Текст
                    AMDI(j
\.
Vishy Anand
y =est ames
of Chess
Learn by studying the games and comments of the
hottest chess player of the 1990s1


First published in the UK by Gambit Publications Ltd 1998 Copyright @ Vishy Anand and John Nunn 1998 The right of Vishy Anand and John Nunn to be identified as the authors of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. This book is sold subject to the condition that it shall not, by way of trade or otherwise, be lent, re-sold, hired out or otherwise circu- lated in any form of binding or cover other than that in which it is published and without a similar condition including this condition being imposed on the subsequent purchaser. A copy of the British Library Cataloguing in Publication data is available from the British Library ISBN 1 901983 00 5 DISTRIBUTION: Worldwide (except USA): Biblios Distribution Services, Star Rd, Partridge Green, West Sussex, RH13 8LD, England. USA: BHB International, Inc, 994 Riverview Drive, Totowa, New Jersey 07511, USA. For all other enquiries (including a full list of all Gambit Chess titles) please contact the publishers, Gambit Publications Ltd, 69 Masbro Rd, Kensington, London W14 OLS, England. Fax +44 (0)171 3711477. E-maill00561.3121@compuserve.com. Edited by Graham Burgess Typeset by John Nunn Printed in Great Britain by Redwood Books, Trowbridge, Wilts. 10 9 8 7 6 5 4 3 2 1 Gambit Publications Ltd Managing Director: GM Murray Chandler Chess Director: GM John Nunn Editorial Director: PM Graham Burgess Assistant Editor: GM John Emms German Editor: WPM Petra Nunn 
Contents Introduction 5 1  Anand -  Inkiov, Calcutta 1986 7 2  Anand - K. Ninov, World Junior Championship, Baguio City 1987 12 3  Anand - S. Agdestein, World Junior Championship, Baguio City 1987 16 4  Anand - J. Benjamin, Wijk aan Zee 1989 23 5 M. TaI-  Anand, Youth vs Veterans, Cannes 1989 29 6  Anand - B. Spassky, Youth vs Veterans, Cannes 1989 34 7 M. Kuijf -  Anand, Wijk aan Zee 1990 40 8 M. Petursson -  Anand, Manila Interzonal 1990 43 9  Anand - I. Morovic Fernandez, Novi Sad Olympiad 1990 49 10 A. Beliavsky -  Anand, Munich 1991 54 11 A. Karpov - V. Anand, Candidates match (6), Brussels 1991 59 12  Anand - G. Kasparov, Tilburg 1991 67 13 G. Kasparov - V. Anand, Reggio Emilia 1991/2 74 14  Anand - E. Bareev, Dortmund 1992 81 15  Anand - R. Hfibner, Dortmund 1992 87 16  Anand - I. Sokolov, SWIFT rapid, Brussels 1992 90 17  Ivanchuk -  Anand, Match (1), Linares 1992 92 18  Anand - G. Kamsky, Alekhine Memorial, Moscow 1992 98 19  Anand -  Ivanchuk, Linares 1993 102 20 B. Gelfand - V. Anand, Linares 1993 106 21  Anand - E. Bareev, Linares 1993 110 22  Anand - F. Izeta, Madrid 1993 115 23  Anand - L. Ftacnik, Biel Interzonal 1993 119 24 L. 011- V. Anand, Biel Interzonal 1993 126 25 M. Adams -  Anand, European Clubs Cup Final, Hilversum 1993 130 26  Anand - A. Beliavsky, PCA Qualifier, Groningen 1993 136 27 J. Benjamin - V. Anand, PCA Qualifier, Groningen 1993 142 28  Anand - G. Kamsky, Linares 1994 148 
4 VISHY ANAND: My BEST GAMES OF CHESS 29 V. Anand - J. Polgar, Linares 1994 153 30 V. Anand - G. Kamsky, PCA Candidates (3), Las Palmas 1995 157 31 V. Anand - G. Kamsky, PCA Candidates (9), Las Palmas 1995 164 32 V. Anand - G. Kamsky, PCA Candidates ( 11 ), Las Palmas 1995 171 33 V. Anand - J. Timman, Tal Memorial, Riga 1995 176 34 V. Anand - G. Kasparov, PCA World Championship (9), New York 1995 184 35 V. Anand - B. Gelfand, Wijk aan Zee 1996 189 36 V. Anand - J. Polgar, Amber Rapid, Monte Carlo 1996 198 37  Anand -  Topalov, Dortmund 1996 202 38 V. Anand -  Ivanchuk, Las Palmas 1996 207 39 V. Anand - A. Karpov, Las Palmas 1996 211 40 V. Anand - J. Lautier, Biel1997 216 Combinations 220 Solutions 230 Index of Opponents 239 Index of Openings 240 Symbols 240 
Introduction I don't want to spend too much time on biographical details, because this is a book about my games, so I will be content with a brief sketch. I was born on 11 th December 1969 in Madras, and learnt chess at the age of six from my mother. A year later I joined the Tal Chess club in Madras. A couple of years later I went to Manila when my father had an assignment there and got caught up in the Philippines chess fever resulting from the 1978 Karpov-Korchnoi World Championship match. Returning to India, my fasci- nation for chess took off, but it was not until 1983 that I achieved a real break- through. In that year I won both the national Sub-Junior (under-16) and Junior (under-19) titles, and qualified for the (adult) national championship. The championship itself was held the following year, and I finished fourth. I had a real success in 1984, winning the Asian Junior Championship, which was held in Coimbatore, a victory which I repeated in 1985. It was also in 1985 that I gained my International Master title; at the time I was the young- est Asian ever to achieve this distinction. In 1986 I won the National Champi- 0nship and became India's youngest champion. The following year, 1987, was special. After having played three times be- fore in the World Junior Championship, finishing 10th, 5th and 7th, I finally won this' title. In December of the same year I became the world's youngest grandmaster (at that time) by making consecutive norms in New Delhi and Coimbatore (FIDE actually awarded the title in April 1988). The next couple of years featured steady progress, but nothing really spec- tacular. Then, in July 1990, I broke through the 2600 barrier. In the same month, I qualified for the Candidates from the Manila Interzonal. The first round of the Candidates was held in January 1991, and I won my match against Dreev, but in the quarter-finals I lost narrowly to Karpov in a match that hinged on the final game. After this disappointment, I concentrated on tournament play with some success: I won outright at Reggio Emilia 1991/2 (ahead of Karpov and Kasparov) and in Moscow (November 1992). These successes pushed my rating up to 2700, and I became only the eighth person ever to reach this level. The following year, 1993, was important because, after the formation of the PCA by Kasparov and Short, there were now two Candidates cycles. The 
6 VISHY ANAND: My BEST GAMES OF CHESS qualifying events for these were to be held in Biel during July (for the FIDE cycle) and in Groningen during December (for the PCA cycle). I was perhaps somewhat fortunate to qualify from Biel, but at Groningen I reached the PCA Candidates in more convincing style, finishing joint first in the quali- fier. In 1994 I was successful in the PCA cycle, defeating first Romanishin and then Adams to reach the Final of the Candidates. In the FIDE cycle my for- tunes were mixed; in the first round I beat Yusupov (who then became my second!), but then I lost to Kamsky after having been two games up with three to play. Round about this time I was also quite successful in quickplay events, winning both the Amber tournament in Monaco and the PCA Grand Prix in Moscow (ahead of Kasparov). In early 1995 I faced Kamsky in the Final of the PCA Candidates, and I managed to reverse the result of the previous FIDE match and gained the right to face Kasparov for the PCA World Championship in New York. As everybody knows, I started well but faded in the second half. However, losing this match was not the end of the world and in 1996 I was back on the tourna- ment trail, finishing joint first with Kramnik at Dortmund and beating Kas- parov in the Final of the Geneva Quickplay However, the most important event of 1996 had nothing to do with chess: on June 27th I married Aruna in Madras. So far, 1997 has been a very successful year. In April I finished joint first with Kramnik in the Category 19 tournament at Dos Hermanas and in May I became the first person to win the Amber tournament in Monaco twice. There followed a win in the Frankfurt Rapid, in which I defeated Karpov in the Final. The summer months saw a second place in Dortmund and outright first in Biel; in this latter event I received the prize for the most beautiful game (see Game 40). At the moment I am working hard preparing for a heavy schedule of tournaments in the first half of 1998, so I will sign off now and leave readers to enjoy the games. Vishy Anand Collado, Spain October 1997 
Game 1 V. Anand - V. Inkiov Calcutta 1986 Sicilian, Richter-Rauzer 1 e4 2 lbf3 3 d4 4 lbxd4 5 lbc3 6 i.g5 7 'iWd2 8 0-0-0 9 lbb3 In 1986 this move was in vogue thanks to the efforts of Mikhail Tal. 9 ... as 10 a4 d5 11 .i.b5 A move first played in Tal-Sis- niega, Taxco Interzonal 1985, even though Tal ascribes the move's in- vention to Vitolin. Tal won that game and scored an even greater success when he beat Korchnoi with it at the 1985 Montpellier Candidates Tour- nament. I figured that I could do worse than to follow in Tal's foot- steps. 11 ... ltJb4 Sisniega played 11.. .lbxe4 and Korchnoi 11...dxe4. 11....i.b4 is an- other possibility; after 12 exd5 exd5 13 'ir'f4, followed by exchanges on c3 and f6, both sides will end up with fractured pawn structures, but c5 d6 cxd4 lbf6 lbc6 e6 .i.e7 0-0 it will be difficult for Black to get at White's c-pawns, whereas White will attack the d5-pawn. The move 11.. .lbb4 had been played before, in Vitolin-Inkiov, Jurmala 1985. 12 :he1 (D) I."'. ... ... .. B_ _ _ _ . ..- . i..'.  _ d d _ - .. . . . . _e.z.J _ a   D  u_ ,u . : - -  - . 12 ... dxe4 After 12... 'iWc7 13 e5 lbe8 14 lbd4! Black is in a bad way. Due to the weakness of the e6-pawn, Black can't play ...f6 and his e8-knight is badly placed. 13 'iWxd8 Really the first new move of the game. Vitolin played 13lbxe4lbxe4 14 'iWxd8 .i.xg5+ 15 'iWxg5lbxg5 16 h4, regaining the piece with an un- clear position. 
8 VISHY ANAND: My BEST GAMES OF CHESS 13 ... .:xd8 After 13...i.xd8 14 llJxe4 i.e7 (14...ltJxe4 15 .i.xd8ltJxf2 16 :d2 is very promising for White; Black has no reasonable defence to the threats of 17 i.xa5 and 17 i.e7) 15 ltJxf6+ .i.xf6 (15...gxf6 16 i.h6 is similar to the game) 16 i.xf6 gxf6 17 :d6 Black can play neither ...b6 nor ...e5, and so has serious problems devel- oping his queenside. 14 ltJxe4 ltJbd5 (D) XB.t._ .._ .. ... W  _ _ Ymu , ?ff$f . . //.//// :r: / /'l ;/} ' ;-;'/1'0 w% ; /./" ;;/..% .i..."  u,u // $ ;// 8 . .tlJ  i%. . F 0 t.t.J _ ;%// %/, "8. 8"        %':f :    /  %   ;0:>   , " '/ ,'/,.: 15 c4 A positionally ugly move, but the main thing is to exploit White's lead in development. To this end, White must keep the d-file open. 15 ... ltJc7 15.. .ltJb4 occupies the 'hole' cre- ated by White's previous move, but then 16 :xd8+ i.xd8 17 :dl .i.e7 18 ltJd6 e5 (18...b6 19 i.e3! also nets a pawn) 19ltJxc8 :Xc8 20 ltJxa5 wins a pawn. 16 .:xd8+ .i.xd8 17 :d1 Everything with tempo. 17 ... i.e7 (D) . /j: .t. ti'% {:;;/d% . 0i; 01;::,/ Z.-: · /';/; '; ;;;. %%"i %;; ;;? { W .. '...   ,," Y, ,,, //:/,' ////:/:;,:/ . . "// ;:-/: , // ,  $}?''/, // /.j 'l. ,', '/0; . _  f . '  if}&  %  ,  ,  , .   ... ;;( :'%ifif 0  Y""  -;//// ;.:: all '" 8R8.tlJ. . .. 0:$  Bt.t.JB %_ O  ' ;;;"lf/  /' ':// A  'iJ'::j   o / / /'  , .  ".  W$Z /;   :j;;;/,//,;:;/-;;; 18 ltJxf6+ ?! White retains the advantage after this move, but Black could have put up more resistance. It was probably stronger to play 18 ltJd6!. I would not claim that thi s is a forced win (Larsen once observed that all long variations are wrong!), but Black is definitely in real trouble: 1) 18...b6 19 i.c6! (not 19 i.e3 because now Black can reasonably play 19...ltJxb5! with a satisfactory position after either 20 cxb5 ltJd5! or 20 axb5 transposing to line 2a be- low) 19...:b8 20 .i.f4! and White's pieces occupy dominating positions. 2) 18.. .ltJx b5 and now there are two possible lines: 2a) 19 axb5 b6 (best; 19...a4 20 ltJa5! and 19.. .h6 20 i.xf6 .i.xf6 21 ltJc5 are more promising for White) 20 .i.e3 a4 21 ltJa 1 ltJd5! 22 cxd5 
ANAND -INKIOV, CALCU1TA 1986 i.xd6 23 dxe6 i.e7 24 exf7+ cj;xf7 with fair compensation for the pawn. 2b) 19 cxb5! (this concedes the d5-square, but Black can't make full use of it) 19...b6 20 ltJc4 (20 i.e3 ltJd5! as in line 1) 20.. .J:b8 21 ltJd4 (intending ltJc6) 21...i.b7 (21...i.d7 22 ltJe5 is also good for White) 22 ltJxb6 i.xg2 23 ltJc4! (although Black has the two bishops, the mass of pawns on the queenside is the most important factor in the posi- tion; 23 ltJd7 is less accurate since after 23...J:c8+ 24 cj;bl i.e4+ things are getting quite messy) 23...J:c8 24 b3 (White only needs to play cj;b2 and ltJxa5 to decide the game with his queenside passed pawns; how- ever, Black can try to win a pawn) 24.. .i.d5 25 cj;b2! i.xc4 26 J:c 1 ! (when I checked this position with Fritz, it said that Black was winning! Surprised, I looked to see why) 26...J:c5 (this is the reason, but after some thought I found a solution) 27 ':xc4 ':xg5 28 J:c8+ i.f8 29 b6! (Black is helpless) 29...J:c5 (after 29...ltJd7 30 b7 ':d5 31ltJc6 Black has no defence to both 32ltJe7+ and 32 ':d8) 30 b7 ltJd7 31 ltJb5! and Black cannot meet the threat of 32 J:d8. 18 ... gxf6 19 i.e3 ltJxb5 20 axb5 f5! Black fights back. This gives his bishop some air and prevents White forcing the exchange of bishops by 9 i.c5. 20...e5 is worse, as after 21 i.c5 i.xc5 22 ltJxc5 i.g4 23 J:d5 J:c8 24 b3 Black's queenside pawns are in trouble. 21 ltJc5 (D) The alternative 21 f4 ambitiously attempts to squash Black, but he can free himself by sacrificing a pawn: 21.. .e5! 22 i.c5 (22 fxe5?! i.e6 23 ltJd2 a4 prevents White supporting his c4-pawn by b3, and then the c4- pawn itself is attacked by ...i.b4 and ...:c8) 22...i.f6 (not 22...i.xc5?, when 23 ltJxc5 gives White exactly what he wants) and now 23 fxe5 fails to 23...i.g5+ 24 cj;c2 i.e6 and Black is doing well as White cannot hang on to all the pawns (for instance 25 cj;c3 J:c8, foIl owed by some combi- nation of ...a4 and ...b6). ...i.. M.' ... ..& B   _ ;; A ;0' /'%/ & %' /{ £i A : _ l\  / .. '! &  .  _ 0 ,;f A /' //u,,!I; / .//;;:: W-X:;' )%-;/c. //,/:'/ //;;:;':/. //.;/ 1 l\ i. ;ff/ .;: »}; 0 ;J!/Z; ./;: ;tr r l' ? /2 // // '/; .  /'i, :/j;',<// /// ,   &ffiK  l\ i Q ./:;; } 0 :  //?'/f; .. / .: 0 . / . '/. . % .  . ': ;'/.:-; . //J/// .%;,;   i 21 ... e5? The best defence was 21...f4! 22 i.d4, and now: 1) 22.. .e5 23 i.xe5! i.xc5 24 J:d8+ i.f8 25 i.d6 and White wins. 
10 VISHY ANAND: My BEST GAMES OF CHESS 2) 22...a4 (by not allowing ltJa4, Black gets some counterplay) and now: 2a) 23ltJd3 i.d7 (23...f6? is very good for White after both 24 i.c5 and 24 lL\xf4 e5 25 lL\d5) 24 lL\xf4 :c8 25 i.c3 i.e8 and Black has suf- ficient counterplay. 2b) 23lL\e4 f5 (23...e5? 24 i.xe5 i.e6 25 c5) 24 i.c5 and White re- tains some advantage. 3) 22...f6 (a solid and sensible defence; Black prepares .. .e5 fol- lowed by ...t7) 23ltJa4! e5 24 i.c5 f7 25 i.xe7 xe7 26 lL\b6 :b8 with a slight advantage for White. Thus 21...f4, while not equaliz- ing, would have restricted White's advantage. The importance of driv- ing the bishop to d4 becomes clear after the text-move. 22 lL\d7! From this excellent square the knight virtually paralyses Black's whole army. The game is already al- most over. 22 ... f4 Too late, as now the bishop need not block the d-file. 23 i.b6 f6 24 i.c7! Since 25 ltJb6 is threatened, Black can no longer delay capturing the knight; the result is that White's rook occupies the seventh rank. 24 ... i.xd7 25 .:xd7 i.c5 26 i.d6 i.xd6 White also wins after 26...i.xf2 27 :xb7 :d8 28 :b8! (not 28 c5? i.xc5 29 i.xc5 :c8) 28...:xb8 29 i.xb8 t7 30 i.c7 and the threat of 31 b6 forces Black to approach with his king, whereupon 31 i.xa5 gives White an overwhelming mass of passed pawns. 27 .:xd6 (D) .B B B.B B B'B B B' .   . ;M_. 8  . -  - - 88-  . ..-- - - . .  . "8P u . u u .  . . -- Black's position is lost. The active rook, combined with White's queen- side pawn majority, guarantees a straightforward win. 27 ... :c8 28 b3 a4 29 b2 axb3 30 xb3 7 Just abandoning the b-pawn, but 30.. .f5 31 :d7 :b8 32 c5 is also hopeless. 31 :d7+ e6 32 .:xb7 e4 33 :&7 e3 34 fxe3 fxe3 35 c3 :d8 
ANAND - INKIOV, CALCUTTA 1986 11 36 .:a2 37 b6! 38 b7 39 b4 e5 e4 :d3+ :d2 40 b8'ir' 41 'iib7+ 42 'ir'd5+ 43 c5 .:xa2 d3 e2 1-0 The 1987 World Junior Championship was a breakthrough for me. It was already my fourth World Junior Championship; I had played in every one since 1984, finishing 10th in my first appearance, 5th in 1985 and 7th in 1986, a fairly zigzagging pattern. I started the 1987 event in fairly good shape but I drew 3 out of my first 4 games, and at that point I didn't imagine that I might win the tournament. However, the following game was a turning point. 
Game 2 V. Anand - K. Ninov World Junior Championship, Baguio City 1987 Sicilian, Kan 1 e4 c5 2 ltJf3 e6 3 d4 cxd4 4 ltJxd4 a6 5 .i.d3 .i.c5 6 ltJb3 .i.a 7 7 ltJc3 ltJc6 8 'iVe2 d6 9 .i.e3 .i.xe3 As a result of this game, players with Black started delaying this ex- change so as to leave the white queen on e2 and thereby prevent the .i.e2 line mentioned in the note to Black's 13th move. In time, this led to the development of a whole new branch of opening theory. 10 'iVxe3 ltJf6 (D) . %: . }.:; J.  // . ' . %<j: . f //>/ 'I - /,ij2" / 'l; 0/;;% :%  j)'l;;i;; /J / / / /  %:;::f .& JJJ%f/: /'/;f!t.&  .& W .. c &:; ..  .. I....I. W;/0: //////  /:/:/-;'; '«:: %: . // . " /, / / //////  x/))/;; ///;'l 9jj;J  ;A //2&;/ / . /: '/ffi / . ;; . ''/// . .  Pjj!f}; t2J ; rLj ;// 'c;  M1 // // /// / / + / / / /./ %/  A  A O{Q;O /i  / / 'l //; 1'; : // -;'  / / . ' iJ /''>  A 1j 'A / / 'O //  "/ / / / v .  .. d 'j/. //;/.::: 1: / ///.// '/' >", / / '/.  11 g4! I found this innovation over the board. Previously White had contin- ued 11 0-0-0, but after 11...0-0 White either has to prepare this advance with :hg 1, or play g4 as a pawn sac- rifice. The idea behind the immedi- ate g4 is quite simple: if you play it when g7 is undefended, then it isn't a sacrifice. 11 ... b5 Black thought for a while, and then decided simply to proceed with his queenside counterplay. However, the tempo White saves by missing out :hgl is quite important, and he should have tried 11. ..ltJxg4. After 12'iVg3ltJf6 13'iVxg7 :g8 14'iVh6 .i.d7 Black will play ...'iVe7 and both sides will castle queenside. How- ever, when Black castles queenside in the Sicilian, White is better unless there is some mitigating factor, which is not the case here. Indeed, after 'ir'e3 by White there is an awkward weakness on b6 and the d6-pawn it- self may become vulnerable later. 12 0-0-0 0-0 13 g5 ltJe8 A poor square for the knight, but after 13...ltJd7 14 i.e2 Black has 
ANAND - NINO V, BAGUIO CITY 1987 immediate problems with his d6- pawn, due to the exchange of dark- squared bishops and White's extra tempo. Usually White doesn't go af- ter d6 in a Sicilian, but that doesn't mean you should forget about the possibility altogether! After the text-move, there is no point to :hgl, which would throw away the advantage gained as a re- sult of White's innovation, so White has to come up with an alternative attacking plan. One possibility is h4-h5 followed by J:dgl. The arrangement with rooks on gland hI is very desirable, be- cause they will break through almost any kingside defence, but it is very time-consuming to set up. Unlike many similar positions in the Sicil- ian, Black's queenside counterplay is rather slow here, which is the only reason White can consider this plan, but in the end I decided on a more conventional approach. 14 f4 b4 15 ltJe2 15 ltJa4 would be a more posi- tional formula. White stops .. .a5-a4 and threatens to invade on b6. After 15...J:b8 White can continue 16 e5, in order to clear the c5-square for the knight. However, I preferred to play for the attack, and for that the knight is needed on the kingside. 15 ... as 16 ltJbd4 ltJxd4 17 ltJxd4 (D) 13 White has the advantage. Nor- mally in the Sicilian, Black's the- matic queenside pawn advance gains time because it hits minor pieces on b3 and c3, but here White has evacu- ated these squares quite quickly. Moreover, the knight on e8 is very badly placed for supporting the at- tack. Just about the only useful thing it can do is to shore up the kingside by ...g6 and ...ltJg7. . r?# ...  1&\ . . p} ... 0:;;;,//.JL i" /j /. ,/   ' // /. ' /. :rft!: ' I. " /.  I / 1. // / B m; './%0    x /;/ , . /. . / . /' . ;:' . '. .  / I. w//, /-/:/ . ;: :;   !t;' ;; ///'l/ ' /., ,0- //:;://;;; -(!; @) [ffj % _ :fi!:: /i0if   /"U/ ;;//% , ;%;0/ /.u y; . % ' fiJ  {fJ //:J / /:; . ;:,//,;   /. " ,;;:, /uu, /." f:,  ::/f:;; % , :# , ;/ ' i.  /. '//.:;::/: , ;y :/S' +;, / j/:«/; / %;; //  /- ;f/:! /  ';'" ;.. A t;'  ,j' :ff L) , L) <' /Q/ : : 17 ... 'iVb6 At the time I felt that Black should have gone in for 17...a4, but now I don't think so. The line that worried me was 18ltJc6'iVc7 19ltJxb4 a3 (if Black doesn't play this, then White plays a3 himself, followed by <it>b 1 and c3, and Black will never break through) 20 b3 (threatening to cen- tralize with 'iVd4) 20...'iVc3 (after 20...J:b8 21 ltJa6 xa6 22 xa6, followed by c4, White's queen- side position is solid since the poorly placed knight on e8 cannot displace 
14 VISHY ANAND: My BEST GAMES OF CHESS White's bishop), but in fact 21 'ir'el! 'iiD2+ 22 d2 is clearly better for White - Black has surprisingly little compensation for the pawn. These were still the pre-computer days, when players were much more intuitive. Nowadays everybody goes home and checks everything with Fritz. The use of computers has made people more sceptical and now they are more prone to go pawn-grabbing unless there is definite compensa- tion. 18 e5 .1b7 19 :hO (D) .I. .a_.. B."'. .._. mz .. . - -   / .   %  U U r  D . fi  , ,!/;. . Pill1i . - ..... . . . " u _ , u . :.:p -  - - 19 ... dxe5 Or 19...a4 (19...:d8 20 f5! is simi- lar) 20 f5! dxe5 (20...exf5 21 e6 is very good for White) 21 fxe6! (not 21 'ir'xe5 :a5!) 21...exd4 22 'ir'h3 g6 (22...f5 23 :xf5! gives White a win- ning attack) 23 exf7+ :xn (23...h8 24 'ir'h6lbg7 25 :f6 wins) 24 :xf7 xf7 25 'ir'xh7+ winning the black queen. 20 fxe5 :d8? (D) Allowing a pretty finish. 20...g6! was best, when White could con- tinue 21lbf5 'ir'xe3+ 22lbxe3 with a pleasant endgame - his knight could head for either f6 or d6. Still, Black has some chances by playing his knight to f5, and he should certainly have gone for this. . _a_.. ....R ../. w _  _ ,!Ii Ii ... _ '/ · D 8   @ %  % %!/; w !/;    . . _   d . .i._ . 88. .  U .  U . :.:. _ w _ _ 21 .1xh7+! xh7 22 g6+ Once again Black suffers because of his miserable knight position. Here it prevents Black from playing 22...fxg6. 22 ... g8 Or 22...xg6 23 'ir'd3+ (stopping ...h7) 23...f5 (23...h6 24 'ir'h3+ g5 25 :gl+ f4 26 :de1 with mate next move) 24 exf6+ with a de- cisi ve attack. 23 'ir'h3 lbf6 23...fxg6 loses to 24 :xf8+ xf8 25lbxe6+. 24 exf6 (D) 
ANAND - NINO V, BAGUIO CITY 1987 . . ... .. .. B."'. _ _ illU .&I\. _ ..UO / N.    . . . '/  J .    W'  - - - -- 8 R 8. . P u _ _ U . :.:. . * . . 15 It isn't every day you see two at- tacking pawns on f6 and g6! When you get a position like this, you go away feeling very pleased and have a warm glow for the next few games. 24 ... fxg6 25 fxg7 1-0 Since 25...xg7 26 lbxe6+ and 25...:'xfl 26 'ii'h8+ f7 27 :'xfl+ are decisive. After this game, which was from round 5, there was a free day and the players went on an excursion. I still hadn't broken free from the pack. The following day, however, saw another good result. I was facing Agdestein, who at 2565 was the highest-rated player in the tournament, although he was not the only grandmaster (Ivan Sokolov was also participating, although he was not very successful). I prepared as well as I could because Agdestein is very unpredictable in the openings, and sat down hoping to playa good garnet 
