20 . . . tt:ld7?! 2 1 .tt:l f5 is a complete collapse. 2 1 . . . tt:l b6 22. tt:lxh6t! i.xh6 23.E!dxh6 tt:lxc4 24.E!h8t is an important little line to see.

a

l 8 . . . tt:l c6?

b

c

d

e

f

g

h

l 9.tt:lxe6 fxe6 20.tt:lg4 gives an

unpleasant position for Black. I am familiar with such bad positions from playing the Najdorf. So after l 8 . . . i.xd4 1 9 .cxd4 White is marginally better. It is plausible that Black can then equalize by playing many accurate moves, but it would never become truly pleasant for him, I think. Still, given the alternative, this was what he had to do.

16.1Wxd6 1Wxe4 17.tl:ie3 1We6 18.0-0-0

White has a significant time advantage. It will be a long time before Black is able to

19.�c4 1Walt 20.@c2 1Wa4t The sixth queen move in a row, after which he is totally lost. Obviously this is not meant as criticism of my opponent. Firstly, it was not such a serious game, being blindfold. Secondly, I would hate it if people were to equate my level of play with my worst performances in rapid, not to mention blitz games! If you wanted, you could quickly make the best players in the world look pretty stupid then.

2 1 .@bl b5 Provocative, but then again, it was a blindfold game. Black's only chance is that a big blunder is committed.

B oris Gelfand - Dynamic Decision Making in Chess

1 56

We have reached a typical moment in an attacking game, where the pieces cannot be improved more, and it is necessary to commit in one way or another. Certainly there is no point in retreating the bishop. It requires care and attention. We should never forget that it is quite easy to go wrong, even when the position is as overwhelming as it is here.

22.:Sxh6! Once again, a move in the style of Mikhail Tai. Black is unable to take all ofWhite's pieces.

8 7 6 5 4 3 2 1

a

b

c

d

25.:Sxe6! fxe6 26.�c7

e

f

g

h

Another solution was 22.tll d4 bxc4 23.tll ef5 , when there i s n o defence against t"Llxh6t and mate to follow.

Black resigned. 27.We7

22 ... bxc4 23.:Sdhl :Se8

I think I played this game pretty well, bearing in mind that it was a blindfold rapid game.

23 . . . Wa5 would require a small tactic.

8 7 6 5 4 3 2

For example:

26 . . . Ei:fS

1-0

The fight for the initiative

The following game was played at a higher level, mainly because it was played with a full time control. It illustrates the theme of time really well. I played one of my little pet ideas, bur my opponent reacted with great intensity, ignoring recapturing a pawn, bur instead pushing forward in the centre. a

b

c

d

e

f

g

h

24.Ei:hSt! �xh8 2 5 .Ei:xhSt �xh8 26.WxfSt �h7 27.Wxf7t �h8 28.WfSt h7 29.Wxc8 and White is totally winning.

24.t"Llxg5 .ie6 The finish is pleasing.

The moment I was able to take the initiative was when he lost the momentum. He did not make the best use of his time . . . Boris Gelfand - Dimitry Bocharov Dagom ys 2007

Bocharov is from Novosibirsk. He is quite a good player, ambitious and plays very creative chess. He never plays for a draw, as far as I can see. He recently did very well at the 20 1 5 World Rapid Championship in Berlin, where

Chapter 6 - Time he beat many good players, Svidler, Anand and yours truly included, and finished 5th. There are many players who are little known outside Russia, but who can play really good chess; Bocharov is one of them.

I .d4 c!£if6 2.c4 e6 3.c!£if3 .ib4t

a

4.cl£ibd2

b

c

d

e

f

g

The main move, but lately 6 . . . d6 has come into fashion after Radjabov played it against me in the 20 1 3 Candidates tournament in London. It is a way to get a complicated game without having to study a lot of theory. Black wants to play . . . e5 or . . . c5 at some point. It seems playable. I won two games against it recently, against Antipov in Moscow 20 1 5 , and Morozevich in Zurich 20 1 6.

h

4.id2 is more popular, but I have always blocked with the knight, with the hope of having a two-bishop advantage. It sometimes works, as you can see in My Most Memorable Games, where I have included games with this variation.

4... 0-0 5.a3 fi.e7 But nowadays it is popular to retreat the bishop in this fashion. Black loses a tempo, but on the other hand White has placed his knight on the less harmonious square of d2, and you can debate how useful a2-a3 is. If Black manages to play . . . d7-d5 on the next move, his position will be good.

6.e4 So for this reason, White probably has to play this move if he wants to fight for an advantage.

6... d5

1 57

7.V!ic2 7.e5 is the main continuation. After 7 . . . lll fd7 there is a lot of theory. For years I have experimented with different continuations, mainly 7.id3 and the game move. I had two interesting games back in 1 988 against two Bulgarian players, which we will look at below. But in this game Bocharov reacted correctly, I think, making it undesirable ever to repeat this variation with White.

7 c5! 8.dxc5 dxe4 9.c!Lixe4 c!Lixe4 IO.V!ixe4 cl£ic6! ..•

Black fights for the initiative. After something like 1 0 . . . ixc5 l I .b4 ie7 1 2.ib2 White has a slight advantage. It would be possible for me to add a long analysis, but it would be pretty pointless. White has space

Boris Gelfand - Dynamic Decision Making i n Chess

1 58

and better development, while Black has no obvious weaknesses. There is no need for variations to determine that White has a slight edge.

1 1 .i.f4 If I had played 1 l .b4, Black would have replied with 1 l . . . f5 1 2.Wc2 e5 when I think Black is already better.

8 7 6 5 4 3 2

1 2.�d l ! An important intermediate move. 1 2 . . . Wa5t 1 3.id2 Wxc5

8 7 6 5 4 3 2 a

1

a

1 1 .. hc5?

c

e

g

.

This move looks very natural and innocent, but it loses the momentum and allows White to fight for the initiative with active play, making Black's task much harder. The strategic view of the position is that White has the better pawn structure, so Black will have to fight for activity. Or in other words, White is playing for long-term advantages {structure) and Black should play for short­ term advantages (time) . The instructive thing is that this can change in an instant, as we shall see later. There were two alternatives I was considering during the game. 1 1 . . . e5?! I did not feel that this was working, but I knew I had to look at it.

b

c

d

e

f

g

h

l 4.ic3 1 4.b4 f5! 1 5 .Wd5 t Wxd5 1 6.cxd5 llJ d4 1 7.llJxd4 exd4 1 8 .ic4 is better for White, but not as convincing. 1 4 . . . ie6!? The following sequence is not necessarily very strong for Black, but if he has to play placidly, White is better. So I checked the most active line. If Black has to play 14 . . . if6, White has an edge; the black pieces are misplaced. 1 5 .b4 f5!? 1 6.We2 Wb6 1 7.c5 We? 1 8.ixe5 White snatches the pawn. 1 8 . . . llJxe5 l 9.Wxe5 Wxe5t 20.llJxe5

Chapter 6 - Time The right move was 1 1 . . . f5! immediately.

�/ � �;. J� .J�� , : �,1.,.� %� %� ·� � 5 �-/,�%·'� �� 3 :/j%�':;��� ���-ltJ':;���-� 2 fH:l�� �J�ii 8

1 59

1 3.id2 Wc7 1 4.b4 e5 again looks as if White is trying to lose. 1 3 . . . Wxd2t 1 4.ixd2 e5!

4

a

b

c

d

e

f

g

h

If White plays 1 2.We2, Black has 1 2 . . . if6 with . . . e6-e5 prepared. For example: 1 3.id6 e5! and it is not hard to imagine that a careless move could cause White to lose a miniature. The engine claims that the position is roughly balanced after this exchange sacrifice, but I seriously doubt that this is a realistic evaluation. There are positions where people have conflicting views and both of them are correct, but in this one, the initiative is very powerful and White's long-term problems with his king are a great obstacle. Very few grandmasters would choose to be White here. But of course there are crazy people in every profession. I am not talking about anyone in particular, and it is certainly not meant to be derogatory. And knowing a bit about psychology, I would guess that a person crazy enough to choose White here, would only be flattered to be told he was crazy! Still, 'original' is a better word . . . S o the main line runs: 1 2.Wc2 Wa5 t This i s simple enough if we are trying to prove that Black is perfectly fine, but I am not convinced that my opponent would not have gone for the more aggressive l 2 . . . if6, seeking the initiative. 1 3.Wd2

1 5 .ic3 1 5 .b4? e4 1 6.lligl a5 and White is in a bad way. The pawn matters little when none of the pieces are able to get out. 1 5 . . . e4 1 6.llid4 ixc5 Black has no problems whatsoever. Because of the above, I have abandoned 7.Wc2.

8

7

6

� m .i.� �i.%•� . . . ,/,_ .,v,., ,,_, . . ?.m:m :m- - ,?,�

� m � � : .11t-.1. �� 3 � � ltj� . . . %�/%� �� � � 2 �t!J - �t!J% � t3J �lflj� %

1

�- - - %- �mi•� a

12.:adl!

b

c

d

e

f

g

h

I d o not know if m y opponent somehow discounted this move as a blunder. It's possible, but I j ust don't know. I did not seriously consider 1 2.b4, when Black could fight for the initiative with either

Boris Gelfand - Dynamic Decision Makin g in Chess

1 60

12 . . . eS or 1 2 . . . fS 1 3.Wc2 e5. Both moves look good. You could work out if the inclusion of ... f7-f5 improves Black's position or not, but from White's perspective, all you have to do is establish that the game is going in the wrong direction. The move I chose in the game required quite a bit of calculation. Black has missed his chance to play energetically ( 1 1 . . .ixcS lost a tempo) so it is White's chance to play energetically now.

12 ...ft'b6 1 2 . . . ixf2t was of course possible, but after 1 3 .©e2 (but not 1 3.c;!;>xf2? ft'xd l 1 4.id3 f5!) 13 . . .Wb6 leads to the game. During the game I also considered 13 . . . fS , but found that White had the advantage after:

a

13.b4!

b

c

d

e

f

g

h

The key move. 1 3 .�d2? e5 is obvious wrong. During the game, I looked only briefly at 1 3 .id3. The line I saw was 1 3 . . . ixf2t 1 4.c;!;>fl f5 1 5 .We2 e5! 1 6.ixeS tll x e5 1 7.tll xeS We3 and Black is not worse. Also Black can play 1 3 . . . fS 1 4.We2 e5!, when 1 5 .tll xeS tll d4 1 6.Wd2 tll b 3 1 7.We2 tll d4 is a typical engine draw.

a

b

c

d

e

f

g

h

l 4.�xd8 fxe4 l 5 .�xf8t c;!;>xf8 l 6.id6t The accurate moment to give the check, which I did not see during the game. Also, it is not something I would want to see. All I need to know to make a decision is to see that White is the stronger side in a variation. I can work out the details later, if it actually happens. I need only enough information to make a decision. Acquiring more is potentially a waste of time. Obviously it happens all the time, but the least we can do is not to celebrate it. 1 6 . . . c;!;>f7 White has a choice between 1 7.tll g St and 1 7.c;!;>xf2, both leading to a long-term advantage.

At the board, when I looked at 1 3.id3 and 1 3 .b4, I felt strongly that 1 3 .id3 would lose momentum. Of course I checked a few lines to see what was happening, but it did not attract me in the slightest.

13 ...Lflt Black has to take.

14.c;!;>e2 The white king looks stupid on e2, but this is a small price to pay for all of Black's pieces being uncoordinated. For a start, c4-c5 is a threat. Black is absolutely forced to do something right away.

Chapter 6 - Time

15.�c2

8 7

8

6

7

5

6

4

5

3

4

2 1

14 . £5

161

3 a

b

c

d

e

f

g

h

..

1 4 . . . e5 is met with 1 5 .c5.

2 1

a

c

b

1 5 ... e5?!

d

e

f

h

g

Black is not putting up the toughest resistance. When the game is not going your way, you have to fight hard to change the flow of the game, or your opponent will face too easy a task. 1 5 . . . id4 j ust loses a piece to 1 6.b5. 1 5 ... tli d4t! This was the best chance. Black had to find the following trick: 1 6.tlixd4 ixd4 1 7.c5 Wfb5t 1 8 .'it>el White looks to be winning, but Black has: 18 ... Wfe8! 1 9.�xd4 e5 Winning back the piece. A very nice tactic that allows Black to stay in the game. So 1 5 ... Wfc? is a must. The simplest is probably 1 6. �xf2 exf4 1 7.ic4 with �he l , �d6, tlig5 and so on, coming. White's position has great potential; it's not too dissimilar from what happened in the game. I also saw that I could try 1 6.tlixe5, which is very complicated, but might also lead to an excellent position. Again, during the game, I did not feel the need to work out the details. It was enough for me to see that the general trend was in my favour.

a

b

c

d

e

f

g

h

1 62

Boris Gelfand - Dynamic Decision Makin g in Chess

But although this line was Black's only try, White is still better here. There are two ways for White to prove an advantage, which he can choose from accordin g to his taste:

22.W/d4 ig4t 23.©f2 ixd l 24.ic4! Followed by 2 5 .E:xd l with an overwhelming position.

17.©xfl exf4 18 .lc4t ©hs •

a) 20.ic4t ©h8 2 1 .E:d6 exf4t 22.W/e2 id? 23.Wfxe8 E:axe8t 24.©f2 ic6 2 5 .E:hd l with a very pleasant endgame. b) Or 20.ib5 Wfxb5 2 1 .Wfb3t ©h8 22.i.xe5 id? 23.©f2 E:ae8 24.E:e l with a sharper position, where White is also better. I have a feeling that at the time I would have chosen the ending - it is more stylish and human, but now I am less sure. Opposite­ coloured bishops favour the stronger player, something I have only discovered recently. So a pleasant choice, but not easy.

16.c5 Y!fc7 A really important line to work out is: 16 ... ixc5 1 7.bxc5 W/c7

8 7 6 5 4 3 2

White is still a pawn down, but I can assure you that I was happy at this point. But being happy is not enough. In a situation like this, where the advantage is purely dynamic, it is easy to let it slip with one or two unenergetic moves. So, it is really important to keep focus, and continue to find good moves. There are many tempting possibilities here, which also makes it easy to go wrong.

19.Y!fc3!?

a

b

c

d

e

f

g

h

Of course I considered 19 .!!he I !, but I felt that after 1 9 . . .id? we would simply transpose after 20.W/c3. I did not see the following nice combination: 20.llig5 h6

If White were to play something like 1 8.ic l , Black would reply 1 8 . . . e4 and 1 9 . . .f4, with compensation for the piece. But White has a direct refutation of this sacrifice: 1 8 .W/b3t ©h8 I 9.llixe5! lli xe5 20.W/e3 Based on 20 . . . E:e8 2 l .E:d5! and Black is caught in a web of pins. He can try active play with: 20 . . . W/f7 2 1 ..ixe5 f4 But White wins easily: a

b

c

d

e

f

g

h

Chapter 6 - Time 2 1 .gd6!! hxg5 (2 1 . . .gf6 22.'Li f7t @h7 23.gxf6 gxf6 24.Wfc3 also wins) 22.Wf c3 White is winning because of gh6# or Wfh3#. During the game I looked for such possibilities all the time, but I could not make it work.

19 ...i.d7 The reason I chose the move order I did was che following neat trap: 1 9 . . . gds 20.'Lie5! gxd l

8 7 6 5 4 3 2

1 63

with other tempting options available, I played something strong and simple. Noe because I wanted to keep it simple necessarily, but because I saw that it worked. In the line above, 22 . . . 'Lie5 is the best defensive cry, after which I doubt chat White has a bigger advantage than in the game, so my intuition did not lee me down, even if my calculation did.

2 1 ...gxel Again, one of the reasons I played as I did, was chat I saw a crap. After 2 1 . . .i.c8 22.'Lig5 gxe l White wins in one move with 23.gh6!. Black is mated.

22.°!Wxel gf6

a

b

c

d

e

f

g

h

2 1 .'Lig6t hxg6 22.Wfh3# Sadly I was not given the chance.

20J�hel gae8 21.gd6 Again I cried to gee 2 1 .'Lig5!? co work. During the game, I could not find a way against 2 1 . . .gxe l 22.Wfxe l h6, but again:

a

b

c

d

e

f

g

h

23.gd6!! hxg5 24.Wf c3 works. If this had been the only way co win the game, I believe I would have found it. Bue

Maybe the way co keep the game going was 22 . . . ges 23.Wf d2 i.c8 , but after 24.Wfxf4 White should win. With equal pawns and a much beccer placement of the pieces, White's advantage is overwhelming.

23.�g5 h6?! A final inaccuracy. Black could have resisted more, but probably not saved the game. 23 . . . gxd6 was the lase cry. I was planning to play 24.'Lif7t @g8 25.'Lixd6t @f8

a

b

c

d

e

f

g

h

26.Wf d2, when Black's position is simply bad.

24.�f7t � h7 25.gxf6 gxf6 26.°!Wdl !

Boris Gelfand - Dynamic Decision Makin g in Chess

1 64

To sum up this game: Black lost momentum on move 1 1 and White took the initiative when he had the chance. This made Black's task difficult, and after a few mistakes he got run over by the unstoppable energy of the white pieces.

8 7 6 5 4

I promised to look more deeply into the 7.Wi'c2 idea and the two games I had with Bulgarians. By what could be pure coincidence, the games followed the same path till move 1 8 .

3 2 1

a

b

c

d

e

f

g

h

Boris Gelfand -Alexander Delchev

The final touch, not only attacking down the d-line, but also eyein g the h5-square.

26...�cS 26 . . . i.e8 is eliminated most simply by 27.Wi'h5 i.xf7 28.Wi'xflt Wi'xf7 29.i.xf7 and White will quickly collect a few pawns and win the endgame. What I calculated durin g the game was 26 . . . llJe5, but after 27.llJxe5 fxe5 28 .Wi'd5 Black has to play 28 . . . Wi'd8 to avoid mace, when 29.Wi'xb7 is simply winning.

Arnhem

1 9 88

I .d4 �f6 2.c4 e6 3.�f3 �b4t 4.�bd2 0-0 5.a3 i.e7 6.e4 d5 7.�c2 dxe4 8.�xe4 �bd7 9.i.d3 �xe4 I O ..ixe4 �f6 l l .�d3 b6 12.i.e3 .lb7 13.0-0-0 �c8 To play for . . . c5 is natural, but not very energetic. Today, it seems to me chat Black has an attractive option in 1 3 . . . b5!?.

27.�d6 @g6 28.b5 8 7 6 5

a

4 2

1-0

c

d

e

f

g

h

The main idea is chat after l 4.cxb5 c6! Black gets a lot of compensation with the open files.

3 1

b

a

b

c

d

e

g

h

Also l 4.c5 i.xf3 l 5 . gxf3 c6 looks reasonable for Black. It is harder for White to create an attack without knights. I think this is the way chat Petrosian would play.

Chapter 6 - Time

14.l3hel c5 Another game saw Black playing even more passively; and being punished for it. 14 ... Ele8 ?! 1 5 .tll e 5 c5 1 6.g4! Taking on c5 first is also strong, but the general idea is to attack before it is too late. 1 6 . . . cxd4 1 7.ixd4 h6 1 8.h4 i':!:d8

1 65

8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

a

c

b

d

e

f

g

h

The task for White now is to properly organize his pieces, which is a lot harder than it looks. It is only now, 28 years later, that I know how to do this. h

1 9.g5! I 9.ic3!? was less direct, but in some ways simpler; I do not see any black counterplay. Still, I like Gurevich's energetic play. One of the more difficult aspects of handling the initiative is: when should you play as energetically as possible and when would a simple retreat make all of your threats stronger? But in this case, both moves win. 19 ... i':!:xd4 20.gxf6 i':!:xd3 2 1 .fxe7 Elxd l t 22.i':!:xd l 'We8 23 .i':!:d7 Gurevich had to see this move on move 1 9, but he did, so all is OK. 23 . . .ic6 24.i':!:c7 'Wb8 25.i':!:xc6 'Wxe5 26.°Wd3 1 -0 M. Gurevich - J. Horvath, Budapest 1 987.

1 5.�e5 cxd4 16 ..ixd4 gds 17.J.c3 hg2 A very risky decision. It was played not once but twice against me that year in junior championships by Bulgarian opponents. The coincidence seems overwhelming, although it is not clear from the main game that they shared information about the opening.

The first time I had this position was in the 1 988 World Junior Championship in Australia. At that point I played a natural-looking move, which did not get the most out of the position: 1 8 .i':!:gl 'Wb7 l 9.f3 Necessary, otherwise Black plays . . . ie4. l 9 . . . ixf3 20.i':!:dfl

a

b

c

d

e

f

g

h

20 . . . i':!:xd3! This is obviously the right move. Black removes a key attacker and gains time to block the open files toward his king. 20 . . . ih5 2 1 .ixh?t! tll xh7 22.tll x f7! would win for White. Strictly as an item of curiosity, I want to

Boris Gelfand - Dynamic Decision Makin g i n Chess

1 66

mention that 20 . . .ie4 was also possible, when 2 l .E:xf6 E:xd3 22.E:xf7 at first looks absolutely decisive:

When we were analysing the game afterwards, a trainer from the Qatar delegation suggested a more dynamic regroupin g of the white rooks.