Game 3 V. Anand - S. Agdestein World Junior Championship, Baguio City 1987 Ruy Lopez 1 e4 llJc6 A provocative move. At the time it was felt that this move had little in- dependent significance because so long as White knew he should play 2 ltJf3, Black had nothing better than 2...e5, thereby transposing into stan- dard king's pawn openings. Subse- quently it was discovered that Black could wind White up even more by playing 2...d6, but I don't trust this for Black! 2 ltJf3 e5 3 .i.b5 a6 4 .i.a4 b5 5 .i.b3 ltJa5 Agdestein is fond of offbeat sys- tems. I didn't know much about this one, although it is popular amongst Norwegian players. I could only re- member a game between Spassky and Taimanov (in fact from the 1955 USSR Championship, held in Mos- cow). Nevertheless, I was quite happy to see it on the board. White can play natural moves and there is not much risk even if he commits a slight inac- curacy - a pleasant situation when facing the top seed! 6 0-0 7 d4 d6 ltJxb3 8 axb3 f6 9 llJc3 .i.b7 I recalled that Spassky had played 10 ltJh4 in the above-mentioned game, with dxe5, iff3 and J:[dl fol- lowing in some order, and that later Spassky sacrificed a piece by means of llJxb5. 10 ltJh4 ltJe7 (D) K   4 Wff0 /, W .t   /-/ & j% . /  . % /% .. //  j};;, /  / // / . & /Y1  # . (/:: . /y/  ,    }"  ( .. y; , ' . / fij /) %J  ,/////-' :;://:/,-;/  'l/',//:';,/ / / // , / / // ; :' . /;:/' /   'l " ;}/::' ;.J '/ ;:0:;'% ;, , -'/////:/;,   G //, // '/  /: if  :;, +'/ : /'  ;.//;;/ ;,i''l'.,; '//////' fIj / ' A ftj ;;  ; .  % 0  / '/// / 0 ,, Wi 11 dxe5! dxe5 HBlack plays 11...fxe5, then White can strongly reply 12 f4, opening the position up while Black's king is still stuck in the centre. However, taking back with the d-pawn retains control of g5, so that f4 can be met by ...exf4 followed by the fork ...g5. 12 iff3 
ANAND - AGDESTEIN, BAGUIO CITY 1987 White avoids the exchange of queens as most of his chances lie in exploiting Black's poor develop- ment and centralized king. 12 ... 'iVd7 13 J:tdl 'iVe6 By now I had worked out that this was indeed what had happened in the Spassky game, and that he now continued with 14 .i.e3. For a mo- ment I wondered what Taimanov had played that allowed the sacrifice on b5, then I realized that it was 14...g5. After 15 ltJxb5! axb5 16 'iVh5+ 'iVf7 (16...ltJg6 17 ltJxg6 'iVf7 18 J:xa8+ .i.xa8 19 'if g4! also wins) 17 J:xa8+ .i.xa8 18 J:td8+ <it>xd8 19 'ifxf7 gxh4 20 'iVxf6 White had a winning posi- tion. However, I was worried by the idea of 14...h5!?, intending ...'iVg4. I mulled over this for some time, but couldn't see an easy answer. If White plays 15 h3, then 15...g5 is now possible because White has no queen check on h5. If 15 ltJd5, then Black just castles queenside. It ap- peared to me that if White was going to play ltJd5, then it would be better to do it straight away, before Black had time to set up the threat of ...'iVg4. 14 ltJd5 ltJxd5 If 14...0-0-0, then 15 c4 and White is already starting to make Black's king feel insecure. 15 exd5 'iVf7 16 c4?! (D) 17 A slight inaccuracy which gives Black the chance to sideline the knight on h4. 16 ltJf5 would have been more accurate, because White can play c4 at any time - Black can't prevent it. Then 16.. .g6 17 ltJh6 .i.xh6 18 .i.xh6 0-0-0 19 c4 J:[d7 would have led to a position in which White has an edge, since Black has no really constructive plan. . . % .. /  t _ 00/ B."'. .1fV.. .. . .. %3 .. :;;::}J A  . / . :r@i; / w . ;{'0: 0   ';j; / -;, / ; / / / / / /:/",;;:;/h,; /;i:/"'/// /// // /////, A ///ffi// /;/:/,  . 0 ;i8: ;0I/ ;  .  'i.I;i  ?iJ fff%    'Jf// )    .   / '/; ; % % /,,/ / / '/////// /, 16 ... .i.e7? Black misses the opportunity he has been given. 16...g6! would have left the h4-knight misplaced. It is true that after 17'iVe2 .i.g7 18 f40-0 19 f5, followed by 'iVe4, White has a good centralized position and Black's bishops are hemmed in. However, after 18...0-0-0!? Black would have chances of putting his two bishops to work. The b7-bishop functions as a kind of 'Dragon' bishop, in that it makes it hard for White's attack down the a-file to strike home prop- erly. Moreover, playing cxb5 will 
18 VISHY ANAND: My BEST GAMES OF CHESS leave d5 hanging. White is certainly not worse, but Black would have far more counterplay than he obtains in the game. 17 lbf5 (D) .. ... - . \Wti B d"'. .-. iB . . B BiB8.tDB 88. B _ .8. .... p . P8      :.      The exchange of inaccuracies has favoured White, because Black's in- accuracy is actually quite serious - his bishop doesn't belong on e7 at all. Now White is clearly better. 17 ... :d8 17...0-0 is impossible because of 18 .i.h6. 18 .i.e3 g6 (D) More or less forced, because oth- erwise Black lacks a constructive move. 18...0-0? still fails to 19 .i.h6 while after 18....i.c8 White has a choice of promising lines: 1) 19 :ac 1 0-0 20 lbxe7+ (not 20 cxb5 .i.xf5 21 'it'xf5 axb5 22 :xc7 :xd5! and Black escapes) 20...'it'xe7 21 cxb5 :fe8 22 b6 cxb6 23 .i.xb6 :d7 24 'it'd3! .i.b7 25 d6 gives White a clear advantage. . 2) 19 c5! (even more forcing) 19.. ..txf5 (19.. .0-0 20 d6 wins a piece) 20 'it'xf5 :xd5 21 'it'c8+ and after 21.. ..i.d8 22 'it'xa6 or 21.. .:d8 22 :xd8+. .i.xd8 23 :dl 'it'e7 24 'it'xa6 White wins a pawn while re- taining a positional advantage. . -.. - 8.t/ .8i wd y;._ i8 . i8 d _ . _ .i.8 _ _ .e.zJ_ .8. _ . .88 _...  . 8R  d    a:.      19 lbh6! White could exchange on e7, but then Black's king could castle or move to f7 and he would have fair de- fensive chances. White's queenside majority is an asset, but it will not win the game by itself because the c-pawn is tied to the defence of d5. Instead, I wanted to keep Black's king on e8, while I broke through on the c- and d-files. 19 ... 'it' g7 Intending ...f5, when the knight might be in trouble. 20 'it'g3?! This move, introducing lbf5 ideas, is not so strong as I imagined during the game. Sometimes, when you 
ANAND - AGDESTEIN, BAGUIO CITY 1987 have a good position, the temptation is just to play easy, comfortable moves and wait for the position to win it- self. The correct move was 20 cxb5!, when 20...axb5 fails to 21 :a7. Dur- ing the game I rejected it because of 20...f5 21 bxa6 .1a8, with the threat of ...f4. However, with ten years' hindsight I don.'t see Black's com- pensation for the two pawns. In par- ticular, once White has played a7 the bishop on a8 will effectively be dead, since taking on d5 will always allow a combination involving the promo- tion of the a-pawn. One line is 22 'ii'h3! f4 23 .1d2 .1g5 24 g4 h5 25 xe5 and wins. It is worth noting that 20 h4 was playable. After 20....1c8 we trans- pose into the game, but White has saved a couple of moves. 20 ... .1c8 Threatening ...g5, so White's next move is forced. 21 h4 .1d6 (D) If Black attempts to play actively by 21...f5 22 .1g5! f4 then: 1) 23 'ii'c3 b4 24 'ii'd2 (24 'iff3 .1xg5 25 hxg5 'ife7 26 g4 'ii'xg5 27llJxe5 O-O! is unclear) 24....1xg5 25 hxg5 'ii'e7 26 :el f8! and Black has counterplay. 2) 23 'ii'h2!. A paradoxical move which nails down Black's kingside. Now White simply threatens :el followed by doubling or even g3 (meeting ...f3 by g4). However Black continues, his e5-pawn will come 19 under fire and his king is trapped in the centre. The point of'ifh2 is that after 23....1xg5 24 hxg5 'ife7 White can defend the g5-pawn and main- tain his kingside bind. -...-.- - _.mu w. _ . .. ..  . . - -  B'.8 B - - - _8_ _ D .8.   . . - -  . 8. u dud  B:B  - - - * 22 'iff3 Other moves are inferior, for ex- ample 22 c5? .1xc5 or 22 :acl e4 23 .1f4 'ifxh6! 24 .1xh6 .1xg3 25 fxg3. In this latter line White keeps a slight edge as the opposite-coloured bishops (which normally have a drawish influence) make Black's de- fence a bit more difficult, when there are still a fair number of major pieces on the board. However, I thought that keeping the queens on offered even more. The move 'iff3 is mainly to pre- empt the threat of ...e4. Now that White has got the knight firmly en- trenched on h6, he can proceed with :acl followed by c5 and d6. The manoeuvre 'iff3-g3-f3 may appear odd, but Black has also wasted time 
20 VISHY ANAND: My BEST GAMES OF CHESS with the manoeuvres ....i.b7-c8 and .. ..i.f8-e7 -d6. 22 ... .i.e7 Black would like to play .. .J:f8, but the immediate 22...:t"8 allows 23 c5, because after 23....i.xc5 24 .i.xc5 the rook on f8 is hanging. So Black returns to e7, still preventing c5 and getting ready to play ...J:f8 next move. The net effect of both players' oscillations has been to add the moves h4 and ....i.c8, which benefits White. At this stage I felt that my posi- tion was very comfortable, but I had (and still have) a great deal of re- spect for Agdestein and I wasn't counting on victory yet. 23 :acl Retaining the option of either cxb5 or c5 followed by d6. 23 ... bxc4 Agdestein finally decides to re- move the cxb5 option. 24 bxc4 :f8 25 c5 f5 26 .i.g5 .i.xg5 27 hxg5 'ilie7 28 'ilig3! (D) White had a more complex alter- native in 28 d6 'ilixg5 29 'ilic6+ .i.d7 (29...J:d7 30 _a8! J:d8 31 J:el should win for White) 30 'ilixc7 'ilixh6 31 c6. This would also have been quite promising, as Black would have to return the piece for one pawn, still leaving White with a an- gerous passed pawn. However, given that an effective, solid alternative ex- isted, I preferred to play safe. Mi.... . .   81. BB . _ _ &. . .& a. _ pa . _'D  . . . ; 0/,j: /W// M$; . [ .  . :r . /' . .& . * ID1t W ;: J} _ ' . . //  fff% i _ .  ".  wgj; ? u.... rf,}   In a way it is quite strange that White is not winning already, since Black's king is irrevocably trapped in the centre - not only can it not castle, but even f7 is denied to it. However, the reason is the knight on h6. It is of course doing a wonderful job, but when it comes to landing a killing blow in the centre, White is effectively a piece down. 28 ... f4 29 'ilih4 e4 30 d6 'ilie5! The best chance. After 30...cxd6 31 cxd6 'ilie6 (31...'ilie5? 32 J:xc8!) 32 J:c7 e3 (32....i.d7 33ltJg4) White can continue 33 fxe3 fxe3 (after 33...'ilixe3+ 34 <it>hl there is no de- fence against 35 J:e 1) 34 'ilic4! J:f2 35 'ilixe6+ .i.xe6 36 J:e7+ <it>f8 37 J:xe6 with an extra piece. 31 dxc7 .:xdl+ 32 1hdl (D) 
ANAND - AGDESTEIN, BAGUIO CITY 1987 ..t.... . . .. /';-/1/1:: ::;;::; .  - !%    )f /:  /' . //' //  '0: //  ;, y.  /",    ,/, //,  . . &  I /  : '/:  (n:l A  y. :;; y. '/ ;/m: './:0/:0 ;/// / Y,;' /   f;  )ff/f .%  :    %t{/ f4 (tJ /% ', A    _ O_ / u Y,;' ,///;;/j /, / Y,;'  r ..  . h_ , . " B _  .. . , 32 ... e3? Black decides to counterattack, but now White's win is fairly straight- forward. Black's only chance was 32...'iixc7, when we can see that it is not so easy for White because his knight is on h6. If it were on any nor- mal square then, for example, ltJc3 or ltJc4, heading for d5 or d6, would be decisive. Nevertheless, after 33 lbg4 (heading for f6) 33...xg4 34 'iixg4 J:f5 35 b4! (White needs to pause for this move; it strengthens c5 and indirectly supports J:d6; after 21 35 'iie2.:xc5 36 'iixe4+  White's attack has got a bit stuck) White has a large advantage. He intends 'iie2 and J:d6, taking aim at e4 and a6, when Black's exposed king causes continuing problems. If 35...e3, then 36 fxe3 fxe3 37 'iie4+ picks up the e3-pawn (37...J:e5 38 'iia8+ q;e7 39 'iih8 is even worse). 33 J:d8+ q;e7 34 ltJg8+ .:xg8 35 .:xg8! e6 Or 35...e2 36 'iixh7+ e6 37 J:e8+ and wins. 36 c8'ii 36 'iixh7+ f7 37 J:e8+ q;xe8 38 c8'ii+ q;e7 39 'iib7+ q;d8 40 'iih8+! is a prettier win, but I preferred the prosaic text. 36 ... exf2+ 37 'iixf2! This game has been published in some magazines with the move 37 xf2, but that allows mate in two!! 37 ... xc8 38 .:xc8 1-0 This win over the top seed put me well on the way to becoming World Jun- ior Champion. My run of wins continued with further victories over Klinger, I vanchuk and Blatny, after which I was in clear first place, a position I held on to until the end of the tournament. Winning the World Junior was my big breakthrough. Normally a player from India would have to waste a lot of time playing in mediocre open tour- naments, gradually improving his Elo rating and hoping to get some invita- tions. However, the two factors of my World Junior title and my GM title, which I gained shortly after this event, enabled me to short-circuit the pro- cess. I got an invitation to a pleasant open tournament in Lugano (but after 
22 VISHY ANAND: My BEST GAMES OF CHESS my play there, I didn't get invited again!). Then I was asked to be a commen- tator at the World Cup event in Brussels (April 1988). In Brussels I was able to meet some tournament organizers and this led to my first top-class invita- tion - to Wijk aan Zee 1989. I had a tough year in 1988, just after gaining my GM title. I lost about 40 rating points in two tournaments at Biel and Blackpool- I still don't under- stand why, although many GMs have told me that they had similar experi- ences after gaining their title. I had a break after Biel and returned to active play near the end of the year at the Thessaloniki Olympiad, making 8 1 /2 points out of 12 games. This score was sufficient to regain 10 of the Elo points I had lost, but later I was surprised to discover that FIDE had not given me any points at Thessaloniki. The reason, I found out, was that the Indian team had arrived late, after the first round, and had therefore lost one match by default. FIDE had counted this as a normal loss, thereby wiping out my ten point gain from the rest of the event! Later on I had my ten points rein- stated. I then scored 6/9 at the GMA Open in Belgrade and subsequently I played at Reggio Emilia, another good invitation resulting from my World Junior success. This tournament started well with wins against Ivanchuk and Sax, but after that I lost some games and finished on 4/9. However, this event was very useful for me; it was the first time that I had played such a strong field. My next event was the long-awaited trip to Wijk aan Zee. This was my first really big event and I was thrilled to be there. I had a zigzag course in the tournament. I won my first two games, just as at Reggio Emilia, then lost to Tseshkovsky (who used to be a nemesis of sorts!). I also lost to Van der Wiel, but then won against Ivan Sokolov. I was still on '+ l' when the following game was played in the penultimate round. 
Game 4 V. Anand - J. Benjamin Wijk aan lee 1989 Sicilian, Richter-Rauzer 1 e4 2 ltJf3 3 d4 4 ltJxd4 5 ltJc3 6 .1g5 7 'iid2 8 0-0-0 9 ltJb3 Two of the first four games in the book reach this position! I have done pretty well with this line so far. 9 ... 'iib6 By the time this game was played it had become clear that 9...a5 wasn't that great a move. 10 f3 c5 d6 cxd4 ltJf6 ltJc6 e6 i..e7 0-0 ':d8 (D) ..J._ ... i .  'ii w_  ,, _ JIL\/ & . /.,   . _..  A , . . . . - - .- . .  .  l\ . _ _ 0  l\ ' l\   l\  O O_ .ou _ 1t:.J..: 11 bl There is an interesting story relat- ing to this move. I used to get a lot of my theoretical information from Ivanchuk. For example, during the 1985 World Junior Championship in Sharjah he had shown me a tremen- dous idea in the Dragon, which turned an existing evaluation upside down; I was amazed (and thankful!) that he was so open and generous about showing his ideas to me. I would try to give him some titbit in return, but unfortunately my novelties were not that good! At the Reggio Emilia event the month before Wijk aan Zee, there hadn't been much to do in the eve- nings, so one day I went to Chucky's room. By this time he had recovered from his first round loss to me and was on a respectable score. He was also feeling bored; we went out for a walk and then returned to his room. I asked him "Why does everybody play the Catalan - it seems such a boring opening." He replied that it was not boring at all and proceeded to show me an interesting idea. The introductory moves were 1 d4ltJf6 2 c4 e6 3ltJf3 d5 4 g3 .1e7 5 .1g2 0-0 6 ltJc3 dxc4 7 ltJe5 c5 8 dxc5 and 
24 VISHY ANAND: My BEST GAMES OF CHESS now Vassily said that 8..:iic7 was supposed to be a good move. He ex- plained that after 9 ltJxc4 'fixc5... (D) a-i.. /.. _ d /. u/ /. _ w.'. M'.' . _'11 _  ill11 . . . . .  BltJ_ . . .  .  / -  - /'/  %' ."    . U /!//  m 'iV  .: / /. / " /. /,  , ...everybody plays 10 'fib 3 , but he had found a very strong idea: 10 b3 ! . The point is that after 10....:d8 White can play 11 i.a3 :xd 1 + 12 ':xd 1 'fic7 13 ltJb5 trapping the queen. I was very impressed by this line, which I had never seen before. He explained that this was the reason why... 'fic7 was not good in this par- ticular line of the Catalan, but was good in the similar lines in which Black plays. ..i.b4 and later retreats the bishop to e7 in response to a3. The reason, of course, is that the pawn on a3 prevents the move i.a3. A nice idea, but apparently not much use to me as I didn't play the Cata- lan. I hope Chucky will forgive me for revealing this piece of analysis! Returning to the Benjamin game, at this time I didn't really study the openings too deeply. I looked at all the theory, but didn't really go be- yond that. After 10....:d8 I suddenly didn't feel very happy with my posi- tion and couldn't find a continuation I felt comfortable with. Then I sud- denly brightened up, because I saw a little trick after 11 bl. 11 ... d5?! (D) If Black plays 11. ..a6, then 12 .te3 'i6c7 13 'fif2 and White gains time owing to the threat of i.b6, so bl is really to provoke ...a6. Later on it was discovered that this doesn't really matter, because Black's coun- terplay consists of .. .a6, .. .ltJd7 and ...b5 in any case, but we didn't know that in 1989! After 11 bl Benjamin looked surprised, because he didn't under- stand the point of the move. Perhaps he was thinking 'Vishy doesn't know that Black's threat is ...d5'. Anyhow, when he played ...d5 I checked my idea carefully, although there isn't much choice because after anything else White is clearly worse. ...t.. ... . /0 &   .  &  & W  .0?;;;  .f_. /n u / ; nu roa JIL\ Wi% & -  ... f&:?/ . _ /w; ;;0') , ' // , > , '/ & /.//" , /" . . {.. ?+ //./ /;:-/' ,./ '/ // , /////--;:;/.'  ///" /' , :/;//' / 'W% ;;':X";;/;m/ //;};: :": //;::/< ;'/>//-; ;//'/)/ x{//;/; >//: /; /-%// ,'/;////' . J":'\ m ;; l\ Jt _  ;, @:i/ 0 :?j;  D '   /. '--; :;//  D ;/    '1!!0   ,   :':/.;;; '/    / <it> w/ . /'1. : f'l . ;/ . ;/;   . / . ' :  ;;; i!JJ:; + 0;;:;';: 4 ;,;:/% ; ;/: 
ANAND - BENJAMIN, WIJK AAN ZEE 1989 12 xf6 dxe4? Benjamin falls for it hook, line and sinker. He could still have bailed out by 12...xf6 13 exd5 xc3 14 'iixc3 exd5, although after 15 'iic5 or 15 .id3 White has some positional advantage. However, as he admitted after the game, he simply hadn't seen the idea at all. 13 xe7 Not 13 d4ltJxd4 14ltJxd4 e5. 13 ... .:xd2 14 ltJxd2! (D) .1."'8 ... ...  .. B '///// . ,,;  illaJl1\.&B . .... A  /£%:% {;;;{j 0  f/   r  ;:f/;  /;: ;/:1;'; /' /"; ;(d ///% %;, '0 ) ,;, // . B'. . B  .. " ?;j ."   u B<&t>.:Bi..: 14 ... exf3 After 14.. .ltJxe 7? White wins by 15 ltJc4 'iic7 16 ltJb5 and Black's queen is lost almost exactly as in Iv- anchuk's Catalan idea. It isn't trapped here as it was in the Catalan, but the threat of mate on d8 means that it amounts to the same thing. At the next tournament where I met I van- chuk, he came up to me and said "I see you used my idea in the Cata- Ian!" . 25 After the game Benjamin said that he wanted to resign at this point, but decided to playa few more moves; however, I then started to play so bad I y that he couldn't bring himself to resign any more. 15 gxf3? Here's the bad move. Later I real- ized that 15ltJc4! would have won on the spot. 15...'iif2 loses to 16 ltJe4, and after 15...'iic7 16 .id6 fxg2 17 xg2 'fid8 18 g3 'fie7 19 J:hel all White's pieces are active and his knights are going on a queenside rampage (ltJb5 or ltJd5, coupled with ltJd6). The fact that White has a lost a pawn is irrelevant. 15 ... eS 16 h4?! (D) Another mistake. 16 .ia3 would have been much better, keeping con- trol of key squares such as c5 and d6. In this case White should still win in the long run, although thanks to White's previous error it is likely to be a laborious process. a.i.. ... ... ... B '///// _ _  illU JI1\ . ;  ....  -   /j ;  . /{' . ' ,  . ; .  . _ 'fi?: _ ;:: /W"-;;:/ " '//// /" ,, ////:.  8 . ::; {$;?  /?;?:; %fi'X ;;0Z :, , ;  m .1\. .  y do. D ''    . :  '  ;;% '%J  '/ ,/ / / //  .'it>.:.i..: 
26 VISHY ANAND: My BEST GAMES OF CHESS 16 ... e6 17 lbde4 White already has to take care: not 17 c4? xc4 18 lL\xc4 'iWb4. 17 ... lL\d4 (D) z. . ... .. ... W_ _ _ _ illi1 - J. - - - . . . . .  . - - - - . llliJ. _ -  .. . . n u   u .<t>.:..: 18 g2?! After this, the position becomes rather murky, although it is possible that White can still retain an advan- tage by accurate play. 18 f2 was a much safer way to prove that White is better. 18 ... :e8 Suddenly I couldn't find a line that worked for White. The threat is ...f5 followed by ....:xc3, and it isn't easy to find a good defence. Finally I found a line that seemed to work for White but I was very nervous be- cause Black has all his pieces aimed at White's queenside. 19 f2 f5! 20 f4! Not 20 lL\g5? :xc3 21 lbxe6 'ifxe6 22 xd4 (22 bxc3? 6+ and Black wins) 22...exd4 23 bxc3 dxc3 24 al 'iWe2! (24...6 25 :bl 'iWf2 26 fl 'ii'xc2 27 c4+ followed by :Xb7 wins for White) 25 :hgl 'iWxc2 26 :bl 'ii'd2 27 a3 and Black is even slightly better. 20 ... fxe4 21 fxe5 (D) _z. ... .. . -. B_ _ _ _ _ .J.. . . . 0 _ . II'. . .  . . .  U  _ u .<t>.:. .: 21 ... :e4 21...:d8! would have made life much harder for White: 1) 22 lbe2? xa2+ 23 xa2 'iWa6+ wins for Black. 2) 22 :hel?! g4! 23 ':d2 'iWh6! 24 .1e3 (24 :xd4 :xd4 25 xd4 'iWd2 favours Black) 24...'iWh4 25 .1f2 with a draw by repetition. 3) 22 :d2! f5 (not 22...lbb3? 23 axb3 and White wins) 23 .1e3! 'iWg6 24 :hdl and, surprisingly, there is little Black can do to prevent fl followed by taking on d4. While 21...:d8 might not have been any better than the text-move against perfect play, 22 :d2! and 23 
ANAND - BENJAMIN, WIJK AAN ZEE 1989 27 i.e3! aren't easy moves to find over the board. 22 ':he1? (D) Missing 22 ltJe2! ':a4 (Black's sacrificial attempts fail, for example 22...'ifxb2+ 23 <li>xb2 ':xc2+ 24 <li>alltJxe2 25 i.el or 22....:xc2 23 i.xd4 'ifc6 24 ltJf4 and White wins in both cases) 23 b3! ':xa2 (if 23...xb3, then simply 24 axb3) 24 i.xd4 and wins. . - ... 'i. . 'i B  . d  II ..i.. . . _ B _ ...i. . .  . . -  - - 88. i." u d _ u ..:a . 22 ... ':b4? Benjamin was in time-trouble by now, and commits another mistake. The best line was 22...i.g4! 23 ':d2 (not 23 ':xd4? J:xd4 24 lbxe4 ':xe4 25 i.xb6 Jhel#) 23...'ifh6! 24ltJxe4 (or 24 ':xd4 ':xd4 25 i.xd4 'ifd2) 24...ltJe6 and Black has avoided los- ing a piece. Despite Black's slight material advantage, I don't think White is worse, since Black's king is fairly exposed and his pieces dis- jointed. 23 <li>c1! It would also have been good to play 23 b3!, which looks a bit para- doxical as Black has so many pieces ready to sacrifice on b3. However, after 23...i.xb3 24 axb3 ':xb3+ 25 <li>c 1 White should win. 23 ... i.g4 If 23...ltJb3+ then 24 axb3 'iWxf2 25 ':d8+ <li>f7 26 ':fl picks up the queen. 24 ltJd5 'ifc5 25 ltJxb4 i.xd1 26 <li>xd1! (D) Not 26 ':xdl? ltJe2+ and Black wins, nor 26 ':xe4 ltJb3+! 27 <li>xd 1 'ifxf2 28 axb3 'ifxg2 29 ':e2 and only Black can be better. After the text-move it suddenly dawned on me that I was completely winning. I had been struggling to contain Black's queenside initiative for so many moves that the realiza- tion caught me by surprise! - . i. . B_ _ _ . . . .   . . _ U .  _&   _.  . . . . 88. i. u  _ U . _a _ -.- i - - . 26 ... e3 26...'ifxb410ses to 27 ':xe4. 
28 VISHY ANAND: My BEST GAMES OF CHESS 27 :xe3 ltJfS 28 .1dS+ An important intermezzo. 28 :f3 is a mistake because of 28..:iixb4 29 :xf5 'if g4+ . 28 ... f8 Forced; 28...<ith8 fails to 29 :c3. 29 :f3 'ifxb4 30 J:xfS+ e8 31 e6 1-0 After 31...'ifxb2 32 c5 it is all over. Before the last round I was in a tie for first place with Nikolic, Ribli and Sax. I had the feeling that a draw in the last round would be enough to main- tain the status quo, and indeed when my game with Douven ended in a draw my feeling turned out to be justified, as the other leaders also drew. Thus we ended up in a four-way tie for first. It was my first success in one of the world's major international events. In 1989, Bessel Kok, the chairman of SWIFT and a great chess benefactor, organized a Youth vs Veterans event in Cannes not long after Wijk aan Zee. It was held in conjunction with the annual Cannes Games Festival and was a very enjoyable event. The veterans were Tal, Spassky, Andersson, Csom and Larsen; the youth team consisted of Lautier, Renet, Adams, Miralles and my- self. As the only GM in the youth team, I was accommodated in an excellent hotel; indeed, the conditions at this event were the best I had experienced un- til then. 
Game 5 M. Tal - V. Anand Youth vs Veterans, Cannes 1989 English I met Misha Tal for the first time when I visited the World Cup tournament in Brussels in 1988. He was the most popular player by far, captivating every- one with his personality and his brilliant chess. This was my first game against my childhood hero and I was obviously quite excited. lc4 cS 2 ltJf3 ltJc6 3 ltJc3 ltJd4 How do you explain a move that violates the rule not to move the same piece twice in the opening? Well, I can't really find a general principle that justifies it, but it does seem to work! One possible expla- nation is that when White recaptures on f3 with his queen, he has lost con- trol of the important d4-square. Then Black can bring out his other knight via h6 and f5 to fight for d4. In any case, 3...ltJd4 is a provoca- tive move which leads to a more un- balanced type of position than is usual in the Symmetrical English. 4 e3 ltJxf3+ 5 'iixf3 g6 (D) 6 b3 Or 6 d4 g7 7 dxc5 (after 7 'iidl ltJf6 White will have to play d5 and then lose a tempo with e3-e4 in order to get his customary space advan- tage) and now: ., tij;   . J&\ £ / .& '/:::.JL_ _/ / ii.i.i W //// ij'//.; Yu///   . jjj{ :$ ,  ::!$ff & . p ¥%  A . . . . . :;;:::; /uU/ 0::://' ///// / %1 l\.  i?M  . ;,//; / 0 ;:;, ;0  ///: // 'l";;0/ 0;(///.;-: /j/// . rm \Wr. .  6.  fj ?'  '  ;  B / /  0       / .;  /.; rf uu :; /// u %I // .; ? '   Wj% . " /  ;;+ / ...  /-,  0%/; 1) 7... 'iia5 8 e4! 'iixc5 9 ltJd5! (stronger than 9 d3 xc3+ 10 bxc3 d6) with a slight advantage for White. 2) 7...xc3+!? 8 bxc3 'iia5 9 e4 'iixc5 10 d3 d6. Normally Black should not give up his bishop like this in an 'Indian structure', but due to White's doubled c-pawns this po- sition should be compared with the Nimzo-Indian rather than the King's Indian Defence. Admittedly Black has already played ...g6, but it's not clear how White can exploit this. 
30 VISHY ANAND: My BEST GAMES OF CHESS 6 ... g7 7 b2 d6 8 g3 ':b8 9 g2 lbh6?! Two years later, I beat Karpov with 9...lbf6; the tactical justifica- tion is that 10 lbd5 g4! equalizes immediately. Actually, it is much more natural than 9...lbh6. Whatever can be said about the objective mer- its of 3...lbd4, I scored 2/2 against Tal and Karpov with it! 10 'ifdl 0-0 11 0-0 d7 To support the ...b5 advance. 12 a4 After 12 d4?! Black's plan of ex- erting pressure on d4 comes to frui- tion: 12...cxd4 13 exd4lbf5! 14 d5 (14 lbd5 b5! gives Black the edge) 14...b5! with good counterplay. 12 ... c6 13 d4 xg2 14 xg2 (D) .mu.. - - .*. .. .. B_ . _ _ .  ..- M _ M _ - - - . 8.8 . . MUM M .8 n    U U  .  _ . u'&'u a .if_:. 14 ... ':c8 After 14...cxd4 15 exd4 lbf5 16 d5 a6 17 ':bl! the position is slightly better for White. He has a space ad- vantage and a weak black e-pawn to play against. On the other hand, his pieces on the queens ide are awk- wardly placed. 15 'ifd3?! Tal later mentioned the possibility of 15 d5! and this move does seem to offer White a stable if minuscule edge. Black will hardly be able to achieve ...b5 and playing ...e7-e6 would create weak pawns on e6 and d6. 15 ... cxd4 16 exd4 lbfS Now Black is fine. 17 dS (D) After 17 lbe2 d5 18 c5 a5 Black is slightly better. White's queenside pawns are crippled and the pawn on d4 is weak. White's best line was probably 17 lbd5! e6 18 lbe3, just playing for the exchange of knights and equality. .. -.. .. '.. B_ d _ - . ... B _88.. 888_ _ _ .8if_ 8  . "<itt _  U u a _ 8:. 
TAL - ANAND, CANNES 1989 17 ...  The reason why 'ii'd3 was bad - b3 is undefended. 18 lbdl i.xb2 19 lbxb2 e5! Now Black is justified in playing actively. The open f-file counts for more than Black's slightly weak cen- tral pawns. 20 dxe6 Forced. If White doesn't do this, Black can aim for ...e4-e3. Moreover, his knight can settle comfortably on d4. 20 ... fxe6 21 :adl .:t'6 Preparing to double rooks on the f-file. 22 :d2 e5! (D) ... ... W '/._ . .. _ ' ',,/, _ o '/ _ %1  &  . z     .. "  ...!I:i ,, . . -.. .. - - __.._ D "  <ifi>   U U . . .:. 23 _d5+ After 23lbdllbd4 Black wins a pawn as White has to meet the threat of 24... 'ii'c6+. 23 ... g7 24 5 'ii'c7 31 25 c5 There is no other defence against 25...lbd4. 25 ... 'ii'xcs 26 'ii'xb7 + :c7 27 'ii'd5 'iib4 Not 27...'ii'xd5+? 28 :xd5 :c2 and White can hang on with 29 c4. 28 .:t'd1 :c5 Exploiting the weakness of b3. 28...lbd4 is less clear after 29 lbc4 :c5 (not 29...lbxb3? 30 lbxd6!) 30 'ii'a8. 29 'ii'a8 If 29 lbd3? then 29...lbe3+! 30 fxe3 'ii'xd2+ winning the exchange. 29 ... 'ii'xb3 30 lbd3 (D) After 30 'ii'xa7+:f7 31 'ii'a6 (31 'ii'a8 :c2 is similar, e.g. 32lbd3 _c3 33 :xc2 'ii'xc2 winning material) 31...:c2! 32 lbd3 :xd2 33 :xd2 'ii'd5+ 34 gl e4 White loses the pinned knight. ... - - . ' . . . B  . . _ . .../ .. -  - - .  IA\. . - .--. . - . . .._. D .  rlh   U1SiU . B:B _ 30 ... :c2?! 
32 VISHY ANAND: My BEST GAMES OF CHESS 30...':c7! would have been bru- tally effective - Black simply main- tains his a-pawn and threatens both 31..:iixa4 and 31...ltJd4. 31 'ii'e4?! White could have made life harder by playing 31 1ixa7+ ':f7 32 1ia5 (D) (32 1ia8 1ic3! wins). I don't know if Tal spent much time consid- ering this apparently greedy con- tinuation. I must admit that it was only after some time checking the position with Fritz that I began to re- alize that Black's task was not so easy. Still, Black does have a way to WIn: . . . - B . / . .I' - .  - .  ...   . - 1UU . .IA\. _ . /U,... . . . . ._.ttJ_   . · g /'    ' '/j; rlh  D  / .A   '&'   / /"  /,,/ . .:8 . 1) 32...e4?! and now: la) 33 ':xc2 'ii'xc2 34 ':d2 1ic6 35 ltJb4 1ic3! wins for Black. Ib) 33 ltJf4 'ii'f3+ (33...J:tc5 34 1id8 is unclear) 34 <it>gl e3 looks strong, but after 35 fxe3 1ixe3+ 36 <it>h 1 there is nothing clear-cut, for example 36...1if3+ 37 <it>glltJd4 38 :tel. lc) 33 'iib4! 1ixb4 34ltJxb4 saves White. 2) 32...1ib7+?! 33 <it>gl :Xd2 34 1ixd2ltJd4 35ltJel and White avoids disaster. 3) 32...':xd2 33 'ii'xd2 (33 ':xd2? 'ii'b7+ 34 <it>gl 1ibl + 35 <it>g2ltJe3+ wins) 33...1ixa4 wins a pawn, but there is still a long way to go. 4) 32...ltJe3+! 33 fxe3 1ixd3 34 ':xc2 1ie4+ 35 <it>g 1 (35 <it>h3 :f5! and there is no way to stop ...':h5#) 35...1ixe3+ 36 <it>g2 1if3+ 37 <it>h3 1ixdl and White's king is hope- lessly exposed. 31 ... ':c4 32 1id5 1ic3 Besides the extra pawn, the differ- ence in strength of the knights can be seen. 33 ':b2 34 ':b7+ 35 5 (D) 35 1ig8 loses to 35...1ic6+. :d4 <it> h 6 ', ///   .:::.. ". . -. B ;a ?i .    _ %i / &  . //.A'0 -   .  @; /uu, , / " .\WJ4i . JIl\ . $./ / ,/,..%    / . / n '/;//0/ /,///c;/:'    ;   }%;/ ;@) »;;   / ;';';0,/:(:/,;;:: ////; "" / '/>///,/ /////// f%/ .  / '/ ttJ x;; ' .   /' :::  /  /t /;//;;, /j /j,/;: :- /  % '.' / :rt]  ' <it> fj   ,@/ ;c.;    'l ,/;" ; ///,; /,/  /' " /    1: ): fjjjM: t%: '/";/;:C; -j/////,-'/ -J//'// , //// 35 ... ltJe3+ 
TAL - ANAND, CANNES 1989 33 36 '1t>gl Or 36 fxe3 'iic2+ 37 '1t>h3 J:th4+! 38 gxh4 J:tf3+ 39 '1t>g4 'iig2#. 36 ... 'iic2 37 J:tn .:xd3 0-1 After 38 'iid7 Black mates by 38...'iixf2+! . I was very happy to win in Tal's own style, although I recognized that his poor play in this game was due to illness. In fact, he withdrew from the tour- nament after the first half for medical treatment. Still, I am proud to have played at least one game against Misha Tal. 
Game 6 V. Anand - B. Spassky Youth vs Veterans, Cannes 1989 Ruy Lopez, Breyer 1 e4 e5 During the course of his career, Spassky has played just about every opening there is, but in recent years he has tended to stick to dual king pawn openings, so this was not a sur- . pnse. 2 f3 c6 3 .i.b5 a6 4 .1a4 ltJf6 5 0-0 .1e7 6 :el b5 7 .1b3 d6 8 c3 0-0 9 h3 ltJb8 The Breyer Defence, for decades Spassky's main weapon against the Ruy Lopez. 10 d4 ltJbd7 11 c4 Round about this time there had been a modest revival in this old move. I felt that Spassky's knowl- edge of the main lines would be much greater than mine, so I tried, with some difficulty, to find a rela- tively unexplored continuation. 11 c4 enjoyed popularity in the 1960s and was probably no surprise, but at least it didn't have such a huge body of theory as 11 ltJbd2. 11 ... c6 12 a3 This was an idea I came up with over the board. At the time I had no idea if it had ever been played be- fore, but I have since found a couple of earlier games. The immediate 12 ltJc3 is met by 12...b4. 12 ... bxc4! The best reply. After 12....1b7 13 ltJc3 I think White is slightly better, since White's structure is a bit more comfortable in this type of 'Old In- dian' position. Black can of course continue with ...ilc7, ...:ac8 and so on, but White can gain more space with an eventual .1a2 and b4. 13 .1xc4 (D) . ill _ . ..i.. .*. . ...'/. B _ . _ ,, 'B'_ (I . - . - . .i.D. . o . .lD.  . R. U . U . .  -- -  
ANAND - SPASSKY, CANNES 1989 13 ... d5 Black is not worse after the text- move, but 13...ltJxe4 would have equalized straight away: 1) 14 dxe5? d5 15 .1xd5 cxd5 16 1Wxd5 :b8 (16...ltJxf2 may also be good for Black) 17 1Wxe4 lbc5 18 1We2 ltJb3 19 :a2 .1e6 and White loses the exchange by force. 2) 14 :xe4 d5 with two possibili- ties: 2a) 15 ltJxe5 (Spassky suggested this, but Black has an escape route) 15...dxe4 (not 15...ltJxe5 16 .1xd5 cxd5 17 .:xe5 .1d6 18 :el and White can be happy because, compared to the Marshall Attack, his piece devel- opment is easier as ltJc3 is possible) 16 ltJxc6 1We8 17 .1d5 .1d6 18 ltJe7+ 1Wxe7 19 .1xa8 ltJb6 20 .1c6 .1c7 (the exposed bishop on c6 gives Black's queen a free tempo on its way to h2) 21ltJc31Wd6! (21...f5 22 1Wb3+ <it>h8 23 1Wb4 is annoying for Black) 22 .1xe4 f5 23 .1f3 1Wh2+ 24 <it>fl ':e8 25 g3 f4 and White is cer- tainly not playing for the advantage - indeed, Black is probably slightly better. 2b) 15 :el dxc4 161We2 (after 16 dxe5 lbc5 the outposts at d3 and b3 give Black the edge) 16....:e8 with equality, as Black's weak pawns are balanced by his active pieces. In fact this had been played as long ago as 1971, in the game Htibner- Lengyel from Wijk aan Zee. 14 exd5 35 14 dxe5? ltJxe4 transposes to line 1 of the previous note. 14 ... cxd5 15 .1a2 e4 16 ltJe5 .1b7 (D) ..  -.. ..t.IA\.. w .._ _ .. . (I . . .. . . ".. .  U _ .  . . .8 u..'. i." . 8. uu_ g '  . / %\WJ a /' =  /.;.;   .  / / './ / /  Here I was quite happy, as it seemed to me that my pieces were well placed to exert pressure on d5. While there is some truth in this, Black has so many pieces available to defend d5 that the inconvenience is not serious. 17 ltJc3 ltJb6 By protecting d5, Black sets up various threats based on moving the f6-knight. The first is to expel the e5-knight by ...ltJe8 followed by ...f6, in which case he would proba- bly be better. The second is the sim- ple ...ltJfd7, when Black can ex- change on e5 without having a knight attacked after dxe5. Hence White must react quickly. 18 f3! ':c8 19 .1b3 
36 VISHY ANAND: My BEST GAMES OF CHESS White could have maintained some slight pressure by 19 i.g5, as Spassky pointed out after the game. 19....1xa3? 20 bxa3 :xc3 is bad af- ter 21 .1d2! :c8 (21...:xa3 22 i.b4 favours White) 22 .1a5! with a very awkward pin. 19 ... i.a8 20 .1g5 :c7 Black's position is basically safe; for example he could have played 20...exf3 21 1ixf3ltJc4 with equal- ity, and indeed this would have been the natural way to make use of ....1a8, which protected the bishop. 21 :c1 ltJfd7 22 .1f4 .1g5 (D) .,. . Wit  .  .JLit _ _ .  . 1&\  .  ..  . / .. W   -a\'"  .. /0;/; "   ////  . "//. . ;/; . i/. : . . . %A . /;'   ..  j1 //  ' . ,../.. 0 W.y ""  .  . /, n / ..  d .  . /  %;,.. -  /.U/  /" /  i.  ./jE/j  /. ////.%  f/j/ . : ' D . g/j. . B \Wr g  wffi: ?  a  /. / /. //%:::/ // /0// / ////-j ,, 23 .1xg5 After 23 ltJxd7 the line 23.. ..1xf4 24 ltJxb6 .1xc 1 25 ltJxa8 .1xb2 (25...:xc3 26 bxc3 is also very good for White) 26 ltJxc7 .1xc3 27 ltJxd5 i.xel 28 1ixel exf3 29 gxf3 clearly favours White. However, Black can improve by 23...:xd7! 24 .1xg5 1ixg5 25 fxe4 dxe4 26 ltJxe4 'iVg6 and he regains the pawn since 27 .1c2 fails to 27...f5. 23 ... 'iVxg5 24 fxe4 dxe4 Black had a good alternative in 24...ltJxe5 25 dxe5 dxe4 and now: 1) 26 'iVd4 ltJd7! (not 26...e3 27 J:e2 attacking b6 and e3) 27 e6 (not 27 ltJxe4?? .1xe4! and Black wins) 27...ltJe5 28 exf7+ <it>h8! 29 'iVd6 :cc8 and Black has dangerous king- side threats. 2) 26 e6. During the game, I be- lieved this was good for White, but Black simply continues 26...'iVc5+ 27 <it>h 1 fxe6 28 .1xe6+ <it>h8 and the position is just unclear. The game continuation is also roughly equal. 25 'iVg4 'iVxg4 Black cannot get away with 25...'iVd2, when White must decide how to take on f7: 1) 26 .1xf7+ (obvious, but this is in fact an error) and now:' 1 a) 26.. .:xf7 27 ltJxf7 ltJf6 28 'iVe6 'iVxd4+ (28...J:xf7 29 'iVxb6 e3 30 ltJe4! wins) 29 <it>hl :xf7 30 J:cdl and White is winning. 1 b) 26...<it>h8! 27 ltJxd7 (not 27 :cd I? ltJxe5 28 dxe5 'iVxb2 and Black wins) 27...'iVxd4+ 28 <it>hl ltJxd7 with an unclear position. 2) 26 ltJxf7! ltJf6 (26...'iVxd4+ 27 <it>h 1 ltJc5 28 J:cd 1 'iVf6 29 J:fl ! wins, while after 26...'iVxb2 27 'iVe6! Black is in considerable difficulties) 
ANAND - SPASSKY, CANNES 1989 27 ltJg5+ <it>h8 (27....1d5 28 1idl! with a clear extra pawn) 281ig3 and White is clearly better. 26 ltJxg4 g6? Black's first step downhill. After 26...<it>h8 27ltJe3 f5 Black activates his kingside majority, which should provide enough counterplay to main- tain the balance. 27 ltJf2! White now wins a pawn, although in view of the reduced material this does not necessarily guarantee win- ning the game. 27 ... 28 d5 29 ltJfxe4 J:te8 rl;g7 ltJxd5 (D) .t :;:,.;/ ;//.  W;0 ;;;;:0 :: ffi _ ., f /;mt/  .aW  . - . - w. ...... .. B B'. B... _ . .lb. . "i. . .8 u  _ _  . .8. .  n ;0 w;  < At first I couldn't believe this move; it looks as though White must win material after 30 .1xd5 .1xd5 31 ltJd6, as Black will end up being threatened with ltJxd5 and ltJe8+. However, Spas sky had worked out a defence. After I had calmed down, I didn't see any way to force a decisive 37 material gain, so I just went for a pawn. 30 ltJd6 30 .1xd5 .1xd5 31 ltJd6 is an- swered by 31...J:txel + 32 J:txel J:tc5, meeting both of White's threats. It is easy to miss that the c3-knight is no longer protected twice and therefore b2-b4 is impossible. White can try 33 J:te7 .1e6 34ltJxf7 <it>f6 35 ':xd7 ':xc3! 36 bxc3 .1xd7 37 ltJd6 <it>e5, but although he has won a pawn, Black's king becomes too active and he draws easily. 30 ... .:xe1+ 31 .:xe1 ltJ5f6 32 J:te7 J:tc6 33 .:xf7+ <it>h6 34 ltJc4 ':e6 Black has managed to get some counterplay; White's f7-rook is sur- rounded and Black can activate his king via g5. 35  36 .1c2 37 ltJd2 <it>g5 .1c6 h5(D) - . . . w. ....:. ...t.a... . . . II' . . . . "   .  . 8 u    i.  8  .  U1flJ  . . . . 
38 VISHY ANAND: My BEST GAMES OF CHESS 38 b3?! This would have been the moment to settle for 38 ':g7! e5 (38...lbe8 39 de4+! <it>h6 40 ':h7+! <it>xh741 g5+ leaves White a clear pawn up) 39 ':a7 and White's rook gets out of the trap! In fact this is quite awkward for Black as a6 is weak. Black's mass of pieces appears menacing, but White has all the critical squares controlled. The sacrifice 39.. .lbeg4+ doesn't work here: 40 hxg4 xg4+ 41 gl (but not 41 g3? h4+! 42 <it>h3 ':e3+ 43 f3+ .1xf3 44 gxf3 ':xf3+ and Black wins) 41....:el+ 42 fl defends. I thought that the text-move was even stronger, but I hadn't really seen Black's 39th move - in fact, I'm quite lucky that it doesn't cause seri- ous damage! 38 ... 39 :&7 e5 eg4+! (D) . . . . wa . . . ...i...II.. . . . -. - - - - . . ... " . .8 u   . 8. u    . . . . This works because White can't play fl any more. xg4+ lbe3+ ltJxc2 .1d7 40 hxg4 41  42 <it>gl 43 .:xa6 44 .:xe6? Though this sets up the nice fin- ish to follow, it is a mistake. Two knights can't win by themselves and there aren't that many pawns left! White should have played 44 ':a5+!, followed by ':d5-d2. By keeping the rooks on, White has much better chances of pushing the queenside pawns and in fact he still has excel- lent winning prospects. 44 ... .1xe6 (D) . . . . w. . . . . ..i.... . . . -. - - - - . . . .  . . U _ _ .. .8. u d .  . . . = 45 lbc5 .1c4 46 a4 4? One reason why I exchanged rooks earlier was the combination which now occurs, but in the interim I had realized that 46...<it>f5 prevents it: 47 a5 b4 48 b3 .1f7 49 d3 xd3 (a forced move, as 49...a6 50 b4 .1c4 51 lbb2 is good for 
ANAND - SPASSKY, CANNES 1989 White) 50 a6 .1e8 51 d5 e6 52 c7+ d6 53 xe8+ c6 and the king catches the a-pawn. Therefore White has to settle for 47 f2, but Black has good drawing chances. 47 as Now we get to see an elegant fin- ish - a lone knight dominating two minor pieces! 47 ... 48 b3 49 d3+!! Spassky had only seen 49 a6?? xa6 50 xa6 .1xb3 and White has no winning chances since he has only one pawn left. 49 ,.. 50 a6 lbb4 .1f7 lbxd3 .1e8 39 51 d5+ 1-0 After 51...e5 52 e7 (D) the position deserves a diagram: . ..tB . /. ////,  /,//////  "?: B. '.  . . _ _.. _ R II R' . . - - -_.- - . - --  . .  E,/* Black can't stop the pawn - a nice bit of domination! In Cannes I made the best score in the junior team, 6 1 /2110, but the 'Senior' Andersson made the best score overall, with 7 1 /2 points. The next game is again from Wijk aan Zee, the year after my success in the 1989 event. 
Game 7 M. Kuijf - V. Anand Wijk aan lee 1990 Ponziani Opening 1e4 e5 Round about a month before I had beaten Kuijf in the tournament at Groningen. That game had been a Closed Sicilian, and at one stage it could have been very dangerous for me. This time I didn't want the same 'excitement' so I decided to play more solidly. Hence my choice of first move. 2 ltJf3 ltJc6 3 c3 A real surprise. 3 ... ltJf6 I played this move instantly and while he was thinking about his re- ply I had to spend a few minutes try- ing to remember the name of the opening! 4 d4 ltJxe4 5 d5 ltJe7 I don't really understand what Kuijf was aiming for with his choice of opening. It is harmless and only useful if White is aiming for a draw. 6 ltJxe5 ltJg6 7 .1d3 However, this indicates that White is not aiming for a draw, towards which he could have made substan- tial progress by 7 'iie2 'iie7 8 'iixe4 (8 ltJxg6?! hxg6 is a little better for Black) 8... 'iixe5 9 ltJd2 (or alterna- tively 9 'iixe5+). 7 ... ltJxe5 7.. .ltJxf2? isn't even a difficult trap: White wins by 8 .1xg6ltJxdl 9 .1xt7+ q;e7 10 .1g5+ q;d6 IlltJc4+ Q;c5 12.1xd8. 8 .1xe4 .1c5 9 'iih5 d6 10 .1g5? (D) 10 h3 was a much better move, al- though even in this case Black can play for an advantage. z..t' :..  - ,, . - B _._ .._. .  . . -..  . ' A . / u// % m ; \WI .   0 ' 0 /, e  ' / , , / . 8.t. . . . . . - ", ;; /'/: / A D  ' %J fj / , ;0 '% A D  % O    o  /,!/ "/'// /'0 Y'/ uu   ltJ. ;;; '.: / , ,/ - .  /;/ 10 ... .1g4! Black can play 10...'iid7 11 0-0 'ii g4, with boring equality, but I had seen that the text-move is much 
M. KUIJF - ANAND, WIJK AAN ZEE 1990 stronger. Technically, it may be a novelty, but I am reluctant to call it that. To my mind, novelties should be at least a little bit difficult to find. If you play the most obvious move and then discover that by an accident of history nobody has played it be- fore, I am not sure that it deserves any special appellation. 11 'iVh4? Losing on the spot. The lines 11 'iVxh7 'iVxg5! and 11 'iVxg4ltJxg4 12 ..txd8 ltJxf2 13 ':fl ltJxe4 14 ..txc7 ':c8 15 ..ta5 b6 16 ..tb4 ..txb4 17 cxb4 ':c2 also offer White no hope. This leaves 11 ..txd8 ..txh5 12 ..tg5 (12 ..txc7 ':c8 13 ..ta5 b6 14 b4 .txf2+ 15 xf2 bxa5 and 12 ..th4 0-0 followed by 13.. .':ae8 are also very promising for Black) as the only realistic way for White to play on. Even here Black has a range of tempting options. He could simply play for the two bishops by 12.. ...tg6 13 ltJd2ltJd3+, but 13...f6 is proba- bly stronger. Then 14 ..tf4 is impos- sible, 14 ..th41eaves the bishop shut out on the kingside and 14 ..te3 al- lows Black to shatter White's pawns. Black could also consider 12.. .f6, with similar ideas. 11 ... t[Ci Now White has no reasonable continuation. 12 ..tcl After 12 .td2 'iVe7 13 0-0 g5 White has the unpleasant choice be- tween: 41 1) 14 'iVh6ltJf7 15 'iVg7 'iVxe4 16 ':el ..te2 17 b4 ..tb6 18 ..te3 e7 19 :Xe2 (19ltJd2 'iVd3) 19....:ag8 wins. 2) 14 'iVg3 f5 15 h3 (the only chance, or else ...f4 traps the queen) 15.. .f4 16 'iVh2 ..td7 and although material is even, White is playing a whole queen down for all practical purposes. After ...0-0-0 and a subse- quent ...g4 the attack should over- whelm White. If White retreats his bishop to e3 the lines are even simpler: 12 ..te3 g5 13 'iVh6 (13 'iWg3 f5 14 f4 ..txe3 15 fxe5 ..tf4 and Black wins) 13...'iVe7 14 0-0 ltJf7 15 'iVg7 0-0-0 16 ..txc5 ':dg8 17 ':el ..td7 and White loses his queen. 12 ... 'iVe7 (D) .. . . . ... //////.  /.//</ ' 0f/ illIi w.l.. _  .  / . '%  '/   --  y""  Ym" . 8 . d _ _ . . ...t. .   - .'.. .    A ' U  ./  /; ' A {% O  _  o  y,  / /, " a /. % . ;  = ,/.A//.  . "   / % /.... "  , , ,; , /m  /:i  _/';&/ .- Y,,,,, . . 13 0-0 This is forced to meet the threat of ...ltJd3+, because 13 f3 is met by 13....1xf3. 13 ... g5 14 'iVg3 f5 
42 VI SHY ANAND: MY BEST GAMES OF CHESS The threat is .. .f4, so White re- 17 ltJd2 ':g8 sorts to desperation. 18 .:&el h6 15 .1xf5 19 .:xe5+ dxe5 15 h3 f4 16 'iih2 .1d7 shuts the 20 'iixe5+ d7 queen completely out of play and 21 .1e3 ':&e8 Black wins by ...0-0-0 and a pawn 22 'iff4 'ifxd5 push on the kingside. 23 ltJf3 .1d6 15 ... .1xf5 24 'if84+ b5 16 .1xg5 'iff7 0-1 This game was finished while some of the others were still in the opening; I spent ten minutes and my opponent a little under half an hour. It was a nice miniature to play in the first round, but the rest of the tourna- ment didn't go as well as the year before. I lost in the second round to Nunn and continued unevenly throughout the event, finishing on 50%. Shortly after Wijk aan Zee, I played an open tournament in Rome. This started well, but I lost a miniature to Miles in the penultimate round and fin- ished with 6/9. Then I went back to India for some rest before travelling to the Zonal tournament in Qatar. I was by far the highest rated player, but even so I was happy to win the tournament convincingly. As a result of this and some other tournaments I gained quite a lot of rating points, and on the 1st July 1990 list I stood at 2610 - I had broken through the 2600 banier. I took part in an open tournament in Manila, went back to India and then returned to Ma- nila for the Interzonal. Based on my new rating I was certainly a potential qualifier, but in an Interzonal you cannot take anything for granted - an Inter- zonal is a tough tournament even for the top seeds. My results in the Interzonal followed a fluctuating course: I won in the second round, lost in the third, won in the fourth and lost in the fifth. After a draw with Chandler in the sixth round, I faced the Icelandic grandmaster Margeir Petursson in round 7. 