18J�e3! Instead of being left with a static rook on the cl-file that was unable to do any good, White is aiming to double the rooks on a file where real damage may be caused.

18 Y!fb7 ...

a

b

c

d

e

f

g

h

But Black has a cool defence: 22 . . . g6 23.E:xe7 Vflxe7 24.tlixd3 E:d8! Edge-of-your­ seat survival tactics. 2 5 .Vfie2 ixd3 26.Vfle5 h6 It seems miraculous that White does not have a direct win. Still, in practical terms, Black's king is too unsafe for his position to be viable, although the engine thinks White is only marginally better. 2 l .Vflxd3 ie4 22.Vfih3 22.Vflg3 ig6 and Black is j ust better. 22 . . . if5 23.E:xf5 exf5 24.Vflxf5 g6 25.tlixg6 hxg6 26.ixf6

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

1/2-'h Gelfand - Dimitrov, Adelaide 1 988.

I was lucky that my opponent accepted the draw. After 26 . . . ixf6 27.Vflxf6 Vfle4 28.Vfic3 E:c8 29.E:d l Vflf4t 30.@b l Vflxh2 Black is definitely playing for a win.

a

19J�grn

b

c

d

e

f

g

h

This was the idea back then, and it looks perfectly natural. White gets strong pressure down the g-file, and Black will have to reply accurately in order to resist it. But if we think a bit deeper, prophylactically, we can see what Black is trying to do as well - we need to prevent his best defensive option, which is taking on d3 and then rerouting the bishop to g6 via e4. Today, with the help of engines, we can see that White has a subtle idea in 1 9.!!de 1 !! with the direct threat of llixf7, although f4-f5 also becomes relevant in some lines. White has a serious attack. The best line for Black might very well be: 1 9 . . . iffi 20 .E: g3 E:xd3 Again this is forced. 2 1 .llixd3 tlih5 22.E: g 5 if3 23.tlie5 ie7 24.E: gg l tli f6 2 5 .llixf3 Vflxf3 Black does not have quite enough for the exchange, although the game is far from over.

.

1 67

Chapter 6 - Time

19 ... �JS? 1 9 . . . g6? would be met by the devastating 20.i.xg6! and 2 1 .lll xg6, when the floodgates open.

Black would be fine if his knight on f6 was not overloaded.

2 1 .lll g4! White wins a piece.

Black should have replicated Dimitrov's idea and played 1 9 . . . E:xd3!.

21 ...ig6 22.ixg6 hxg6 23 ..hfU gdc8 Black could not recapture. After 23 . . . gxf6 24.lll x f6t White has a winning attack. For example with 2 5 .Wc3 and all kinds of threats are appearing.

24.@bl b5

a

b

c

d

e

" 1a �. r••'� :� ·�!. 3 � � � �� �� ����,,,,%��-2 " ��� ��,;;-;. -�� - - - Y-���-'0, - - %���-'0�v 1

g

f

s � m � � �-� % �-Im m_,�, 6

h

I was intending to play 20.lll xd3 i.e4 2 1 Ylie2 i.f5 22.lll f4!?, which the engine thinks makes no sense at all. Black's counterplay down the c-file is too dangerous. Probably 20.Wxd3 with roughly even chances is more prudent, but Black appears not to be worse at all. This can be hard to see at first glance, as Black only has a pawn for the exchange after all, which is maybe why Delchev tried to get by without giving up the exchange.

20.gg3 ie4

"� ;;-;

a

25.lll h 6t!

%

b

'.0

�

c

d

e

%

f

'.0

g

h

This also worked on the previous move, but when you are completely winning, the main thing is not to throw the point away needlessly, so I was cautious.

25 ... gxh6 26.gxg6t fxg6 27.�xg6t ig7 28.ixg7 �f7 29.�xh6 �f5t 30.@al @f7 31 .ic3 gra 32.gg5 �e4 33.gg?t @es 34.�h5t @ds 35.ia5t 1-0

a

b

c

d

e

f

g

h

One part of fighting for the initiative, trying to take over the momentum, is gambits. One player gives up a pawn or more in order to spend his time developing, while the opponent is wasting his time cashing in.

1 68

Boris Gelfand - Dynamic Decision Makin g i n Chess The Shabalov-Shirov Gambit

Alexander Shabalov is slightly older than me, and I have known him since the late 70s. I have a game in my database that we played in 1 98 1 , but I have a feeling that I am missing an even earlier game. We played a lot of interesting games against each other. Two of them were included in My Most Memorable Games (from Vilnius 1 988 and the game with Black from Bermuda 2004 - the same tournament as the game below) . Alex was one of my main rivals in the 1 98 5 Soviet Championship. We played a marathon game that was adjourned three times, where he managed to hold this position:

©f6 55J�f3t ©e7 56.tlixe5 tixa3 57JU7t ©d6 58.tlid3 tlic4t 59.©c3 tie3 60.©d4 tlic2t 6l .©c3 tlie3 62.gxg7 ga3t 63.©d2 tlic4t 64.©c2 tlie3t 65.©d2 tlic4t 66.©e2 ga2t 67.©el tie3 68.g3 tic2t 69.©dl tlid4 70.tlic5 gh2 7l .gd7t ©e5 72.gh7 tlie2 73.gxh5t ©d4 74.tixe6t ©xe4 75.gc5 tlixg3 76.tig5t ©d3 112-112

The game resumed several times, while of course we were also playing the other rounds. le cost him a lot of energy and while I kept on winning, he fared less well and lost some games, and fell out of contention for the title. I think he also had university exams at the same time, which certainly can add some pressure!

Boris Gelfand - Alexander Shabalov

Yurmala 1 98 5

Shabalov is a very creative player, and originally from Riga. It is not difficult to see the influence of Mikhail Tal . From quite an early age, Shabalov and Shirov were fortunate to analyse with Tai. They worked on the Botvinnik variation together, for example. Bagirov helped the young players develop a systematic work ethic, but the stylistic influence came from Tai. Boris Gelfand - Alexander Shabalov Bermuda 2004

l .d4 d5 2.c4 c6 3.tic3 tif6 4.tlif3 e6 5.e3 tlibd7 6.Yfc2 i.d6 7.g4!? You can see that Black i s a solid pawn down at this point. Shabalov, however, then played better than me. Although he later was three pawns down, he created enough counterplay to hold the draw.

45JU3t ©g5 46.©g3 ©h5 47J�d3 ga8 48. ©fl ©g5 49. ©e3 ©f6 50. ©d2 ©e7 51 .h4 h5 52J�g3 ©f6 53.gf3t ©e7 54.gg3

This is the Shabalov-Shirov Gambit. le is not known which of them thought of it first, but it certainly does not matter who managed to put it on the board first, as they analysed it together. The special thing about this game is of course that I had the rare chance of playing a variation against its inventor! This variation was popular in the early 1 990s, but then for some years it was neglected. It came back in the early 2000s, when I also played it against Dominguez (see page 1 09) .

Chapter 6 These days it is less popular, as it is difficult to play such a risky variation against computer preparation. Still, Nakamura won a game earlier this year in Gibraltar against Mariya Muzychuk. But it remains risky, although not refuted. I think it has most value as a surprise weapon these days.

-

Time

why it was possible for a computer to do it. 1 0 . . . llJh5 l I .i.e3 0-0 1 2.0-0-0 was played in Kasparov - Deep Junior, New York ( I ) 2003. The computer has played badly, but the rest of the game was even worse, so bad that it cannot be included in a quality book like this. Let's just say that Garry won very convincingly.

a

10.e5

8 ..ixc4 b6? This is a very rare move. As we shall see, it does not really work, but in 2004 this was not obvious.

9.e4 The most natural move. Black sometimes plays 8 . . llJd5, so 9.g5 would make little sense, forcing Black to play 9 . . . llJd5 under what are preferable circumstances. .

9 ...i.b7 This is the point of 8 . . . b6, and it makes no sense to consider anything else, as no one would play it. 9 . . . e5 would be a considerable concession, as after 1 O.g5 the knight is kicked to a bad square. It makes no sense to play like this, which is

1 69

b

c

d

e

f

g

h

I had looked at this idea at home. The game was played at a point when the level of preparation was already quite high, although not as high as today. So I had looked at the variation till a certain point. But I think it is important to make a distinction when we talk about preparation. Top players do prepare very deeply, but there are so many possibilities that you just cannot prepare for everything. And everyone is using the same computer programs. Everyone knows the main games and has analysed them already. So when 1 5 moves land on the board instantly, the players are following known paths, which they have both analysed and formed their own opinions on. The cases where a player manages to catch another player in preparation to a very deep point are rare. In most games, something slightly different than expected happens, and the players are following basic ideas just as much as long lines of memorized preparation.

1 70

Boris Gelfand - Dynamic Decision Makin g i n Chess

In the present case the line was pretty forcing, so I knew it a few moves ahead. After this game some commentators claimed that the game had been analysed by me all the way to the end, because it followed the first line of their 2004 computers all the way to the end. But correlation is not causation. If you look at the moves after 8 . . . b6, you will not find many chances to play anything else before move 1 8, which is the point I had prepared to, and probably also the point Shabalov had prepared to. Such exaggerations about my preparation have happened many times. See for example my game with Amonatov on page 1 43.

10 ...c5 l 1 .exf6 .ixf3 l l . . . Wxf6 l 2.i.e2

does not work on account of i.xf3 l 3.g5 and White wins a piece.

12.fxg7 �g8 13.�xh7 tlif6 14.ib5t ©e7 15.ig5 if4 Without this move Black is lost. After 1 5 . . .cxd4 White wins with either 1 6.lll e4 or l 6.Wh6! i.e5 l 7.lll e4! i.xe4 l 8.f4 and Black's position falls apart on the dark squares.

16 ...ixhl 17.ixf4 �xd4 18.�g3 When I looked at this variation as far as this point in my preparation, I felt I had a good position, so I moved on to other things.

18 ... c!lJe4 A curious line goes 1 8 . . . i'!xg7 1 9.i'!d l i'!xg4 and White has to play very accurately to prove an advantage:

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

20.i'!xd4 i'!xg3 2 l .i.d6t c;t>dg 22.i.xcSt i.d5 23.i.xb6t axb6 24.hxg3 and White has an extra pawn. 20.i.d6t c;t>dg 2 I .Wh3! ? is also possible.

19.�h4t?!

8

This is pretty decent, even strong, but today we can see that I had a stronger option.

7 6

l 9.lll xe4

5 4 3 2 a

16.�h3!

b

c

d

e

f

g

h

As I said, I am still in preparation. After 1 6.i.xf6t 'ktixf6 1 7.lll e4t i.xe4 1 8.Wxe4 ' g7 23.tll g4 and White wins. 22.'Wxe4 I saw this far, and believed that White's attack should be winning. After: 22 . . . g6 23.ig3

h

a 1 ) I should point out that the immediate 1 8 . . . f5? does not work. White plays 1 9.tll xd6 'Wxd6 and either 20.Elc6 or 20.ie?, winning. a2) A more critical line would be: l 8 . . . tll c5?! Against this, it was my intention to sacrifice the exchange: l 9.Elxc5! bxc5

a

b

c

d

e

f

g

h

This is indeed the case. 23 . . . 'kt> g7 is 'best' when 24.tll d? leaves Black a pawn down for no compensation whatsoever. And after: 23 . . . ixg3 24.fxg3! The knight is lodged on f6 and Black has no reasonable defence against 'We3, when . . . 'kt> g7 is met by 'Wc3 with death and destruction. a3) The move I was afraid of was 1 8 . . . 'kt> h8! with the idea ... f5.

20.tll f6t! Black cannot accept the sacrifice. After a

b

c

d

e

f

g

h

1 79

Chapter 6 - Time a3 l ) My idea was to play 1 9 .g4, but Black is in time with 1 9 . . . llJc5!. I was trying to make 20.E:xc5! ? .ixc5 2 1 .g5 work, but I did not trust it. The computer shows a very nice defence for Black: 2 1 . . .¥Md8! 22.¥Mh5 ixd5! 23.gxh6?! The most natural move and it looks winning, but in fact Black is ready for it. (For this reason it is better to play 23.llJxc5 bxc5 24.E:d l , when Black has to give up the queen. 24 . . . .ixb3 25.E:xd8 E:axd8 The position is quite unstable, but neither side seems to be for choice! 26.gxh6 g5!? 27.ig3 f6 The engine says the position is close to equal, but White's position looks ugly to me.)

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

23 . . . g5!! A surprising move, even when you are up close. 24 . .ixg5 E:g8 2 5 .E:d l .id4 The complications continue and do not become any less confusing. But the feeling, and the computer evaluation, is that White's attack is running out of steam and, at best, the attack will be enough to counter Black's material advantage.

a

b

c

d

e

f

g

h

Black looks well placed. But it turns out that I had a very strong move in this position that changes this assessment: 23.f4! I did not consider this move. Analysing with the computer, it is easy to find and easy to understand, but during the game it was not so, I can promise you. I was worried about this position. If I did not find a move like 23.f4, it would be easy to lose my way, and have no advantage whatsoever. 23 . . . ¥Mc7 24.ixfB E:xf8 2 5 . fxe5 llJ xc6 26.¥Mc2!? Threatening !!c l . 26 . . . ¥Md7 27.llJxf7t E:xf7 28.E:xf7 ¥Mxf7 29.¥Mxc6 VMf4 30.e6

a32) And after 1 9.a5 f5 Black is in time. a33) I also considered this tactical solution: 1 9.E:c6!? Here we get this forcing line: l 9 . . . .ixc6 20.dxc6 llJc5 2 1 .llJxd6 llJxb3 22 . .ie7 I was suspicious of my position on account of: 22 . . . llJ d4

b) Instead of 1 7.llJde4, I also considered playing: 1 7 .a5!?

1 80

Boris Gelfand - Dynamic Decision Making in Chess

8

17.ltlcxe4 I spent nine minutes on this move. The thing about 1 6 . . . e4 is chat if it works, Black is absolutely fine, so I had co make sure chat I reacted in the most accurate way. I was sure chat I had co take with a knight on e4, but I wanted co work out if there was a difference between taking with the c3-knight or the d2-knighc.

2 a

b

e

c

g

f

h

b l ) 1 7 . . . b5 would allow White to show his idea in full after 1 8.ltJde4 ltJxe4 1 9.lDxe4 �h8 20.g4! and che advantage is obvious.

I worked out pretty quickly that there was no big difference. Black should cake back, and this was definitely his intention. Bue it was worth the time co make sure anyway. You never know where the subclecies are hidden, which is why we call chem subcleties, and hidden . . .

b2) The most natural reply is: l 7 . . . bxa5 After 1 8.lDde4 White has many ideas. We have seen the g4-idea already, but in this specific variation, we also gee possibilities with t.a4, putting Black under serious pressure. The computer suggests 1 8 . . . t.e? 1 9.�e l t.d8 as one of the choices, which is an indication chat White is doing quite well. It is maybe only 'plus over equals', but it is very pleasant co play.

I d o not wane c o spend much time on Dominguez's decision co play 16 . . . e4. It is difficulc enough for me co work out what happens from my side of che board. I chink he must have missed something in the long variations chat follows. Bue this is chess. If we saw everything, there would be no reason co play.

If he had played 1 6 . . . t.b?, I would have had co make a big choice between l 7.lDde4 and l 7.a5 . It seems likely chat I would have sacrificed the pawn. What is certain is chat I would have thought for a long time, before making chat choice. Even now I am not sure which move offers White the better chances.

By the way, I should say that I never considered playing 1 7 .h3 or something like chat. I started considering the critical move and it worked. Why then would I wane to spend time on a time-wasting move? If you can take, you should cake. It also feels very natural chat it works: White is better developed and the pawn on d5 is limiting my opponent's freedom somewhat.

17 ... �xe4 I had worked out pretty quickly chat Dominguez planned co play as he did in the game, with l 9 . . . W'f4. You do not play a move like 1 6 . . . e4 with the intention of defending a lousy position. The purpose is co counteratcack. I did not spend any real time on it, buc lee's have a look at the only plausible alcernative anyway: 1 7 . . . t.xh2t 1 8 .�h l t.e5 1 9.�e l would give a

b

c

d

e

f

g

h

Chapter 6 White a strong initiative. Something horrible could happen. For example: l 9 . . . ixb2 20.ig3! and Black is losing something.

7

s

6

. . %;..t �, �� �-,Y,-· �%1" ,,,J . . %�.i"' • �"• i f.I' �� fi. I>".

: !.•i!.a'� 3

2

��-�j_���if ��;����)"r�[�8'"��� a

b

�""" �

"'"�

��

c

d

e

f

g

h

We can look at 20 . . . ie5 , when after 2 1 .llixf6t gxf6 22.'Wg4t @h8 23.'Wh5 @g7 24J'k4! Black is quickly mated. For example: 24 . . . llic5 2 5 .!!xe5!

- Time

181

fine. For a start, he threatens to play . . . 'Wh4t followed by . . . lli g4 with mating threats. No thank you!

20 ti'xe4 •••

He has to take. Nothing else works. a) 20 . . . l:!e8 2 l .d6 is depressing for Black. The bishop is stuck on h2 and White dominates the position. The combination 2 l . . . l:!xe7! ? had co be anticipated, of course. It looks highly speculative, but this is almost always the case with fantastic combinations. 22.dxe7 'Wh4

18.�xe4 .ixh2t 19.hl ti°f4

a2) 2 l . . . ib7 22.id5 looks simple; White is all over his opponent. All you have to do is spot the following amazing variation: 22 . . . l:!xe7!

20 .le7! •

I do not know how it came into my head, but I think this is the accurate way to play. Maybe it was simply chat I saw chat after 20.ig3 ixg3 2 I .lli xg3 llie5 Black is more than

a

b

c

d

e

f

g

h

1 82

Boris Gelfand - Dynamic Decision Makin g in Chess

23.g3! Wff5 24.ixb? Wi'h3 2 5 . lLi f6t!! lLi xf6 26.ig2 and the only piece Black is not about to lose is the knight. I saved a bit of time by not considering the combination in this version.

24.Wi'h5! I like this line. Maybe it is not the only way to refute the attack, but it is simple and pleasant to look at. b3) Finally, the simple version: 22 . . . g5 23.Wi'h5!

b) Another line I saw was 20 . . . lLie5 2 1 .ixffi (this is not necessarily forced) 2 1 . . .Wi'h4, which looked dangerous.

2 U�e1 I have to play this move. Athough two of Black's pieces are hanging, I cannot take either of them. At this point 2 1 .ixffi? loses to 2 1 . . .Wi'h4, when after 22.ie? g5 White is lacking knight checks.

2 1 . ..�f4

b l ) 22 . . . Wi'xe? 23.'it>xh2 Wi'h4t 24.'it>gl lLi g4 25 .l:!e l and White is a rook up for nothing. Not a very exciting variation, but you need to be in control of these things; it's the bread and butter of calculation. As we can see elsewhere in this book, mistakes in calculation are often very simple things, not j ust flashy moves. b2) 22 . . . f6 23.ixf6 gxf6 It is possible to take on ffi, but after 22.ixffi lLixffi things are much less clear. I do not want to analyse it here, as it has no relevance to the game. I took my time earlier and saw that I had 22.l:!c4, which is a clear improvement. This was easy to determine by comparison, as winning this tempo had no visible downside.

22 ...�xfl

Chapter 6 I expected this move as well, as there was no plausible alternative.

-

1 83

Time

technical challenge ahead. But the white pieces are so good, and the black pieces so bad, that I thought it was possible to find more.

The only other way to play I considered (and eliminated) was 22 . . . Wfb8, when after 23.d6 White is obviously winning. We can see the line to the end. 23 . . . ie5 24.ixf8 @xf8 2 5 .l:!c7 I wondered if we let the engine run for a while, would it tell us how many moves it is till checkmate? But Black can give up his queen, and run with the king, hiding from the horizon of the engine, but not from inevitability.

I kept thinking and thinking. Tomashevsky told me that he was really surprised, positively so, that I was not satisfied with having an advantage, but was ambitiously looking for the best move. And I found it, but spent too much time. I had to pay dearly for it later (incidentally, the game had no increments before move 6 1 ) .

23.@xh2 .tb7

24 ..ic2! This is the best move, playing for piece activity. Among other things, it threatens to trap the queen.