Game 8 M. Petursson - V. Anand Manila Interzonal 1990 Queen's Pawn 1 d4 d6 2c4 e5 This was part of my usual open- ing repertoire at the time. I liked 1.. .d6 because it is such a complete system in itself. IT you play the Pirc then you have no reason to fear 2 e4, and 2 f3 can be met by 2...i.g4. Since then, however, White has found ways to keep some pressure. As a re- sult I lost faith in the system and had to learn a decent defence to 1 d4. 3 lbc3 exd4 4 'ifxd4 lbf6 An important finesse. If Black plays 4...lbc6, then 5 'ife3+ is a little annoying as 5...i.e7 6 lbd5 snares the two bishops. 5 g3 lbc6 6 'ifd2 g6 White's pieces are not badly placed - for example, the queen is quite useful on d2 if White intends to play b3 and i.b2. On the other hand, Black has gained a tempo because of 'ifxd4 and 'ifd2. 7 i.g2 i.g7 8 h3 0-0 9 lbf4 White is aiming to establish a grip on d5. He has played the opening accurately, for example by not play- ing b3 too early, which sometimes allows a tactical ...d5! by Black. 9 ... as (D) . . .t.  .  .   - -  .i' .ii w_ , _ _  . . -i .   - -  '/ .  .   .  , d   ..  .  .    .  "    U "  U  UU - . ': - - *  This is designed to prevent White from easily developing his queen's bishop, for if 10 b3, then 10...a4 11 ':bl (11 lbxa4 lbe4 wins the ex- change, while 11 i.b2 may be met by 11...a3 12 .1c 1 .1f5 or 11...lba5) 11...axb3 12 axb3 ':e8, followed by ...i.f5, with active play. 10 0-0 ':e8?! A slight inaccuracy. 10...a4 was better, simply preventing b3. 11 ':el?! White misses his chance. 11 e4 was correct, and after 11.. .a4 12':b 1 
44 VISHY ANAND: My BEST GAMES OF CHESS ltJe5 13 b3 White has secured his queenside pawn structure. Of course the immediate 11 ':bl is answered by 11.....tf5. 11 ... a4 12 ':bl ltJd7! A good move. By the time White has played b3, Black has arranged to occupy active squares with his knights. 13 b3 axb3 14 axb3 (D) ....t.SI..... ...../. B    /" '/ B_ B.. B B . . ..  r :// £tli;  /;im . . " -- . "i." . _ U U  .  "  m %; D /. /  '0 _....    /j  :- , , ,,/ '/ ,,/ /, Y;: 14 ... ltJce5! If 14...ltJde5!? (intending to play 15...ltJd4) then 15 ltJb5! is slightly better for White. It is the c6-knight which belongs on e5, so that the other one can occupy c5. 15 'ifc2 ltJc5 16 b4? Premature, because with ener- getic play Black now gets the better posi tion. After the correct 16 ..tb2 c6 we reach a position typical of the Fianchetto King's Indian. White has the centre, whilst Black has open lines for his rooks and active squares for his knights. Chances are about equal. 16 ... ..tf5! An important intermezzo. 17 e4 After 17 ltJe4 ltJa4, followed by 18...ltJc6, Black has excellent control over the long dark-square diagonal. 17 ... ltJe6! (D) Margeir had missed this tactic. ... ...... .. ... W   _ . _.B.B . . M.t.. ".  . U // /'/ :0;; .  . /  /. ' . /  ... "i." d . U U  . " %; a  = ,,/  ....    /. /j  / " '/ /. /, Y;: 18 ltJxe6 Black's play is tactically justified by the variation 18 exf5 ltJd4 (at- tacking the queen and threatening 19...ltJef3+) 19 'ife4 (19 'ifdlltJef3+ 20 .1xf3 ':xe 1 + 21 'it' xe 1 xf3+ and Black wins) 19... gxf5 ! (the queen is trapped) 20 'ifxb7 (20 'ife3 c2) 20....:b8 with a clear advantage to Black as White must now give up his queen. If 18 ..te3, then 18.....tg4 19ltJxe6 ':xe6 and f3 is very weak. 
PETURSSON - ANAND, MANIlA INTERZONAL 1990 45 18 ... ..txe6 Now Black is much better; his pieces are active and the c4-pawn is weak. 19 ltJd5 19 c5 dxc5 is very good for Black after 20 bxc51Wd3! or 20':d l1Wf6!. 19 ... b5! Winning the c-pawn and forcing White to search for some sort of compensation. 20 ..tb2 (D) Not 20 cxb5? ..txd5 21 exd5ltJf3+ wInnIng. .B ii.... .  ... B.  . _ . _.i.... B.RltJ. . "8.88 . u W ;; /;///// . _ . D .. ".1" /, "Y . U U .:.   . . -  20 ... :a2! I prefer this to the line 20...ltJxc4 21 ..txg7 <ifi>xg7 221Wc3+ f6 23ltJf4 c6! (23.....tf7 24 e5 ltJxe5 25 ..txa8 '6'xa8 26 ':e3 ':c8 intending. ..c5 is unclear) 24 ltJxe6+ ':xe6 25 ..tfl ltJe5. Here Black is a pawn up, but White has some compensation due to the weak queenside and the open 7th rank, which would enhance the strength of any rook penetration by White. Note that 26 f4 can be met by 26...1Wb6+ followed by 27...ltJg4. 21 ltJc3 ':xb2! 22 .:xb2 ..txc4! Not 22...ltJxc4 23 ':a2 with an un- clear position. After the text, Black has extremely active pieces and strong dark-squared pressure in re- turn for his small sacrifice. The im- mediate threat is 23...ltJd3. 23 ':e3 (D) After 23 ':d 1 'ii'f6 24 ltJd5 ltJf3+ 25 ..txf3 1Wxf3 White is pretty close to being lost. . IB.. . . ..'. B. uu _ / , , ,; '{(if;!) .  . / W!.&  . 0   /:@(}j    .  ;;///, //// / {ffi-;{f  % . /" . 'l//X ;%1 . "//   / 2X A   _ Wd /'/;  .i..8. .   .. /   // _ A   /  / ,    \Wr. .1"  R U U . . . /0 -  - /,  23 ... ..th6? A really awful move whereby, in one stroke, Black throws away all his advantage. The problem was that I was so excited about the way all Black's pieces were working well together that I forgot White could still develop counterplay. The cor- rect line was 23. ..ltJg4! 24 ':f3 (24 ':e 1 ..td4 25 1Wc 1 1Wf6 26ltJd 1 ..txb2 wins) 24.....td4 251Wd2 c5, followed 
46 VISHY ANAND: My BEST GAMES OF CHESS by 26...lbe5, and Black has a large advantage. 24 f4! More or less equalizing. 24 ... lbg4 25 :f3 i.g7 26 'it'd2 26 ':b 1 !? was also possible. 26 ... 'iVa8 27 h3 'iVal+ (D) The tactics don't work for Black after 27...'it'a7+ 28 <ifi>hl 'iVaI + 29 ':bl .1xc3 30 'iVc2! 'iVa2 31 ':xc3 'it'xc2 32 ':xc2 .1d3 33 hxg4 .1xc2 34 ':cl .1xe4 35 ':xc7 .1xg2+ 36 <ifi>xg2 ':e4 and the resulting rook ending is probably a draw in view of White's active rook position. I was quite surprised that Black had noth- ing better in this line; with a King's Indian bishop and a knight rampag- ing around, you expect something to work but in this case there was noth- ing. . ..... . // ... w_ / _ _ . ,' ... ... . - P.t..8.. u _ u _ . " .:8 .  _ U   .i..  - . .  . .  - . .  28 .1n .1xn 29 .:xn 'iVa3 Once again White can hang on af- ter 29...'iVa7+ 30 <ifi>hl lbe3 by 31 ':cl! lbc4 32 ':a2 'iVb7 33 'iVd5!. 30 .:f3 lbf6 31 ':e3 (D) . ..... .  ... B_._. . . _.. ... B .  .8 . u _ u _ _ ' '//,   \ a ' D    % '  8 /.  / / / %-    . .  -  . . - . = 31 ... d5 Black can regain the exchange by 31...lbd5?! 32lbxd5 'iVxb2 33 'iVxb2 .1xb2, but after 34 lbxc7 Black is suffering, for example 34.. .':b8 35 ':d3, 34....:c8 35lbxb5 ':b8 36 ':b3! or 34....1d4 35lbxe8 .1xe3+ 36 <ifi>g2 .1d2 37 lbxd6 .1xb4 38 lbxb5 and White is better in every line. 32 lbxd5 White should avoid 32 e5 d4!, which gives Black a clear advantage after 33 'iVxd4 'it'xb2 34 exf6 ':xe3 35 'iVxe3 .1xf6. However, 32lbdl! 'iVa7 33 e5 was a perfectly good al- ternative to the text; after 33...lbe4 the position is unclear. 32 ... 'iVai + 33 ':el 'iVa7+! (D) 
PEIURSSON - ANAND, MANIlA INTERZONAL 1990 47 I could have forced a draw by 33...ll)xe4 34 ':xallbxd2 35 ':xd2 ..txal 36 lbxc7, but even though things had not gone according to plan, I was still hoping to win. Sometimes it is a mistake to play under the influ- ence of your former advantage, but in this case Black's optimism proved justified. a aIa.a illU  &_& w_ _ _-.- a . lI.a ..alb. a  .8 . u . u  . . . P8 _ _ _ U R m . . - - . . . .    . - * 34 h2? Petursson returns the favour with this blunder. The alternatives were: 1) 34 hl?! lbxe4 35 ':a21Wb7! is also bad for White. 2) 34 'ife3! 'ifxe3+ 35 lbxe3 lbxe4 36 ':c2 xg3 37 g2 ':d8 38 ':xc7 lbf5 with a near-certain draw. White has a nominal material advantage, but Black's pieces are well coordi- nated and White's pawns are dis- jointed, so he has no chance of putting it to use. It often happens that a player who has fought back from a bad po- sition with a long series of accurate defensive moves subconsciously re- laxes just when his troubles appear to be over, and commits a further er- ror. In this case time-trouble proba- bly also played a part. 34 ... lbxd5 35 :a2 'ifd4! 36 'ifxd4 ..txd4 Petursson had overlooked that there was no way to catch one of the minor pieces, even though they are temptingly lined up on the d-file. 37 ':dl fails to 37...lbc3 and 37 ':d2 to 37.....tc3. 37 :as lbc3 38 e5 ..tb6 39 :a3 lbd5 40 ':b3 (D) . .I... B a _ a.B. _ B B.a ....n . _ _ u _ p . n . u  u . B:B B 98 . . .  d    a . a a 40 ... xb4! Simplest. Now if 41 ':xb4, then 41.. ...ta5 42 ':eb 1 ..txb4 43 ':xb4 ':b8, and 44...c5, when the connected passed pawns will romp home. 41 ':dl c5 42 ':d7 ':b8 
48 VI SHY ANAND: My BEST GAMES OF CHESS 43 g4 44 :a7 45 :a6 .1a5 ltJc6 ltJd4 46 :a3 47 <ifi>g3 .1d2 b4 0-1 This game put me on + 1, but although it put me in the right direction, it was a late winning streak in which I beat Lautier, Miles and M. Gurevich in consecutive rounds that made me a qualifier for the Candidates. My next major event was the Novi Sad Olympiad towards the end of 1990. I have always enjoyed playing in Olympiads. Dubai 1986 and Manila 1992 were my favourites; in both cases the organizers went all-out to make the players feel comfortable. However, the Novi Sad Olympiad was also pleas- ant, despite the cold and dismal weather. In general I enjoyed playing events in the former Yugoslavia because of the great public interest, which always ensured a good turnout of spectators. I started the Olympiad with a good win against Olafsson, but lost a rather silly game to Bouaziz in round 3. Later on things started to get better, and I was satisfied with my final score of7 1 /2112. The following eventful game was played in round 7. 
Game 9 V. Anand - I. Morovic Fernandez Novi Sad Olympiad 1990 Sicilian, Maroczy Bind le4 c5 2 ltJf3 ltJc6 3 d4 cxd4 4 ltJxd4 g6 5 c4 .1g7 6 ..te3 ltJf6 7 ltJc3 0-0 8 .1e2 d6 9 0-0 ltJd7 (D) I had played the white side of the Maroczy Bind before, including a good win against Larsen from the Cannes event mentioned previously. However, in that game Larsen played 9....1d7 10 'it'd2ltJxd4 11 .1xd4.1c6. Morovic adopts a rather unusual move, re-deploying his knight to the queenside. XB.t.  .B i..ii W .  _ B._ BiB . . . . ... UAA. . .oo_ _ '   ,w:P/  /j '1 / +  @;?' . / ; , . A  D ' u -   /' A   / O _...o /  ;r /,  /'/   .iV.:   -  10 ':bl I was just casting around for a logical move, and since Black in- tends to play ...ltJc5, it seemed rea- sonable to prepare to meet it with b4. It perhaps looks a little odd to put the rook on bl rather than cl or (after 'it'd2) dl, but in fact White often puts his rooks on bl and c 1 in this varia- tion, in order to support a queenside pawn advance. Typically White con- tinues ':cl-bl to play a2-a3. Black responds with ...1Wb6-b4. White then plays ':fc 1 to support the c3-knight. Play then revolves around White's ability to get a3 and b4 in, and Black's ability to stop it. The idea behind the text-move is to save time by going to bl directly. 10 ... lbc5 When he played this anyway, I wondered what the idea was, as he seemed to be running into b4. Then I saw that after 11 b4 ltJe6 12 ltJxe6, he could play 12...fxe6! (12....1xe6?! 13ltJd5 is clearly better for White, to be followed by 'it'd2, ':dl and possi- bly c5) 13ltJb5 e5. However, 14 c5 would then give White a slight ad- vantage, so this would have been a valid alternative to the text-move. 
50 VISHY ANAND: My BEST GAMES OF CHESS 11 'ifd2 A flexible alternative; White keeps open the possibility of b4 while de- veloping his pieces. 11 ... xd4 Black decides to prevent b4 by playing ...a5. If immediately 11...a5, then 12 db5; the plan is f3, fol- lowed by d5, and White will play b3, a3 and b4 at his leisure. Black's c5-knight can eventually be kicked away, while the knight on b5 is there permanently. Hence Black's deci- sion to swap knights before advanc- ing the a-pawn. 12 .1xd4 as 13 b3 White is aiming for an eventual b4, but the immediate 13 a3 allows 13...a4. 13 ... .1xd4 14 'ifxd4 (D) .B"'_ _.B B BiB .iBi .  .i. - - - -  - . . - - - - B.B B B B B . .!LP . . U u .:. .: - - - * 14 ... b6?! There are some lines of the Mar- oczy Bind in which ...b6 and ....1b7 is played, but normally only where Black is aiming for central and king- side play by ...h5 and ...f5 - it's a completely different type of plan to that Black has adopted here. In this position... b6 is an inaccurate move, which reduces Black's options. His queen can no longer occupy b6 and in some lines where Black plays ...e5, it is useful to have ....:a6 de- fending the weak d6-pawn. 14....1d7 was probably better, al- though White has the interesting line 15 e5 e6 16 'ife3 dxe5 17 ':bdl! and Black still faces difficulties (if 17...f6, then 18 .1f3). 15 :reI Now White has a very comfort- able position. He has a space advan- tage and Black has no chance of playing ...b5 or ...d5, the two breaks which normally give Black counter- play in the Maroczy Bind. Moreover .. .a4 is no danger, as Black cannot back it up by .. .1i'b6, so White can always reply b4. The only question is how White arranges to play h4-h5 to step up the pressure on Black's poorly defended kingside. 15 ... .1b7 16 ':bd1 The rook has done its duty on bl, inducing Black to weaken his queen- side. Now the need is for play in the centre and in the changed circum- stances White reacts by moving his rook to the half-open d-file. 
ANAND - MOROVIt FERNANDEZ, NOVI SAD OLYMPIAD 1990 51 16 ... f6 17 i.g4 A nice move, activating White's bishop. He doesn't mind exchanging bishops, because in the Maroczy Bind White doesn't need many mi- nor pieces to prosecute his advan- tage. Black can normally hold a pure major-piece position, but even one pair of minor pieces can be enough for White to exploit his space advan- tage. 17 ... i.c6 18 h4 1ic7 19 1ie3 (D) 19 h5!? is possible, but I didn't want to commit my pawns to light squares too quickly; Black might still set up some sort of dark-squared blockade by continuing ...g;g7, ...g5 and . ..h6. .. . -.. . mu  Ba B. _ _ ... i. .. - - -  / . . . N _ _ _ .8.8.i.8 .8"  . .  -  8. . 88. . .:    M  19 ... 7?! Black should play 19...G;g7!, when after 20 h5 g5, followed by .. .h6, Black has set up the sort of blockade mentioned above. In this case White would have to prepare h5, for exam- ple by playing lDd5 or f4. It's only a small point, but in such positions you have to put as many obstacles in your opponent's path as possible. 20 h5 White can push his h-pawn with- out more ado, because ...g5 can al- ways be met by h6, both creating a permanent danger to Black's king and making f4 much stronger. 20 ... :f7 (D) Now it is too late for 20... G;g7 as White can play 21 h6+ G;h8 22 f4, preventing ...g5. .. . ... W .... '/... i. .. - - -   - . .8 - - - - .8.8.i.. .8  .   -  8. . 08. . .:  - - -  21 lDd5 Preparing to step up the pressure on g6 by lDf4. 21 hxg6 would be premature; Black can defend after 21...hxg6 221ih6 :g7. 21 ... g5 22 h6 As intended. Now the g5-pawn is vulnerable to f4, and Black has to 
52 VISHY ANAND: My BEST GAMES OF CHESS look after e7, as the rook can always be driven away by i.h5. 22 ... cit>h8 23 a3 Black's position on the kingside is creaking and now White inconven- iences him further by resuming his queenside play. 23 ... 24b4 25 axb4 26 exd5?! A mistake, not because it doesn't win but because 26 bxc5! was much simpler. If Black takes on c4 or e4, White obtains a tremendous passed pawn on d6, while after 26...d4 27 J:txd4 dxc5 28 J:td6, followed by e5, Black's king will come under a dev- astating attack. 26 ... e6 axb4 exd5 i.a4 (D) . . . - .. . . - w..B B.B' _ 0 / /% _ 0 / 0  / /        // / //// !/,: %-/ /  //  . 8.  ___/,i "'08B BiLB . . ruu . . . - . B _ 8B . .:  - 27 bxcS? Seduced by beauty, I allowed my- self to be distracted from the process of gaining a point! I had seen the winning move 27 J:tal!, whereupon 27.. .li:Jd7 28 J:txa4 J:txa4 29 'ii'e8+ ltJf8 (29...J:tf8 30 'ii'e7 forces mate) 30 'ii'xa4 is hopeless for Black, but I unwisely decided to 'win' in more flashy style. 27 ... 28 c6 i.xdl 'ii'e7! (D) .B B B , wB _ ...  . / /% A . / _ %/       /, , Y,.,,, %: /,,"' , . .8M . A __ .. .a..  r ;' '/ - . .  . . . 88. . ....   , w; / / ,- // Somehow this move had escaped my attention. White still has the ad- vantage because of his powerful pro- tected passed pawn, but of course I had spoilt my winning position. 29 1ic3 White has a slight advantage after 29 J:txdl 'ii'xe3 30 fxe3 f5, but the pawns appeared so strong that I thought I could play on with a queen against two rooks and a bishop. 29 ... 'ii'xel + 30 'ii'xel i.xg4 31 'ii'b4? Another error. 31 'ii'e4! i.h5 32 'ii'e6 wins the d-pawn (32...J:td8? 33 c7 wins) without allowing Black to 
ANAND - MOROVIt FERNANDEZ, NOVI SAD OLYMPIAD 1990 53 take the c4-pawn in return, which should be enough for a clear advan- tage. 31 ... 32 'ii'xd6 ..te2 ..txe4 (D) .B / , . .  $$.- rff{@  _._ & w . .AB. /01i . :'-: /% 8  /. '// I  /  '/:1 '  /     ///// , /// // ///  P .8.  I- R.i.. . . % / _ ;@0 ... . . W!J8B rt.  %:$; h' /  If 33 e7?! 33 'ii'e6 is a better try: 1) 33..J:tff8? 34 'ii'e7 wins im- mediately. 2) 33...g8 34 c7 ..txd5 (or 34.....ta6 35 d6 ..tc8 36 'ii'e8+ 1:tf8 37 d7 ..txd7 38 'ii'xd7 and wins) 35 c8'ii'+ (35 'ii'xd5 1:te8 36 'ii'd8 is also effective) 35...1:txc8 36 'ii'xc8+ 1:tf8 37 'ii'd7 ..tf7 38 'ii'e7 wins. 2) 33...:af8! 34 'ii'e4 ..ta6! (a dif- ficult move to see; after 34.....txd5 35 'ii'xd5 Black can't immediately double rooks against the pawn, for example 35...1:tc7 allows 36 'ii'd6) 35 d6 i.c8 36 'ii'd5 g8 37 d7 ..txd7 38 cxd7 1:td8 39 'ii'e6 <it>f8 and Black draws. 33 ... 1:te8?? A time-trouble blunder. Black could force a draw by 33...1:taf8! 34 'ii'c6 ..txd5 35 'ii'xd5 1:txc7 36 'ii'd6 1:tcf7 37 'ii'xb6. 34 'ii'e6! Grabbing my chance. 34 ... :iTS 35 e8'ii'! A pretty win. 35 ... :Xe8 36 'ii'e7 1-0 In April 1991 I played in a tournament held in Munich. While I very much liked the city, I have less happy memories of the chess. In fact, my greatest pleasure was the blitz tournament held at the end, which I won with 14/15, 2 1 /2 points ahead of the next player. In round 1 I lost to Nunn, and in round 2 I was fortunate to defeat Zsuzsa Polgar. In round 3 I met Beliavsky, and the result was one of the few good games I played at Munich. 
Game 10 A. Beliavsky - V. Anand Munich 1991 Pirc Defence In our previous encounter (Linares 1991) I had built up a totally won po- sition only to perpetrate a form of hara-kiri. I was very pleased to get revenge in this game! 1 d4 d6 2 e4 lbf6 3 lbc3 g6 4 f4 i.g7 5 lbf3 0-0 6 i.e3 b6 (D) .-  . B ..... - . -  -i-i w. . . _   - . . . .i B B _ . B ft8ft . . U U . _  rlliJ. 8 ft 8B .8 ft u d . u II ..=..: When I was studying this line from White's point of view, I won- dered why 6...b6 wasn't a more popular reply. Black forces through ...c5, and if White is to try for an ad- vantage he has to push all his pawns forward in the centre, which is very committal. If the pawns eventually turn out to be weak, the weakness will probably be serious. 7 e5 lbg4 8 i.g1 c5 9 h3 White avoids a little trap: 9 dxc5 bxc5 10 _d5 'ir'b6 11 'ir'xa810ses to 11...i.b7 12lbd5 'ir'xb2. 9 ... lDh6 10 d5 i.b7 (D) ._ mt1  . . -. . - . . i-i w..... . _   Bi- . . . . . 8 . . . U . . . 0 . _ [1 .liJ.8 808. .8.  Bli': - .  - I knew the theory, but at this point I decided to ignore it and just look at the position. It seemed to me that Black could play very natural moves. The point of this one is to play. ..e6 and completely destroy White's cen- tre. After the resulting exchanges 
BEUAVSKY - ANAND, MUNICH 1991 Black may be left with a weak pawn (for example, on e6) but it doesn't matter because Black has generated so much active play for his pieces. While the specific move 10.. ...tb7 was thought up over the board, I had looked at these lines before and the ideas I had during this earlier analy- sis germinated into this 'innov- ation' . 11 1id2 After 11 1ie2 ..ta6 (11...a6 and 12...b5 is also possible) 12 1if2 ..txfl 13 1ixfl lDf5 Black equalizes comfortably. If White tries to shut the h6-knight out by 11 g4, then Black plays 11.. .dxe5 12 fxe5 e6 13 i.c4 exd5! (13...b5 14 ..txb5 exd5 15 ..txc5 favours White) 14 ..txd5 with an unclear position. The knight on h6 is bad, but the e5-pawn is weak and gl-bishop is also oddly placed. 11 ... lDf5 12 ..th2 After 12 ..tf2 die5 13 fxe5 e6 White doesn't have time to castle long owing to 140-0-0 ..th6. Hence the text-move. 12 ... 13 fxe5 14 0-0-0 Other moves are ineffective, e.g. 14 g4lDh4 is very bad for White, 14 d6 lDd7 leaves e5 collapsing and fi- nally 14 ..tc4 ..th6 (14.. .exd5 15 ..txd5lbc6 and 14...b5 15lDxb5 exd5 16 ..td3 are also possible, in both cases with an unclear position) 15 dxe5 e6! 55 ..tf4 ..txf4 16 1ixf4 exd5 17 0-0-0 d4 18 lDe4 lDd7 leaves White with- out enough for the pawn. 14 ... exd5 15 lDxd5 lDc6 16 c3 (D) Underestimating Black's initia- tive. If White had time for ..tc4 and ':hel then he would have a clear ad- vantage, but Black's counterplay is so fast that he has no time for the necessary consolidation, e.g. 16 ..tc4 lDcd4 17 lDxd4 cxd4 attacking the d5-knight and, when it moves, creat- ing the possibility of .. .lDe3. .. - -.. .t.. ... B. B B _ JI1\. ... ."'B B B . RJI1\.  .'tJU"' . . . . . R ..8 B U B'tJB 8 R  .8 U _ . _ . =:..i..: 16 ... lDcd4 17 lDf6+ 17 cxd4 1ixd5 18 bl ':ad8 is very good for Black. 17 ... ..txf6 18 cxd4 18 exf6lDxf3 19 gxf3 1ixf6 gives Black a safe extra pawn. 18 ... ..tg7 19 d5 (D) 
56 VISHY ANAND: My BEST GAMES OF CHESS .B  . .. %0  U/ /  .t. ... B _ d _ . . ... //// ?W . /  /.. - - //  - - . - . . . .ltJ. A  8 / /  ; //   /J: A 0/ +% Q    .Q0  /,. /  //.  /' / / / / / . :.i..: _  _ ;I/&i 19 ... c4! Essential. If White could play .i.c4, then Black would be in some trou- ble. Now Black locks in the bishop on fl and frees the c5-square for the manoeuvre ...%lc8-c5. All Black's moves are very natural; he has the initiative and must make use of it be- fore White completes his develop- ment, so he need only consider active moves. 20 .i.e2? In the critical position White goes wrong. He should play 20 <ifi>b 1, with the possible continuation 20.. .%lc8 21 d6 (21 .i.e2 ..th6 22 .i.f4 .i.xf4 23 'ii'xf4 ..txd5 transposes to the game) and now Black can choose between 21...'ii'd7 and 21...b5. Black has a lot of trumps, not least of which is his lead in development. The position is unclear but I prefer Black. 20 ... %lc8 21 <ifi>bl 21 g4 li:Je7 22 d6 li:Jd5, threaten- ing 23...c3, favours Black. 21 ... .i.h6 22 .i.f4 After 22 'ii'c3 Black can also safely take the d5-pawn. 22 ... .i.xf4 23 'ii'xf4 .i.xd5! (D)  ill11 _ .  a.. _   W{$ f.&  . / .& - < @.... W ///U/ : Iffj% $ . ;/- / ;.  /. / .&  . . /   ..- /"u  ::::-/$ // //// W  8.-\"1&\. & ;&.JLU"_ w% . illit W' j'. - . . . _ltJ. D .iL._ ,,- ." '&'.. d 24 h4?! There is no way White can exploit the d- file pin, so he tries for a king- side attack. The alternatives were: 1) 24li:Jd4 c3 (24...li:Jg7 25li:Jb5! is bad, but 24...'ii'h4 25 .i.g4 li:Jxd4 26 %lxd4 %lcd8 is a reasonable alter- native for Black) 25 bxc3 'ii'h4 26 ..tg4 li:Jxd4 27 %lxd4 ..txg2 28 %lgl .i.xh3 29 .i.xc8 'ii'xf4 30 %lxf4 %lxc8 is excellent for Black. He has two connected passed pawns for the ex- change and White's remaining pawns are weak. 2) 24 %ld2 c3 (alternatively, after 24...%lc5 25 %lhdl 'ii'a8 Black keeps his extra pawn) 25 bxc3 %lxc3 26 %lhd lli:Je3 and Black is clearly bet- ter. 