8 7

At first I looked at 24.l:!e2 Wff5 2 5 .ic2 Wxd5 26.l:!d4 Wh5t 27.@gl tll e5, which was not at all obvious to me. The computer gives: 28.l:!h4 tll f.3t 29.gxf3 Wxf3 30.Wd3

6 5

4 3 2

1

a

b

c

d

e

g

h

I saw this far from move 1 7. The variation was quite forcing up till this point, where my conclusion in advance was that White is doing really well. It is also a good moment to stop and think. White has a wide range of choices, one of them being the back-up move 24.ixf8, winning the exchange. But he is still ahead in development, and it is possible that more can be squeezed from the position. After the game I had an interesting discussion with Alex, and also Evgeny Tomashevsky, who played in the tournament and had followed the game out of the corner of his eye. So, White already has the advantage, though after taking on f8 he is only better, with a big

a

b

c

d

e

f

g

h

White has won a piece for three pawns, with good winning chances. But to go into such complications, I would have to have a greater incentive than getting a technical advantage, which I would be able to get after taking on f8. It turns out that after 24.l:!e2 Wf5 White has an "invisible move", as my countryman Yochanan Afek likes to call them: 25.@g l ! Black i s unable to take o n d5, as there is no check on h5 anymore.

1 84

Boris Gelfand - Dynamic Decision Makin g i n Chess

�6 Z f� ��:Jfl

���� ,� � , � 5 �,f."''-� �� , , , '•�'� 32 '·'-""' � � ��-�'0. ..t• �� ��• �8 �-�0 ' " "�� �� �� �-- 4

a

b

c

d

e

f

g

a

h

After 25 . . . lll c5 26 . .ic2 Wxd5 27.gd4 Wc6 28.b4 lll e6 29.Wd3 g6 30.gd6 White's attack crashes through. Honestly, I don't chink there is any chance I would see a move like this unless I really needed to.

b

c

d

e

f

g

h

29.Wg3 lll d3 and it is time to transpose into an easily winning ending with: 30 . .ixffi gxffi 3 I ..ixd3 There are other lines, but they are not much different. White has a technical win; too many potential passed pawns.

Also 24.gg4!? is winning. 24 . . . gae8 25 .ge2 Wf5 26.Wd4 f6 27.ge3 gives a winning attack. But the following line is not at all obvious: 24 . . . gfe8

26.a5!

25.gfl ! Wxb2 26.gxf7! with a winning attack.

I am very proud of this move. White gives up a pawn, but forces the black queen further away from the battleground.

24... b5

26 ... Wxas 27.Wd3 f'5 2sJlf4 gf7

This is the only move, as 2 5 .ge2 is about to win the queen.

It is obvious at this point that White is winning, but somehow I had spent too much time earlier, and here I started to play really inaccurately and allowed my opponent to come back into the game.

a

b

c

d

e

f

g

h

25Jle2 Wb6 25 . . . Wxe2 is the engine's suggestion, but no human would take it seriously. White has too strong an attack: 26.Wxe2 bxc4 27.We4 g6 28.Wg4 lll e 5

Chapter 6

-

Time

1 85

is winning here too. Probably I was being overambitious, still thinking dynamically. This line would have been a simple win.

3 1 . J.d5 ••

With a bit more time to think, I would have played 29.d6, keeping the queen out of the game, and threatening ib3. White has total domination. 29 . . . llic5 loses in many ways, for example to:

32.b3 I decided to win the queen, but again I had a simpler win. After 32.id3 E:xf5 (32 ... Wc6 33.E:c2 Wb6 34.Wg3 is a disaster as well. Next comes E:c7.) 33 .ixf5! ll'i f8 34.E:c2 Wfl 3 5 .!:!:!1 Wc4 36.E:f4 Wa2 37.ixfB E:xf8 White can win with 38 .d?, 38 .ih?t and almost any other move. All these finesses could be found with five minutes on the clock, but with one minute, it is another story.

32 Yfxc2 33Jtxc2 gxf5 •••

The position is obviously winning; Black's pieces are very poorly placed. Almost any move wins.

29 Yfc7t •••

The only move. It is not a defence, but at least the game goes on, and White gets the chance to make more mistakes.

30.d6 Yfc4 31 .Vfe3 The technical choice was to play 3 1 JM7 Wxd3 32.E:xg?t @xg7 33.ixd3 and White

34.Yfd3 gh5t 35.@gl E:e5 36.E:e2 36.E:c? ie6 37.Wg3 E:f5 38 .Wg6 was a possible win. It would be time to resign. The quality of the game is heavily influenced by the few seconds we both had for the remaining moves.

1 86

Boris Gelfand - Dynamic Decision Makin g in Chess

36 .. Jlh5 37.tfg6 i.fi 38.tfg4 �e5 39.tfe4? At this point 39.l'!xe5 wins immediately. I regret more that I did not play 39.VBg3! tll d 7 40.VBg4 tll e 5 repeating the position. With one hour on the clock, it is easy to see 4 1 .l'!xe5 l'!xe5 42 . .if6, but with ten seconds it is possible to miss it.

39 .. Jks

order to combine this with threats of pushing the pawn. It is the basic principle of two weaknesses, although I did not consciously think of it like that at the time. A good question is if I should have played b3-b4 somewhere, but I wanted to regroup my pieces first. The feeling now is that Black cannot really improve his position much, so I should have played b3-b4. But these things are very difficult to say with any certainty.

42.tfe4 This is a natural move, centralizing the queen.

42 ... ©fi 42 . . .ixb3 loses to 43.g4.

43.i.h4 a

40J::l c2?

b

c

d

e

f

g

h

This is also a bad move, but my flag was almost falling. The last few moves have been made truly randomly. I did see the option of 40.g4, but did not realize that it won immediately: 40 . . . l'!cl t 4 1 . 'kt> g2 l'!d l 42.gxh5 id5 43.Wxd5t l'!xd5 44.l'!xe5 l'!xe5 45 .d7 and the endgame is simple.

40 .. Jhc2 41 .tt'xc2 We have passed the time control and I was faced with a big challenge to readjust. I knew I had played a great game and then spoiled it, so I needed to win it all over again.

41 ...i.e6 It was time to think deeply again. I had to come up with a strategy to win the game. What I came up with was to bring the bishop back to c3, to attack g7 and thus the king, in

a

b

c

d

e

f

g

h

Continuing to regroup the pieces. I think there is a style issue here. Some players would be taking it more calmly, while instead I went for a quick regrouping. Of course, one of the questions is if Black's fortress holds. Does White have all the time in the world or will Black organize counterplay? At the board I was not sure at all whether the position was winning. My feeling was that it

Chapter 6 was 50-50. So I thought I had a good chance of winning, but I was not certain that the position was winning against the best defence. And probably this determined my attitude. If you believe that the position is winning no matter what, you will probably be calmer. But I was not that sure, and thought that if he coordinated his pieces better, it would be more difficult to break through. So I thought I had to set some goals and aim for them quickly.

..

I think this is not the best. During the game I was more concerned about: 44 . . . !!f5 I feel I should warn the readers that the analysis that follows is long and involved; it's a difficult position, and giving j ust a couple of moves would not get close to telling the truth about it.

� �� �n��r� f �!w. .-�

� �t.�,,. : �,. � ��;' �� �� �� ,;, , 3 � � �2 �� �� �1 ! �� �� �; �;.� �

@g8 48.°%Ve8t @h7 49.ig3!? is also interesting) 47 . . . !!e4 (47 . . . if5 48.°%Va8 ie6 49.b4 !!d5 50.Wxa6 tLl f6 5 I .id4 and White has made progress) 48.°%Ve7t @g6 (48 . . . @g8? 49.id4) 49.We8t @h7

8 7 6 5 2

43 �d7 44.Wf3t i>g6?!

�

1 87

Time

4

43.b4 was worth considering.

i

-

3�.r�

a

b

c

d

e

f

g

h

50.id4! !!xd4 5 1 .°%Vxe6 tLl f6 52.°%Vf5t @g8 53.°%Vc8f @h7 (53 . . . @f7 also allows 54.Wxa6 and after 54 . . . !!d l t 5 5 .@h2 l:!d3 Black's mating dreams are foiled by 56.g3) 54.'%Vxa6 !!d5 5 5 .Wc6 The b-pawn is dropping. c) 45 . . . !!c5 46.Wd8! At last separating Black's king from his other pieces. 46 . . . if5 (46 . . . !!e5 transposes to line b) 47.°%Ve7t @g8 48.°%Ve8t @h7 49.Wf7 l:!e5 50.if2 ie6 5 I .We8 with good winning chances.

4

a

b

c

d

e

f

g

h

I intended to play: 45 .°%Va8 The idea is to come to e7 via d8. 45 . . . !!e5! This is the only move. Black has to keep his eye on the e7-square. The alternatives are not so good. a) 45 . . . ixb3?? 46.Wd8 and White wins. b) 45 . . . !!d5 46.Wd8 !!e5 47.if2 (47.We7t

a

b

c

d

e

f

g

h

At this point I looked for a long time, and had a few friends look at the positions as well, but no one has been able to come up with a way to break the fortress. I will give my variations below and offer it as a challenge to

Boris Gelfand - Dynamic Decision Makin g in Chess

1 88

the readers. Please see if you can find a way for White to overcome the resistance. 46.b4 I do not see a way to avoid this move. Actually, it appears that the other lines are less challenging. 46.Wfxa6 .ixb3 47 . .ig3 !!f5 48.Wfc8 .ie6 49.Wfd8 lll f8 50 . .ie l 'it>g8 and I think Black is holding. 46 . .if2 !!d5 47.Wfxa6 lll f6 48 ..id4 id? 49.ixf6 gxf6 50.Wfb6 'it>e6 looks like a dead draw to me too. 46.Wf d8 if5 47 . .if2 !!d5 48 .Wfe?t 'it>g8 49.Wfe8t �h7

47.Wf d8 l:!e5! 47 ... .ifS? fails to a nice little lever: 48.Wfe?t 'it>g8

49.g4!! White wins after 49 . . .ixg4 50.Wfe8t and either 50 . . . lll ffi 5 1 .ie? or 50 . . . 'it>h? 5 1 .Wfe4t. 48 ..if2 !!e4! The accurate move, with the idea . . . lll f6. 48 . . . lll f6? 49.id4 and White wins after Wfe?t, ixf6 and d6-d7.

a

b

c

d

e

f

g

h

Black is ready for . . . lll f6, when he is very well coordinated and has counterplay against the white king with . . . lll g4 and . . . !!d l . My idea was to get the bishop to g3, but it is not happening. If we get to such a position as this, Black might have . . . !!d l t and . . . lll e4t, winning the bishop. It is not even about breaking down a fortress; White has to careful. 46 . . . !!d5 Accurate again. 46 . . . 'it>g6 47.Wfxa6 !!d5 48.Wfc6 'it>h7 49.Wfb? 'it>g6 50 . .ie l and White has made serious progress. Or after 46 . . . .ifS 47.if2 'it>g6 48.Wfxa6 !!d5 49 . .ic5 lll f6 50.Wfxb5 Black's position has collapsed.

8 7 6 5 4 2 a

b

c

d

e

f

g

h

The following line is rather forcing in nature (but not necessarily forced!) . It is an attempt to push White's advantage to the maximum. 49.Wfe?t 'it>g6 50.Wfe8t 'it>h7 5 l .id4 !!xd4 52.Wfxe6 lll f6 53.Wff5 t 'it>g8 54.Wf c8t @f7 5 5 .Wfxa6 l:!xb4 56.Wfa?t 'it>g6 57.Wfa l !!g4 58.Wfb l t @f7 59.Wfd l !!g5 60.d? lll xd7 6 1 .Wfxd?t 'it>g8

1 89

Chapter 6 - Time

8 5 4 3 2 b

a

c

d

e

g

f

h

This position is a fortress with no leaks. White cannot successfully put Black in zugzwang.

b) But it seems that I can get an improved version of the game after: 47.Wfe4t �f7 48.ixg7 tlJ f6 48 . . . l:!xd6 49.Wff4t �e7 50.ixh6 should be a trivial win. The black king is in danger and White has a passed pawn. 49.ixf6 � xf6 The key difference is of course that the pawn is on h6 instead of h 5 . This gives White some possibilities. 50.Wff4t �g6 The first point is that Black cannot interpose the bishop on account of 5 1 .g4. 5 1 .Wff3!

45 ..lel ! I was o f course happy that I gained the necessary time to regroup. White's position has improved a lot.

45 .. J:�d5 46.Ac3 8 7 6 5

4 3 2

1

�� �� �l}j.,. � �%�,'Al% �%� r� A � r�.i. � •rt � l�� -i�� %-"" ��" � � � �� IW,�� .,. ���.'l. � � �� � � ,,

a

46 ... h5?

b

c

d

e

f

. . , 7,�p�

g

h

We will see below the problems with this move. Note that Black is not in time to play: 46 . . . a5? a) I could play 47.Wfg3t l:!g5 48.Wfe3 � f7 49 .ixa5.

a

b

c

d

e

f

g

h

With the king banished to the g-file, White can start making direct threats. The first is Wfxd5, winning in one move. 5 1 . . .i:!d4 There are no other squares. 5 2.Wfe3! l:!d l t 53.�h2 Black has lost his coordination. The black king cannot go to the f-file because of the check on f3. And after 53 . . . ig4 White has 54.Wf e8t �g7 5 5 .Wfxb5, when we can clearly see that it was a mistake to move the a-pawn. was expecting 46 . . . i:!g5 , when my intention was to play 47.Wfa8 �f7 48.Wfd8 and continue to attack the king and support the d-pawn. (48.b4!? on this move or the previous one is a possible improvement. Black could quickly end up in zugzwang.) 48 . . . tlJ fS 49.ib4 id7 I

Boris Gelfand - Dynamic Decision Makin g in Chess

1 90

White has achieved an improved version of the game. Black cannot protect everything/ anything.

47... 'i!i>f7 48 ..ixg7

a

b

c

d

e

f

g

h

Black has been pushed back a lot. It is not clear if the fortress holds, but Black had to try it.

47.Ve4t Black's previous move is bad not only because of what happened in the game, but also because of: 47.b4! Alex pointed out chat this is a real zugzwang.

� �� 1� �� 6 ��

· , � ��'· � 5 , �, �� ..�, �� � � � .... 3w.� �--� � � ���� � ���-� ;-0;.

4 3

2

a

b

c

d

e

f

g

h

White wins after: 47 . . . h4 This is forced, which cannot be a good sign. 47 . . . tt:\ f6 48.ixf6 :!! f5 Otherwise White cakes on d5. 49.Ve4 'itixf6 50.Wd4t White wins the bishop and, in the long run, the game. 47 . . . :!'!g5 48.We3 ig4 49.i.xg? also does not work. 48.We4t @f7 49.i.xg?

48 ... �f6! I had missed this defensive idea. I don't know if he had anticipated this and I had fallen into a clever trap. I had only looked at: 48 . . . 'itixg? 49.Wxe6 :!! d i t 49 . . . tt:\ f6 50.Wxd5 is j ust over. 50. 'itif2! tt:\ f6 5 1 . 'itie2! The rook is running out of squares on the d-file. 5 I . . .:!! d4

Chapter 6 52.g3! A nice little improvement of the position, based on 52.We5?! l:!d5 53.Wxd5? lll xd5 54.d7 lll f4t and Black is in time. 5 2 . . . c;t>g6 53.b4!? The cleanest. 5 3.We5 wins the game with 53 . . . l:!g4 54.d7 lll xd7 5 5 .We8t, but in the main line, White wins a rook, not a knight. 53 . . . l:!xb4 54.d7 l:!d4 5 5 .We8t White wins. What I was hoping for was something like 48 . . . l:!xd6? 49.Wf4t c;t>e7 50.Wg5t cj;>f7 5 1 . .ic3, when it is clear that the position is bad for Black.

-

191

Time

In Grandmaster Preparation - Endgame Play Jacob Aagaard described fortresses in a slightly different way from what people had done before. Instead of talking about something as a fortress if it holds, he described it as a technique. Whether it keeps the enemy out or not, it is still a fortress. His other point was that the way fortresses most often fall is not by brute force, but by zugzwang. The fortress strategy is to hold by doing nothing, but in chess we always have to do something. It is an interesting way to look at it.

55 .. J!d7 56.Yfe5t gd6 57.©fl ©c6 58.g4

49.hf6 ©xf6 The idea of the trap is clear: White loses the d6-pawn.

50.Yff4t J.5 50 . . . c;t>g6 would allow White to keep the d-pawn.

5 I .Yfh4t ©e5 52.Yfxh5 ©xd6 53.b4 .ie6 54.Yfhs ©c7 55.Yfg7t The computer suggests the following line: 5 5 .Wc3t c;t>b7 56.We3 l:!d6 57.g4 8

a

b

c

d

e

f

g

h

Another time scramble is in full swing, and I knew I had to push the pawn.

58 ...J.c4

7

Dominguez anchors the bishop on c4, from where it keeps a watchful eye on the g8-square.

6 5

4 3

2

a

b

c

d

e

f

g

h

But after 57 . . . .ixg4 58.We4t l:!c6 59.Wxg4 l:!c4 Black has a simple fortress. White cannot force his queen to a8, so there is no zugzwang.

I have to admit that during the game I did not pay a lot of attention to Black giving up the bishop here, but it's a serious option: 58 . . . .ixg4 59.We4t c;t>b6 60.Wxg4 l:!c6 60 . . . a5 does not work on account of 6 1 .Wf4! and White wins a pawn. (6 1 .bxa5 t c;t>xa5 is very close to holding, but the tablebases show that White wins in the long run. Black is not able to get the rook to c4 in time to

Boris Gelfand - Dynamic Decision Making in Chess

1 92

create a fortress. The white king makes it to b7 and the b5-pawn eventually falls.) 6 1 .VMd4t @b7 62.YMd5 @b6 63.@e3 �c4

�� �� �� : �-� · � � %�� ��� � � · ·� � : ��11 � . . . �· 0 � � 32 �� -�������� �� �� �� 8

a

b

c

d

e

f

g

66. VM d2! Black is in zugzwang, with 66 . . . @a4 67.YMa2# a likely possibility. 65.VMd4t @b7 66.YMd5t @b6 67.YMd6t @b7 White now has the chance to cross the 4th rank with his king. I cannot see how Black can avoid this. 68.@d4 �c4t 69.@d5 @a7 70.@e6 @b7 7 1 .@d7 @a7

�� =� �� : -,,�.� ii ��� �� 5 r� �- · 7,��� ��� �� . . . %�r �� �� � �-�.�--� �� �� �� 8

4

h

My understanding of this position was that if Black makes it to a7 with the king, we have an unbreakable fortress, as described above. But it is White to move, and Black is unable to gain the necessary tempo. 64.YMd7 The engine wants 64.VMa8 �xb4 65 .Wd8t ©c5 66.YMfBt @c4 67.YMc8t @d5 68.YMxa6, but this is not a human solution. At this point I would have preferred to avoid a queen versus rook and pawn ending. It is a win, but in a game it is easy to do something wrong. More about this later. 64 . . . �c7 There are other moves, but they are not very different. 64 . . .�xb4 is a fun line: 65 .Wd6t @a5

b

a

c

d

e

g

f

h

And here it is important to get the black king away from a7. 72.We5 This is a simple way to do it. White wins after 72 . . . @b7 73.YMc5! as well as: 72 . . . @b6 73.Wb8#!

59.�eSt rtlc7 60.�e7t gd7 8 7 6 5

4 3 2 1

a

b

c

d

e

f

g

h

a

b

c

d

e

f

g

h

We have passed the second batch of time trouble, but here I went wrong again.

Chapter 6

61 .YlYeSf?! After the time control, you go to wash your face and then have 1 0- 1 5 minutes for the rest of the game, plus half a minute a move.

-

Time

70.©f4 was probably the most accurate move, in order to cross the 4th rank immediately. 8

6 1 .11Nc5t Wb7 62.g5 was better. We get the same position as in the game, but he does not have the . . . a5-break.

7

61. .. ©b7 62.gS a5!

4

Of course he grabbed this chance co improve his king and eliminate one of my pawns.

63.bxaS ©a6

5

3 2

��7, �� �lm'%% �������� ������ �j ��� �� �� .! �, �� �� �s �a �� ����,,� � ��

·

a

70 ... ©b6

-

b

c

d

e

--

f

g

h

The reason for the previous move being inaccurate is chat Black had 70 . . . i.a2!? here.

7 6

The key point is that 7 1 .1!Nxa2? does not work: 7 1 . . .!%xa2 72.g8=11N !%a4 If White was able co bring the king co b7, he would win. Bue even getting it to d5 (which is unlikely co happen) is not enough.

5

4 3 2

1

6

1

8

1 93

a

64.YlYc3

b

c

d

e

f

g

h

So my intention was not to go for chat ending, but to try to create some mating threats. 7 1 .%Yd6t Wa7 72.11Nc5t Wa6

As I mentioned, time was lacking at this point, and as I had worked out a possible plan co win the game, I decided co give up the a-pawn in order co push the g-pawn. 64.YNa l ! ? was a potential improvement, co hold on to the a-pawn for as long as possible. But I doubt the ending would be significancly different. After analysing it, it seems chat I gee the game, with a few extra tempos. But when your opponent is relying on a fortress, this matters litcle.

64 .. J�d3 65.YlYb4 gb3 66.YlYel ga3 67.g6 gxas 68.g7 ga2t 69.©f3 ga4 70.YlYd2

73.We3! This is the correct path. The win is not only entertaining, but also very instructive. White has all the time in the world.

1 94

Boris Gelfand - Dynamic Decision Making in Chess

73 . . . .ib3 73 . . . .ic4 74.@d4 is similar to the game. 74.Wi'c6t @a5 75 .Wi'c3t l:!b4 The pin looks bad, but this is not really the problem for Black. The problem is that the white king is coming over to trap the bishop. 76.@d3 .ig8 77.Wi'a3t l:!a4 78.Wi'd6 .ia2 79.@c3 l:!c4t 80.@b2 l:!a4

Black is completely tied down. 76 . . . @b6 77.@e5 @b7 78.@f6 @c6 79.@g6 .id5 80.@h7 Black has to give up the b-pawn, when the win is elementary. 74 . . . l:!b8 leads to the same stuff after 75 .@e5 l:!g8 76.Wd6t @b7 77.Wf8 .

75.lYd6t ©b7 76.©c5 :Sa6 77.lYe7t ©b8 78.lYd7 ©a8 79.lYxb5 8 7 6 5

a

b

c

d

e

f

g

4 h

The fortress falls due to one little detail: 8 1 .Wi'd8t @a6 82.Wi'a8t @b6 83.Wi'xa4!

71 .©e4 :Sa8 72.Wd6t ©b7 73.Wd7t ©b6 74.©d4 :Sa4 This probably does not offer the most resistance. But White is winning no matter what. Possible was 74 . . . l:!g8 with the intention of being able to check the white king away if it comes to c5, but White can play for a different plan: 75 .Wd6t @b7 76.Wi'fB!

7

3 2 1

a

b

c

d

e

f

g

h

This gives queen against rook, which at this stage was a pleasant change. I did not play it very well, but luckily I was not playing against a computer.

79 .bb5 80.g8=lYt ©b7 8 I .%Yf7t ©c8 82.©xb5 gd6 83.©c5 :Sa6 84.lYe7 ©b8 85.©b5 ga7 86.Wf:Bt ..•

The simple win is 86.Wi'd8t @b7 87.Wd4 and we get the game continuation.

86 ©c7 87.9f4t @cS 88.©b6 ge7 89.Wf5t ©d8 90.©c5 ©c7 91 .Wd5 :Sd7 92.lYe5t ©b7 93.©b5 :Sc7 94.lYd5t ©a7 ..•

5

94 . . . @c8 9 5 .@b6 l:!e7 offered tougher resistance. White is winning, but it is not so simple.

4 3

2

a

b

c

d

e

f

g

95.©a5 h

Chapter 6

-

Time

1 95

There are many ways to win a game of chess. If you want to be a strong player, you need to be able to handle whatever type of advantage you get at a high level. One of these skills is to sense when you have to play with great urgency; another is to feel when you have enough time to do everything you want. Studying games like those in this chapter is a start in the direction of developing this feeling.

8 7 6 5

4 3 2 a

95 .. J�b7

b

c

d

e

f

g

h

He should have played 95 . . . l:!h7, which is a theme we know from the tablebases. The rook has to go far away from the king on a different­ coloured square. In the late 1 980s Artur Yusupov showed me the first ever tablebase, which contained queen against rook, and he told me to go away and practise winning this ending. It was not easy at all. A few years ago Alex failed to win it in rapid, and I managed to hold with the rook against Svidler in the knock-out championship in 200 1 . Morozevich failed to win it against Jakovenko. It is by no means easy to win. The fortunate thing is that it is not easy to defend either.

96.'!Wd4t But now it is easy.

96 ... @bs 97.@a6 1-0

Chapter ? Dynamic Masterpieces

More Khanty-Mansiysk 2009

-

The Trophy!

Diagram Preview

On this page you will find a few diagrams with critical moments from the coming chapter. If you want to compare your thinking with the games, you have the possibility. Take as much time as you need or want. This is not a test, but a chance to 'think along' with the grandmasters in the games.

4 3

2

a

b

c

d

e

f

g

h

a

What is the direct way to the full point? (see page 220)

a

b

c

d

e

f

g

h

a

Time for a deep think: what i s White's best move? (see page 207-8) 8

7

6 5

4 3

2

� � � �i Ii,;�-�• � %,, � � � ,,,,,%� �• ��� � .t. .,,,, � '""� � � ��-0 � 8 �-� z� �� � �� .1 ��� � �� --;-�"'-8'" � • � � � -;�� ' � � � 6. ,, , , ,

a

b

c

d

e

f

,r , ,

· g

b

c

d

e

f

g

Has White messed up? (see page 2 1 5)

a

b

Can you find the forced win ? (see page 22 1 )

8

c

d

e

f

g

h

Calculate accurately what happens if White takes the piece. {see page 23 1 -2)

h

c

d

e

f

g

h

g

h

A lovely tactic. (see page 234)

8

7

7

6

6 5

4

----

h

b

2 1

a

b

c

d

e

f

g

h

This is one of my nicest combinations. Can you see all the details? (see page 224)

a

b

c

d

e

f

What is the correct way for Black to strengthen his attack? {see page 237)

Chapter 7

-

Dynamic Masterpieces

In this chapter I want to go over three games I am immensely proud of. Two of chem were played in games where the stakes could hardly have been higher. In the first, the prize was a World Championship match, as it was played in the lase game with a standard time control in the Candidates final. I am very proud chat I was able to show my best chess on this occasion. Boris Gelfand -Alexander Grischuk Kazan (6) 2011

This was the most important game in Kazan, which is often the case for che lase game of a match or, as in this case, a series of consecutive matches. This match was very hard on me. The first five games were drawn, but I was definitely worse in three of them. We will see some of these moments elsewhere in this book. If this game had been drawn as well, the match would have gone co a tiebreak, starting with rapid games and maybe continuing with blitz. Public opinion was that I should do anything to avoid a ciebreak, as Grischuk had eliminated Aronian and Kramnik in tiebreaks. Some even believed that he had consciously aimed for the tiebreaks, though I chink he j ust struggled in the matches against very strong opposition. I also think he was overly intimidated by Kramnik's opening preparation, thinking he had no chance co get a game with White against him. I was fully aware that this was not his usual strategy, and chat against me he would play very ambitiously, as indeed happened in our match. My attitude cowards the ciebreaks was far from desperate. I was White and I wanted co play a normal game; to try to see if I could get an advantage and apply some pressure. So I was not planning co do anything desperate in order to avoid a tiebreak. I am not sure if I was consciously influenced by the way Topalov had lost Game 1 2 of the

1 99

World Championship match with Anand only a year before, but I never miss the chance to learn from other people's mistakes, and I was aware chat what I needed co do was play my own game. My general attitude co chess is chat it is not decided by statistics. The player who shows the best chess on the day will prevail, no matter what the pundits make of previous resulcs. And if you have no belief in your abilities, you really have no place in the Candidates tournament. I already gave my view of Grischuk in Positional Decision Making in Chess (page 1 53) .

l .d4 ltlf6 2.c4 g6 3.lll f3 Ag7 4.g3

i. • -*-� i•�

8

�

'·%�·-,lfi" , , , %�· ----- %.-,��· - - - %� � �� z , , � 5 �W,ia �% �Y/. �-� �8� � �W,ia�- - 3���- % �� 8 - 8�-- �

7

6

-�

4 3

2

I

�lE�lm:f•j a

b

c

d

e

f

g

h

This variation of the Griinfeld was popularized by my friend Boris Avrukh in his two volumes on l .d4 in the Grandmaster Repertoire series, published in 2008 and 20 1 0; it was very trendy at the time. le will be interesting to see if Boris is bringing it back for his second trip down 1 .d4, which started in 20 1 5 and will be four volumes this time, if I understood him correctly. My reason for choosing this variation in this game was that I had found a new idea chat I wanted co cry out.

4 ... d5

Boris Gelfand - Dynamic Decision Makin g in Chess

200

This was a popular variation at the time and it still is, though I also had to be prepared for 4 . . . c6 as well as umpteen versions of the King's Indian. Recently Grischuk has made 4 . . . c5 his main move here.

5.cxd5 llixd5 6.i.g2 llib6 7.llic3 llic6 8.e3 0-0 9.0-0 s

7

6 5

4 3 2

.

� � .tS �� ·

i:�if, , ,, %��,, iA% & y,W� %·'i)- %�f · �� � • %% %%%% %% %%·%%·%%·%%· ��� --,, %����Kr � '�J %'l_j% «LJ %'l_j% ��f: �,,,,,;� mY:, ,,:;��o1 ffef�J::;��0M �0% � ,,,, %Di•!�--,, (� A Y-A% A%

,

a A :W: A

,,,,,

,, ,

� 0 a

9 . . i;es

b

c

d

e

f

g

a

b

c

d

e

When we have a lot of theory, bur as I said, nothing convincing. But there are so many other possibilities here chat it would change the nature of this book entirely if l were to go into chem all. For now I will just mention char the recent trend is going towards: 1 0. lt:l h4!?

h

9 . . . e5 is the most natural move, bur after 1 0.d5, as Boris explains in his book, Black has some difficulties. I do nor want co go into detail here; it can all be found in Boris's book, and not much has changed since it was published. Without goin g into excessive derail, 9 . . . E!:e8 is intended to prepare . . . e5 under better circumstances.

10.i;el A simple and useful move, which was very popular around the time when this game was played. The key problem is chat both sides find it hard to prepare for . . . e5 in a useful way. So this type of micro-improvement feels natural. 1 O.d5 is possible, but not too convincin g. Black plays: 1 O . . . lt:l a5 l 1 .l2l d4 i.d7

a

b

c

d

e

f

g

h

This makes it possible to play f2-f4, stopping . . . e5 altogether (as actually happened in our main game) . Another idea is chat the d5-pawn is protected a bit more than usual, meaning that Black cannot gain tempos by attacking it. The knight might look stupid on h4, bur it cannot be exploited in any way. This subtle idea was invented by Aronian for his game against Ragger, and was subsequently copied by just about everyone, basically immediately. I mean, Sargissian played it in the same match between Armenia and Austria on Board 3! Aronian and Sargissian work closely

Chapter 7

-

20 1

Dynamic Masterpieces

together, so it is equally possible the Aronian was copying Sargissian and not the other way around.

10 ... a5 This was also popular at the time. Black is making another useful move, avoiding 1 O e5 l l .d5, where he is worse. Again, this is covered in Boris's book. Little has changed since it was published. . • .

But the most fashionable at the time of our game was: l l . . . ie6 This had received a bright start at the Khancy­ Mansiysk Olympiad half a year earlier. The first person to be hit by this brutal idea was Scottish Grandmaster Colin McNab, who after proofreading Avrukh's book decided for once not to play some ancient line of the English . . . 1 2.l:!d l ic4 1 3.Wfc2 lll b4 1 4.Wfb l e5!!

1 1 .Yfe2 Again there are not so many useful moves in this position, though this is not the only one of course. I am still following Avrukh's book, forcing my opponent to fight against this thorough preparation. Incidentally, I chink I blundered a pawn with something like l l .d5 ixc3 1 2.bxc3 Wfxd5 against Grischuk in a blitz game at the Tai Memorial . White has compensation for the pawn and the game ended in a draw, but it was certainly not an attempt to rewrite the opening books! 8 7 6 5 4 3 2 1

a

1 1 ...lg4

c

d

e

f

g

.

Black has a few moves here. l l . . .e5 is the most popular, and is still played a lot.

8

7

6 5

4 3

2

a

b

c

d

e

f

g

h

This is the big surprise. (The most obvious tactical point of Black's play is l 5.dxe5 ? Wfxd l t 1 6.lll xd l id3 and Black wins.) I t is not often chat it is possible to come up with such brilliant moves in the opening, so lee's have a quick look at how the first two games played in this variation went: a) 1 5 .b3 This is not very energetic, but probably quite sensible nonetheless. 1 5 . . . exd4 l 6.exd4?! Probably the way to keep the balance for White after 14 ... e5 is in this line, when after 1 6.bxc4 dxc3 1 7.l:!xd8 l:!axd8 1 8 .Wfb3 c2 1 9.lll d4 l:!xd4 20.exd4 l:!e l t 2 1 .ifl ixd4 22.a3 .ixa l 23.ih6 ig7 24.axb4 ixh6 25 .Wfxc2 axb4 26.c5 lll d5 27.c6 active play keeps White alive. 1 6 . . . .ie6 1 7.if4 if5 1 8 .Wfb2 lll d3 1 9.Wfd2 lll xf4 20.Wfxf4 a4

202

Boris Gelfand - Dynamic Decision Makin g in Chess 1 8 . . .axb4 1 9.tll d2 cxd2 20.ixd2 l:!a8 2 1 .l:!xa8 l:!xa8

, ,. .I ��- �. � ; 7 �� · · �� � 6 .. . . . • .,'. ,. . . .� �� �� �� �� : �- ,� � � �:. . .. ?.. � � ef;· . .�Jw·0 �m.·.. .ef'. · . . . �w�.i.� ·0 � · 2 �w �·� . ... � 8

.

a

b

c

d

e

f

g

?.

3

h

Black is simply better. Now White collapsed. 2 l .bxa4?! lll xa4 22.tll b 5 ? c6! 23.lll d6 tll c3 24.@h l tll d 5 2 5 .lll xf7 @xf7 26.tll g 5t @g8 0- 1 McNab - Djukic, Khanty-Mansiysk (ol} 20 1 0. b) A few days later the young Italian/American grandmaster had a chance to show that he too had analysed this variation carefully: 1 5 .a3 exd4 1 6.axb4 dxc3 1 7 .i:!xd8 l:!axd8

.

.....z

b

a

z

.

c

d

e

f

·

·...

g

h

Now Leitao cracked under the pressure. 22.ixb7? 22.h4 l:!a2 has been played in a few games. Black is better. 22 . . . l:!a l t 23.ic l ? b3 24.Wd l ? ixb2 0- 1 Leitao - Caruana, Khanty-Mansiysk (ol} 20 1 0. But at the end of 20 1 0/start of 20 1 1 , the computers started to shout that l 1 . . . ig4 was a more accurate move order. At first I believed the computer was simply insane, but then I realized there are lines where Black gets in . . . tt:l b4 (after lll d2) and . . . Wc8 with tempo, making it possible to play . . . c5.

12.h3 This is forced in order to prevent . . . e5. a

b

c

d

e

f

g

h

1 8 .Wc2 1 8 .bxc3 l:!d l t 1 9.ifl l:!xfl t 20.@g2 axb4 2 1 .Wxb4 l:!d l 22.tll d 2 id5 t 23.e4 ic6 24.f3 tt:la4 25.l:!xa4 l:!xc l 26.l:!a2 l:!xc3 has been played in a number of correspondence games. Black is better, but not surprisingly White was able to hold most of the time. It is strange there are so many correspondence games, as you would expect people to be aware of the dangers here by simply consulting their databases during the game.

12 ....le6 8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

Chapter 7

-

Dynamic Masterpieces

So I started to analyse this variation, thinking it was illogical, as White will often play h2-h3 unprovoked in this variation. Basically, Black is giving White a tempo on account of one tactical idea. I felt that it was too much, but it was not so easy to prove it.

203

8 7 6 5 3 2

4

13.b3 I developed this idea together with Maxim Rodshtein, with whom I was working closely at the time. The idea is to dominate the light­ squared bishop in both directions. If you want to know more about my general approach to such positions, you could check out two chapters in Positional Decision Making in Chess, the Squeeze and Space Advantage. Actually, this game was for a brief moment to be found in the file for the latter of these two, but only long enough for me to come to my senses and defer it to this volume. The variation mentioned above is this one: 1 3.lll d2 lll b4 1 4.l:!d l °1Wc8!

a

b

c

d

e

f

g

h

It is important to note that I would not have played in the same way if Black had played l l . . .ie6 and not given me that extra tempo. The difference is most apparent here, where Black could greatly inconvenience White with . . . if5 ! and after e3-e4?! he would have . . . ig4 with a better game. So against l l . . .ie6, I would have played. . . something else! It would of course be nice to have a unifying idea; to be able to play in the same way against all move orders. But fortunately chess is not such a simple game. There will be a price to pay for such a blunt approach. I want to point out that at this point White has succeeded in his first ambition in the opening: to prevent Black from playing . . . c5 or . . . e5.

15 ..Wfc8 16.©h2 .

a

b

c

d

e

f

g

h

Black is fine. White can do nothing against . . . c5. My opponent was clearly 'out of book' here. We cannot analyse everything and when the computer says that Black is doing well, it can easily be neglected.

13 ... a4 14.i�bl axb3 1 5.axb3

The computer might suggest that there are other ways to deal with the threat to the h3pawn, but for a human, this is the only move that makes sense. We looked a bit at this position when we found this 1 3.b3 idea, simply to check that White's game is indeed easier. We did not anticipate Grischuk's next move, so this was the end of my concrete preparation, and from here on I had to find all my moves on my own.

204

Boris Gelfand - Dynamic Decision Making in Chess

Grischuk told me the following story at the closing ceremony: Peter Svidler was Grischuk's second during the Candidates tournament and following the game, with the engine running, while it was played. After the game he told Alexander that he had had a great position and it was a pity he had not made more of it. Grischuk rejected this claim, stating that I understood the position on a deeper level and outplayed him. Svidler referred to the computer assessments, so was forced by Grischuk to take on the black pieces. Despite repeated attempts, he did not manage to get a satisfactory position even once. I do not want to give the wrong impression. I am sure that when Peter writes his best games collection, I will be there as an ignorant lamb led to the slaughter on at least one occasion! The fact that he got it wrong shows that even such a great player can be misled by looking at the computer assessment.

8 7 6 5 4 3 2

1

a

16 J�a5!?

b

c

d

e

f

g

h

..

Grischuk thought for a long time here. The previous moves were all natural, but here he had to decide what to do. He could have played quieter moves, such as 1 6 . . . �d8 or 1 6 . . . ltJa5, when l 7.ltJd2 f5!? would have led to a more positional type of

game, although I am not sure it would solve all of his problems. I am not claiming it is anything special , but I like my position. But instead Grischuk went for the most critical move. And I have to say that I found 1 6 . . . �a5 posed real challenges for me. He wants to play . . . �h5, obviously. This is not a surprising idea at all. We see it often in the g3-system of the King's Indian Defence, though there the rook comes via e8. It was introduced by Yurtaev, a very inventive player, and I tried it myself a few times in the late 1 980s. This is a really interesting posmon. White has a clear choice. . .. �h5 is sort of a threat. During the game I was not sure if I had analysed it previously or not, nor did it affect my thinking. At the board I always think more deeply and am more concentrated, so I will at times change my mind and not play what I have analysed at home, but instead go with my feeling on the day. Another example was move 1 2 of Gelfand - Anand, Moscow (7) 20 1 2, which can be found on page 1 92 of Positional Decision Making in Chess. After the game I told Grischuk that I thought I had the idea as played over the next few moves in my files, but when I returned to the hotel and checked, there was nothing there. Later Rodshtein confirmed that we had never looked at this idea at all. Maybe this is something for a psychologist to explain? l 7.g4 is possible. Even though I did not see anything wrong with it, I did not seriously consider it. It is a classical way of chinking, chat I do not want to weaken my position unless it is absolutely necessary. Whether there is a concrete reason is less important; long­ term factors have a tendency to only matter long term . . . So the choice was either to allow Black to play . . . �h5, as I did in the game, or to prevent it with l 7.ltJ b5.

Chapter 7

-

Dynamic Masterpieces

17.gdl I seriously considered playing 1 7.tll bS.

205

The move in the game is ambitious and critical, and I still think I made the right choice. As I mentioned on the previous page, the engines of the day did not like White's position much. A friend of ours called my wife during the game and told her that I was crazy to allow Black to go for the attack, and that I was going to lose. Live commentator Alexander Khalifman was also not very optimistic about my chances.

a

b

c

d

e

f

g

h

The first idea is that 1 7 . . . �.fS l 8.�d2! is a nice counterattack and White keeps a slight edge. What I disliked was: l 7 . . . tll b4 1 8 .�d2 c6

But these days the computer already favours White's position when we get to move 1 9. My point is that if Black does not show anything now, then White gets in e3-e4 and d4-d5 and his position will be overwhelming.

17 ... ghs The natural follow-up. It is very hard not to play this, as it was the intention of the previous move. It makes it harder to think about intermediate moves. Most people will continue in the same gear and the same direction unless something makes them change it.

a

b

c

d

e

f

g

h

looked at 1 9.tll a7 Elxa7 20.�xb4 tll d5 2 l .�d2. It felt like the wrong direction. In this position White has a slight space advantage and when you have a space advantage, if you can, you should avoid exchanges (moreover, Black gets the option of . . . tll c7-b5 with a stable position for the knight) . With fewer pieces on the board, the defensive task becomes easier. See Chapter 3 of Positional Decision Making in Chess for more on this theme. Later I of course realized that l 9.tll c3 is possible, but I don't think it is anything special.

During the game I seriously considered: l 7 . . . Eld8 I still think it makes a lot of sense. Black improves his worst piece; the rook has no purpose on e8 anymore, as it is clear that . . . e5 will not happen. At the same time, White is prevented from going e3-e4, which is his main objective. So now White does not have so many active moves. Black says: you improve your rook, I improve mine. I wanted to play: 1 8.b4 But it seemed to me that the inclusion of the two rook moves is in Black's favour. The c4-square is still weakened of course, and all of Black's pieces are playing. But White is better all the same.