BEUAVSKY - ANAND, MUNICH 1991 3) 24 g4 '£Jg7 25 1Vh6 '£Je6 26 h4 c3 27 bxc3 (27 '£Jg5 c2+ 28 <it>al cxdl1V+ 29 :xdl '£Jxg5 and Black wins) 27.. ...te4+ 28 <it>b2 1Vc7 fa- vours Black. White's pawns are weak and his king exposed. 24 ... c3 25 bxc3 .:xc3 26 h5 '£Je3! Black's attack is much faster than anything White can muster. 27 '£Jg5 57 After 27 hxg6 fxg6 28 1Vh6 1Vc7 Black defends h7 and White will be mated, while after 271Vh6 '£Jxdl 28 hxg6 (28 ltJg5 ..txa2+ forces mate) 28.....te4+ 29 <it>al ..txg6 Black wins easily. 27 ... 28 1Vc7 J:tb3+! '£Jxh7 0-1 It is mate after 29 axb3 1Wc2+ 30 <it>al 1Vc3+ 31 <it>bl 1Wxb3+ 32 <it>al '£Jc2# . A few days after qualifying from the Manila Interzonal, I received an invi- tation to play in Linares (1991) from Senor Rentero. On my way to Linares (which was a couple of months before the Munich event mentioned above) I stopped off for a couple of days in Amsterdam. There I received the news that FIDE had made the pairings for the quarter-finals. Instead of the simple 1 vs 8, 2 vs 7, etc., they had changed the system such that anyone in the top half could be paired against anyone from the bottom half. Karpov, Timman, Yusupov and Short were in the top half of the draw, with Gelfand, Ivanchuk, Korchnoi and myself in the bottom. I was paired against Karpov, who at that time was a formidable opponent. Just at that moment I was quite annoyed by this pairing, but later I took the view that you couldn't become World Champion by avoiding people - you just have to take oppo- nents as they come. I was heartened by the fact that I beat Karpov quite easily in Linares after he misplayed a promising position. Between then and the match in August neither of us had produced any inspiring results. My own performance in Munich was not very satisfactory, while Karpov had drawn a match 2-2 with Agdestein. In the first game of the match itself I played an insipid system; to be honest we (my second in this match was M. Gurevich) knew that it gave White noth- ing against best play, but we decided to try it anyway, because Karpov had failed to find the correct solution in a previous game. He got a bad position with an isolated pawn and suffered a lot, but defended very well and, indeed, outplayed me completely. However, he threw away all his good efforts by misplaying the ending. 
58 VISHY ANAND: My BEST GAMES OF CHESS In the second game I outplayed Karpov completely, but then went wrong and had to acquiesce to a draw. In the third game I could have mated him in a few moves, but somehow just didn't see it. Then I lost the fourth game. In both. game two and game four I had played the Meran Defence, which I pre- pared especially for this match; I felt it was a dynamic opening and that Kar- pov wasn't particularly good against it. In games two and four I had played, after 1 d4 d5 2 c4 c6 3 lDf3 lDf6 4lDc3 e6 5 e3 lDbd7 6 .i.e2 .i.d6 7 0-0 0-0 8 'ir'c2 dxc4 9 .i.xc4, a plan involving ...a6. In game four Karpov had found a pretty good line against this system, so in game six I decided to switch to 9...'ir'e7, which also formed part of my preparation. I should add that game five was unfinished when the following game was played, but I was winning the adjourned position. 
Game 11 A. Karpov - V. Anand Candidates match (6), Brussels 1991 Semi-Slav 1 d4 2 c4 3 lDf3 4 lDc3 5 e3 6 _c2 7 .i.e2 8 0-0 9 ..txc4 10 h3 In game eight he finally found the right recipe, which is to play 10 a3. 10 ... c5 In game four I had played 10...a6. 11 dxc5 .i.xc5 12 e4 (D) d5 c6 lDf6 e6 lDbd7 i.d6 0-0 dxc4 _e7 .BJ.B ..B &.mu&_& B_."_._. _ Bill B B _ B B BB8B B B  BtbB8 8.B 8B u dud   .:  - .  12 ... .i.d6 13 lDd4 To avoid the exchange of knights after ...lbe5, but it is not a very ambi- tious move. 13 ... lDe5 14 ..tb3 ..td7 15 ..te3 (D) .B _ ..B B _iLl_i_i B _ill B _ B 1.1 B B 8_ _ _ . _8 8 W \WJB W8B uwd u d  B B: - - - * 15 ... lDg6 15...:td8? is a loss of time; White continues 16 f4lbc6 17 e5 lDxd4 18 ..txd4 i.c5 19 :tad 1 (after 19 i.xc5 _xc5+ 20 _f2 _xf2+ 21 :txf2lDe8 22 :dl ..tc6 23 :fd2 :Xd2 24 :xd2 f8 Black should hold on) and now: 1) 19.. ...tc6? 20 exf6 :Xd4 (White also wins after 20....i.xd4+ 21 :xd4 'ii'c5 22 lDe2!) 21 fxe7 :xdl+ 22 h2 :xfl 23 _d3 ..tgl+ 24 hl and White is winning. 
60 VISHY ANAND: My BEST GAMES OF CHESS 2) 19.. ...txd4+ 20 :xd4 ltJe8 21 :fdl and White is clearly better. After the text-move Black has more or less equalized. 16 :&d1 :rd8 17 lbf3 ..tc6 18 :rei :dc8 This looks a little odd, but I felt it was important to inconvenience White's queen by playing a rook to c8. The a8-rook can't go there be- cause the a7-pawn is hanging, so it has to be this one. White must move his queen because 19.. ...txe4 is a threat. 19 1fb1 (D) .aKa ... .. . ' I /' B uu!/j  _ ///u!/j . i.' . "//..  - .  . . . . . .. . .ii. mltJ. " . ". U  U  .... :  /; . -   19 ... ..tb4 Everything is based on threats against the e4-pawn. 20 ..td2 After 20 ..td4 e5 the e-pawn would be in serious trouble. In this position only White can be worse, because of his exposed e-pawn. 20 ... :d8 The rook returns now that White has been forced to block the.d-file. 21 a3 ..tc5 22 ltJa4 ..td6 (D) Black is more or less committed to this piece sacrifice, as 22.. ...txa4 23 ..txa4 would give White a slight advantage based on his two bishops. However, I had no objections as I felt that it was promising for Black. .. - ... E.. ... W %"U  _ uu .A&_1&\. ..JL ;-...& ... .  . fI; . . A p$% t  _  Q iJ};;:/, ::;;:f'$ Dii. .ltJ6  %" . (5 &' A r /. ,:;; /, ,:;; Q /;;:p % ,// , / /, . // ;. /;/ \Wr/  /0 .B.:r   23 ltJc3? After this cop-out, Black's pieces are more harmoniously placed. The critical line is 23 e5! ..txf3 and now: 1) 24 exf6 (bad) 24...'ii'xf6 25 gxf3 ..tf4! and now: 1 a) 26 ..tc3 'ii' g5+ 27 h 1 (27 fl 'ii'b5+) 27...'ii'h5 28 g2ltJh4+ and wins. 1 b) 26 ..txf4 ltJxf4 also wins. lc) 26..tb4 'ii'g5+ 27 fl 5+, followed by 28...a5, with a distinct plus for Black. 2) 24 exd6 (best) 24...'ii'xd6 25 gxf3 h4 and now: 
KARPOV - ANAND, BRUSSELS CANDIDATES 1991 61 2a) 26 .tc3?! ltJxf3+ 27 <itg2 ltJh4+ 28 <itfl (28 <itg 1 'ii'c6 is dead lost) 28...'ii'a6+ 29 J:e2 'ii'c6 and Black should win. 2b) 26 .te3? ltJxf3+ wins. 2c) 26 J:e3 (D) and now: .,  .  p . .  .& 0//// _ W//% _ i. .ii B . _  . illlia- . . .-. . . . . . lb. . .  - j2 /%% - Dil..  .   /  rJj  / ; . :    /I'/. - / V ,  ;'"  / .\WJ"p  da..... * 2cl) 26...'ii'f4?! 27 J:d3! ltJxf3+ 28 <itg2ltJh4+ (28...ltJxd2 29 J:lxd2 'ii'g5+ 30 <itfl 'ii'xd2 31 ':xd2 J:xd2 gives White an edge) 29 <itfl 'ii'h2 30 .tg5! (30 .tc3 ltJd5 31 .txd5 ':xd5 32ltJc5 is unclear) 30...J:xd3 31 'ii'xd3 'ii'g2+ 32 <ite2 ltJg6 33 'ii' g3 favours White. 2c2) 26...'ii'c6 27 .tc3ltJxf3+ 28 <iti>fl b5 with a final branch: 2c21) 29 J:xd8+ J:xd8 30 .tdl ltJd4 (30...ltJd2+ 31 .txd2 J:xd2 32 .tf3 defends, while 30...ltJh4 31 .txf6 gxf6 32 ':g3+ <ith8 33 ltJc3 'ii'hl + 34 <ite2ltJf5 35 J:d3ltJd4+ 36 <itd2 is at least equal for White) 31 .txd4 ':xd4 32ltJc3 'ii'hl + 33 <ite2 ltJh5 34 J:f3 'ii'h2 35 .tc2 ltJf4+ 36 <ite3 ltJg2+ 37 <itxd4 'ii'd6+ 38 <ite4 'ii'c6+ 39 <ite5 and this exciting line ends in a draw by perpetual check. 2c22) 29 ltJc5 ltJh2+ 30 <ite2 'ii'xc5 31 J:xd8+ J:xd8 32 .txf6 gxf6 33 'ii'gl + <ith8 34 'ii'xh2 'ii'cl is unclear. The conclusion is that White may be able to hold the balance by ac- cepting the sacrifice, but it would have been very difficult for Karpov to find all this at the board! 23 ... 'ii'c7 (D) If we look at the total effect of the last five moves, White has played .td2, 'ii'bl and a3, while Black has achieved ...'ii'c7. Thus White's queen and bishop have been pushed back, while Black has improved his queen position and now controls e5 - Black's manoeuvre can be counted a success. Now Black is slightly better be- cause he can expand on the queen- side, while in the meantime White's pieces are only crawling back to their earlier positions. .. - ... i .ii w. _ .  .J. / i .. W /,,,y _ _ . . . . . .. . 0J/Ji.  .lb.    - . " . fj .  "y . .jV.: /0 . . _ u 
62 VISHY ANAND: My BEST GAMES OF CHESS 24 i.a2 It's hard to say what this move is for. Perhaps he was worried that af- ter .. .i.c5, there might be a threat of ...i.xf2+ and ...Wb6+, but it's only a guess. 24 ... a6 2S i.e3 b5 White's 24th move is revealed as a mistake. It may have solved one problem, but now the bishop has to return in order to meet ... 'fib 7 by i.c2. 26 i.b3 i.b7 27 :cl 'fie7 (D) I was very proud of this game, be- cause I felt that I had outplayed Kar- pov in the manoeuvring phase of the game. The idea now is .. .ltJd7 fol- lowed by either ...ltJc5 or ...ltJde5 heading for c4. z_ _ _.. B.t. ii w. . _ _ i. -i-.. - - - - .i. . . - -. . i." B   . p . R.  .  . B"   . - -  28 i.b6 :dc8 29 i.d4 Now 29 e5 i.xf3 is good for Black, as 30 exf6 'fixf6 leads to variations similar to those in the note to White's 23rd move, while 30 exd6 'fixd6 attacks the bishop on b6 so there is no win of a piece. 29 ... ltJd7 30 :cdl Over the last few moves White has only been moving his pieces backwards and forwards, simply re- sponding to Black's various threats, while Black has gained space on the queenside and created an outpost at c4. Nevertheless, the symmetrical nature of the position exerts a draw- ish tendency. Black's position is more comfortable and easier to play, but one cannot say more than that. 30 ... ltJge5 31 ltJxe5 ltJxe5 (D) z_z_ ... ..mum w..t. _i_i i. _i. . .i. (I . . -. . "i. . .   . .  . ". U .  . .\WJ.:  ... -  32 ltJe2? An error, overlooking the reply. 32 ... 1i'h4 33 f4 After 33 ltJc3 Black continues 33...ltJf3+ 34 gxf3 'fixh3 35 :d3 
KARPOV - ANAND, BRUSSELS CANDIDATES 1991 63 .th2+ 36 <ithl .tf4+ 37 <itgl :c6, and the possibility of playing ...e5 and switching the rook to the king- side gives Black an easy win. The al- ternative 33 lbg3 is similar; then 33...lbf3+ 34 gxf3 i.xg3 35 <itg2 .tf4 doesn't lead to a winning at- tack, but White's pawn structure is significantly damaged. The text-move is therefore forced, but it is a move White certainly doesn't want to play with Black's bishops pointing at the kingside. 33 ... lbc4 34 .tn .tc5 35 .txc5 .:xc5 (D) .a a a.a wa.i.a a._. .a a.a a a._ a a .Jk\.  d..d u _ Di.a _ _  ... u _vu_ _ .\Wr.:  dwd _  36 :cl? 36 :d4 was a better defence, when 36...lbxb2? loses to 37 :fllbc4 38 .txc4. However, Black can continue 36...:ac8 with a clear advantage. 36 ... lbd2 37 'ir'd3 lbxe4?! One of my weaknesses during this match was my inefficiency in converting technically winning po- sitions. Too often I played moves which just maintained my advan- tage, instead of pressing it home forcefully. On the other hand, Kar- pov, although he only gained a large advantage twice, pushed it home both times, and this effectively de- cided the match. Here 37...:d8! 38 'ir'e3lbxb3 39 'ir'xb3 .txe4 would be a much im- proved version of what happens in the game. 38 'ir'e3 .:ac8 39 :cdl h6 39.. .g6 would have been slightly more precise. 40 :n 41 lbc3 lbf6 1i'h5 (D) ..a ..a . .i. . .. W_ _ _ _ .. ..-  d d _ _ a._ . .... _ . D . "  . u"' _ .  . .. u _ _ _ . .:.: - - -  Black has a clear extra pawn, and despite White's stern resistance, the win should only be a matter of time. 42 :d6 :5c6 43 .:xc6 .txc6 44 m lbd7 
64 VISHY ANAND: MY BEST GAMES OF CHESS 45 :d2 'fic5 I decided not to try to find a mid- dlegame win, but just to liquidate down to an ending. 46 'fixc5 ltJxc5 47 .tc2 <itf8 48 <ita rt;e7 49 <ite3 lbd7 50 g3 ltJf6 51 .tdl lbe8 (D) At some point I should play ...g5, so that if White plays h4, then Black can reply ...gxh4 followed by ...:g8, when White has the possibility of a g-file penetration to worry about. BZB. B . . '/.;, ' W  . /.  n  .8.t....  -    B'B B . . .  . ' '/       %   ,  /. ,.  , , " " . . U _ .  . .i.. . 52 .tc2 lbd6 53 .td3 .tb7 53...g5 would still be good, but this is the last chance as White now prevents it (of course, he could have played h4 earlier himself). 54 h4 Now any attempt to prepare ...g5 will just lead to a lot of simplifica- tion. 54 ... ltJc4+ ?! A real mistake. Again, it -doesn't give away Black's advantage but it makes the win fairly complicated. In endings with bishop and knight each, exchanging dissimilar pieces (i.e. a bishop for a knight) has the effect of increasing the defender's drawing chances. The superior side should therefore only exchange simi- lar pieces. 55 .txc4 .:xc4 Now White has much better pros- pects of setting up a dark-squared blockade. 56 ':d4 :c5 (D) . . . . E.t.. '0,  W . ,  . .   . . .       .1._ . .  . a ' 'i D '     %  /. f:f$; /  /'n  /",    %'0 D '  /. . . / %  %  /. ,  /'/n  - - . . . . . 57 a4 At this point I decided to sit down and really calculate, because I could see that the win was no longer going to be trivial. 57 ... e5 Over the next few moves Kar- pov's resistance starts to weaken. 58 ':b4 
KARPOV - ANAND, BRUSSELS CANDIDATES 1991 65 Already 58 fxe5 ':xe5+ 59 <iti>f2 was more accurate, when Black has a long way to go to create a passed pawn on the kingside. 58 ... .tc6 59 axb5 axb5 60 ll)e2? (D) This was White's last chance to play 60 fxe5 ':xe5+ 61 <iti>f2. . . WJffi . /" all////  ;i ffi: . i B.  _ _ -.... . . ;i / .    /  a u  %' /  $j    w, "'/ . .  " . . * u 8 BtlJ. .   %/;0;:: ;ffi  // . ' -0% / (i' f: ;'/ }$;/ $'.- 60 ... f6! Now Black retains control of the key dark squares d4 and f4, and the white knight will feel a lack of good squares. 61 ':b3? Another bad move. White should have tried 61ltJc3. 61 ... <iti>e6 62 .:aJ ':c2 Now Black is winning. 63 fxe5 fxe5 64 .:&6 The only line I needed to calcu- late was 64 ':c3 ':xc3+ 65 ll)xc3 <iti>d6 66 b4 <iti>e6 67 <iti>d3 <iti>f5 68 <iti>e3 <iti>g4 69 <itf2 e4 70 ll)dl <iti>f5 71 <iti>e3 <iti>e5 72ll)c3 g6 (but not 72...g5? 73 hxg5 hxg5 74 g4 and White escapes with a draw) 73 g4 g5 74 h5 .td7 and Black wins. 64 ... <iti>d6 65 b4 65 ':a7 ':xb2 66 ':xg7 b4 is also an easy win. 65 ... 66 .:&7 67 .:&6+ 68 ':g6 69 ':d6 70 ':d5 ':c4 i.d7 <iti>e7 <iti>f7 .tg4 ':c2 0-1 As the knight is trapped after 71 ll)g 1 ':g2. This game was finished after the conclusion of the fifth game; the two ad- journments were played on the same day. Unfortunately, in the adjournment of the fifth game I first made a simple win rather complicated and then, when I needed to calculate a long forced line in order to win, I made a mistake that allowed him to gain a tempo. The game ended in a draw. Here Karpov's expe- rience also played a part. He had two lost positions (games five and six), but he concentrated all his efforts on finding resources in game five (where he had a better chance to save the game) and eventually managed to salvage a half-point. 
66 VISHY ANAND: MY BEST GAMES OF CHESS I missed a forced win in the seventh game and then he won the eighth game to take the match. Certainly he defended very well in this match, but he didn't demonstrate the form which would have taken him to the world champion- ship, and he subsequently lost to Short in the semi-finals. The two matches I played in this Candidates cycle were virtually the first matches I had ever played. Before, I had only played a not especially serious four-game match with Levitt. When the next Candidates cycle came around, the experience of match play that I gained against Dreev and Karpov turned out to be very useful. The 1991 Tilburg tournament was a double-round event. Prior to this tour- nament I had only played Kasparov once, the game ending in a draw. In the first cycle at Tilburg I lost, but gained my revenge in the game immediately following. I won again in Reggio Emilia (see Game 13) but, to date, that was the last I saw of a plus score against him! 
Game 12 V. Anand - G. Kasparov Tilburg 1991 Sicilian, Scheveningen This was quite a pleasing game. In fact, I hadn't really bothered to pre- pare for this game - I decided that whatever I did, it would be inade- quate. He'd played this line so many times I couldn't hope to out-prepare him, so I preferred to concentrate on keeping a clear head for the game. Although the strategy worked well on this occasion, it would be easy to exaggerate its advantages! 1 e4 c5 2 lbf3 d6 3 d4 cxd4 4 lbxd4 lbf6 5 lbc3 a6 6 f4 e6 7 .1d3 lbbd7 I later discovered that Kasparov and Nikitin's book on the Scheven- ingen gives 'The position after 8 0-0 iib6 9 .te3 'ir'xb2 10 lbcb5 axb5 11 ltJxb5 :a5 is interesting for analy- sis', but during the game I wasn't aware of this. At the board I was considering 8 0-0, and my analysis ran '8 0-0 'iib6 9 .1e3 'ir'xb2 and there must be something strong.' Af- ter I played 8 0-0, he instantly flashed out 8...'iib6 and I thought for a while, during which I saw that it was not so easy. 8 0-0  9 .te3 'ir'xb2 Here I realized that I had to sacri- flee on b5, or else I would just be a pawn down for nothing. 10 lbdb5 (D) 10 'ir'd2lbg4 is clearly better for Black. ...t. .   . d _ _ .'.Jk\.''' B. ....   .. -.- . .tb_ B _ q B8' . ; ?' ",  ; . i. . .  -  I\I\. .I\ o_o .oU  .\WJ.: _ ..W  10 ... axb5 11 lbxb5 .:as The line which had convinced me to play 10 lbdb5 was 11...'ir'b4 12 lbc7+ <itd8 13 lbxa8 'ir'a5 14 lbb6 ltJxb6 15 'iiel! 'iia7 16 a4 lbg4 17 a5 and White wins. However, Black can improve on this by 12...<ite7! 13 
68 VI SHY ANAND: My BEST GAMES OF CHESS ltJxa8 'ii'aS which leads, just like the game, to a complex and unclear po- sition. Other moves are bad, for example 11...d8? 12 a3! (threatening 13 .td4) 12...:a4 (12...d5 13 :bl 'iia2 14 :b3! followed by lbc3) 13 c4! and wins, or 11...:b8? 12 :bl 'iixa2 13 :al 'iib2 14 .td4 'iib4 15 c3 trap- ping the queen. 12 :bl Better than 12 a4 :xb5 (12...d5 13 :bl 'iia2 14 :al is a draw) 13 axb5 'iic3, which is better for Black, e.g. 14 :a8 .te7 15 'iiel 'iic7 and 16 b6 ltJxb6 17 .txb6 fails because Black takes on b6 with check. 12 ... .:xb5 Not 12...'ii'xa2? 13 lbc3 'ii'a3 14 :b3 and Black does not get enough for the queen. 13 .:xb2 .:xb2 (D) ..t..  . - - - wB'B..'_' n . . .-; - . B B . 88" . W{{.U_ _ Bi.m . 8.8. B8B . ...: - - -  14 'ii'al I was quite surprised that Kaspa- rov went in for this, but I was sure it was still analysis as he was playing very fast. 14 ... :b6 Not 14...:b4? 15 'iic3 and wins. 15 .txb6 ltJxb6 16 'iic3(D) This move was the product of long thought. I realized that any other move would allow Black to re- group his b6-knight to c5. Then it is hard to judge whether the queen or the three minor pieces would be bet- ter, but Black's solid pawn structure gives him a head start. I therefore de- cided that it was necessary to keep the knight fixed on b6 as a target, and force the other knight to occupy d7. ..t. .   . . - - B.'. _.r'. . ' '/ & II /.j ":/.  . W4  . ' - . /,  /'!@d B . . . A .  - . A D ' '/    _ / 0   - ;;/ /$h    . i.. . ---; 8_8. .80 . . .: -- 16 ... .te7 Alternatively: 1) 16...d817'iiaSlbfd718:bl and now: la) 18...c7 19 .tb5 b8 (White wins after 19.. .ltJc5? 20 .1e8) 20 .1xd7 ltJxd7 21 'iid8 g6 (21...d5 22 
ANAND - KASPAROV, TILBURG 1991 fl dxe4 23 ':b3 lbc5 24 ':c3 is again winning) 22 hl! .i.g7 23 'ir'e7 c7 24 ':dl and White is clearly better. 1 b) 18...d5 19 exd5 exd5 20 .i.f5 .1c5+ 21 fl e7 22 .1xd7lbxd7 23 'ir'c7 f6 24 ':dl d4 25 c3 also fa- vours White. 2) 16...lbfd7 17 ':bl d5? (Black can transpose into the game by 17....1e7) 18 ':xb6 and White wins after 18....1c5+ 19 fl d4 20 'ir'c4 or 18...lbxb6 19 'ir'c7. 17 ':b1 (D) B"'_.. _ ... '... B.  , , /,oo . '  '/ .  . / , /,o   - - - - . .' B _ _.18 B BB _D B:. . = 17 ... lbfd7 Black can get castled at the cost of his d-pawn by 17....1d8 18 'iid4 lbbd7 (18...lbfd7 19 'iixd6 also fa- vours White) 19 'ii'xd6 .1e7 20 'iic7 0-0 21 <itfl, but White retains a slight advantage. 18 'iixg7 Now 18 'iic7 fails to 18...0-0! 19 :xb6? .1d8 20 'iixc8 .1xb6+. 18 ... .1f6 69 Now that Black cannot castle, he finds it hard to coordinate his pieces. 19 'iih6 (D) R...... : BB'...'.' . /. ' /:  .  /% .  ' . /. ' , //. , %j&  / , /, ., " ; /.,  . B . .  . n  . .U . 8 8.1. . 8. . .:-   - - -  19 ... e7 The alternative is 19....:g8 20 e5 dxe5 and now: 1) 21 .1xh7 with two lines: la) 21....:h8 22 ':xb6 lbxb6 (if 22...e4, then 23 ':b4 e7 24 ':c4 wins) 23 'ir'xf6 ':xh7 24 'ir'xe5lbd5 25 c4 lbe7 26 a4 with the plan of 'ir'c7, c5-c6 and pushing the a-pawn forward to queen. Of course, matters are not so simple as this, but White has the advantage. Ib) 21....:g4! 22 ':xb6 ':h4 23 ':xe6+ fxe6 24 'ir'g6+ e7 25 .1g8 ':xf4 26 'ir'f7+ d8 27 'iixe6 is un- clear. 2) 21 ':xb6! e4 22 ':b4 (22 .1b5 .1d4+ 23 fl .1xb6 24 'ir'xh7 f8 25 'iixe4 is unclear, but 22 ':xe6+ fxe6 23 .i.xe4 should be slightly bet- ter for White) 22...exd3 23 'ir'xh7 with advantage to White. 