206

Boris Gelfand - Dynamic Decision Making in Chess 2 5 .Wxd8t iig7 26.i.d2 Wxd2 27.Wxd2 i.xd2 28.i.xb?t favours White of course, and I would seriously consider it. 25 . . . lll e6 26.i.g4 Black is in trouble. One look at the rook on h6 will make any grandmaster despise his position. White is considering e3-e4 and Black will have to take urgent action in order not to have to give up the exchange.

l 8 . . . l"i:h5 l 9.lll h4 i.f6

a

b

c

d

e

f

g

h

20.i.f3! Here this is necessary. 20.f4 does not work in these lines, as White no longer has d4-d5 . 20 . . . l"i:h6?! This appears to be the critical line. The game is very forcing (obviously I had no feeling for this during the game; we are on the very outskirts of what I saw then) . Black's best chance is to give up the exchange for a pawn and some compensation: 20 ... l"i:xh4! 2 l .gxh4 i.xh3t It is a game, but White's chances are slightly better. 2 1 .d5 i.xc3 22.dxe6 Wxe6 23.l"i:xd8t lll xd8 24.Wd3 Wa2

a

1 8.lii h4!

b

c

d

e

f

g

h

I spent a lot of time on 1 7.l"i:d l - 47 minutes - during which I also decided how to meet Black's last move. So, when it came up, I played it quickly. White does not want to weaken his position with l 8.h4 of course.

1 8 ...i.f6 19.f4! This is a rather elaborate positional concept, but the strategic concept is clear. White keeps control over the centre and gets control over the dark squares.

19 .. J!! d S! 25.c;t>g2! This is a very difficult move to find. Even with the assistance of an engine, it took me some time to realize that this was the best way forward.

What I found most difficult during the game was to deal with the concrete aspects of my plan. To be specific, I was spending a lot of time calculating:

Chapter 7

-

Dynamic Masterpieces

l 9 . . . ixh4? ! 20.gxh4 l"!xh4 20 . . . tll d 5 2 1 .tll x d5 l"!xd5 22.ixd5 ixd5 23.e4 'i!Ne6 24.l"!e l lll xd4 2 5 .'ifNf2 c5 26.f5 ! and White wins the bishop, leaving him a rook up. There is still a lot of resistance left in Black's position, but with good technique, White should be winning. 2 1 .Wg3 I was happy to find this move, and became optimistic.

21 . . . l"!xh3t! I considered this to be bad during the game, but when we were analysing the game for this book, a big blow hit me right between the eyes. The thing I was really happy to find during the game was 2 1 . . .l"!h5 22.d5 ixh3 23 .if3! and it appears that White is winning. I now notice that White is also much better if he gives up the queen, but if3 is the strongest and is what I had planned. 22.ixh3 ixh3 23.d5 8

7

6 5

4 3

2

�,-�i�a �'1:J�,,� �� �� , ,,� �.=,.:��� � �'�' I&

a

�.i.

- -Z� zg,· · uwu� l�l� � b

c

d

e

f

g

h

207

I believed this would be close to crushing. 23 . . . e5!! This is the big surprise provided by the engine. When I realized it was possible, I simply went: "Woooow!" I have to admit that I missed this completely. In hindsight it is easy to explain this with strategic concepts. White has relied on his space advantage, and Black on dynamic play. If Black cannot solve his problems with dynamics, but scans to get pushed around, he will almost certainly be worse. I like to find such new options when I analyse my games. I like the idea that chess still holds a lot of mystery for me, even after forty years! It is easy to make superficial assessments. It is one of the few things in life that can really rub me the wrong way. I do not know why, but it can really get to me. Another aspect of this move that pleases my aesthetic sensibilities is that White has been trying to prevent Black from playing ... e5 since move 9. And suddenly, when I was certain chat my position was finally really better, Black revives this old forgotten idea and creates diabolical counterplay, seemingly out of nowhere. 24.dxc6 24.f5 would avoid exposing the queen, but the tactics work out well for Black. 24 . . . ixf5 25.dxc6 ixb l 26.cxb7 Wfxb7 27.tll xb l I think the position is maybe easier to play with White, but I am not that sure about it either! 24 . . . exf4t 8

7

6 5

4 3

2

a

b

c

d

e

f

g

h

208

Boris Gelfand - Dynamic Decision Makin g in Chess

25.'it>xf4 2 5 .@h2 .ig4 26.cxb7 Wf5 and Black is OK. For example: 27.Wfd3 Wh5t 28.@gl i.8 29.e4 Wfg4t 30.@fl f5! and it is White who has to play accurately to hold the draw:

��-% �i �a . .. : � �� ,� 7 �� � �� �� �� : �� !,,� � � !�l t . . � �� '/. :� �� �� r�;� r�=� 8

6

',

b

a

c

d

e

f

g

h

3 1 .Wfb5! �b8 32.lll d5 Wfh3t 33.@f2 .ixd l 34.i.xf4 Black has to give perpetual check with 34 . . . WBt 3 5 .@gl Wg4t 36.@f2 WBt and so on. 25 . . . Wfe6 26.@g3 26.cxb7 Wi'e5 t is also OK for Black. 26 . . .if5 27.cxb7 We5t 28.@f2 Wxc3 29.i.b2 Wfc6

��-%��i �a . . . :: �� �,� 7 �r � �� � �� : �� ���W�� % �� �� � �i �-��. �� 1� 1---. %� 8

6

',

a

b

c

d

e

f

g

h

Black has full compensation for the exchange. The game would be played for three results. It was only when I returned to the position once again that I decided that 23 .e4!? is probably stronger.

a

b

c

d

e

f

g

h

The following line is natural: 23 . . . h5 24.�d2 i.g4 2 5 .Wf e3 e6 26.i.b2 I guess that White's position is preferable due to the potentially strong bishop on the long diagonal. However White will need to open up the game with d4-d5. It can be difficult to do this under favourable circumstances, as it would most likely also weaken the safety of his own king, and provide Black with plenty of chances for active counterplay. Chess is an inexhaustible game as far as I am concerned, at least in practical terms. This is illustrated by a discovery by Jacob Aagaard, who went back over these lines and came up with the cool 2 1 .Wffl !!.

i �i �a � 67 �-· . � �� ·� �'l . :. -��% �z� ·« @w·« �z--- - z �"ti tj J [j, � J : �! ' , . . : z �w� . . . � .� � 2 r�;/� r� ,,� 8

� , . t� • � ·

5

..

a

b

c

d

e

f

g

h

The thinking behind the move is rather simple: the h3-pawn is defended and compared to 2 l .Wf f3?! there are no . . . i.g4 shenanigans. At first it might seem that White has sacrificed a pawn for not very much, but if you look further

Chapter 7

-

Dynamic Masterpieces

ahead, you will see that the centre pawns will start to roll forward, and Black has no good pieces at all. His counterplay has entirely dissipated and White has a powerful initiative. It is always hard to estimate the size of White's advantage in such a position, but both from a practical and an objective perspective, it can be said to be large.

209

20 . . . .id7 2 1 . b4

a

b

c

d

e

f

g

h

White has a clear initiative. As I do not believe Black can live without his dark­ squared bishop, I think he should play either 2 1 . . .e6, where White is a bit better, or: 2 l . . . Ei:xh4!? 22.gxh4 .ixf5 White needs to continue with great vigour: 23.b5! a

20.'?;lff2

b

c

d

e

f

g

h

The idea of this move is to keep the tension and to support the d4-pawn. Black has to keep pressure on it in order to prevent e3-e4. The computer suggests: 20.f5 ! I did n o t consider this a t all during the game, as it was far from the strategy I was pursuing. Maybe if you play with no memory of what happened on the previous move, you could consider this, but it did not seem relevant to me. The idea is simple. If Black takes with the bishop on f5 , his counterplay on the kingside is immediately in the past, and White can start to put pressure on him in the centre and on the queenside. Besides, the rook on f5 is going to be trapped, and White will have an extra exchange: 20 . . . .ixf5 2 1 .ti:'ixf5 Ei:xf5 22.b4! For this reason, Black should retreat the bishop.

8

-'i¥�� �f!;:.f';{

· �� J�iT� : �f: 5 � � • m.t.. • • � ,;, ,.]� 3 �.,,,m � . ;; ;, � � � 2 0

(,� ;;;

4

a

�

� �if�..t�

,�,- �.�- -b

c

d

e

f

g

h

23 . . . ti:'ixd4! ? This seems the best chance. 23 . . . .ixb 1 ?! 24.bxc6 .if5 25.cxb7 'W'b8 26.h l I don't like Black's position. His pieces have no purpose. 23 . . . ti:'ia5 24.e4 also gives White a strong initiative. 24.exd4 .ixb 1 2 5 .ti:'ixb 1 .ixh4 26.ti:'ic3 White has the advantage. I can easily see him starting a mating attack in the near future.

210

Boris Gelfand - Dynamic Decisio n Making i n Chess 22 . .ia3! is very unpleasant for Black. I was expecting Grischuk to play: 20 . . . lll d5!? 2 1 .lll xd5 l:!hxd5! I intuitively rejected 21 . . . l:!dxd5?! nonsense.

as

total

8

...

a

20 .txh4?

b

c

d

e

f

g

h

A bad decision. Black will not only feel pressured on the dark squares; he has also reduced his pressure on d4 considerably. White will be able to support the d4-pawn sufficiently to push e3-e4, leaving Black in serious trouble. Actually there is not a single variation in this game where this capture on h4 works out well for Black. I think Grischuk was under the belief that he would be able to sacrifice the exchange and create a blockade on the light squares. He completely underestimated the J.b2 idea. I considered two alternatives during the game: 20 . . . lll b4?! turns out to be weaker than I believed. 2 1 . lll e2! keeps the pressure up, as 2 1 . . .c5?!

�·r,!� �� �p �· -- · 76 ��---1 �. -,� l� � ���•!B�. M... .•��I 4 �• re3 � ���-- - �ft.. ..""�ft'"" l � 2 -- � z.�� �m�·�., lI WYD.)- � w lI s

3

...

b

a

d

c

e

- -�.

f

g

h

White is indeed much better if he plays 22.J.b2! l:!b5 23.J.f3, which has some similarities to the game, though without Black having given away his strong bishop. But during the game I had intended to play 22.J.f3?!, when Black holds the balance with this unimaginable line: 22 . . . .ixh3! 23.e4 ixd4! 24.We2 l:!xh4 25.gxh4 l:!c5 26.b4 l:!c3 27.b5

·, �� �· i � 0 , .... . . 76 ��.-.. . ,,..,� 5 .� -�.!,� �� 4 � .. . .. � . ... . � �� l. � � 1 � 2 "a . %� � �lI mB·0 �lI � �-� W7'b''0 8

%

%

%

%

3

a

b

c

d

e

f

g

h

27 . . . lll e5!! 28.fxe5 .ixe5t 29.c;i?g l l:!xf3! 30.Wxf3 J.g4 3 1 .Wfl J.xd l 32.Wxd l Wh3 And the computer shows that Black has enough counterplay to hold the game. This would of course have been hard to find, even

Chapter 7

-

Dynamic Masterpieces

for a world-class player like Grischuk. And White runs no risk of losing at all, so the practical challenges are still overpowering. At this point I of course did not intend to win the exchange. 22 . .tb2! gbS 23.�e2 ghS 23 . . . gas 24 . .ic3 ga6 2S.lll f3 gb6 26 . .!ll d 2 also leaves White with better chances. Black has no pawn breaks and White keeps his space advantage.

21 1

White has a serious initiative. Black will be fighting for his life. 27 . . . .ixh3 28 . .if3 l:!h6 29.dS .!Li a? 30.l:!a l ± Black must say goodbye to his knight, although I do not want to pretend that there is no resistance left in his position. The computer suggests 20 . . . gaS!?, but no one would play it. You j ust moved the rook to hS and after a few moves, you go back?

0 �; �• x•�• � �··· ""'"�-� �� 6 ���a""'"��,-� 54 ��• �w.-ift!i- -0- � ��-�- -0� �"m��- n �� �� �---:r�-� 8 7

�

3

2

a

,� ,� ���-- b

c

d

e

f

g

24 . .tc3!;!; This elegant move underlines that the rook on hS is not strong but weak, by taking away its retreat route. 24 . .if3 gas 2S . .ic3 ga6 would leave the knight stranded on h4. 24 . . . Lh4?! This is still not recommendable, but I j ust want to show why not: 2S.gxh4 gxh4 26.'it>g3! Not the first time we have seen this idea. 26 . . . ghs 27.e4

a

b

c

d

e

f

g

a

h

h

b

c

d

e

f

g

h

I did not consider it at all and I doubt Grischuk did either. 2 1 .lll f3 .!lids 22 . .td2 leaves White with an edge. It is far from overpowering, but it is an easier ride.

21 .gxh4 � d5 22 . .!ll xd5 22 . .id2! ? is also strong, but again a slightly different concept.

22 .. ghxd5 .

Boris Gelfand - Dynamic Decision Makin g i n Chess

212

23.J.b2! I was happy that I found this move. White manages to push the centre forward and keep control at the same time.

in serious trouble, though I was not totally confident about how bad it was. It is not inconceivable chat Black would be able to consolidate and get a defensible position.

I did not consider for even a second to win the exchange. After 23 . .ixd 5 ? .ixd5, if I was given a choice, I would choose Black here, though the position is probably just equal. The computer also suggests 23.e4?! :Sxd4 24.:Sxd4 :Sxd4 25.f5 .id?

After the game I looked at: 24 . . . @f7 2 5 . .if3 Wf d7 26.Wfh4 :Sg8 27.:Sg l @e8

a

b

c

d

e

f

g

h

26.h5 f6 and claims chat White is better. Bue strategically it is unsound. Things could easily go wrong in this double-edged position. We can leave this type of chess to the computer tournaments and people sitting in front of the screen following the games. (I wonder how riveting the post-game interviews chat have become so popular in tournaments over the last decade are in chose events?!)

23 .. J�b5? This move totally surprised me. I was ready to give up everything in order to play e3-e4, so I did not chink he would waste time attacking the insignificant b-pawn. I was expecting only 23 . . . f5 , when after 24.h5 I was very happy with my position. Black is

a

b

c

d

e

f

g

h

And became happy when I saw 28 .:Sa 1 ! with a deadly attack. In general it seems that Black is not in time to consolidate. Obviously there is nothing forced in this variation, and Black can cry to defend in many ways. 23 . . . :S5d6?! is a computer move chat makes little sense to me. 24.h5 .ixb3 25 .:Sg l is very dangerous for Black. White keeps the stability in the centre, while rerouting his pieces to the kingside.

Chapter 7

-

Dynamic Masterpieces

213

This is a good moment to stop for a selfie. White has achieved all of his strategic goals. The black pieces are completely uncoordinated and White has full control over the centre. The position is simply won, but the stakes on this game were high, and it influenced the players; the rest of the game was not played with peak accuracy.

8 7 6 5 4 3

27 ... b6 28 ..ieS cS 29.dxc6

2 a

24.�e2!

b

c

d

e

f

g

h

Grischuk had missed this unpleasant move. Even after 24.e4 ixb3 25 .l'!dc l his position would also be desperate. But the inclusion of Vfie2 and . . . l'!h5 seems like it is in White's favour as well.

At this point I also considered 29.Vfib5 and it appears that the position is simple winning. It is always possible to add variations, but this would give the wrong impression. Black has no tenable moves whatsoever. But the game continuation is also strong, so nothing is spoiled.

29 ... f6 30 ..ial gcS 3 1 .gxcS bxcS

It seems he was already panicking a little bit, being very short of time as usual.

24... ghs 25.e4 hb3 25 . . . l'!xh4 26.iii g3! and White wins a piece. This king move is a recurring theme in this game, so there is no chance I would ever have missed it.

32 ...�c7? Grischuk is collapsing under the combined pressure of a bad position and no time on the clock.

a

b

c

d

e

During the game I had anticipated: 32 . . . ia2 33.l'!b2 Vfic7 34.e5 ie6 3 5 .Vfib6 f

g

h

Boris Gelfand - Dynamic Decision Making i n Chess

214

a

b

c

d

e

f

g

h

Here I had only considered 3 5 . . . :B:c8 36.Wfxc7 Wfxc7 37.:B:b7! and White wins. But instead he can take on b6 and offer more resistance: 35 . . . Wfxb6 36.:B:xb6 :B:c8 White still has to work in order to win.

a

b

c

d

e

f

g

h

39 . .if3!! This is an amazing resource. I seriously doubt that I would have found it. 39 . . . tD b3 40.i.h5t ®f8 4 1 .:B:xe6 lD xa l 42.exf6 :B:xc7 43.f7 tD c2 44.:B:h6 tD d4 45.:B:xh7 tD e6 46.:B:h8t ®g7 47.:B:g8t ®f6 48.ffi=Wft tDxffi 49.:B:xffit ®g7 50.:B:xf5

8

a

b

c

d

e

f

g

h

An extremely beautiful idea is: 37.f5 ! Although I would most likely have taken twice on f6, which should also be good enough. 37 . . . gxf5 37 . . . .ixf5 38 . .ic3 is the point. 38.c7 This leads to a long, relatively forced, variation: 38 . . . @f7

� � � %"'" �-0 �-0 -·· -� • -····!< • �--- � 76 ��� • � � � � � ·� �-.., � � -···-� •�� 5 ��� . : •� � � � � � � 4 � � ��· -· � �� -� �� -� �� �� �� --· · a

b

c

d

e

f

g

h

Without rooks this would of course be a draw, but with the rooks on, White still has serious winning chances.

33J�xb3 �xc6 During the game I hoped that I would be allowed to play the following trick: 33 . . . Wfxf4t 34.:B:g3 lD xc6 3 5 .Wfxc6 :B:d3 I missed that the check on e6 would allow the queen to go to g4. But I saw another nice trick instead: 36.Wfe8t ®g7 37.Wfxe7t ®h6

Chapter 7

�

-

Dynamic Masterpieces The Imperfect Masterpiece

�� �- �� , �� �� �-...��f% �� -0 ���� �.. : ��-0 illfj ..... !:• �� 4 �� •� 8 �m� � � 3 r �� �----��8 ···; ·Y.� ·0 � � �� 2 /�� �� �� �i w

� �. .

b

a

c

d

e

g

f

Boris Gelfand

-

Evgeny Alekseev

Dagomys 2008

This game was played in the Russian league, which I wrote about in Positional Decision Making in Chess on page 27.

h

38 . .ie5 !! and White wins. Amazingly, this is the only winning move in this position!

34.e5 �d4 8 7 6 5 4 3 2 1

215

My opponent in this game is one of many strong Russian grandmasters. At the time of this game, he was performing really well, winning the Russian Championship in 2006, the Aeroflot Open in 2007 and taking 2nd place in Dortmund, ahead of Anand and Leko. He was over 2700 for a few years, but more recently he has mainly been just below this mark. His best result in recent years was sharing first place in the 20 1 3 European Championship, though he lost out on the tide on tiebreak. His main strengths seem to be in defensive play, with counterattacking intentions and good technique.

l .d4 d5 2.c4 c6 3.�f3 �f6 4.�c3 e6 a

35.�c4t

b

c

d

e

f

g

h

Usually Alekseev would choose the Queen's Indian Defence, but for this game he chose to surprise me with the Semi-Slav, which I have played with Black for most of my career.

Game, set and match.

1-0 The following game is probably more interesting for grandmasters than it is for amateurs, but I hope that everyone will learn something from it.

5 ..lg5 h6 8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making i n Chess

216

6.J.h4 In a game half a year earlier, Alekseev played 6.ixf6 against me.

6 ... dxc4 7.e4 g5 8 ..ig3 b5 This is the big tabiya for the Anti-Moscow Gambit. It was one of the most important tabiyas between 2005 and 20 1 0. It occurred repeatedly in my rapid match with Leko, but after this, for some reason, it went out of fashion. 8 7 6

.i �mi: ?� A itii�tda.W� �'=i i �m� -=-w � � , , , , . , , , ,, _ ,,,_ .•Y. �a,, , ... , ,%� �%' "//, · · �� �� ·� ef�6

��

��/,'" ,

,,, , , , , ;

•

""

·

��.. , ,% �' :3 ��,!ti!··· � 2 1

/ ; � �d� 8w, , %�""' . %� ·=�-� • •

a

c

b

9.tll e 5!?

d

e

f

g

h

Looking through my files, I found chat the first time I played the Anti-Moscow was in a private training game, presumably a rapid game, in my house against a friend, International Master Eduard Raisky. 9.ie2 ib7 1 0.Wc2 ll'i bd7 l I .h4 g4 1 2.ll'ie5 h5 1 3 .f3 Wb6 1 4.0-0-0 ih6t 1 5 .@b l gxf3 1 6.ixf3 E:g8 l 7.if2 Wa5

a

b

c

d

e

f

g

h

l 8.ll'ixf7! @xf7 l 9.e5 b4 20.exf6 bxc3 2 l .Wh7t @xf6 22.d5 ig7 23.id4t e5 24.ixc3 1-0 Gelfand - Raisky, Minsk {training) 1 987. Improvements exist for both sides, obviously, but I do not want to spoil anything with computer analysis.

In 2007, during the World Championship tournament in Mexico City, 9.ie2 was still the main line, bur cowards the end of the tournament I had the idea that 9.ll'ie5 was an interesting attempt to save a tempo in an important line {which is relevant for this game) . I asked my seconds at that event, Pavel Eljanov and Alex Huzman, to analyse it and we continued to do so after the tournament. This game was one of the first chances I got co play it. I played two rapid games with it, one with either colour. But soon the theory developed immensely.

9 ... .ib7 10.h4 g4 1 1 .c!ll xg4 This is what White wanted to achieve without having played ie2. The advantage of this will be clear on move 1 3.

Chapter 7 - Dynamic Masterpieces main theory with 1 2.ie2, but this move is not in the spirit of the position. The critical line is 1 1 . . .lll xg4 1 2.1Mfxg4 1Mfxd4, which has been played in hundreds of games.

12.tlixf6t c!lixf6 It was my feeling during the game chat Alekseev was not so well prepared for this game. 1 2 . . . 1Mfxf6 is more natural. The knight is usually very well placed on d7, while not very effective on f6, where it does not fight for the vital e5- and c5-squares. I still managed to gee an advantage in my game against Najer, who opted for the queen recapture.

217

14 ..ie2? This is a mistake. I really did not like how my opponent had handled the opening, and somehow I relaxed and was a bit careless. (Actually, the first pare of this sentence is a lie. I was very happy with how he played - I have the advantage!) 1 4.B:d 1 was correct, as prophylaxis against a rather simple idea. It is natural to put the bishop on e2, but there is no reason to allow Black to get counterplay in the centre.

14 ... a6? This clumsy move is not in the spirit of the position. Black is dreaming of playing . .. c6-c5 under luxurious circumstances chat do not occur in the real world.

13.1Mff3! I like this move, developing the pieces naturally.

13 ... ggs 1 3 . . . WI xd4? 1 4.B:d 1 obviously does not work. Black also cannot cake the pawn after 1 3 . . . b4 1 4 .lll a4 1Mfxd4, as after 1 5 .i.e2± the rook is coming to d l , creating big problems for Black. One of the main points is chat 1 5 . . . i.e? does not work on account of 1 6.e5 lll d7 1 7.B:d l , trapping the queen.

8

7

6 5

8

4

7

3

2

6 5

a

4

b

c

d

e

f

g

h

Nevertheless, here are a few variations co demonstrate the above assessment:

3 2 1

Black should have reacted energetically with: 14 . . . b4! 1 5 .lll a4 c5!, when the position is a mess. For Black, this is a good thing - creating a mess is one of the main goals of his opening variation! In practical terms, it is not necessary to analyse this position deeply; suffice co say chat Black should play like this, and White should not have allowed it.

a

b

c

d

e

After 1 6.i.xc4 ixe4 ( 1 6 . . . lll xe4? 1 7.d5!) l 7.i.b5t cJ:;e? Black has a good position. f

g

h

218

Boris Gelfand - Dynamic Decision Making i n Chess

The king might look awkwardly placed on e7 to the uninitiated, but for those with experience in this opening, it is not a matter for special concern. 1 6.d5 exd5 1 7.0-0-0 lt:lxe4 1 8 . .ixc4 Wd7 is unclear but good for Black. White should try 1 9.1'%xd5! .ixd5 20.1'%d l Wg4 2 1 ..ixd5 Wxf3 22 . .ic6t We7 23.gxf3 lt:lxg3 24 . .ixa8, where he would have to fight for a draw. Thus best is: 1 6.lt:\xc5 .ixc5 1 7.dxc5 lt:lxe4 1 8.Wf4 Wa5 1 9.0-0 Wxc5 20.wh2

more flexible to play 1 5.0-0. White has to make both moves on the next two moves, so in reality there is no big difference. One of the overriding principles of being practical (and this book discusses practical decision-making, not theoretical fantasies) is that during the game you should not get bogged down in small finesses, which you can only determine after hours of analysis. This analysis might be very useful for understanding chess better for the future, but not for evaluating how you should have approached the position at the board. I played the easy move, and did not think about it again until we started working on this book. It turns out that the two potential problems are not problems at all. White is much better after both l 5 . . . b4 1 6.e5! bxc3 l 7.exf6 cxb2 1 8.1'%ab l and 1 5 . . . lt:\ g4 1 6.e5! (Black's idea was 1 6.!%ad l Wf6±) 1 6 . . . Wxd4 1 7.lt:\e4 with a dangerous initiative. For example: l 7 . . . c5 l 8 . lt:\ f6t lt:lxf6 l 9.Wxb7 Wd5

8

7

6 5

4 3

2

b

a

c

d

e

g

f

h

With chances for both sides.

s i, � �•m .1. �� ,, , , /,� ,,,,,�� �� ,, � ····· � � 5 �� �?ii % �� �"'"' � ��-�J� �,,lJ�-ii � 2 � 8 H% _ _ _ _ _ ;_ lDii�� � � ,, � m,,,, %� :a:

: TiT� T,�. 4 3 1

, ,

a

ISJ�dl

b

c

d

e

f

g

h

After realizing that I should have played this on the previous move, it was hard not to play it here. Having said that, it was, in principle,

a

b

c

d

e

f

g

h

20.Wc7! 1'%xg3 2 1 .fxg3 Wd7 22.Wb6 lt:\d5 23.Wa5 and Black does not have enough compensation. For a start .ih5 is a big threat. But this is of course a rather long variation to calculate when it is not necessary. In reality, I j ust felt I should make both moves, and chose one without thinking too much about which was the more accurate.

Chapter 7 - Dynamic Masterpieces

1 5 ... tlid? 1 5 . . . tli g4 has the point chat the queen cannot go to e3 on . . . \Wf6. But after 1 6.e5! the knight is j ust poorly placed on g4.

219

A major alternative i s 1 8 . . . c5 l 9.d5, when Black must decide between the following two options:

16.0-0 I also considered the prophylactic move 1 6.\Wf4, with the idea of not allowing the queen to come out to play, underlining how bad Black's position is. It is j ust another way of playing the position. All in all, I was not too scared of him castling queenside.

16 ... �f6 17.�e3 0-0-0 During the game I also thought chat 1 7 . . . c5 was possible, but after 1 8.d5 Black's position looks horrible. For example, 1 8 . . . e5 1 9.a4 or 1 8 . . . 0-0-0 1 9.e5 \Wf5 20.i.f3 and Black's position is simply bad.

a

b

c

d

e

f

g

h

a) 1 9 . . . b4?! During the game I was sort of expecting him to play this. Black is trying to prove chat his position is OK, but it is too ambitious. 20.dxe6 fXe6 2 1 .tll a4 c3 22.a3

18.b3 A very natural move. The king is on the queenside, so we should open lines in that direction.

� -�- � i. � �� �� ,,,,,/,� ,, , , ,

8

: f"a!����-��t·�, : �i&fd �%, , , ,0 ,,,,%z, , 3 8 ID � m �8%0'" ""0�0'-��� � �-�� '' ·

2 1

a

18 ... cxb3?

b

c

d

e

f

g

h

At this point I think he offered me a draw. I did not hear it, but some of my teammates told me so after the game. Maybe I blocked it out subconsciously. . .

a

b

c

d

e

f

g

h

I believed I had excellent chances. There is really no reason to look any further during the game. When you have something you like, and there are no immediate tactics, there is nothing to calculate. Looking further now, it appears that White is indeed very close to winning. The engine considers 22 . . . bxa3 a realistic move, which it certainly is not. And the normal move is met by a direct attack: 22 . . . a5 23.axb4 axb4

Boris Gelfand - Dynamic Decision Making in Chess

220

a

b

c

d

e

f

g

h

24.�xd7!! The engines can be useful at times. This move wins in a pleasing way. 24 . . . xd7 24 . . . �xd7 is of course possible. But after: 25.lll b6t �d8 26.lll xd7 �xd7 27.�d l t c8 28.Wd3 We7 29.ie5

a

b

c

d

e

f

g

h

White is completely dominating and has sacrificed nothing. We do not need the support of an engine to see that this is winning. 25.lll xc5t �c8 25 . . . ixc5 26.Wxc5 leaves the black king totally exposed. The threat of ib5 t as well as 26 . . . �e8 27.ih5 t simply wins. 26.lll x b7 26.ic4 is also strong. The forcing line is nice, but not mandatory. 26 . . . xb7

a

b

c

d

e

f

g

h

27.ia6t! �c6 27 . . . �xa6 28.�a l t leads to mate. 28.We2 �d7 29.Wb5t �e7 30.Wb?t �d7

This variation is very forcing and long. Although it might seem a bit impractical, I still might have tried to work it out had the position at move 23 arisen at the board. White's position is very promising and it is natural that the attack should win, so why not go for it? b) Instead Black should play: 1 9 . . . cxb3 20.axb3 (20.e5 �xg3! gives some counterplay) 20 . . . e5!

Chapter 7 - Dynamic Masterpieces

%��· B. . Y.� •.:,. '.�.1 �· ,�""�· a� �·ef,'" 3���ifi-� 0 ��t• •�>! 5 � �l. -8 �•�%. ,. ,,3, .8%• �· · ef �· � � � ���8� ,. . . .�� "£W/J' 8 ·� >! � � : � :%m

8

8

7

6

�

4

�

3

2

a

b

��

:

c

d

Y,

....

e

7

6

5

••.•

4

%. �

�,

f

g

�

22 1

3

2

a

h

b

c

d

e

f

g

h

After which I was planning to either continue the attack with 2 1 .:B:a 1 or prepare it a little more with 2 1 .�h2. In either case White is better, but it is certainly a playable position for Black. The engine is very much on White's side, but I think it overestimates the exposure of the black king.

But this is all thinking after the fact. During the game I did not consider it as a serious option for Black to open the c-file with . . . cxb3 and then play . . . c5.

19.axb3 fa.e7

20.l:kl

l 9 . . . c5 makes no sense now.

· • .1 · �• • 15 �' ,�. ., a�A�6�m· � . J� • • � � 0 ���· >!· ., ..,J��r�• �� � ,��.J� � �1���iwtJ�� �� r��,, 8

6

22.:B:xd7! ixe3 23.:B:c7t �b8 24.:B:xf7t :B:xg3 25 .:B:xf6 ixf2 26.liJd l ! and White wins a piece.

This is a natural move.

20 .. �h8 .

It is hard for Black to come up with anything.

. .

4

.

a

b

c

d

e

f

g

h

White has several tempting ways to play. To be honest, it looks to me as if White should win in a few moves. We could probably analyse a more or less forced win with 20.e5 , but perhaps the strongest move is the human 20.�h2!, stepping away from the tactics on the c5-g l diagonal and asking Black what he wants to do. For example: 20 . . . cxd4 2 1 .:B:xd4 ic5 (2 l . . . e5 22.:B:d5! and Black's position looks untenable)

After the game I was unhappy that I had not played 2 1 .:B:fd 1 , when Black has no

222

Boris Gelfand - Dynamic Decision Making i n Chess

counterplay whatsoever. If we have achieved this, we are already doing well. White should win if he does not blunder badly. During the game I also considered 2 1 .ll'lxb5 axb5 22 . .ixb5 , but did not like it after 22 . . . �xg3, when the position has become unnecessarily messy. The computer points out that 2 1 .ll'ld5! just wins. It is a really beautiful move and I do not want to ruin it by attaching long variations; it is quite clear that the black king's position is falling apart. I think I would have found it if it was the only good move in the position, but the availability of other attractive options means that we sometimes look less hard for particularly attractive moves like this one. The real disappointment is that I did not play 2 1 .�fd l , which would be a simple and sound positional move. We cannot rely on the inspiration that finds a move such as 2 1 .ll'ld5, but in order to play good chess, we need to make sound decisions. And the decision in the game was not good.

21. ..Vg7 22..ta e5 Black should not have been allowed to do this.

23.�d5 J.d6 The most realistic move. Taking the pawn on h4 was not compelling: 23 . . . .txh4

I intended to play 24.dxe5 ll'l xe5 25 . .ixe5 Wfxe5 26.Wf a7, when it is not easy for Black to make a move. The direct threat is 27.�xc6t .ixc6 28.ll'lb6 mate. But after 26 . . . �g6 27.�fd l (threatening 28 . .ig4t! with mate} , Black plays 27 . . . h5! and although his position is horrible and no doubt lost, there is still not a direct punch to end the game. A funny point is that he would most likely find 27 . . . h5, even without seeing the threat. There are simply no other moves. So for this reason it was actually simpler to play 24.Wf c3 'tti b 8 25.ll'lb4 with a winning attack. For example: 25 . . . �c8 26.�a l and Black is facing a catastrophic invasion. 8 7 6 5 4 3 2 1

a

24.Vc3!

b

c

d

e

f

g

h

I was happy when I discovered this move. It is flexible, preparing both ll'l b4 and ll'l e3-f5. At the same time the queen is heading for a5, after which all sorts of combinations become possible. A neat little point is that 24 . . . exd4 25 . .ixd6 dxc3 26.ll'le7 is mate. At the time the game was played, the computer did not even consider this move, even at a high search depth. But from a human perspective, this move is very attractive. It gives a lot of

Chapter 7

-

Dynamic Masterpieces

options and, compared to 24.Wd2, it does not allow 24 . . . Wf8 (with the idea of playing . . . exd4) because of 2 5 . dxe5 . I also anticipated the combinations that became relevant in the game. Not that I saw every detail, but I discovered that it was very dangerous for Black and that this theme was decisive in many lines.

223

So instead I should play 26.tLl d3 f6, when White has a big advantage, but the game goes on. I expected that he would play like this, and try to hold. I also wondered if he could defend with: 24 . . . f6 2 5 .Wa5 e8 33.YMcBt llid8 34.¥Md7t leads to mate, as well as winning material , satisfying all urges.

Boris Gelfand - Dynamic Decision Making in Chess

226

8 7 6 5 4 3 2

33.Vcst dle7

1

a

b

c

d

e

f

g

h

At this point something horrible happened. I wanted to play 34.tll d 5t 'it>f7 3 5 .i.h5t and it's over, but somehow I touched the queen. I have no idea how it happened, but you can imagine my horror at the thought of ruining such an amazing game because of touch-move. Luckily, checking with the queen also wins the game.

34.'!Wd7t?! @f'8 35.Vxe6 I could not believe my luck. White simply wins.

35 ...Ve7 After 35 . . . Wfg6 the simplest is 36.lll d?t 'it>g7 37.Wfxg6t @xg6 38 .exf5t, winning a piece.

36.Wfxh6t gg7 37.tll cs

1-0

a

b

c

d

e

f

g

h

The big point I want to make using this game is that a combination such as 26.:B:xc6t! is wonderful, but not something you need to calculate accurately in advance. Black has so many possible moves that it makes no sense to be prepared for all of them. Calculation is about seeing as much as is needed in order to make a good decision. Not about seeing everything and having a reply ready to all of your opponent's moves. Often it is enough to see the general possibilities and leave the calculation till the moment when it will be rewarded. For example, in the line with 30 . . . 'it>eS 3 I .Wfxc6t I did not look further. There are so many possibilities . . .

Chapter 7 - Dynamic Masterpieces As good as it gets

Sergey Karjakin - Boris Gelfand World Cup, Khamy-Mansiysk (6.1) 2009

The World Cup is a very interesting event. Originally it was the knock-out World Championship. The way I see it, there was a time when we had no reputable World Championship, and the World Cup substituted for it in that period. Now we are in the same situation as in English football, where we have a main championship that is decided over a long period of time, and a cup system that is played over a brief intensive period. In my opinion it is great that we now have both. The more good chess tournaments the better. I have played in a lot of these knock-out tournaments and have not really fared especially well, but Khanty-Mansiysk 2009 was an exception - my Glory Days if you like. Some of the days I played excellently, as the reader will have a chance to see in these books, while on others I played not so well at all, and I was rescued by something as simple as dumb luck. All in all, this is the formula for success in the World Cup: you need to play well overall and be lucky when you play less well. . . Starting from 2005 there were a couple of chess tournaments in Khanty-Mansiysk and since then it has become one of the chess capitals of the world. Many people will be surprised to learn that this city, located in the middle of Siberia, far away from everyone and everything, is a modern, prosperous city with good infrastructure. The tournaments held there are always organized very well, and I have been fortunate to play there many times. This game was played in the semi-final after a lot of tough matches. For example, I had a tiebreak with Jakovenko in the quarter-final,

227

where I played some very good games. But because I did not have a rest day, it was universally decided that Karjakin was the big favourite. So-called experts called my wife to tell her that I would be home in a few days, and she would not have to miss me for much longer. The media also jumped on the bandwagon, declaring that I was too old, and that somehow my brain had stopped working well due to senility or something. . . I have heard these stories since 1 998 and have grown entirely numb to them. The only way they will ever manage to hurt me is if I should pull a stomach muscle laughing at them. There was one cliche that was in my favour - I was rated 30 points higher than my teenage opponent. I am not one of those who think that rating points help us find good moves at the board, or tell us the absolute truth about a player's strength. But ratings were irrelevant, as the media's thirst for burying old players is absolute. This narrative has grown a bit softer in recent years, mainly due to the successes of Anand and myself, but to those who think it is gone, I suggest that you go back and look at the statements made by some chess journalists and retired chess players about the 20 1 2 match. It was decided in advance that the match for the World Championship was boring and irrelevant, because of the age of the players. And as we failed to do the only thing that could have changed this narrative - to get younger during the match - this is still how many people described the match afterwards. If you ever get a chance to talk to anyone who was there, you will know that they have an entirely different story to tell. . . A recent example i s the status o f Levon Aronian before the 20 1 5 Sinquefield Cup. Before the tournament he had 1 8 months of bad results, sliding from a solid Number 2 spot in the world all the way out of the Top 1 0. According

Boris Gelfand - Dynamic Decision Making i n Chess

228

to the rating system, he should have been used for target practice by the other participants. But in reality he won the tournament very convincingly. Anyway, I approached this semi-final like any other match. If I played well, I knew I would get my chances.

encyclopaedic, so I decided to play something simple, and j ust have a game.

3 ... �c6 In the previous rounds I had played 3 . . . c6, which also includes my only loss in the tournament, against Judit Polgar in the third round. I mentioned this game on page 1 50.

1 .e4 e5 During the event, and in those years, I played the Petroff Defence a lot. It served me well, especially in this event. One of the points of the opening is that if White does not have precise knowledge of the opening, and maybe a new idea, Black gets an even game.

2.i.c4 So in this tournament, especially in the tiebreaks, people resorted to this non­ threatening system against me. It is important to have confidence in your openings, especially in knock-out events, so I felt I had already won a little moral victory.

2 �f6 •••

4.�a i.e7 4 . . . ic5 is the more popular move, but would require more knowledge. So I chose obscurity on purpose.

5.0-0 0-0

8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

I had some experience with this system from the early 1 980s, when it was a part of my white repertoire. This was obviously before Karjakin was born. I had also played it a few times with Black against Morozevich and with White against Sutovsky. In both those games 6.1'%e I d6 7.a4 was played, which was fashionable at that time. a

3.d3

b

c

d

e

f

g

h

I realized that when Karjakin played this, he intended to improve on my games with Polgar or Vachier-Lagrave. But I was sure that his knowledge of these lines was not

6.i.b3 This is also what was popular in the 1 980s.

6 d5 ..•

I decided to open the game, simply because I wanted a more open game!

Chapter 7 - Dynamic Masterpieces I had a game in the 1 980s against Viacheslav Dydyshko, who managed to become Belorussian Champion many times and played in the Olympiad team repeatedly, without ever being full time in chess. He wrote an interesting book and has trained a number of promising young players. He had a very good understanding of chess and I learned a lot from playing three or four games with him when I was young. My game against Dydyshko in this line went 6 . . . d6 7.c3 lLi d7, while a 2006 blitz game against Navara continued with 7 . . . 'it>h8.

Another vananon I remember from my childhood goes like this: 9 . . . i.xf3 1 0.Wxf3 lLi d4 1 1 .Wxd5 If White does not cake here, Black has equalized. 1 1 . . .Wxd5 1 2.i.xd5 lLixc2

7.exd5 l£ixd5

8 7 6 5 4 3

2 1

played the not-very-dangerous 8.�a4 against Alexander Cherepkov, who was the trainer of Yudasin and many other strong players. I had cwo amazing endgames against him that are still vivid in my mind.