70 VISHY ANAND: My BEST GAMES OF CHESS 20 i.b5? Missing a stronger possibility: 20 g4! (while not winning, this gives White an advantage) 20...:g8 (not 20...i.d4+? 21 <it>fl c5 22 :b4, nor 20...c5? 21 e5 dxe5 22 g5 and White wins in both lines) 21 g5 i.g7 (21.. .i.d4+ 22 <it>fl f8 is possible, but still better for White) 22 1Wh4 (D) (not 221Wxh7?? i.d4+ 23 <it>fl :h8 and Black wins) and now: ..i.. BI.  . .  . Jk\' . B  .._ _ . . .   .  - -   . . . D  .   . p    U _ . .ii.B . B. . r  .  : .   .   -    1) 22...h6 23 <it>fl and now both 23...hxg5 24 1Wxg5+ and 23...a4 24 e5! dxe5 25 fxe5 favour White. 2) 22...a4 23 <it>hl (23 :b4 ac5 24 1Wxh7 <it>f8 251Wh4 is also good for White since he has the e4- pawn well supported) 23...ac5 24 g6+ (24 1Wxh7 <it>f8 25 'ir'h4 :h8 26 1W g3 b6 is only slightly better for White) 24...i.f6 25 gxh7 :h8 26 1Wh6! (now the manoeuvre :gl-g8 will also introduce the possibility of 1Wf8+) 26...e5 (26...xd3 27 cxd3 f8 28 e5 dxe5 29 fxe5 .1xe5 30 1Wg5+ i.f6 311Wc5+ <it>d8 321Wd6+ lbd7 33 :gl is very unpleasant for Black) 27 :gl (27 .:fl f8 28 fxe5 .1xe5 is less clear) 27...exf4 28 :g8 e6 291Wxf6+! xf6 30 :xh8 and wins. 3) 22...e5 23 <it>hl exf4 and now: 3a) 24 g6+ .1f6 25 gxh7 :h8 26 1Wh5 i.e5! 27 1Wg5+ f6 28 :xb6 :xh7 29 :b5! (29 i.e2 f3 30 .1xf3 :xh2+ 31 <it>gl :xc2 is not easy for White - Black's pieces are very ac- tive) 29...:h5 30 :xe5+ dxe5 31 1Wgl .1d7 32 i.e2 (321Wc5+ <it>e8 33 i.b5 .1xb5 341Wxb5+ <it>f8 351Wxb7 <it>g7 is less clear as Black has coun- terplay with ... g4 or with his f- pawn) followed by either 1Wc5+ or 1Wg5, with advantage to White. 3b) 24 e5! (also strong) and now: 3bl) 24...dxe5 25 g6+ i.f6 26 gxh7 :h8 (26...:f8 271Wh6 d5 28 :gl i.h8 29 :g8 5f6 30 :xf8 xf8 311Wg5 g6 32 .1xg6 fxg6 33 1Wxe5+ wins) 271Wh5 (271Wh6 d5 28 :gl is also promising for White) 27...a4 (27...d5 28 .1c4lbe3 29 1Wxf7+ <it>d6 30 :b6+ is winning) 28 :gl ac5 29 .1c4 e6 30 .1xe6! <it>xe6 31 :g8 and White should win. 3b2) 24...i.xe5 25 g6+ .1f6 (the line 25...<it>e8 26 gxh7 :h8 27 :b5!! lba4 28 :xe5+ dxe5 291Wg5 is deci- sive) 26 gxh7 :h8 271Wh6, again in- tending :gl-g8, and Black faces serious problems. Kasparov pointed out many of these variations after the game. My 
ANAND - KASPAROV, TILBURG 1991 inaccuracy arose because I failed to appreciate that this was really a criti- cal position in which White had to continue very precisely, and not just play natural-looking moves. 20 ... :g8?! (D) Missing a simple chance: after 20...e5! (not 20...c5? 21 e5 dxe5 22 fxe5 .txe5 23 1Vg5+ d6 24 :dl+ ltJd5 25 c4 and wins) 21 :fl (after 21 f5 c5 Black has wrested control of some dark squares; 21 a4 ltJc5 22 a5 bd7 is also fine for Black) 21...:g8 22 fxe5 .txe5 23 'iixh7 :g7 24 'iih4+ ltJf6 Black has regrouped his pieces with a solid po- sition. B-*-a a-a wa'Ba_'B' II '  '/ _  '  iH /. ' . '/'   '0    , I'  Hi-B _ B . .A" . _ /ou _ .. - f)WR .. .u B:a . = 21 :dl! e5?! After 21...c5 White can gain the advantage: 1) 22 :xd6?! is inferior after 22.. .:g6 23 1i'h5! ltJxe4! (23... xd6 24 e5+ c7 25 exf6ltJbd7 261Vxh7 :xf6 27 g3 is slightly better for White) 24 :d3 d5 and Black's 71 active pieces provide him with suffi- cient counterplay. 2) 22 e5! dxe5 23 fxe5 .txe5 24 1Ve3 and now 24....td6 251Vd4 :d8 (25...d5 26 c4 wins for White) 26 1Vh4+ f6 27 1Vxh7+ f8 28 1Vh8+ e7 291Vg7# is mate, so Black must play 24...bd7 25 .txd7 xd7 26 :xd7+ .txd7 271Vxe5 .tc6 28 g3, when White has fair winning pros- pects. The best move is 21...:g4! (D), which was Kasparov's original in- tention, but when he was about to play it, he saw a hole in his analysis. However, it appears that the move is playable after all: a-*-a _ H .'.JI1\_'.' w. .... . - '. mu -  - - Hi-B a _ _ _D-a . . . . _/%/ AA. AP 0.0& .0U a _:a = 1) 22 g3? e5! 23 a4 exf4 24 a5 fxg3 25 axb6 gxh2+ 26 h 1 .te5 fa- vours Black - the h2-pawn is very strong. 2) 22 e5 dxe5 23 1Vh3 :xf4 24 'iia3+ e8 (not 24...d8? 251Vf8+ c7 26 1Vd6+) 25 'iic5 .td8 26 a4 and now: 
72 VISHY ANAND: My BEST GAMES OF CHESS 2a) 26...:f5? 27 :fl :g5 (White wins after 27...:xfl+ 28 xfl as Black cannot defend against a5) 28 h4! is very good for White. 2b) 26...e4 27 :fl! (now not 27 a5? :f5 281i'b4ltJd5 291i'c4ltJe7 and Black is better) 27...:g4! 28 h4! (28 as :g5 291i'f2 f6, intending 30 c4 ltJa8, is slightly better for Black) and now Kasparov goes into some spectacular lines, but he missed 28...e3!, which looks fine for Black, e.g. 29 a5 (29 1i'xe3 ltJd5 is very good for Black) 29...e2! 30 :el .tf6 31 c31i>d8! with advantage to Black. 3) 22 .te2!? (probably objec- tively best) 22...:g8 and White may have nothing better than to repeat moves. 22 f5 (D) 81.. ... ..... ..... .&._&.& B ...._... . "/./'/;  .  //  % ii /.' '/"-: /::1    /. V/ /. /. , ' %:i   / /.  /. /uu  " / , .i.. '. /. _ /uu/ 0;$/ . .. . . . . . .. .D . .:. / /. / %:0: _ _ ,,//. 22 ... ltJc5? It is now too late for this move. I was rather relieved when he didn't play 22...:d8!, after which White is still better, but the position remains complicated. The analysis runs 23 g4 c5 (23...:g8 24 h3 is a simple win) 24 g5 ltJxe4 (not 24....th8? 25 f6+) 25 gxf6+ ltJxf6 (Kasparov ana- lysed this in excruciating detail and decided it favoured White; however, when you go for your best practical chance, you shouldn't be too fussy!) 26 1i'e3! (26 c4?! :g8+ 27 fl .txf5 28 c5 dxc5 291i'e3 ltJbd7 and 26 :fl :g8+ 27 h 1 :g4! are less clear; Black has real counterplay in both cases) 26.. .ltJbd5 (26.. .ltJbd7 27 1i'a7! ltJc5 28 :xd6! :g8+ 29 fl xd6 30 1i'b6+ wins) 271i'b3 ltJc7! (27...ltJf4 281i'c4 heading for c7) 28 1i'c4 ltJxb5 29 'iixb5 :g8+ and Black can play on, but White has excellent winning chances. Other 22nd moves are bad, for ex- ample 22...:g4? 23 1i'd2 d5 24 i.xd7 ltJxd7 (24.. ..txd7 25 1i'b4+ and 24...xd7 251i'e2 are also win- ning for White) 25 1i'xd5 wins, or 22...ltJa8 23 .txd7! .txd7 24 :xd6 i.g5 (24...xd6 2S1i'xf6+ c7 26 1i'xt7) 25 'iixh7 and White wins. 23 .:xd6! Black's position crumbles. 23 ... .tg5 23...xd610ses to 241i'xf6+ c7 25 1i'xe5+. 24 1i'xh7 Now all the tactics work out for White. 24 ... ltJxe4 Or 24...xd6 25 1i'xg8 .te3+ (25...ltJxe4 26 'iixf7 and wins) 26 
ANAND - KASPAROV, TILBURG 1991 <ittfl ltJxe4 (26... <itte7 27 f6+ <ittxf6 28 'ii'd8+ and 26.. ..td7 27 .txd7 ltJbxd7 28 'ii'xf7 are decisive) 27 'ii'xf7 with too many passed pawns. 25 .:xb6 ':d8 25....te3+ 26 fl :g4 (26...:d8 27 'ii'h4+) 27 f6+ f8 (27...d8 28 'ii'xf7) 28 e2 would be fatal for Black. 26 .td3 .te3+ 27 n (D) 27 ... .txb6 27...ltJd2+ 28 e2 .txb6 29 'ii'h4+ <itte8 (29...<ittd7 30 xd2) 30 .tb5+ .td7 31 'ii'h8+ <itte7 32 'ii'xe5+ <ittf8 33 'ii'd6+ decides the game. 73 .t . ::'1 / /. "/@-/. '/;://);   ;:/ '0?JJ   / //'- // /1"" :f(:;-/ ' '/ ;{ .. f; _.. ;&, \Wr B .. %;5/ _ ..  .  J( '(;,;)   //  5@' %( ::.tJf:, ! A ;/<j:</ %i:: _ 0 r -//:/;///// ////u!0 "/:/"//// ... f P'% /:?:-:-f/ ii. . ;;::% /% : %'$// + /  :" % : ,  8.8. .8.'  . / . ':0 . . %: '// ..   :{    /:;;   / @ '@:}{ /:.:// @ ' // .',' :.o//- '.// './ ./;..0; O'/(-/:.//. , '/// //:::>x; ;> ,/,:j 0/,,0:;-: 28 .txe4 ':d4 29 c3 1-0 In view of 29.. .':xe4 30 f6+ <ittxf6 31 'ii'xe4. The result of this game was a pleasant surprise for me. At the time Kaspa- rov already had a huge lead over the other players, but now I had visions of overhauling him. I already had a winning adjournment against Kamsky in the bag, then I won this game, and the following day I had a winning position against Karpov. However, I lost the game against Karpov and then I blun- dered into a mate in two in the Kamsky adjournment, so to win the tourna- ment Kasparov didn't have to do anything other than to watch my mistakes. 
Game 13 G. Kasparov - V. Anand Reggio Emilia 1991/2 French Defence This game was played in the New Year tournament at Reggio Emilia, which at the time was the strongest tournament ever held and was the first to reach category 18. Nowadays this has become par for the course, but at that time it was something special. What was also special about this event was that I was the only player in the tournament who didn't speak Russian, the other nine participants all being from the Soviet Union (af- ter the fifth round, they were from the former Soviet Union!) - the event was effectively the last Soviet Championship. 1 e4 e6 I chose this because I didn't want to challenge Kasparov again in the Sicilian. He had already shown in Tilburg how well prepared he is for the Sicilian and I didn't see the point of provoking him again. 2 d4 d5 3 ltJd2 In Tilburg, he had started playing 3 exd5 and 4ltJf3 against the French and it was partly in the hope that he would repeat this insipid system that I chose the opening. 3 ltJd2 was a bit of a surprise. 3 ... c5 4 exd5 'iixd5 I had studied this line fairly exten- sively for the Dreev match. 5 dxc5 A very surprising move, which I had never seen before. It seemed quite unlike Kasparov to step out of theory into unknown territory. 5 ... i.xc5 6 ltJgf3 ltJf6 (D) _II"'... _ '.. ... w%  . _ B.'. . . \Wr. . d .-d d - - - - - - -- 8R8 R8 R      \WJ.: - _..... 7 i.d3 Here I realized that he wanted to get the type of set-up that often arises in the Rubinstein French (i.e. 3...dxe4) - White castles long and 
KASPAROV - ANAND, REGGIO EMIUA 1991/2 75 has attacking chances on the king- side, but Black has an extra centre pawn. 7 ... 0-0 8 1ie2 ltJbd7 Better than 8...ltJc6 9 ltJe4 .i.e7 10 0-0, with an edge for White. Here it is better to have the knights con- nected, and in some lines with ltJe4 and ltJxc5, Black can reply ...ltJxc5 and hit the bishop on d3. 9 ltJe4 b6 10 ltJxc5 1ixeS (D) 10...ltJxc5 is also fine for Black, e.g. 11 .tc41if5 12 .i.e3 .tb7 with equality. All Black's pieces are in play and the advantage of the two bishops is purely academic. .-...- -..  JI1\'' w_ .. _  .,- . -  -  - II _ _ - - - - _ _i.__ 8088_080    .: _ _ w d 11 .te3 White's idea is to put the bishop on the long diagonal and aim for ltJe5 followed by f4. If he could achieve this then he would have an advantage, but there just isn't time for it. The quiet 11 0-0 may be better. 11 ... 1ic7 12 .i.d4 .i.b7 13 0-0-0 Thanks to the loss of time with .te3-d4, Black can safely meet 13 0-0 by 13.. .ltJc5. The exchange on f6 is not dangerous as White's pieces are not active enough to achieve any- thing before Black plays ...h8 and ....:g8. 13 ... ltJc5! (D) .- - -.- .t .,, w _ . _ / ,- .  ,  - -  - II _ _ - - - -    .T:'\r . ....'U. 8"8B\WJP8" u dwu u . :. -: _ w  _ 14 .i.e5 After 14 .i.xf6 Black has a range of satisfactory options. 14...xd3+ 15 ':xd3 1if4+ 16 bl 1Wxf6 is the simplest possibility, but Black can even play for the advantage with 14...1if4+ (more accurate than the immediate 14...gxf6 15 1ie3 g7, although that is also playable) 15 bl gxf6!? and the active queen on f4 immobilizes White's queen, while the f6-pawn prevents ltJe5. 14 ... ltJxd3+ 15 .:xd3?! 
76 VISHY ANAND: My BEST GAMES OF CHESS 15 1ixd3 was better, leading to an unclear position after 15...1ic6 16 .txf6 gxf6 17 ltJd4. 15 ... 1ic4 16 ltJd4 Other moves are not very impres- SIve: 1) 16 bl?! .te4 17 ':e3 1ixe2 18 ':xe2 .txf3 19 gxf3 ltJd5 with a clear endgame advantage for Black. 2) 16 .txf6 1if4+ 17 bl 'iixf6 is at least equal for Black. 3) 16 ltJd2 1ig4! (16...1ixa2 17 .txf6 gxf6 18 1ig4+ h8 19 'iih4 ':g8 20 1ixf6+ ':g7 21 ':d7 is too dangerous for Black) 17 f3 1ig6, threatening 18....ta6 or 18....:ac8, again with a comfortable position for Black. 16 ... .te4! (D) I thought for some time about 16...1ixa2!? I couldn't see anything definitely wrong with it, but it ap- peared more prudent to have the bishop on e4, from where it could de- fend the kingside. After 16...1ixa2 17 .txf6 Black can play: 1) 17...1ial+ (this move is defi- nitely too risky) 18 <iti>d2 1ia5+ 19 b4 1ixb4+ and now: la) 20 cl gxf6 21 'iig4+ h8 22 1ih4 (22 ':h3 1ia5! is favourable for Black) 22....:g8 23 1ixf6+ ':g7 24 ':g3 with a likely draw. Ib) 20 c3! 'iib2+ 21ltJc2 gxf6 22 1ig4+ h8 23 'iih4 ':g8 24 'iixf6+ ':g7 25 ':el! with a dangerous at- tack. 2) 17...gxf6 18ltJb3 (I don't see a direct try that works, so this sensible move seems best - it stops ...'iial + and temporarily shuts out the queen; 18 1ig4+ h8 19 1ih4 'iial+ 20 d2 1ia5+ and 18 ':a3 1id5 19 ':dl 'iie5! favour Black) 18... 'iia4 (after 18...h8 19 1ig4 the enemy queen is totally isolated) 19 ':d4! 1ic6 20 l1hd 1 is a hard position to evaluate, but White certainly has compensa- tion. ZB _ _._  . .. W B R _  .. . -,-- B B  . _..'.t_ _ . .:. .  0ii n /. UU  'W,   .'  .. - / /; :/;' :,//:/ ... /    j/:x;//a?; 17 ':e3 1ixa2! Not 17...1ixe2?! 18 ':xe2 .txg2? 19 ':gl and White wins. 18 .txf6 18 ':xe4 doesn't work because of 18...1ial+ 19 d2ltJxe4+ 20 'iixe4 1ixhl 21 1ig4 f6 22ltJxe6':f7 de- fending, for example 23 .txf6 1ixh2 24 .txg7 ':xf2+ 25 c3 1ixg2 and the attack collapses. 18 ... .tg6! (D) I felt much more secure with my bishop placed on g6, safeguarding 
KASPAROV - ANAND, REGGIO EMIUA 1991/2 77 the kingside. 18..:iial +?! is risky: after 19 d2 Wi'xh 1 20 ':xe4 gxf6 21 'iig4+ h8 22 'iih4 Wi'xg2 (22....:g8 23 'iixf6+ ':g7 24 ':g4 ':ag8 25 ltJf3 wins) 23 ':g4 Black has to give up his queen. .B . ... //  .  ... w/  _ _ '  . . . .t .  /,  -  . . . . . -$-   a ' :   . . % ¥ %% Jri%/   / iV".ii D "  w " !/j U .   ... . /"   ... 19 :a3 'iid5 The point of Black's play; White can't defend his knight because 20 i.e5 is met by 20...f6. 20 h4?! A risky try which I hadn't really looked at - White could have settled for sterile equality with 20 .te5 (20 'iie5 is met by 20... 'iixg2! and not 20.. .gxf6 21 'iixd5 with excellent play for the pawn) 20...f6 and now: 1) 21 .td6?! ':fc8! (21...Wi'xd6 22 'iixe6+ 'iixe6 23 ltJxe6 is equal, while 21....:fe8? 22ltJb5 a6 23ltJc7 is good for White) 22 'iixe6+ (22 c3 'iixd4 '23 'iixe6+ .tf7 24 'iie7 'iixf2 and Black wins) 22...'iixe6 23ltJxe6 ':xc2+ 24 dl ':c6 and Black will be a pawn up. 2) 21 .tg3 'iixd4 22 'iixe6+ with a draw. 20 ... gxf6 21 h5 'iixd4 22 hxg6 hxg6 23 .:&h3 fS 24 ':h4 f4! (D) Kasparov had placed many of his hopes on 24 ':h4, based on the line 24...Wi'f6 25 'iie3 ':fd8 26 Wi'h3 f8 27 :h8+ e7 28 'iia3+ d7 29 ':dl+ c6 30 'iia4+. After 30...c7 (not 30...b7? 31 ':d7+ ':xd7 32 'iixd7+ a6 33 ':xa8 and White wins) 31':dxd8! (31':d7+':xd7 32 ':xa8 'iid4 is unclear) 31....:xd8 32 'iixa7+ c6 33 ':h3 :d5 White has a dangerous attack for the pawn and can force a draw whenever he wants. .. . -.. w_ . ... - ..... . . . . - /  ;y@) /;  . ::tf)% / . /' '//. '%d a '" :    'W&      //f) i :  W#  »;@/ . / . % . 1: , ,;}, ;, % .\Wr.  B . .  . .:   , ;( I had pinned my own hopes on 24...f4!, keeping the queen in the centre. My queen functions like a Dragon bishop in preventing mate at h8. Kasparov is very good at long forcing lines, but it is the nature of 
n VHYMD:MYBTGOFCHE such lines that one cannot be abso- lutely certain about them. In this case he was just wrong. 25 11f3? After the game Kasparov pointed out that 25 g3! would have given him excellent drawing chances, for example 25...:ac8 (25...e5 26 :h8+ g7 27 :8h7+ f6 28 :dl 11b4 29 1Ig4 gives White at least a draw) 26 gxf4 (D) and now: .a. ...  . ... B_ _    ..... . - - - . . . . mu  _ U _ - . . . 8.\WJ" .  .u  - = . .: 1) 26...1If6 (my intention during the game) 27 1Ie5 (27 1Ie3 :c5 28 1Ih3 1Ixh4! 29 1Ixh4 :h5 is also drawn) 27...1Ixe5 28 fxe5 g5 29 :h5 :fd8 with a near-certain draw. 2) 26...:c5 27 f5 1If6 28 fxg6 fxg6 29 :h8+ 1Ixh8 30 1Ixe6+ g7 31 1Ie7+ (or 31 1Id7+ g8 with a repetition) 31...:f7 32 1Ixf7+ xf7 33 :xh8 draws. The text-move is a mistake be- cause it gives Black time to bring his rook into play and thereby gain a tempo by threatening mate on c2. Kasparov should have abandoned his winning attempts and gone for the draw, but he decided to 'fish' for a move too long! After 25 1If3?, the game followed his analysis but the position arising favours Black rather than White. 25 ... :tac8 26 .:xf4 Not 26 1Ih3 1Ixf2! 27 :h8+ g7 and Black wins after 28 1Ih6+ f6 or 28 :h7+ f6 29 c3 1Ie3+. 26 ... 1Ic5 27 c3 g7 (D) .a. _ . W  . .. . ?/ . - -   ..... . - - - - - . - . . a _ -  .... " . '8. u  y.. . = - .: By now Black is slightly better. 28 :hh4 Following the game I pointed out that 28 ':fh4! 1Ig5+ 29 c2 1If5+ 30 1Ixf5 might have been a better chance. However, 30...gxf5 (after 30...exf5 31 :d4 it would be very hard for Black to win) 31 :a4 :c7 32:hal a5 33 b4:fc8 34':la3:h8 35 bxa5 bxa5 36 :xa5 :h2 is still quite promising for Black. 
KASPAROV - ANAND . . REGGIO EMIliA 1991/2 During the g with queens on :: .1 felt happier ending and so 1 10 a pure rook the text-move. was pleased to see 28 ... 29 g3 30 c2 31 :d4 :cd8 Now Black 'iVeS pawn has secured h i " S . extra 32 1%hf4 33 _ 3 'fic7 B' e eS orcmg an exch out allowing th ange of rooks with- to d4. e other rook to come 34 .:xd8 35 :e4 (D) .:xd8 . /.' .-, .,,- - . B _p. .&0 . "p . _."   _  _>;.' _i. -  - .- '. - .  _   .0' _:. 0. . R . '.,/,- · rgx u 8 rgx R %    {ul'- 8'. . - .  -  'fie5 'fiel+ 35 ... 1% 36 g4?! dS 36 f4! would ha ducing the nu mb ve been better, re- p o . b er of P SSI l y ex po " awns an d sing the b l more checks. ack king to 36 ... 37 g5 b5 79 White's P l an " IS to cr" I pawn majority, but' Ip e Black's square to Black' It gIves the f5- 37 s queen. 38 f3 'iVd6 39 'iVe2 as 40 'ifh2 'iVe6 Perhaps W hi 'fif5 te could h more resistanc ave put up should be won r :t the position 41 'iVg3 ack. Or 41 'fih6+  and wins. g8 42 b3 :d2 41 .. \WI . .d 42 'iVel (D) 7 42 ... 4 b4' 3 cxb4 · Black also " 'fia4+ 44  l wlns after 43 1%x 5 :x wC bxc3' 4 5 e d5 'iVal+) 45 'iV. bxc3 (45 so the best ch ... f4+ 46 :e3 :d3 avoids an' ance was 43 b3 whi h the lOstant catastro h ' c text-move BI p e. After game by a direct ack decides the 43 attack. ... 'fi84+ 
80 VISHY ANAND: My BEST GAMES OF CHESS 44 b3 The continuations 44 c 1 axb4 45 ':xe5 ':d8 and 44 c3 'ifc6+ 4S ':c4 axb4+ 46 xb4 ':b5+ 47 c3 'iixf3+ 48 c2 'iib3+ are equally hopeless for White. 44 ... 'iia2+ 45 c3 45 clloses to 45...axb4 46 'iixb4 'iial+ 47 c2 'iidl+. 45 ... a4 46 bxa4 'iia3+ 47 c2 'iixa4+ 48 c3 'iia3+ 49 c2 ':d3 0-1 I started Reggio Emilia with my traditional two wins. In the next round I drew with Khalifman but lost in round 4, which allowed everybody to catch up with me. The rest of the tournament was a race between Kasparov, Gel- fand and myself and in the end the three of us tied for first - my greatest suc- cess up to that point, and even today I would consider it one of my best results. 