8.h3 A non-critical move. The attempt co prove an advantage for White would scare with: 8.E:el This can transpose to a kind of Marshall after: 8 . . .i.g4 9.h3 i.h5

229

a

b

c

d

e

f

g

h

This is a very sharp position, leading co an unclear game. Tiviakov played it a few times with White, but it is not obvious chat White is fighting for an advantage. For example, Tiviakov - Hector, Helsingor 2008, continued: 1 3.E:xe5 E:ae8 1 4.E:e2 lLixa l l 5 .i.xb7 and here che game i s still unclear, but essentially Black is fine, I chink. 1 0.g4 i.g6 1 1 .lLixe5 lLi xe5 1 2.E:xe5 c6

8 7 6 45 23

a

b

c

d

e

f

g

h

As played in Kramnik - Kasparov, Novgorod 1 99 5 . I should add chat this is not j ust me looking it up in the database while writing. le was on my mind during the game, although somehow I believed it was a mid- 1 990s rapid game. Spot on, for the year, less so on the game format!

Boris Gelfand - Dynamic Decision Making in Chess

230

8 . . a5 .

I remembered that this move was possible. The idea is chat 8 . . . tll d4?! is dubious on account of 9.tll xe5;!;.

a

c

b

9.a4

d

e

f

17 ... !!a6!+ directly was stronger. 1 8.@h l ? 1 8.f4 Wfh5 1 9.id2+ was necessary. 1 8 . . . Wfd5t 1 9.@h2

h

g

This is of course what Black wanted to provoke. I remembered the following game very well: 9.ia4 tll d 4! 1 0.tll xe5 tll b6 1 l .c3 1 l .ib3 if6 gives Black excellent compensation. 1 1 . . .tll xa4 1 2.Wfxa4 tll e2t 1 3.@h l if6 1 4 .d4? 1 4.Wfe4 tll xc l 1 5 .!!xc l ixe5 1 6.Wfxe5 Wfxd3 and White cannot really be said co be worse, although he needs to be careful.

a

b

c

d

e

f

g

h

1 4 . . . ixh3!! 1 5 .gxh3 Wfd5 t 1 6.@h2 ixe5t 1 7.dxe5 Wfxe5t?!

a

b

c

d

e

f

g

h

22 . . . !!d8? A tactical oversight. 22 . . . !!h6 23.l:!f3 tll g3!! would have made this game a perfect masterpiece. 23.l:!f2! !!d3 24.Wffl Suddenly White has defended everything. 24 . . . !!h6 2 5 .tll f3 g6 26.Wf g2 Wf d5 27.tll g 5 Wfxg2t 28.@xg2 f6 29.l:!xe2 fxg5 30.fxg5 !!hxh3 3 1 .if4 !!h4 32.!!e4 1-0 Shirov - Mozetic, Tilburg 1 993. 9.a3 a4 1 0.ia2 @h8 was played in another Kramnik - Kasparov game, and this time it really was a rapid game, played in New York 1 99 5 .

Chapter 7

-

Dynamic Masterpieces

9 ... �d4 I was already happy. We have a normal game with no theory, and Black has made it out of the opening with a decent position. 9 ... .ie6 is also possible, but less flexible. Still, Onischuk used it to equalize against Tiviakov without too many problems.

10.�xd4 1 0.llixe5? llixb3 l l .cxb3 gives Black more than enough compensation.

come to g6, so the only thing I needed to check was if White took the piece. l l . . . .ie6 with even chances was also possible. 8 7 6 5 4

10 ... exd4

3 2

8

1

7 6 5 4 3 2 1

23 1

a

b

c

d

e

f

g

h

While Karjakin was thinking, I noticed the idea l 1 .lli d2 E:a6! with attacking prospects. So when he moved his rook to e 1 instead, I was initially disappointed, thinking the . . . E:a6 move was now impossible, as a piece would be hanging. Instead I wondered about playing 1 1 . . ..ie6 or something.

I I J�el But once I started to look deeper, I realized it actually was possible, and I got really excited. It is a rare opportunity to play such a move, and I did not want to let it go to waste.

l 1 ..J�a6!! Played after 14 minutes. The point is simple: if White does not take on d5, the rook will

White has a tough choice. At move 1 2 in a two-game match, you already have to decide whether to allow a draw, or to take risks. Time and again, I have seen players being criticized for forcing or agreeing a draw, in a position where they felt that things had already tipped, and they were in danger of being worse if they decided to avoid the draw. I had the same situation in my game in Amsterdam 20 1 0 against Caruana, where he made a mistake in the opening and offered a draw after 13 moves. I declined and won a tough game, which we will see in the next volume in this series. After the game a journalist asked me if Fabiano had "chickened out" by offering a draw. I tried to explain that rather it should be seen as showing a subtle understanding of the position, where there was no realistic way he could fight for the initiative. I believe a lot of the chess public think along the lines of: ''Anything can happen, a bad position can turn around, so you should fight to the end." But in some positions this is not the case. They are played for two results only. In such a situation, a classical player will try to minimize the damage, and try to make a draw. To me this is rational.

232

Boris Gelfand - Dynamic Decision Making in Chess

I heard the following story, which is probably true. Bent Larsen was playing in some tournament against a weaker opponent. The position is heading towards a draw, when Bent decides to make a kamikaze sacrifice. After he wins the game, Petrosian comes up to him, shaking his head, stating: " Bent, you are crazy. . . I go here, here and here, and you are just lost." "Of course," says Bent. " But in this position I make 2/3, while you make three draws." As you will know by now, I went to Petrosian's Chess School, and probably share his sensitivities to a large extent. Also, the level of the "weak" opponents I face today is no doubt significantly higher than the level of the weakest participants in top tournaments 50 years ago. Still, the romanticism of that time lives on with the public, and it is easy to understand why. But we should not forget that Petrosian became World Champion, while Larsen was brutally stopped in his attempt to do the same by Fischer.

12.Vhs I think this moment was one of the most important in Karjakin's career. It helped him to mature immensely. In the games from the 20 1 5 World Cup we can see that he learned the lesson and won with a more "realistic" approach. Then he changed again for the 20 1 6 Candidates and won that too. It will be interesting to see which evolution of Karjakin will face Carlsen a few months after this book has been published. As said, the critical line was: 1 2.ixd5 'Wxd5 1 3.E!:xe7 As 1 3 . . . i.xh3 does not work on account of 1 4.'Wf3, Black needs to swing the rook over: 1 3 . . . E!:g6

a

b

c

d

e

f

g

h

White has only two reasonable ways to avoid checkmate. The first one I calculated was 1 4.f3 .ixh3 1 5 .E!:e2 'Wxf3 1 6.'Wfl , when Black has to play 1 6 . . . .ixg2! 1 7.E!:xg2 E!:xg2t 1 8 .°Wxg2 °Wd l t and Black gives a perpetual. Karjakin said that he had seen this, but thought it was too early to agree a draw. The other question is what happens if White plays: 1 4.g4 I had planned 1 4 . . . fS , which is probably not that clear after 1 5 .c4 °Wd6 1 6.°We2. But instead another move exists: 1 4 . . . °Wh5 !

a

b

c

d

e

f

g

h

I had not seen this move. When you are calculating something like 1 1 . . . E!:a6, you are checking if it can be refuted. In this case I already had a decent option and it was clear

Chapter 7 - Dynamic Masterpieces that with White's position being this open, I would get my chances. This is all I needed to know about this variation. Analysing 14 . . . WhS after the game, realized that this is directly winning: 1 5 .if4 l 5 .�e4 f5 l 6.�f4 Wxh3 is j ust over. 1 5 . . . ixg4! 1 6.hxg4 �xg4t 1 7.@fl �g l t Although materially White is not really behind, his position is devastated. The tactics here are not difficult to calculate. It is all about seeing 1 4 . . . WhS, which I probably would have done once we had reached that position. As my position is clearly promising no matter what, I do not need to see this move in advance.

But strategically my move is also rather logical. White wants the a l -rook to be on e l , while Black wants to put the a6-rook o n g6. This happens in the game, but in a situation where White is able to be a bit better organized, which is the reason I prefer 1 2 . . . ib4 now.

13.tlia3 The most logical alternative is: 1 3.llid2

8 7 6 45 23

a

a

12 ... tlib4!?

b

c

d

e

f

g

h

I played this quickly. But I realized after I made my move that l 2 . . . ib4! was even stronger. After 1 3.�e2 �e6 1 4 .igS We8 l 5 .�xe6 ixe6 Black has comfortably developed his pieces, while White is struggling with the task of getting his knight out. We can say that Black is at least comfortably equal. What I played is by no means weak, but it would have been nice to continue in the aggressive style of ignoring the threat to the d5-knight.

233

b

c

d

e

f

g

h

I think I had nothing planned, as he made his decision rather quickly, leaving me no time to seriously consider my options. A clear point would be that 1 3 . . . �g6?! is met with 1 4. lli f3, when White is more comfortable. But it is likely that Karjakin rejected it on account of: 1 3 . . . �h6! This is rather unpleasant for White. 1 4.WxaS

8 7 6 45 23

a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making in Chess

234

This would be the logical move, as moving to e2 would be a retreat. White's position is still essentially OK, but he would have the feeling of being pushed back, without going to a better square. At this point we have two options. I had previously only considered: 1 4 . . . b6!? This forces the queen to go to the corner. 1 5 .Wa8 l 5 .We5 ? is punished by a lovely tactic:

8 7 6 45 32 a

b

c

d

e

f

g

h

It is possible to analyse this more or less to a draw, but very few would deliberately put their queen in the corner. A key point is that with the queen in the corner, White could easily find himself lost at any moment. Analysing the game this time around, I became aware of an enigmatic bishop sacrifice: 1 4 . . . ixh3!

a

b

c

d

e

f

g

h

1 5 . . . lll x c2! 1 6.Wxe? Forced. ( 1 6.ixc2? Ei:e6 loses material) 1 6 . . . Wxe? 1 7.Ei:xe? tll xa l Black has won the exchange and is better. The knight will not be trapped. White does have some activity and will not simply fall over, but Black's chances are preferable. l 5 . . . tll c6� Something like 1 5 . . . Ei:g6 1 6.Wf3 ie6 would be out of the question for me. I had a strong determination to avoid Petrosian­ style sacrifices such as l 7.Ei:xe6!? fxe6 1 8 .We2, where I believe White has quite a comfortable game. This is what I considered after the game. White now has to play something like: 1 6.Ei:e2 The engine will claim that the position is equal, but with many good options for Black and not so many for White. This alone makes it shaky territory to navigate for White.

a

b

c

d

e

f

g

h

This looks incredibly dangerous for White. If he does not accept the sacrifice, Black would play . . . Ei:g6, forcing further concessions around the white king, which is clearly intolerable. So White has to accept the sacrifice: l 5.gxh3! It is better to feel uncertain about the evaluation of the position, than it is to know for certain that you are worse; even if psychologically it can feel better to know (people tend to lean towards some sort of

Chapter 7 - Dynamic Masterpieces certainty, when possible - as well as when it's not possible!) . 1 5 . . . !%g6t 1 6.@fl !%g5!

8 7 6 3

for the exchange with a serious of good moves: 2 1 . . .ih4 22.l:%e2 !%f5 23.'fic4 'fid7! and Black is OK. l 9.tLlxg5 .ixa7

8 7 6

5

4

5

b

c

d

e

f

g

� � �· 8 �r. 7 �� -----%���------,�-�-,��-----%-,

4

23

� � � , �r,/�� lfi'tti� �� . ----� � 8 ,B ·----"� a rm� � '------t� �m� �i� � � � ��� �- ----%� '"'(D ""% 8 �- � . .....

a

b

c

d

e

f

g

h

Somehow the engine can make 20.f3! tLl c6 2 1 .'fic4 hold together, although it looks terrifying to me. (The human 20.@fl ? .ic5 2 1 .'fic4 leads co disaster after 2 1 . . . .ixf2!+ with what looks like a deadly attack.) 1 8 . . . .ic5 ! After l 8 . . . tLi c6 l 9.tLlxg5 .ixg5 20 . .ie6! fxe6 2 l .'fic5 Black has co j ustify his compensation

� i w% ...�'.� .....

.....

>,, , , , , ;

� �6!1·�1 �· § � � 1

h

Forcing the queen to a poor square. 1 7.'fia7 1 7 .!%e5? b6 would lead co material losses. 1 7 . . . Wc8 The attack is certainly dangerous for White, but objectively his position is tenable. 1 8 . tLl f3! 1 8 .'fixd4 'fixh3t 1 9.@e2 Wh5t looks dangerous.

:

� ·;: ''0 ! �• & �% i �� � �� . ��� �% "'� � � • • � �- "'--- --� � '/,,,,,;� � . .. }- - - %�.,� �c-z- - -. . %� .:

4

a

235

a

�

b

c

d

e

g

f

h

If Black had time to play . . . .ic5 or . . . Wd7, his position would be under control, and he would have the advantage. But this is not the way things work in a dynamic setting. White plays: 20.tLlxf7! !%xf7 2 1 .l:%e7 Black has to give a perpetual with: 2 1 . . .Wxh3t 22.@g 1 Wg4t In order not to lose. Again, to be clear, I saw none of this during the game.

13 . . J:�g6

The only move that makes sense in the position. 8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making in Chess

236

14.Af4?! This is the most obvious move, although it was not very successful in the game. White has not made any obvious mistakes, but still it is hard to see exactly how he should secure equality. We are able to do this after an hour with the help of an engine, but for a human during the game, this is not relevant. Another thing I considered during the game was 1 4.J.d2 b6 1 5 .ge4 (with the idea 1 5 . . . J.b? 1 6. gg4) .

such as l 5 .tll c4 could easily get into trouble after 1 5 . . . J.b? 1 6.J.g3 J.d5!, with a difficult position for White. I should maybe mention that I saw this, rather than only pointing out what I did not see! 8

7

6 5

4 3

2

b

a

c

d

e

g

f

h

Black is seriously considering taking on c4, with the trap that 17 .ge2? loses instantly to 1 7 . . . gg5! 1 8.�h4 ge5 1 9.�g4 h5, as I discovered a good while after the game.

a

b

c

d

e

f

g

h

1 5 .J.g3 is possibly the best move, but Black is better.

But Black has 1 5 . . . tll a6! 1 6.gae l J.f6, when there is no clear path for White to equality.

14... b6i I spent 1 4 minutes on this decision. 1 4 ... ie6!? was a serious alternative, but I did not yet have this idea. At this point I was trying to make . . . J.b7 work in various variations, which is the most natural thing to do when you get a rook to g6.

15.Yff3 An easy move to understand. White is

preventing the bishop from going to b7. During the game there was a moment when I believed that I was actually worse here. But just a moment. White's task is not easy. A natural move

a

1 5 ...J.e6!

b

c

d

e

f

g

h

This is a nice move to play, but it was really not that difficult, and did not take a lot of time to find, as there are not many alternatives. What I like about this move is that White has done quite a bit to restrict this bishop. First

Chapter 7

-

Dynamic Masterpieces

he played h3, preventing it from pinning the knight from g4. Then he played 'Wf3 to prevent it from coming to b7. Finally, it comes to e6 and White's position becomes really difficult. 1 5 . . . .id6 1 6 ..ixd6 l3xd6 1 7.l3e5 feels freeing for White, who will soon be able to get the other rook into the game.

16.J.xe6 Not a pleasant move to play, but with the clock ticking away, sometimes all you can find is that you have to play a move in order to be allowed to play another afterwards. The first thing I looked at was of course the positional sacrifice with 1 6.i3xe6 fxe6 1 7.'We4.

1 6 . . . lll xc2!!+ wins the exchange after the forced 1 7 . .ixe6 lll x a l , as 1 7 . .ixc2? .id5 leads to a winning attack. Again, it can be analysed, but during the game it is enough to see that Black has many tempting possibilities.

16 ... fxe6 17.Yfe4

8 7 6 5 4 3 2 1

a

17 ....id6!!

a

b

c

d

e

f

g

h

But then I found 1 7 . . . lll a6!, which gives Black a decisive advantage. Of course I did not know at the time that this was winning; all I knew was that I did not see a decent move for White. 1 6.i3e2 also has a serious drawback. Black has a big tactic:

8 7 65 4 32

237

b

c

d

e

f

g

h

I liked this move very much for many reasons. I looked at 1 7 . . . lll d 5 first, but was not satisfied with it. Then I got the idea that exchanging the bishops would give me control of the f4-square, which is vital for the attack. As Razuvaev taught me: when all the pieces are attacking and none has to defend anything, the attack is probably strong enough to be successful. He called it "coefficient of attack"; you calculate the strength of the attack against the strength of the defence. Here four pieces are attacking and, after the f4-bishop goes, only the queen helps with the defence. Under such circumstances, something decisive usually pops up. So 1 7 . . . .id6 tries to exchange White's prime defender. The knight on a3 is too far away, and the rook on al is not going to do much good either.

18.hd6? After the game it is easy to see that it is not possible to hold the position once the knight a

b

c

d

e

f

g

h

238

Boris Gelfand - Dynamic Decision Making i n Chess

comes to f4, and that for this reason 1 8 . .ig3 was the only way to stay in the game. After 1 8 . . . 'iNg5 it is clear that Black has an attack, but actually White also has some defensive resources. It would be possible to include a few pages of computer analysis here, but I hope the readers will trust our j udgment, based on having done the analysis, which is that Black's chances are extraordinarily excellent! You may ask why I chose to analyse 1 3.tlid2 in as much detail as I did, especially the variation with 1 4 . . . .ixh3, and now here I just give a verbal explanation. The reason is simple: those lines were too beautiful to ignore!

18 ... cxd6

Either 2 1 .Wg2:

a

b

c

d

e

f

g

h

2 1 . . .l:!h4!! and there is no adequate defence against the knight landing with great power on f4, causing widespread destruction. Or 2 1 .Wh l tlie3! 22.l:!fe l , when Black has the equally beautiful:

19.'l'xd4 From a human point of view, Black's position is totally winning, even if it would still be quite a headache to prove the same result against a computer. Still, let me give you a few extracts from the Matrix. 1 9.g3 is met strongly with 1 9 . . . tlid5 20.'iNxd4 tlif4 2 1 .@h2 e5! 22.We3 tli e6 when the knight is coming to d4 or g5 . 1 9 .i"ffl is refuted by a series of brilliant moves: 1 9 . . . tlid5 20.g3 l:!f4! and now:

a

b

c

d

e

f

g

h

22 . . . 'iNh4!! 23.Wh2 l:!f3 with a winning attack. Again, I did not consider any of this during the game, but the line is lovely all the same.

19 ...'l'gS It is always fun to play a mating attack when you have only invested a pawn. If you have sacrificed two pieces, it is easy to become nervous, and the pressure is really high. In no way did I feel like that during this part of the game.

20.g3 '1'5 2 1.g4

Chapter 7 - Dynamic Masterpieces 2 1 . fle6 42.f3! and no matter where the rook moves to, White secures a perpetual check. 37.l:!xf7t l:!e7 38 . .ia5 b4! 39.WfdBt @e6 40.l:!f4

h

Again I had a narrow choice, and I think I made the wrong one. I was certain that the following line did not work for me: 3 1 .Wff4 f5! 32.Wfxf5 ge8 33.l:!f4 l:!b7 34.Wfh7 ge6

40 . . . l:!f7! 40 . . .Wfd7 4 1 .Wf g8t @e5 42.Wfb3 gives White a lot of counterplay. Not only is it difficult

Boris Gelfand - Dynamic Decision Making in Chess

252

in practice for Black to find all the right moves, but even then, White is sort of OK after: 42 . . . E:f7! 43.Wfe3t 'itid6 44.Wfd3t 'itie7 45.Wfe3t 'itif8 46.Wfxc5t 'itig8 47.E:xf7 Wfxf7 48.j,xb4 Although Black has some winning chances, White should hold, as long as the queens stay on. 4 1 .E:h4 Wfd7 42.E:xh6t 'itid5 43.Wf g5t Wff5 44.Wd2t 'itic4 45 .Wa2t 'itib5 46.We2t c4 47.j,xb4 Wd3! 47 . . .Wxf2t 48.Wxf2 E:xf2 49.'itixf2 'itixb4 50.E:xa6 and White holds. 49.Wfe l E:fd7 50.E:h5t E:d5 5 1 .E:xdSt Wxd5

a

c

b

d

e

f

h

g

All of this was forced. Black is winning, but we still have to make moves. My ambition here was to try to create as many traps as possible. I think Ponomariov played pretty decently up to the time control.

35.h4 ©es 36.'flYxh6 YlYe4t 37.©gl ©d7! 38.YlYd2t ©c6 39.hS YlYd5

a

b

c

d

e

f

g

This is a natural and perfectly logical move to make with little time on the clock. h

Black has excellent winning chances, but the task is not at all easy. All in all, this is how I should have played, especially when you take into account all the chances for Black to make a misstep along the way.

If he had more, he would no doubt have spotted 39 . . . Wb l t! 40.'itig2 Wf5!, which is very uncomfortable for White.

31 ..J�e8! 3 l . . . E:b6?? would be a big blunder. White wins after: 32.Wfg4! Wfxf6 33.Wxc8t 'itig7 34.E:g4t

32J�xe8t 'flYxe8 33.'flYf4 The main target is the h6-pawn. White needs something with which to create counterplay. 33.Wg4 does not work because of 33 . . . Wf c8! 34.Wg7t 'itie8 3 5 .Wg8t 'itid7 36.Wfxf7t 'itic6 and Black escapes.