Game 14 V. Anand - E. Bareev Dortmund 1992 French Defence This was my best game from Dort- mund 1992. I also won a nice game against Htibner (see the following game), but I particularly like this one as it is a kind of model game for dark-squared play against the French. 1 e4 e6 Bareev's favourite defence is the French. Although he has also experi- mented with other lines, all our en- counters in which I was White have been French Defences. 2 d4 d5 3 ltJc3 ltJf6 4 e5 ltJfd7 5 f4 c5 6 ltJf3 ltJc6 7 e3 a6 8 1*'d2 b5 9 dxc5 xc5 10 xc5 ltJxc5 11 1ff2 1i'b6 12 d3 ':b8 13 0-0 ltJb4 This is all fairly standard stuff in this opening. Black must develop play on the queenside, whilst White tries to play on the kingside. 14 .:fdl 0-0 After 14...ltJa4 15 ltJxa4 bxa4 16 b3 Black isn't doing too badly, but he has a long evening ahead of him as White tries to exploit the d4-square and Black's bad bishop. However, heading for an ending may be Black's best chance once he has de- cided to play ....:b8 and ...ltJb4. My personal view is that if Black wants to keep the queens on, he should adopt the plan with ...b4, ...a5 and ...a6. 15 ltJe2! d7 The option of .. .ltJa4 has already gone, as now it would just be punch- ing thin air. 16 ltJed4(D)  . .. - - - - B . J..i/i _ W2  nm aillU .a. . .. ?Iiri . .. . g .i . "'1i D . - . // ,%' . "/    /-/   //- /.J  /.J   ;i3% , ' /. .  /. (-/;0; :// ;.'> / !% ///.'/ _ o/0  /  // . {A.t.z.Jd n  WU R 8u8. .8u  .:.  - . -  16 ... ltJbxd3 16...ltJe4 is bad in view of 17 xe4 dxe4 18ltJg5, but 16...a5 is a 
82 VISHY ANAND: My BEST GAMES OF CHESS possibility. From Black's point of view there is no advantage to this move, because White can just reply 17 b3 and Black has nothing better than 17.. .ltJbxd3 18 cxd3 b4, trans- posing into the note to Black's 17th move. However, White can try to refute 16...a5 out of hand by 17 xh7+!? During the game I looked at some of the following lines; the analysis is quite interesting, but the tree of variations is so large that I will only give the basic variations: 17...xh7 18 1*'h4+ g8 19 ltJg5 ':fc8 and now: 1) 201*'h5h821 hl (211*'h7+ f8 22 f5 exf5 23 ltJxf5? ltJcd3+! mates) 21...ltJe4 22 1*'h7+ f8 23 ltJxe4 dxe4 24 f5 ltJd5 (the knight comes back to aid the defence) 25 1*'h8+ e7 26 1*'xg7 d8 27 fxe6 fxe6 28 1*'g8 e7 29 1*'g5+ d7 30 1*'g7+ d8 is unclear. 2) 20 hl!? (D) and now there are three possible lines: ... ... B _ ..t.. i '/  BiB . - - - - i-i  _ _ U   _ ',  /. " fi.; 8 0 ' /. _ / ' ' '/. ,  % % %' ,  - - . a AnA. .An OUO. .OU a a.:. . 2a) 20...ltJcd3? 21 c3 ltJc2 22 1*'h7+ f8 23 ':xd3ltJxal 241*'h8+ e7 25 lIxg7 ':f8 26 f5 exf5 27 e6 xe6 28 ':e3 winning. 2b) 20...f8 21 1*'h8+ e7 22 1*'xg7 ':f8! (22...e810ses to 23 f5, while 22... d8 23 'iixf7 clearly fa- vours White) 23ltJdf3 ! ! (threatening 24 ltJh4; 23 a4 is less effective - the idea is to play ':a3-h3, but 23...lbe4 is quite an annoying reply) 23...lbe4 (23...ltJxc2 24 ':ac 1 ltJe3 25 ltJh4 with a crushing attack) 24ltJxe4 dxe4 25 ltJh4! ':fc8 26 1*'f6+ e8 27 1*'h8+ e7 28 ltJg6+ fxg6 29 1*'g7+ e8 30 ':xd7 and White wins. 2c) 20...ltJe4! (best) 21 1*'h7+  22 'iih8+ (22 ltJxe4 dxe4 23 ltJf5 exf5 24 ':xd7 1*'h6 wins for Black) 22...e7 23 1*'xg7 (D) and now: ... . . ..-rou B. ..i..i. ill11 . .  _ .i. . i.i D  - . ,  . "  "  .D  '   "/'//. /1 / /1  rh _ ',    . Bi.. _h 88. .8" u . d u a _.:_ _ 2cl) 23...ltJf2+ 24 glltJxdl 25 1*'f6+ e8 26 ':xdl ':xc2 (26....:c4 27 c3 is hopeless) 27ltJxf7 ':c6 28 f5 c8 29 'iih8+ d7 (29...xf7 30 f6 wins) 30 ltJg5 c7 31 'iih7+ d7 
ANAND - BAREEV, DORTMUND 1992 32 fxe6 ':d8 33 'ile7! and White ends up well ahead on material. 2c2) 23...ltJxg5 24 fxg5 (if 24 'ilxg5+, then 24...e8) 24....:g8 25 'ilf6+ and now 25...e8 26 ':fl 'ilxd4 27 'ilxf7+ d8 28 'ilxg8+ is good for White, but 25...f8! is un- clear. 2c3) 23...:f8! (this defence ap- pears to hold out for Black) 24 c3 (24 xe4 dxe4 25 c3 ltJd5) 24...ltJf2+ 25 gl ltJxdl 26 ltJh7 and now 26....:g8 27 'ilf6+ e8 28 'ilh4 is probably a draw, while 26...ltJxc3!? 27 bxc3 'ilc5 28 'ilg5+ e8 29 f6+ d8 is unclear. One can see why Bareev decided not to invest a lot of time in working through these variations and opted for the simpler text-move. 17 cxd3 ltJa4 ? A serious error - Black shouldn't allow White to fix the enemy pawns on light squares. Black should play either 17...b4 or 17...a5, which gives him chances of eventually activating his bishop at b5. After 17...a5 18 b3, for example, White is just slightly better. 18 b4! (D) At first it seems that Black's knight can reach a good square by ...a4-c3, but in fact it is not very effective at c3 because it lacks ade- quate support. By contrast, White's knights, after b3 and ltJfd4, have well-supported and useful squares to land on at d4 and c5. Alternatively, 83 White can use the fact that Black has no queenside counterplay to start kings ide operations. - B _.B BB B-*-B'_' .. -.- - B'_'O B JI1\   . ..u  u . _ B8BtlJ_ 8. . mfi8 d . _ u  :.  -   * 18 ... as 19 a3 :rc8 20 ':dcl axb4 21 axb4 :Xcl+ After 21....:c7 22ltJb3 ':xcl+ 23 ':xcl White retains a clear advan- tage. The position is very pleasant for White because it doesn't matter whether he swaps rooks, queens, or both - his advantage persists in any case. The plan of h4-h5 followed by g4 and f5 is effective both in the middlegame and in the endgame. 22 .:xci :c8 23 .:xc8+ i.xc8 White could play for the exchange of queens, but keeping queens on the board allows him the option of start- ing an attack using his space advan- tage on the kingside. 24 'ilc2 25 f2 d7 f8 
84 VISHY ANAND: My BEST GAMES OF CHESS 26 g4 e8 27 e3 (D) 27 g3!? might have been more precise - White's king is safer on g3, and the king would not obstruct White's kingside attack. I intended to bring the king to d2 to complete the domination of the a4-knight, but this is not necessary and in fact al- lows Black some slight counterplay. . ... . . ..i..'' B___  ..m . _ . . d ..../ . - - /  l&\  A. -"U  CO. _ .8=liJ. ... _ r - . . . 27 ... j[e) 28 h4 White could have played more di- rectly by 28 exf6 gxf6 29 g5 fxg5 30 ltJxg5 'iWd6 (the attack on b4 and the threat to play ...e5 look awk- ward, but...) 31 'iWe2! (...White simply abandons the b-pawn) 31...'iWxb4 32 'ii'h5+ d8 33ltJgxe6+ c8 34 'iWe5 'iWel+ 35ltJe2 xe6 36 'ii'xe6+ c7 37 'iWe5+ b6 38 f5 and Black's pieces are practically powerless to stop the f-pawn. When you have a grip suc as White has in this game, it is not easy to decide on the right moment to con- vert the advantage. Once you have opened the position you can't close it again, so a heavy commitment is involved in playing a line such as the above. It mustn't be too early, but you must also avoid enjoying your grip for too long and letting Black gradually free himself. During the game, I thought that exf6 and g5 was still premature, but it turns out to be possible for tactical reasons. Still, I prefer the game con- tinuation; as yet there is no need to hurry. After the text-move White clearly threatens exf6 and g5, be- cause ...fxg5 can be met by hxg5. 28 ... 'iWb8 After 28...f7? White just con- tinues with his plan by 29 exf6 gxf6 30 g5. 29 'iWcl (D) Now 29 exf6? gxf6 30 g5 allows 30...e5 !. illi1 _ . _ .  :. ; - . R.i.. . B_ w _  H .. . - //%@   .'R'D . D %'  % 8 / / A /. D /  '  %J00 /:  /. %  /:  . .8. d W vud . . . .  . ' Ri ,/. '/', %" ..  / /. ;ffj#  ;%$ /,  "/tlfd /w/ 29 ... f8 
ANAND - BAREEV, DORTMUND 1992 30 e2 Imprecise. 30 f2!, heading for g2 or g3, would have been more effi- cient. 30 ... f7 31 'ile3 'ilf8 After 31...ltJc3+ White has to take some care: 32 d2 would allow a measure of counterplay by 32...'ilc7! (but not 32...ltJa2? 33ltJc2! 'ilc7 34 'iid4 and the a2-knight is trapped), but 32 fl! is very good for White after 32...'ila7 33 'ilel or 32...'ilc7 33 f5. 32 f5 'ile8 33 g5 (D) Not 33 exf6 gxf6 34 fxe6+?! xe6 35 ltJxe6 'ilxe6 36 'ilxe6+ xe6 37ltJd4+ e5 38ltJxb5 f4 with a likely draw. When you have such an advantage, simplifying eve- rything just to win a pawn is insane. However, White could also have maintained the pressure by 33 fl, stepping out of the way of some checks. - .,.,. . B. ..t..... . .. . . - - , .''"8" d  u U 1&\  '  .  ..  /..'    _ _ illi1 _ . .8_l2J . .. . . . . . 85 33 ... exf5 34 gxf6 gxf6 35 'ilh6 fxe5 After 35...'ilh8 White wins by 36 e6+ xe6 37ltJe5+!. 36 'ilxh7+ f6 37 'iib6+ f7 38 ltJg5+ e7 39 'ilg7+ d6 40 ltJf7+ c7 41 ltJxe5 (D) . .,.,. . '    . WU B. % .t.. _ . . B _ .'B''. B  /  /.:'j' /'%  B '  '%J:% /,  /..'  , . .8. . _ ;;j;i% ;  . .<&it. _ . . . . White has established complete dominance and should win without difficulty. On top of all his other ad- vantages, he has an outside passed pawn. 41 ... 42 ltJxb5+ 43 ltJd4 44 ltJdf3 45 f2 46 'ile7 Threatening 47 b5 xb5 (or else ltJc6+) 48 'ilc5. 46 ... ltJb6 b8 'ilh5+ e8 f4 ltJc8 
86 VISHY ANAND: MY BEST GAMES OF CHESS 47 _f6 b7 If 52..:iWg7+, then 53 f2. 48 'iWxf4 8 53 'iWxf7+ b6 49 lOg5 6 54 'iWf4 'iWe6 50 g3 lOd6 55 h5 b7 51 'iWf3 c6 56 'iWf6! 1-0 52 lbgf7 lbxf7 A neat finish. Bareev subsequently recovered and had a very good tournament. The above game was played in round three. Over the next few rounds I had a number of sharp draws, including a very exciting one against Shirov, and then in round seven I faced Htibner. 
Game 15 V. Anand - R. Hubner Dortmund 1992 Petroff Defence le4 e5 2 lbf3 lbf6 Round about this time I was very well prepared against the Petroff, be- cause I had played the opening myself for many years, so I was quite happy with Htibner's choice. I had many interesting ideas stored up for White and was hoping that I would be able to use one of them. 3 d4 ltJxe4 4 d3 d5 5 lbxe5 lbd7 6 lbxd7 xd7 7 0-0 'iWh4 8 c4 0-0-0 9c5 g6 An unusual move instead of the normaI9...g5. Htibner had played it once before, in 1983, but I had never seriously looked at it. One reason was that Htibner had annotated the earlier game in lnformator, and had spent about two pages explaining why the move was really lousy! For the moment I decided to con- tinue as if Black had played 9...g5. 10 lbc3 g7 11 g3 The earlier game, Tinunan-Htibner, Tilburg 1983, had gone I1ltJe2ltJf6 12 b4ltJh5 13 b5 and White gained the advantage, but presumably there was an improvement somewhere. 11 ... _f6 Perhaps this is the point behind 9.. .g6; with the pawn on g5 this re- treat would not be possible. 12 e3 (D) ..- . - i.t.ii B_ _  _ .  i. d  _  _ Di. _ B 0._ . . i. p     AR .   o .   a ...:= 12 ... f5? Black should have attempted to make use of the fact that g5 is free by playing 12...lbg5 !? Then White can try: 1) 13 ltJxd5 lbh3+ (if 13...c6, then 14 'iWg4+ wins) 14 g2 (14 hl c6 15 'iWg4+ 'iWe6 and Black wins) 14...c6 15 'iWg4+ (15 xh3 xd5 leaves the king miserably placed) 
88 VISHY ANAND: My BEST GAMES OF CHESS 15...<ifi>b8 16 _xh3 ':xd5 and Black is much better. 2) 13 e2 h3 14 ':el ltJe6 is unclear. 3) 13 f4 and now: 3a) 13. ..ltJe6? 14 lDxd5 wins. 3b) 13...ltJh3+ (a difficult move to make; of cours the knight is rather annoying for White, but it has no way out) 14 <it>g2 :he8 (14...h5? 15 f5! is very good for White) 15 _d2 with an unclear and double-edged position. 3c) 13..._e6! 14 ':el ltJe4 15 .i.xe4 dxe4 16 d5 is another unclear line. Black can try to develop light- squared counterplay, while White can use his pawn-mass for attacking purposes. The text-move is a clear mistake. I didn't realize this immediately, be- cause I was trying to imagine what this position would be like with the pawn on g5. Suddenly I noticed that ltJb5 was very strong, precisely be- cause with the pawn on g6 my bishop has access to f4. 13 ltJb5! h3 After 13...a6 White continues 14 liJxc7! ri;xc7 (14..:iWc6 15ltJa8! and the knight escapes via b6) 15 f4+ ri;c8 16 e5 _c6 (16..:iWe6 is the same) 17 .i.xg7 ':hg8 18 e5 f6 19 i.f4 g5 20 e3 with a safe extra pawn and an attack by b4-b5. 14 ltJxa7+! b8 15 liJbS xn 16 .i.xfi (D) I didn't think long about this ex- change sacrifice or even bother to calculate variations; it's clear that without the black a-pawn, White's own a-pawn can just run all the way through.  - . - -  . - .. .. B_  ?%% _ _ . _'B BliJB'_ . . ".. . dud d . .   _ ;rf$ ',,; U A B / / : .   /%   /  D /;  Q -   !// ' & /!// //!//  ;&:%\WJ.     ?J I!!J 2; .JL  I '/ ,/ ,/ 0:-, //y // 16 ... ':he8 More or less forced, to meet f4 by....:e7. 17 'iWa4 'iWa6 18 'iWb4 'iWa8 White wins after 18...'iWc619 'iWaS f5 (..._a6 is no longer possible be- cause c7 is also attacked) 20 f4 ':e7 21ltJxc7 ':xc7 22 b5. 19 a4! Intending to tuck in her majesty with as-a6. 19 ... g5 Black even has to spend a tempo preventing f4+ before he can play .. .c6. 20 as 21 ltJc3 22 bxc3 (D) c6 ltJxc3 
ANAND - HORNER, DORTMUND 1992 \Wr0 .    .  B       . ,   //.... ;// "/" , '"  ?,,'l/// /////-; B -.. ... --;- ... - . .      '/, . ;  . % '0  ;%J '/. {;/, //.  ,/"  /,' !/; ///;, ("'/" ij '/' :// ' /(//';'/ /w% " '/ , ;'(;j;;;:0% %,    0{W; ;$ 0('   ' m   '/    /)  / / ??{% , "  /, / ,/"   ;r/ W      " . .i.  . .  Black is dead lost. 22 ... h6 23 a6 f5 24 .1h3 :f8 25 a7+ c8 26 'iibl 89 Forcing the g-pawn forwards and so gaining f4 for the dark-squared bishop. 26 g4 27 n d7 28 i.d3 e6 29 f4 :t7 I was tempted to toss in b8, but it isn't really necessary! 30 'iWc2 f8 31 'iWe2+ 1-0 White finishes off by 31...f6 32 'iWe5+ g6 33 'iWe6+ :f6 (33...g7 34 e5+ g8 35 xf5 ':e8 36 h7+! mates) 34 xf5+ g7 35 e5 and mate in three more moves. A nice crisp win, though Black's 12th move made it easy for me. The above two games were my only two wins from Dortmund and I fin- ished fourth with a score of 5/9 - not one of my most memorable results. In July 1992 we had the GMA 'farewell party'. The organization had fallen into difficult times and with the failure of the second World Cup cycle it was recognized that its tournament-organizing days were over. SWIFf sponsored a final rapid-play knock-out event held in Brussels. It was su- perbly organized. 
Game 16 V. Anand - I. Sokolov Brussels SWIFT rapid 1992 Sicilian, Scheveningen This was the second game of a two- game mini-match. I had won the first game and so only needed a draw to go through to the next round. 1 e4 c5 2 ltJf3 e6 3 d4 cxd4 4 ltJxd4 a6 5 ltJc3 d6 6 a4 ltJf6 7 i.e2 ltJbd7 I found this move a bit strange, but 7...ltJc6 would have transposed into one of the main lines of the Scheveningen and I suppose Sok- olov wanted something a bit sharper. 8 0-0 ltJc5 9 i.f3 i.e7 10 g3 0-0 11 i.g2 In a sense White has lost time with this bishop manoeuvre, but on the other hand the e4-pawn is well- defended and so the knight on c5 isn't doing much. Moreover, Black cannot play ...b6 easily and ...i.d7 is met by b4, so he is going to have a lot of trouble developing his queen- side. 11 ... 'i/ic7 12 i.e3 I think that White already has a significant advantage. 12 ... ':b8 13 f4 ':e8 Here I was about to play the stan- dard plan g4-g5, when I noticed that it wasn't really necessary as White can break through right away. 14 e5 dxe5 Forced, since 14...ltJfd7 15 exd6 i.xd6 (or 15...'ii'xd6 16 b4) 16ltJdb5 . WIns. 15 fxe5 ltJfd7 (D) 15...'ii'xe5 loses to 16 i.f4. ..i...... w...._. .. ... - .   . d _ U _ 8.  . . .      _ U 8. .i. U d d U  1i.: - - -  16 .:xf7! A very common tactic when the rook is on e8 and the f-file opens. 
ANAND - I. SOKOLOV, BRUSSELS SWIFT RAPID 1992 91 16 ... xf7 IT Black tries to decline the sacri- fice by 16...l[}xe5 then 17 i.f4 _d6 (17...xf7 18 _h5+ f8 19 i.xe5 wins) and now: 1) 18hll[}cd719:xe7andaf- ter 19...J:xe7 20 l[}f3 _xdl+ 21 J:xdl l[}xf3 22 i.xb8:f7 or 19..._xe7 20 1i'e2 l[}c6 21 l[}xc6 bxc6 22 i.xb8 l[}xb8 White has a positional advan- tage. 2) 18 :xe7! (an even stronger move) 18...:xe7 19 l[}db5 _xdl+ 20 ':xdl axb5 (20...l[}cd7 21 i.xe5 l[}xe5 22 :d8+ f7 23 l[}d6+ wins) 21 i.xe5 :a8 22 i.d6 :d7 23 i.xc5 ':xd 1 + 24 l[}xd 1 :xa4 25 l[}e3 :a2 26 i.d4 with a winning endgame. After the text-move the attack crashes home. 17 'iWh5+ 18 :0+ 19 exf6 20 l[}db5 21 l[}xb5 21 i.xc5+ :e7 22 l[}xb5 also f8 l[}f6 i.xf6 axb5 . WIns. 21 ... _d7 (D) .t..- . _ d _ d ., .\WJ. /., w. .-. ,  . .'E . .llJII . .. . . . . . .  " d    ". .  . -  . . .: . - . * 22 _xh7 A nice move. White can take the c5-knight with check in two different ways, but ignores it. White threatens 23 :xf6+ gxf6 24 i.h6+, so Black has to waste more time. 22 ... _e7 23 Jbf6+ _xf6 24 i.xc5+ :e7 25 'iWh8+ f7 26 l[}d6+ 1-0 I was eliminated by Adams in the semi-finals, so the above game was the highlight of the tournament for me. In September I played a match with Ivanchuk in Linares. It was not part of any cycle, but simply a one-off event arranged by Rentero (the organizer of the Linares tournament) who wanted to hold an event in addition to the an- nual super-tournament. Ivanchuk and I were his first choice for the players, and we each had our own reasons for wanting to participate. Ivanchuk wanted to erase the memory of the loss to Yusupov in the Candidates quarter-finals in Brussels, and I wanted to keep my match experience going between world championship cycles. Subsequently, he organized two more matches: Ljubojevic-Illescas and Lautier-Karpov. 
Game 17 V. Ivanchuk - V. Anand Match (1), Linares 1992 Sicilian, Richter-Rauzer 1e4 c5 2 l[}f3 d6 3 d4 cxd4 4 l[}xd4 l[}f6 5 l[}c3 l[}c6 6 .tgS e6 7 _d2 a6 8 0-0-0 h6 I had prepared this variation spe- cially for the match. I hadn't played it before so I could hardly imagine Ivanchuk having prepared it deeply. Nevertheless, Ivanchuk blitzed out his next few moves and in fact played the whole game at high speed! 9 .1e3 l[}xd4 10 .i.xd4 b5 11 f3 11 bl is more promising. Ivan- chuk played it in the fifth game of the match and gained the advantage, although the game finally ended in a draw. 11 ... _as 12 a3 A new and quite good move. 12 bl was played in earlier games, while 12 _f2 b4 13 .1b6 'ii'g5+ 14 .1e3 _a5 is a popular variation for players who feel like an early din- ner! 12 ... 13 .1e3 14 bl e5 .1e6 .1e7 (D) -. ... - R M . w  _ ,, _ .. ..t-  . /uu/ _ _ : . / / &   /  // .  %.-  % - %  ;;::  U /  ffj;fj/; . _8. .   /   '' 8  // .   / +   /. /. ,  / / y,; ;., '/ '/:0")  ' / A :/ ,,/, W{@'A B %J o Po  //     /  ..:.i..: 15 g4?! Playing l[}d5 is an option which is available to White at virtually every move. However, without any knights White can hardly expect to do any- thing against Black's slightly weak- ened queenside, so playing l[}d5 is an admission that White can no longer hope to gain the advantage. The problem with Ivanchuk's move is that he is soon forced to play l[}d5 in any case, when the move g4 not only fails to benefit White but can even prove a weakening of his king- side. 
IVANCHUK - ANAND, MATCH (1), LINARES 1992 93 After 15 h4 ':b8 16 d5 'ii'xd2 17 ll)xf6+ gxf6 18 ':xd2 f5 the conse- quences would not be so serious for White as h4 does much less damage to White's kingside. 15 ... ':b8 15...b4!? was an interesting alter- native: 1) 16 lbd5 .1xd5 17 exd5 :b8 is unclear. 2) 16 axb4 'ii'xb4 17 d5 lbxd5 18 exd51Wxd2 19 ':xd2 .1d7 with a roughly equal position. 3) 16 a2 d5 17 axb4 'ii'c7 with compensation for the pawn. I wasn't feeling quite awake at this point and so I played the 'solid' 15....:b8. 16 llJd5 'ii'xd2 (D) -.- - ... - . . . wd _ _ _ .. J.  . - -  .../ . / -%  - . .8.8. " . 8. d._. 8 .  U _ . u ..:.i..: 17 xf6+? In conjunction with 15 g4 this is a terrible move. I suspect Ivanchuk hadn't woken up either! He should have continued 17 ':xd2, but after 17...xd5 18 exd5 .1d7 Black is slightly better, because he has the ...f5 break, whereas White has no comparable play on the queenside. After the text-move I was about to make the routine capture 17.....ixf6 (when White could perhaps be a lit- tle better after 18 :xd2 e7 19 h4) when suddenly I noticed that 17...gxf6 might be a good move. After about ten minutes' thought, I decided to play it. 17 ... gxf6!! Now we were both wide awake! 18 .:xd2 h5! (D)  . . .  - . . - . . .. w_ _ _ _ ..  .t  . - - - - ...  .. -  -  . .8.8. o  m 8. // %J& ," % "8 . 0 a _.  .. ..: At first glance White is better, or at least not worse, in view of Black's damaged pawn structure. However, White is actually seriously worse. If White could consolidate his king- side pawn structure by h3 then he would indeed be better, but just at the moment this is impossible. 19 :gl Also after 19 .te2 hxg4 20 fxg4 :h3 Black's rook takes up residence 
94 VISHY ANAND: My BEST GAMES OF CHESS at a most inconvenient location for White. 19 ... 20 fxg4 hxg4 .1c4!! (D) - ... . . . -'B w_ _ _ _ 'B    B  d _ _  ., .  // B -  % - .1..8.8B R .  . u d _ d R8  B   u _  u .<it>. .a Just in time to stop White setting up some sort of fortress on the king- side by h3 and :g3. Black's 'bad' e7-bishop will protect his pawns while he forces pawn exchanges eventually leading to connected cen- tral passed pawns. 21 b3 After 21 .1xc4 bxc4 White again has no time to consolidate his king- side by 22 :g3 (and 23 h3) because of 22...c3. If 22 :d5, then 22...:b5 and again Black is better. 21 ... .1xn 22 .:xn :h3 Black appears to have committed a whole list of positional sins: allow- ing doubled f-pawns, giving White an outside passed h-pawn and ex- changing his 'good' bishop with ....1c4; yet he is better. Paradoxfcal? Yes, but this does not mean that the old positional rules have been sus- pended for the course of this game. Black's play depends on two things. First of all, his long-term aim is to exchange his d-pawn for White's e- pawn by ...d5 and to exchange his f6-pawn for White's g-pawn (either by ...f5 or by forcing White to play g5). Then he will be left with two connected central passed pawns, sup- ported by his king, whereas White will have pawns on c2 and h2 that aren't going anywhere. Secondly, he can only put his plan into action be- cause he has the initiative, and espe- cially as the rook on h3 disrupts White's whole position and leaves both g- and h-pawns vulnerable to attack. Had Black wasted even one move, White would have fortified his kingside and the old positional values would have reasserted them- sel ves. The text-move is more accurate than 22...d7, not because of 23 :f3, when 23...:h4 24 h3 :bh8 wins a pawn, but owing to 23 g5!, which confuses the issue. Then af- ter 23...fxg5 24 :xf7 Black's over- all plan has been disrupted, while 23...e6 24 gxf6 .1xf6 25 ':xd6+! xd6 26 :xf6+ e7 27 .1g5, fol- lowed by h4, gives White plenty of counterplay. 23 ':e2 23 .1gl would have been a better chance, but Black maintains the 
IVANCHUK - ANAND, MATCH (1), LINARES 1992 95 advantage by 23...d7 24 ':d3 ':h4! (24....:xd3 25 cxd3 ':h8 26 :f3 d5 is unclear) 25 :g3 (25 h3 :bh8 26 ':ff3 f5! 27 ':xf5 ':xh3 28 ':xf7 :Xd3 29 cxd3 ':hl 30:fl .1g5 wins material, as in line 1 below) 25....:g8 26 h3 ':gh8 27 ':ff3 f5! and now White cannot maintain his fortress: 1) 28 ':xf5 ':xh3 29 ':xh3 ':xh3 30 .1f2 (30 ':xf7? ':hl 31 ':fl .1g5 32 ':el .1d2 is winning for Black) 30...e6 and White's pawns on e4 and g4 are so weak that he might easily lose both of them. 2) 28 exf5 e4 29 ':e3 (29 ':c3? .1f6) 29...d5 and Black has strong pressure. 23 ... d7 24 g5 (D) . . . . B B .._i. iB '/  B . , - . BiB   . . _ u B BB B .  .. U _ _ _ BB:B 0 B<it>B B:B 24 ... e6 Now this is good as White does not have an exchange sacrifice (see note to Black's 22nd move). 25 gxf6 .1xf6 26 .1d2 26 ':ef2 makes no sense as after 26.. ..1e7 White has to attend to his attacked bishop. 26 ... .1e7! Simplest and best. The alterna- tives 26....1h4 27 .1b4 and 26....:g8 27 :ef2 .1e7 28 ':xf7 :g4 allow White more counter-chances. 27 .1el f6 28 .1g3 White has finally defended the weak h2-pawn, but Black has time for .. .d5. 28 ... dS 28....:d8!? was also possible, al- though in this case Black would have to wony about 29 c4. I pre- ferred to play.. .d5 immediately. 29 exd5+ d5(D) . . . B wB B _ B i. . ' . .iB." B . . . . "B . . U . . _ ..:. R _ _ _ u .<it>B B:B 30 :t5! A good defence, forcing Black to lose some time. 30 ... c6 A forced move, as 30...e6? 31 .1xe5 ':e8 32 ':xf6+ and 30....:b7? 