33 ... 'flYei t 34.'itig2 gc8

a

b

c

d

e

f

g

h

In order to keep the game going, I would have to play 4 1 .Wh6, but the position after something like 4 1 . . .WdSt 42.'itih2 b4 is virtually hopeless.

Chapter 8

-

During the game I expected 39 . . . b4, but it also makes sense to make no big decisions till after the time control.

42.h6 8 7

40.Yfe2 Yff5 41 ..lg7

6

8

5

7

4

6

3

5

2

4

1

3 2 1

253

Dynamic Defence

a

42 ... i>b7? a

b

c

d

e

f

g

h

We have passed the time control, which is to my detriment; my position is still lost.

41 ... c4? Ponomariov did not look at the position in the same way as I did, and fell for my one remaining trick. And it is easy to see why: if White does not have the trick, Black simply retreats the king to b 7 and then pushes the c-pawn all the way to the end. Looking at the position from another angle: Black should want to win the game from here without giving his opponent any counter­ chances. Pushing the c-pawn opens up Black's king position horribly, and makes the win rely more on accurate calculation than it should.

c

b

d

e

f

h

g

Black continues with his plan, but at this point there were still ways to win the game. The computer likes 42 . . . Wf d3, but the most human way to play is as follows: 42 . . . �dS! This is what I was worried about. Black is planning ... �d3 in order to push the c-pawn. The most critical line goes: 43 . .ic3 �d3 44.Wf eSt 'it>b7 4 5 .Wf e?t 45 .h? could be tried, but Black can simply take it: 45 . . . Wfxh?! 46.Wfe?t 'it>c6 47.WfeSt 'it>c7! A bit of triangulation is necessary. 48 . .ia5 t ®b7 49.Wfe?t 'it>c6 50.Wf eSt ®d5 Black escapes the checks and wins. One point is that ... �d l t is a threat, so White is in a desperate state. 45 . . . �d?!

During the game I thought he should play 4 I . . .a5! 42.h6 a4, when I did not see any way to create counterplay. There are some ways to keep the game going for a bit, but it all appears pretty desperate. I did not have any idea of what I would do. The black king goes to b6 and is pretty safe, so what can I do? a

b

c

d

e

f

g

h

Boris Gelfand - Dynamic Decision Making i n Chess

254

46.Wxd7t?! White can retreat the queen to e3 and play on, which is what I would have done. 46 . . .Wxd7 47.h7 Wd8 48.h8=W Wxh8 49.�xh8 a5 50.@fl b4 5 1 .@e2 a4 52.©d2 ©c6 53.g4 c3t 54.@c2 @d5

43 ...t;Vbl t 43 . . . a5!? 44.h8=W :B:xh8 45 .�xh8 Wd5 was possible, of course, but with queens on the board, there is no reason to believe that Black should be better.

44.©g2 t;Vxh7 45.t;Vf3t l:k6 46.t;Vxf7t gc7 47.t;Vdst ©cs 48.t;Vast ©d7 49.t;Vdst ©cs so.t;Vast ©d7 1/2-1/z

Counterplay without queens

Black is in complete control. Either he will play . . . ©c4 and the pawns will win the game, or White will take on c3 and have to go all the way to the a-file to pick up Black's passed pawn. Meanwhile Black will take all the white pawns, queen the f-pawn, dig a grave for his opponent, and run a victory lap around the cemetery. . .

In the second game of the chapter, which was the Game 2 of the 20 1 1 Candidates final, I lost my way while exiting the opening phase, and soon had to look for counterplay in a difficult endgame. There is a big difference between this game and the previous one. In the previous game, I believed that I was worse, and later on losing, and played accordingly, while in the next game I was worse, but never lost hope about my position. As you will see, this difference in perception is at the centre of how I approached the position. Boris Gelfand -Alexander Grischuk

43.h7!

Kazan (2) 201 1

8 7 6 5 4 3 2 a

b

c

d

e

f

g

h

White's counterplay finally pays off. The game will end in a perpetual check.

While Game 6 of this match (see page 1 99), in all modesty, belongs to chess history, we should not forget that the match could not have been won without the previous games. In three of them I was in trouble, with this being the most critical situation. I was also in trouble in Game 1 , which we will see in the third volume of this series. Also in Game 5 I played the opening poorly, but managed to offer enough resistance to save the game. In a match, every game has equal importance, even if the final one, especially if decisive, can attract a lot of the attention.

1 .lll f3 c5

Chapter 8 - Dynamic Defence

a

b

c

d

e

f

g

h

This was fashionable at the time. I played the English a lot at that time, and Grischuk played like this all the way through the Candidates.

2.c4 l£ic6 3.l£ic3 e5 4.g3 g6 s.J.g2 J.g7 6.a3 t£ige7 This is considered to be more precise than 6 . . . a5. 8 7 6

255

through opening books with the intention of following a repertoire, but it is always useful to compare with your own knowledge, and see if you can learn something. Personally, I go through all serious opening books in order to see if there are any important ideas that I can use, or which I should be prepared to face. I especially like the ones where the author has done more than just pressing the spacebar. Often all you get from a book is a hint of a new direction that might be interesting to investigate further, but this alone will be worth it. Obviously I am talking about top-level preparation here. I understand that repertoire books such as those by Avrukh, Marin, Khalifman and Negi will be very useful even for average grandmasters. But we will talk more about openings in a later volume. For now, the main thing to say is that 7.b4 was my idea for this game, and we had analysed it a bit, but it appears Grischuk was at least as well prepared. More about this later.

7 ... d5! This is the most natural move.

8.cxd5 �xd5

5 4

8

3

7

2

6

1

5

7.b4 There are a lot of subtle finesses in this system. I think Marin suggested 7.0-0 0-0 8.E:b l , with the idea 8 . . . a5 9.lli e l d6 1 0.llic2 i.e6 I 1 .llie3 in order to prevent . . . d5. It seems that a number of top players looked at Marin's three-volume series on the English at this point. Obviously top players do not go

4 3 2 1

a

b

9.l£ig5 l£ic7

c

d

e

f

g

h

9 . . . llixc3 was also possible. 1 0.dxc3 Wxd 1 t 1 1 .'it>xd l

Boris Gelfand - Dynamic Decision Making in Chess

256

l 8 . .ixc6! with a big advantage on account of 1 8 . . . bxc6? 1 9.tDe4t. l 6.�a4 °%Yb2 And at this point White has a nice move: 1 7.lDdS! lDe5 1 8.�b4 °%Ya3 l 9.i.xe5 he5 20.�a4 °%Yb2 2 I .lDc5 White is very comfortable.

14.cbfl Ylrd4 a

b

c

d

e

f

g

h

I thought this would give me a playable Catalan-style position .

10.d3 cxb4 ll .axb4 e4!? The most challenging, but we should not ignore that Black has a wide choice on this move as well. 1 1 . . .0-0 was playable, for example. The line was unexplored, so no trends had formed yet.

���- - �� �· %· � �* • -� · � � � � ..... � ��.: � l ��tifi,���ir¥1;��� 54 �� -�� � I��-��tt.J 8 -%��� � �-- --, � 8wdi.wtf ----� �� � � 8 7

6 3

� -- - - - %

�

2 1

a

12.�gxe4 f5 13.Ag5 hc3t The best move.

1 5.�xc3!?

1 3 ... '%Yd4?! This was also possible, although dubious. Our idea was to play: 1 4.e3 °%Yxb4

'

�

c

b

"'

, � �: e

d

f

h

g

We felt that this piece sacrifice was a more tempting possibility. We had also spent some time looking at: l 5 .i.f6 .ixal 1 6.i.xd4 .ixd4 l 7.tDd6t 'it>e7 1 8.lDxcSt �hxc8

8 �,�, �� ·0� �� "

7

6

- ;�' ill·� •%� •� ..... �%� · '% • •

54 ��-!������-�,���r£���� %� � •% 3 �. . . �� �-ef�· - ��-� �li · � 2 �� 8 ef� j_ef �. �

� �� ,� �� 7::

a

b

c

d

e

f

g

h

1 5 . .if6! 0-0 Against 1 5 . . . .ixf6 1 6.ttJxf6t 'it>f7 I had prepared 1 7.°%Yd2!, with the idea 1 7 . . . 'it>xf6

a

b

c

d

e

f

g

h

But it was hard to come to any solid conclusions. The feeling is that Black's position is adequate.

Chapter 8

-

Dynamic Defence

1 5 ...tyxc3 White has sacrificed a piece, but Black is a bit disorganized. If White had castled instead of playing Wfl , he would have the advantage.

16.i.f4

�6 i!ra �� � �.

"� " � �� � •a�.-�� � �n�.� ,�. " R �R ··n= % 2 �--- --- - lwrJJ:f! � �� ·� �•: 3 1

a

16 ... lll b S!

c

b

d

e

f

h

g

The most accurate, giving Black a chance to play for three results. 1 6 . . . ltl e6 gives White a chance to play for an advantage: 1 7.l3c1 'Wxb4 1 8 .l3xc6 bxc6 1 9.i.xc6t Wf7 20.i.xa8 lll xf4 2 1 .gxf4 i.e6 (2 1 . . .'Wxf4 22.°Wa l and White is better) 22.i.f3 a5 23 .°Wa l

8 � � � �-- :5 � � ,� �..�� ..�,,,� � 4 3B � ��-�� �� "�� �� 1. �� ���� 2 � � i� i� , �• ��� 7: �....

�-

3

a

b

c

d

�J; • � -� e

f

g

h

It seems that Black has enough compensation for the pawn, because of the passed pawn. Most likely, the game would end in a draw,

257

but White would be able to pose Black some problems.

17J:kl tyf6 18J�c5 a6 19.i.xc6t bxc6 20 ..ieS tyf8

8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

This is where my preparation ended, but to some extent this is also where Grischuk's preparation started. Of course this is an exaggeration, but it felt as if he was more in control than I. I received the file with this variation about an hour before the game, and did not check it especially thoroughly, as I found it very unlikely that it would happen. Grischuk had not prepared specifically for this game, but had simply prepared this line for the tournament, and now he had a chance to use it in this game. Basically I thought the position was complicated, but highly unlikely to appear on the board. One and a half hours later, there it was. There is an important psychological aspect to point out here: After playing such a long and strange line, you have to adjust to thinking for yourself This time I did not handle the adj ustment well, which was an important lesson for me. It is a common mistake among young players to learn some long forced line, but then have no idea of what to do once it

Boris Gelfand - Dynamic Decision Making in Chess

258

appears on the board. It seems this can happen to experienced players as well. . .

21 .YlYcl This is a natural move, combining pressure down the c-file with options of going up the long diagonal. 2 U�xc6? does not work: 2 1 . . .i.b7 22.%Yb3 W/g8! and Black has the advantage.

1

21. ..i.d7 This is very natural as well. 2 l . . .l:'!g8 looks risky. White would play: 22.l:'!xb5 cxb5 23.W/c6t �f7

� i. .I

.I � ... �

�� ��� - - -"� 1r� : r.1. � �;� :� r� � : �� ��"" � ""-�� � 3 �.8�-8�· 8• � 2 � - · "� �� �� �-: 8

8 7 6 5 4 3 2

�, , , , %

1

a

b

c

d

e

f

g

a

22 ..txhS?

b

c

d

e

f

g

h

Played after 1 6 minutes, this was a mistake. I should not have released the tension. Of course I did not want to take on h8, but I did not see anything better, and I also did not fully appreciate that I was about to be worse already. I considered 22.l:'!xc6?! for a while, but after 22 . . . ixc6 23.Wxc6t �e7

h

White has at least a perpetual with 24.%Yf6t, although it is more likely he would continue: 24.W/xa8 W/xb4 2 5 .W/d5 t i.e6 26.%Yb7t We7 27.Wxa6, when White is running no risks and can play for a win with his extra pawn. Contrast this with the game, where we were playing for three results, and Black's position was arguably easier to play, as he has extra material and thus less to prove (though obviously some players would prefer White) . In my opinion the game continuation offers Black a very easy game.

24.Wb7t �e6 2 5 .%Yc6t does not give a perpetual because of 25 . . . tll d 6, while 25 . . . �xe5 ?? 26.f4t �d4 27.�f2! Wxb4 28.l:'!c l leads to checkmate. The computer suggests: 24.e4

Chapter 8

-

Dynamic Defence

The main point of this move is that White is preparing a perpetual on b7 and d5. But it is not a move I considered during the game. OK, if l got to this position during a game, I would find this move; it is not that difficult. But to go voluntarily in this direction did not make much sense to me. 24 . . . l:!d8 A natural move. 24 . . . 'Wc8 ? 2 5 .'Wf6t c.t>d? 26.'Wf7t c.t>c6 27.c.t>g2 and suddenly everything works for White: 27 . . . c.t>b6

1. �1�� �� �. . �� :� :�·� -0� �� m-��1£ · 5 �·� �W·� � -� · 4 ���� li -- � . . . , %� !� �� �� �� 2 �. �� �� � . . . , %� -: 8

3

a

b

c

d

e

f

g

h

28 . .if4!! A simple move, but not easy to see. For example: 28 . . . l:!d8 29 . .ie3t tll d4 30 . .ixd4t l:!xd4 3 1 .°Wf6t and White wins. 2 5 .c.t>g2 White is apparently fine. But to play like this is not natural, and I would be reluctant to do it during a game. However, I should add that the computer recommends: 22 . . . l:!g8!

259

23.c.t>g2 'Wxb4 And claims Black has slightly better chances. But the computer also points out a much stronger move. 22.'Wc4! There are a few lines that are not at all obvious, but make sense once you look at them. The computer suggests a lot of random-looking moves, and it would be possible to confuse you endlessly by just pressing the spacebar repeatedly, and then printing it. But let's look at one variation that leads to a draw, randomly. . . 2 2. . . h 5 22 . . . l:!g8 ? loses to:

�·�, . ,,%� I. � � � 76 �� �,����. ��� i .A � . �m-0 �� : �-��� �� .. . . >. . .. �� ! 2 �. �.!-� . .. . %� �� �� �� : 8

I.

3

a

b

c

d

e

f

g

h

23 . .if6!! with various threats. 23.l:!xc6 .ixc6 24.'Wxc6t c.t>e? 25.f4! l:!c8 26.'Wb?t c.t>e6 27.'Wxa6t tll d 6 28.'Wa2t c.t>d? 29.'Wa4t l:!c6 30.b5 l:!cl t 3 l .c.t>f2 l:!xh l

Boris Gelfand - Dynamic Decision Making in Chess

260

23.'i'e3t

32.bGt 'it>e6 33.%Va2t With a draw. One of the problems with my approach at this stage of the game was chat I was seeking an early crisis (preferably immediately) . I was unhappy with my preparation, chat my bluff had been called. It would not be unfair to say chat I had lose my psychological composure, which can be seen by the poor way I handled this phase of the game. What is surprising about all of this is not chat I played as poorly as I did, but chat White has so much time to get his pieces into the game. It feels natural to chink chat Black would then be able co get organized, but with such a weak king, this is harder than at first anticipated. There is an important point I want to make, which is at the core of what these books are about: I want to explain the chinking chat has led to my reasonable success as a chess player, and not 'cheat' in the process. It is quite easy co analyse a variation with the engine and then explain why it works. And this certainly has its uses, but to me it is more interesting to talk about how we find the moves in the first place. This is the key to playing better chess.

22 ...'i'xhs

23.h4 Wf6 24.Wg5 does not really work. For example: 24 . . . We7! 2 5 .Wf4 and then a move it is easy to forget is legal: 25 . . . 0-0-0 with an advantage for Black. 23.g2 is also a natural move to consider, but after 23 . . . %Vf6 Black is well-placed again. One example is 24.e4 lll d4 2 5 .:!! e l f7 and Black is very harmoniously placed. These examples show chat it is much easier for Black to improve his pieces, which is why I felt I was in a rush. This is one of the reasons why I cook on h8, ensuring chat I would not be seriously down in material. If Black gets fully coordinated here, I am worse, but at least I do not have to resign. The paradox is chat it was perhaps only after I took on h8 chat this was really the case!

23 ... @f7 24J�e5 8 7

8

6

7

5

6

4

5

3

4

2

3 2 1

The obvious follow-up. It would have made no sense to cake the rook if I did not give the check. Still, it is not obvious chat it improves White's position, so I analysed a few alternatives:

a a

b

c

d

e

f

g

h

24...'i'tlU

b

c

d

e

f

g

h

This seems rather artificial, but in fact it is an

Chapter 8

-

Dynamic Defence

indication of how well Grischuk was playing in this match. A natural way to play was 24 . . . '1Wf6!? and follow up with something like 2 5 .h4 &i'J c7 26.f4 &i'Jd5 when Black is doing very well. But as we shall see, in many lines it is advantageous for Black to be able to take the b4-pawn. This might sound like a simple difference, but to a strong player it is quite a compromise to put your pieces on odd squares. He played this part of the game very well indeed.

25.h4 h5 The most natural move, preventing White from creating any counterplay. 2 5 . . . '1Wd6? would allow White to create counter-chances with 26.hS g5 27.f4! g4 28.h6! with the idea l:!h5-g5/-f5 .

26 1

26 Yfd6! •••

This was his idea all along; Black forces the exchange of queens. After this, his king will be safe and his advantage beyond doubt.

27.®g2 I had no choice; the rook needs to get into the game.

27 ci>f6 28.ge4! •••

Sometimes the best move is the lesser evil. A worse version of the queen exchange would be 28.l:kS '1Wxf4 29.gxf4, when after 29 . . . &i'Jc7 Black is very well placed and White has lost some elasticity on the kingside, meaning he has no counterplay whatsoever.

28 '1Wxf4 29J�xf4 •••

26.Yff4 I had to play this, as the b-pawn was hanging. After 26.@g2 '1Wxb4 27.'1Wh6 l:!g8 28.'1Wh7t l:'!g7 29.'1Wh8 Black can play 29 . . . 'IWIB immediately, or j ust ignore White and his lack of real threats with 29 . . . aS 30.l:'!c l '1Wd6, with full control of the position.

8 7 6 5 4 3 2 1

a

b

c

d

e

f

g

h

This is how I saw my defensive task: 1 ) First of all, I had to be careful that I did not lose immediately. This sounds incredibly obvious, but it still has to be done.

a

b

c

d

e

f

g

h

2) I had to create active counterplay, as otherwise I would lose slowly. I wanted to open the position, play f3 and e4 and g4 to get the rooks playing, and to create targets. The exchange of pawns would also favour me.

262

Boris Gelfand - Dynamic Decision Making in Chess

This is in contrast to the feeling of desperation it would be easy to feel, which would lead to setting traps and other forms of playing the position as if it is beyond salvation. This is often connected with nervousness. Some players find it difficult to withstand the pressure of being continuously worse, and try to get some clarification quickly, so they can get out of an unpleasant situation. My approach was the opposite. I wanted to stay in the game as long as possible. The position is bad, but it is the position I have, and I wanted to do my utmost to defend it. I should add that my opponent has played perfectly up to this point. I do not know for sure if this position is a direct win, or if it is just very close.

29 . .i.e6? .

This was played in order to trap the rook on f4, preventing me from playing !!c4. The downside is that it is slow. We wanted to check if Black would fare better with more direct play, and therefore we analysed the following lines: 29 . . . !!b8!? The aim is to attack the b4-pawn immediately. 30.!!c l 30.!!a l !!b6 3 1 .!!c4 does not work well, on account of 3 1 . . .ie6 followed by . . . ltJ d4, and Black will be threatening a fork on b3 and the pawn on e2. Black also looks very well placed after 30.!!c4 ltJd6 3 1 .!!d4 @e7 32.!!c l (32.e4 ie6 and the rook is trapped) 32 . . . !!b6, when I don't like White's position at all. 30 . . .!!b6 3 1 .!!fc4 itJ d6 32.!!4c2!? This leads to an endgame which has to be considered: 32 . . . !!xb4 33.!!xc6

White has to do this. 33 . . . ixc6t 34.!!xc6 @e7 35.!!xa6

a

b

c

d

e

f

g

h

Is this endgame winning or not? It's hard to determine without spending many hours on analysis. I believe it offers Black good chances. He will play 3 5 . . . !!d4, move the king so it defends the g6-pawn, and then regroup to a perfect set-up. But it would take a while to break through, so it would be a long game. In all of these lines White's position is obviously bad, but we do not have an answer about whether the position is winning for Black with best play.

30Jkl ©e7 3 1 .f3 ©d6 After the game Grischuk suggested that it would have been more accurate to play: 3 1 . . .@d7

Chapter 8 - Dynamic Defence 32.g4! White has no time to waste. If he does not react immediately, Black will regroup and his slight material advantage will count. In short: if White waits, he loses. 32 . . . fxg4 33.fxg4 hxg4 34.E:c5! a5!? This is what the computer gives, but I don't think anyone would play it. But I am analysing it all the same, as White holds in the following variation: 34 . . . E:h8 3 5 .