96 VISHY ANAND: My BEST GAMES OF CHESS 31 .1xe5 e6 (31...fxe5 32 .:cxe5+ d6 33 :e6+ d5 34 :xe7 is no better) 32 -*.xf6+ xf5 33 .1xe7 would lead to a draw. 31 ':ef2? 31 :f3! would have made life far harder for Black as White threatens both 32 .1xe5 and 32 ':c3+ followed by 33 i.xe5. There would be noth- ing better than 31....:h7 32 :c3+ b7 (not 32...d7 33 :d2+ .1d6 34 :cd3 :b6 35 i..n :c6 36 i.c5), but this would represent a success for White. The black king belongs on e6 and while the connected passed pawns guarantee Black an advan- tage, he would have a hard technical task ahead. After the text-move White has a large disadvantage. 31 ... ':h6 (D) Not 31...d5 32 :xf6 .i.xf6 33 :xf6 and White has complicated matters. The text-move prevents any sacrifice on f6. . . . - . .  . w  _  .. .  ir  - - , - ... :. --- . . . . "8. .  U   ,- /  .  A     g %i" D ' o_    /. %"   "'0'   -. . . 32 b2 <itrd7 33 :e2 i.d6 34 :f3 :'e8! Very precise. 34... e6 would al- low 35 ':c3, intending :c6, and again White has some counterplay. 35 .i.el e6 Mission accomplished! 36 :d3 (D) ... . . . . . . B. _ _ _ ..  .    - - - ... . . . . . . .. A." . Uo..... . '08.".  , G- u . . + . 36 ... :h7 37 :g3 .i.c5 Black doesn't even have to push the pawns immediately. He can play to improve the position of his pieces, or try to exchange a pair of rooks to reduce the chances of a blockade. 38 a2 :d7 39 :c3 ':cc7 39....:dl? 40 .1n .i.xf2 41 ':xc8 .1d4 doesn't mate after 42 c3. 40 h4 ':dl 41 .1n .1d6 42 ':g3 e4! After all the fuss about Black's connected passed pawns, he gives 
IVANCHUK - ANAND, MATCH (1), LINARES 1992 97 one of them up! However, it does win the exchange. 43 .:xe4+ 43 J:gl J:xgl 44 .1xgl f5 wins easily. 43 ... .1e5 44 .:xe5+ Or 44 c3 J:d2+ 45 bl :Xf2 win- ning. 44 ... fxe5 45 b2 J:d2 0-1 This match was an important point in my career as I gained a great deal of confidence as a result of beating Ivanchuk 5-3. He was the first really strong opponent I had beaten in a match and I took this as a promising sign for the next world championship cycle. In November I took part in a very strong (and enjoyable) tournament in Moscow. I think it was only the second tournament to reach category 18, al- though this had been achieved by having only eight players. I lost the first game to Gelfand after allowing myself to be swindled in a favourable posi- tion, and after losing the first game in a seven-round event I couldn't really expect that I would win. However, Moscow turned out to be one of my more successful tournaments. I scored 4 1 /2 out of my remaining six games and shared first place with Gelfand. The following game is from the second round and was the first step towards recovering from my first-round defeat. 
Game 18 V. Anand - G. Kamsky Alekhine Memorial, Moscow 1992 Ruy Lopez, Worrall Attack 1 e4 e5 2 l[}f3 l[}c6 3 .1b5 a6 4 .1a4 l[}f6 5 0-0 .1e7 6 _e2 I played this move because I had beaten Kamsky with the same move the previous year in Tilburg, and I just decided to repeat it. 6 ... b5 7 .1b3 0-0 8 c3 d6 9 d4 .1g4 10 ':dl exd4 11 cxd4 d5 12 e5 l[}e4 13 a4 (D) A couple of months before Short's semi-final Candidates match vs Kar- pov (which was in April 1992), he asked if I would like to come to Ath- ens to work with him. This visit was just a one-off arrangement and I wasn't Nigel's second. We analysed the _e2 Ruy Lopez during the week I stayed there, but I had no part in some of the other openings Nigel played in the Karpov match. A strange coincidence occurred in the Linares tournament shortly before the Short-Karpov match. I played the Queen's Gambit Accepted against Karpov, tried out something very unusual and drew without any difficulty. I had no idea that Nigel had prepared the QGA for Karpov until Nigel revealed the fact after my game! - :/.' //, - . - .I. %.   . .  .. B. _ _ _ .... . . .... . _ _ U _ 8. D.."'. .ii.. .lb.  ..8 u  u u lb:.  - - .  13 ... ':b8 One of the points of 13 a4 is the line Short showed against Karpov: 13...bxa4 14 .1xa4l[}b4 15 h3 .1h5 16 l[}c3 .1g6 17 .1e3 ':b8 18 l[}a2!. This last move is one of the discov- eries we made during our analytical week in Athens: it is much more im- portant to eliminate the knight on b4 than the one on e4. The reason is that 
ANAND - KAMSKY, Moscow 1992 the knight on e4 is only potentially threatening (e.g. if Black plays ...c5 and.. .cxd4) while the knight on b4 is critical, as it stops White invading on c6. Kamsky, not surprisingly, does not repeat the line that led to a loss for Karpov. 14 axb5 axb5 15 h3 .1h5 16 .1e3 (D) _ mu _ . _ - - - .  . '/ % . ' & _ 0' & B -  0.. ??  ," , ... . B B'B'B B.1 . I&\. . . U'-'. . .i.B mt2J.8  .."8_ u dud .:.  _ "l...J. .  16 ... 'ii'd7 Not 16...l[}g5 17 .1xg5 .1xg5 18 l[}c3 l[}b4 19 g4, winning the d5- pawn. 17 :c1 17l[}c3 would have been interest- ing, e.g. 17...l[}b4!? (17...l[}xc3 18 bxc3 b4 19 c4 dxc4 20 .1xc4 is un- clear) 18 .1f4 'ii'f5 (18....1g6!? is also possible) and now both 19l[}xd5 l[}xd5 20 .1xd5 l[}xf2 and 19 'ii'e3 l[}xc3 20 bxc3 l[}c2 21 .1xc2 'ii'xc2 are unclear. 17 ... :b6 99 Supporting the weak third rank and the c6-knight in particular. 17...l[}d8 was an alternative. 18 l[}c3 l[}b4 After 18...l[}xc3 19 :xc3 White has a small advantage. 19 :&7 l[}g5 Again the correct choice. After 19...:g6 (19...l[}c6? loses after 20 l[}xd5l[}xa7 21 :xc7) White replies simply 20 h2 and Black has to deal with the threats of 21 'ii'xb5, 21 l[}xb5 and 21 :xc7 (20...l[}c6 fails to 21l[}xd5 !). 20 .1xg5 .1xg5 21 :ca1?! (D) I intended the exchange sacrifice given in the next note, but I had mis- calculated one variation, so the solid 21 :dl would have been better. . . ...  .... B_ _ d _ - . . . ..B.n .1 d d u - .  . . _ U . . .i. _t2J.8  ..8. u dud  . .  _ d   21 ... .1f4? 21...l[}c6! was the right move: 1) 22 :a8? l[}xd4 23 :xf8+ xf8 24 :a8+ .1d8! and Black wins, but not 24...e7? 25 l[}xd4 .1xe2 26 
100 VISHY ANAND: My BEST GAMES OF CHESS ltJxd5+ 'ii'xd5 27 .1xd5 .1c4 28 .i.c6! wInnIng. 2) 22 ltJxdS ltJxd4! (this is the move I had missed; after 22...ltJxa7 23 1txa7 White has good compensa- tion for the exchange) 23 'ii'd3 lbxf3+ 24 gxf3 'ii'xh3 and Black is better. 3) 22 'ii'd3 .i.xf3 (22...ltJxa7 23 ':xa7 wins the d-pawn and is prom- ising for White) 23 'ii'xf3 lL1xd4 24 'ii'xd5 'irxd5 25 .i.xd5 c6 26 ':dl (26 i.e4 is also equal) 26.. ..i.f4 27 :xd4 .i.xe leads to a draw. 22 l[}xd5! This gives White a large advan- tage in every line. 22 ... ltJxdS 23 _e4 (D) - . ...  /.., /, BM .  - . . . .'.D .... . "\WJ . d a-. d B.t. .l2J.8  . 8. a dad  . .  M d d  23 ... .i.xf3?! Now White is winning. Black could still have continued the fight by 23...ltJe7! 24 'ii'xf4ltJg6 25 'ii'e3 .i.xf3 26 'ii'xf3 'ii'xd4 27 :Xc7ltJxe5, but now either 28 ':xf7 ltJxf3+ 29 :xf3+ 'iic4 30 :c3 or 28 .i.xf7+ h8 29 'ii'e3 'ii'xe3 30 fxe3 gives White a clear extra pawn. 24 1i'xf3 .i.xe5 25 dxe5 b4 (D) - - -..  ..,, w_ _ . n - . - . .,.  . _ _ U _ II . . . .i.. ...8  . "8. u . u .  . .  ,  M zt. // 26 'ii'e3?! After the game Gelfand pointed out a much simpler (and quicker!) win by 26 ':a8! ':bb8 (or 26...ltJc6 27 e6) 27 ':xb8 ':xb8 28 ':a8!. 26 ... 'ii'c8 Or 26...'ii'd3 27 .i.xf7+ ':xf7 28 ':a8+ ':f8 29 ':xf8+ xf8 30 'iic5+ and White wins. 27 _e4 Once again missing an easier win, this time by 27 e6 fxe6 28 :c 1, for example 28...:t7 29 'ii'xb6, 28...ltJa6 29 'ii'xb6 cxb6 30 :xc8 ':xc8 31 .i.xe6+ or 28...ltJd5 29 .i.xd5 exd5 30 ':cxc7. 27 ... ltJa6 After 27...ltJc6 28 ':a8 'ii'xa8 29 ':xa8 ':xa8 30 e6 fxe6 31 .i.xe6+ h8 32 .i.d5 White wins material. 28 .i.d5 c6 
ANAND - KAMSKY, Moscow 1992 On 28...ltJc5, 29 'it'b4ltJd7 30.:a8 finishes Black off. 29 'it'e3 ':b7 More or less forced, because af- ter 29...c5 White's bishop is abso- lutely dominant, but now White has a forced win. 30 .1xf7 + 31 ':7xa6 32 e6 (D) 32 ... 33 e7 34 :a8 35 h2 White must still be careful. After 35 ':xbl? Black would slip out by 35..:ii'xa8 36 'it'e6+ h8 37 ':fl h6 38 'it'd7 g8. ':bxf7 .:xn .:xb2 ':e8 ':bl+ 101 ... ...  . / ;  % / . B - /,m  i j /t/%@ c/'/u  .:_..- - ... _ 8 - - - - - . - - R  _ _ U . ..  . . =  - . /./,  35 ... 'it'c7+ 36 g3 ':b2+ 37 gl 'it'd7 38 .:xe8+ 1-0 Since 38...'it'xe8 39 'it'e6+ h8 40 ':fl leaves Black defenceless. 
Game 19 V. Anand - V. Ivanchuk Linares 1993 Petroff Defence 1 e4 2 lL\f3 3 d4 4 .1d3 5 ltJxe5 6 lL\xd7 7 0-0 8 c4 9 c5 10 lL\c3 11 g3 12 lL\xe4 13 .1xe4 e5 lL\f6 ltJxe4 d5 lL\d7 .1xd7 -"4 0-0-0 g5 .1g7 -"3 dxe4 .1b5 (D) ..- - - . ... w. . d _ . B B B .R . ' ....U d  B DJi.B B B . B D'ii' A .   lJ U  U U  \Wr.: _ .-d  This position had been reached umpteen times before the game and the path to equality for Black had been more or less worked out after 14 'ii'b3 and 14 .1xg5, but I had a new idea that I had been waiting to use. 14 .1g2! After 14 .1xg5 ':xd4 15 .1g2 'it'f5 16 'it'b3 c6 17 .1e3 .1xfl 18 ':xfl, Black can continue 18.. .':hd8! 19 'it'a3 ':dl 20 'it'xa7 ':xfl+ 21 .1xfl ':dl 22 'it'a8+ c7 23 .1d2 'it'xc5 24 ..ta5+ b6 25 'it'a7+ d6 26 .1xb6 'ii'b5 27 'it'c7+ e6 28 'ii'c8+ e7 with a draw. 14 ... 'ii'f5 Ivanchuk looked surprised, be- cause it was well known that 15 d5 is bad, but this was not White's idea. 15 .1e3! (D) 15 'it'b3 c6leads nowhere. ..- . - . ..-. B_ _ d _ B B . B .. .. d...u . _ B 0 _ . . .   . . . U A . i.R lJU d U U  .\Wr.: - . - .  15 ... .1xfi 
ANAND - IVANCHUK, LINARES 1993 16 .1xn White intends 'it'a4 followed by :dl-d3, swinging the major pieces into the queenside attack. 16 xfl ?! is inferior as the king is exposed on fl. 16 ... ':he8 Black has several alternatives: 1) 16....1xd4 (the tactics don't work for Black) 17 .1xd4 'it'e4 18 .1xh8 ':xd119 ':xdl with too much for the queen. 2) 16...:xd4 17 ..txd4 ':d8 18 1Wh5 :Xd4 19 .1h3 wins the queen. 3) 16...'it'd7 17 'it'b3 ':he8 18 ':dl, intending d5, with a strong ini- tiative. 4) 16...c6 17 'it'a4 h5 18 'it'xa7 ':xd4 19 ':el ':hd8 20 .1xd4 1-0 Kharlov- T.Christensen, Nrresundby Open 1993 was an abrupt finish. 5) 16...b8 17 'it'a4 c6 18 ':dl ':d7 19 ':d3 ':hd8 20 ':a3, Zar- nicki-Howell, Capablanca Memorial, Matanzas 1993, with fine compen- sation for White. Ivanchuk's move is good and al- lows him to develop counterplay against d4. 17 'it'a4 b8 White wins after 17....:xe3 18 fxe3'it'e4(18...'it'f319':el) 19'it'b3! (not 19 ':el .1xd4 and White's queen is threatened) 19....:e8 (19....:xd4 20 .1g2 also wins) 20 ':el .1xd421 .1g2 .1xe3+ 22 hl and Black's po- sition collapses. 18 ':dl (D) 103 Ii ..B B . ..-. B_ _  _ - B . _ . "  . \WJ  U ._ -B B B B  B   .   n   _ u 8" . " n u  u u B B:B.t= White's plan is to use the bishop on fl to support ':d3-a3, and then to swing the bishop back to g2. 18 ... c6 After 18....:xe3 19 fxe3 'it'e4 (the line 19... 'it'f3 20 ':d3 is similar) 20 ':d3 White consolidates his extra pawn. 19 ':d3 19 .1g2 is premature since after 19....:e7! 20 d5 cxd5 21 c6 d4 the position is unclear. 19 ... 'it'e4 20 .:aJ In view of Black's improvement at move 21, an interesting alternative here is 20 'it'd 1 !?, intending either .1g2 followed by d5, or simply b4, a4 and b5. The queen on dl over- protects d4, preventing a counter- sacrifice by Black on that square. 20 ... a6 21 .1d3 (D) After 21 .1xa6 ':xd4! Black gains enough counterplay to hold the bal- ance: 22 .1xd4 .1xd4 (22...'it'xd4 is 
104 VISHY ANAND: My BEST GAMES OF CHESS also viable) 23 .1fl .1xc5 (but not 23....1xb2? 24 'iVa7+ g;c7 25 ':b3 and wins) 24 'iVa8+ g;c7 25 'iVa5+ .1b6 26 'iVxg5 ':e5! and the pressure on f2 gives Black enough for the pawn. //- .. B / /,_.- B .i. .iili i.i. . . . P;. .  dud. 'if. 8.. .  ..i. "  d _ u 8" B " " u  u u . . . .' 21 ... 'iVg4? This move loses because it leaves the c6-square weak. During the game, I realized that 21... 'iVd5! probably saves Black: 1) 22 .1c4 'iVe4 will be a repeti- tion. 2) 22 ':b3?! .1xd4 and now: 2a) 23 ':xb7+? g;xb7 24 'iVxa6+ cit>b8 25 'Wb6+ cit>c8 26 'iVa7 (after 26 .1f5+ ':d7 27 .1xd7+ 'iVxd7 28 .1xd4 'iVxd4 White has no perpetual check) 26....:d7 27 'iVa8+ g;c7 28 'iVxe8 .1xe3 and Black wins. 2b) 23 .1xa6 ':e7! 24 ':d3 (24 .1xd4 'iVxd4 25 'iVxc6? 'iVdl+ 26 g;g2 'iVxb3 wins) 24.. .':e4 25 ':b3 ':d7 and White's attack runs out of steam. 2c) 23 .1xd4 ':el + (23...'iVxd4 24 ':xb7+ g;xb7 25 'iVxa6+ g;b8 26 'iVb6+ g;a8 27 'iVxc6+ g;b8 leads to perpetual check) 24 .1fl 'iVxd4 25 ':xb7+ (25 'iVxc6? ':xfl + 26 xfl 'iVdl + 27 g;g210ses after 27...'iVxb3) 25...cit>xb7 26 'iVxa6+ cit>c7 (26...cit>b8 27 'iVb6+ is a draw) 27 'iVa5+ cit>d7 28 'iVxel 'iVxc5 and Black is slightly better. 3) 22 .1xa6 .1xd4 23 .1xb7 (23 ':b3 is line 2b above) 23.. ..1xc5! ! (D) (an amazing defence that I spot- ted while waiting for his reply; not 23....1xe3 24 'iVa7+ cit>c7 25 .1a6# nor 23...g;xb7 24 'iVb4+ g;c8 25 ':a7 winning) and now: II ... . w..i.. .i.i .i. . . . ..  d _ d _ ... . . . a . m D "/ "/ . , , ,Y' /  8" . " R U d u u . . .  - - -  3a) 24 ':b3? 'iVdl + 25 cit>g2 .:xe3! 26 'iVa8+ (or 26 fxe3 ':d2+ 27 cit>h3 'iVh5+ and wins) 26...cit>c7 27 'iVa5+ g;d7 28 fxe3 'ii'e2+ 29 g;h3 'iVfl + 30 g;g4 f5+ and Black mates. 3b) 24 'iVa8+ c7 25 'iVa5+ g;xb7 26 .1xc5 'iVd 1 + 27 cit>g2 'iVd5+ and White cannot gain the advantage: 
ANAND -IVANCHUK, LINARES 1993 3bl) 28 f3? ':e2+ 29 <it>h3 _e6+ and Black wins. 3b2) 28 ':f3 ':a8 29 _b4+ (29 'iib6+ <it>c8 and White is lost be- cause of the doomed rook on f3) 29...c8 is unclear, but White can- not have the advantage. 3b3) 28 <it>h3 'iif5+ is a draw. 3b4) 28 <it>gl _dl+ is also a draw. 3c) 24 .1xc6 (best) 24..._dl+ 25 <it>g2 _xa4 26 ':xa4 ':xe3 27 fxe3 ':d2+ 28 <it>f3 (28 <it>h3 :xb2 is just a draw) 28....:xb2 with slight winning chances for White. A subsequent game Gi.Hernandez-Howell, Capa- blanca Memorial, Matanzas 1993 ended in a draw. It's certainly possi- ble to repeat this and try to play for a win in the ending, but White's ad- vantage isn't that great. 22 ':b3! .1xd4? (D) This loses immediately, but the alternatives are not much better: 1) 22.. ..:ct7 23 .1xa6 (not 23 ':b4? ':xe3 24 fxe3 _f3 and Black wins) 23....1xd4 24 _xc6 _dl+ 25 <it>g2 _xb3 26 _xd7 and wins. 2) 22....:e7 is relatively best, but after 23 ':b4!, threatening both 24 .1xa6 and 24 d5, White has a clear advantage in any case. 105  .. . - - . . w_'_ B'_' .-.- - . . " .  .U_ \WJ.  ' .\1Ur. .- ,. -/ .-. .:..t " _ d _ U 8" . " " u d u u .  .  .  .  23 .:xb7 +!! <it>xb7 24 _xa6+ b8 2S + <it> a 8 26 _xc6+ <it>b8 27 + <it> a 8 28 .1bS 1-0 Because 28....:c8 29 .1c6+ ':xc6 30 _xc6+ <it>a7 31 'ii'xe8 _d 1 + 32 <it>g2 .1xe3 33 _xe3 leaves White three pawns ahead. Gurevich and I had the final posi- tion on the board during our work in 1991 - an unusual experience for me! I consider myself reasonably well-prepared, but to have worked out the whole game in advance is rare. It is also quite unusual to catch Ivanchuk out in opening prepara- tion. Several players later asked me: "Gee, you out-prepared Ivanchuk?" If it hadn't been for the following game, which I played a couple of days later, I would have considered this my best game from Linares 1993. 
Game 20 B. Gelfand - V. Anand Linares 1993 Queen's Gambit Accepted 1 d4 d5 2 c4 dxc4 3 e4 c5 4 d5 It)f6 5 It)c3 b5 I had prepared this line for my match against Ivanchuk the previous year, when I was analysing with Wolff, but as.Ivanchuk played 1 e4 throughout the match, I had to wait a year before it came up. During the interim I had realized there were some holes in the earlier analysis, but I hadn't really looked at it thor- oughly, so to some extent the spe- cific line played in this game was improvisation at the board. 6 .1f4 _as 7 e5 It)e4 8 It)ge2 It)a6! 9 f3 After 9 a3 .1b7 10 f3 It)xc3 11 It)xc3 It)c7! Black is slightly better because the d5-pawn is weak. 9 ... It)b4! (D) In the second round I played 9. ..It)xc3 10 It)xc3 .1f5 against Be- liavsky, but he responded very accu- rately: 11 g4.1g6 12 a4! and White gained the advantage. In my earlier analysis I had considered 9...ltJb4, but only with the idea of taking the rook on hI (see the note to Black's 11th move). After the Beliavsky game I looked at 9...lt)b4 again, and discovered the move 11...g6!!. I didn't imagine that I would face this line again, but a few days later Gelfand confidently went down the same variation. I was quite happy as I felt sure that he would not have seen 11.. .g6! !. z..t.. .   . . - -  . .. w_ . _ _ - - - . '8R .  -  . -.... - .  .8. -  - - 8 R ..8R  d"l.Jd   ...i..: - -  - 10 fxe4 It)d3+ 11 d2 g6!! (D) I played this and got up from the board. Gelfand sank into lengthy thought, during which time he seems to have convinced himself that he was completely lost. This is an 
GELFAND - ANAND, LINARES 1993 exaggeration, but I think that Black is already slightly better. White's best lines lead to positions in which Black has something like two pawns and a dangerous attack for the piece. 11...lt)f2?! 12 'ii'ellt)xhl is com- pletely wrong. The knight will be trapped after 13 g3 and Black has surrendered all his pressure. After the text-move Black's im- mediate compensation lies in the paralysing effect of his knight. How- ever, in the longer term Black may play ....1g7 and ...It)xe5, when the queenside pawn-mass, supported by the g7-bishop, will become extremely dangerous, especially as White's king will be floating around in the centre. .I.J..*_ _   .  ,., w_  _ _ . _ _i. '8 .  -   .'.8m . . .- . 88 =lb.8D a .'iI..: 12 b3?! This is just a mistake. The alterna- tives are: 1) 12 a4 b4 13 It)bl (13 It)b5 a6 and the knight is trapped) 13....1g7 with fantastic positional compensa- tion for the piece. 107 2) 12 d6 exd6 (12...e6!? is inter- esting, taking away the d5-square; Black plans ....1g7 and ....1b7 with great play, but maybe the d6-pawn improves White's prospects slightly) 13 a4 and now: 2a) 13...lt)xf4 (this was our post- mortem analysis, but there is a hole in it!) 14lt)xf4 .1h6 15 g3 dxe5 16 c2! (the line 16 axb5 'ii'd8+ 17 c2 'ii'xd 1 + 18 ':xd 1 exf4 favours Black) 16...exf4 (16...b4 17 'ifd6! is good for White) 17 'ii'd6! (the flaw; after 17 'ii'd5 O-O! 18 'ii'xa8 b4 Black has more than enough compensa- tion) with an unclear position. One possible line is 17....1e6 18 'ii'c6+ e7 19 'ifxc5+ f6 20 'ifd4+ with a draw by perpetual check. 2b) 13...b4 14lt)d5 .1g7 15lt)f6+ (15 exd6 0-0 is unclear) 15....1xf6 16 exf6 .1e6 intending ...0-0-0. Ad- mittedly this is all a bit speculative, but Black has real compensation. Both line 2b and 12...e6 offer Black good chances against 12 d6. 3) 12 g3 .1g7 13 .1g2lt)xe5 and Black continues with his plan. 4) 12 e3 .1g7 13 g3 (13lt)cl? It)xf4 14 xf4 .1xe5+ 15 xe5 g5! leads to mate in six more moves by 16 d6 f6+ 17 d5 e6+ 18 xc5 'ii'b6+ 19 b4 a5+ 20 a3 'ifc5+) 13...lt)xe5 and again Black has good play. He will continue with ...0-0 and possibly. ..f5. 5) 12 .1g3 .1h6+ 13 c2 b4! (13...'ifb4 is met by 14 'ifbl! with 
108 VISHY ANAND: My BEST GAMES OF CHESS ad vantage to White, and not 14':b 1 ? 1ib3+! 15 axb3lbb4#) regaining the piece. 6) 12 <it>c2 can be met by 12...b4 or 12.. ..1g7 13 <it>b 1 0-0, intending . ..':b8 and ... b4. 12 ... 13 bxc4 14 lbxf4? 14 cxb5 .1xe5 15 1;'b3 lbxe2 16 .1xe2 0-0 is excellent for Black, but the text-move loses even more rap- idly. .1g7 lbxf4 14 ... 15 ltJfe2 .1xe5 b4 (D) ..J...*. _  / .  / . . . w/ ,  . . ...  8 . - - - . .8.8. . .  . - 8. '0liJ.8" d ,, d u a .ii.i..': 16 1;'a4+ White also loses after 16 ':c 1 bxc3+ 17 lbxc3 .1f4+ or 16 1;'b3 bxc3+ 17 lbxc3 ':b8. 16 ... 1;'xa4 17 lbxa4 .1xa1 18 ltJxc5 O-O! I thought about 18...f5, trying to break up his pawn-chain, but then I realized that after 19ltJf4, the arrival of a knight on e6 would only compli- cate the winning process. 19 ltJd3 as Black has a decisive material ad- vantage, but White still has some po- tential counterplay with his central pawns, so the technical phase is still quite interesting. 20 g3 .1g7 White must try to make something of his pawn-mass, but Black will strike with both a left hook (....1a6) and a right hook (.. ..1h6). 21 .1g2 .1a6! 22 c5 22 ':c 1 ':ac8 23 c5 will be a trans- position. 22 ... .:ac8 23 c6 :Cd 8 Threatening 24.. ..1xd3 25 <it>xd3 ':xc6. 24 ':c1 25 lbef4 25 lbdf4 e5 26 .1h3 f5! wins for Black. 25 ... 26 <it>xd3 27 <it>c4 Again, if 27 .1h3 then Black re- plies 27.. .f5. 27 ... exf4 28 ':e1 fxg3 Black still has to be careful; if White could obtain three connected passed pawns then his counterplay could prove troublesome. For this reason 28...f6? is inferior since after 29 ':fl! g5 (29.. .fxg3 30 ':xf6) 30 .1h6+ .1xd3 e5 
GELFAND - ANAND, LINARES 1993 h4 Black will have to be extremely careful. 29 e5 Now 29 hxg3 f6leads to the block- ade of the pawns. 29 ... .1f4 Here the bishop is well-placed to hold back the pawns. 30 hxg3 Or 30 ':e4 g5. 30 ... .1xg3 31 ':e3 .1f4! (D) Such little finesses make life eas- ier. After 31. ...1h2 White could play 32 d6. .-- ... w. . .... .. ... // ." . ,,!« /. U  '/ <it>.  . ",'  - - . .. . -  . . . .. . . . . 109 32 ':e4 .1h2 32...g5 was also good. 33 .1h3 Now White cannot play 33 d6 be- cause his rook is blocking the long diagonal. 33 ... ':c7 34 ':e2 .1g3 35 ':e3 .1f4 36 ':e4 g5 37 <it>c5 ':e7 This move wins tactically. 38 <it>d4 Or 38 d6 ':xe5+ 39 ':xe5 i.xe5 40 .1d7 h5! 41 c7 ':f8 42 <it>d5 (42 .1e8 .1xd6+) 42....1f4 (or 42....1xd6 43 xd6 g4 and after 44 e7, with the threat of 45 .1e8, Black wins by 44....:a8!) winning as 43 .1e8 ':xe8 44 d7 fails to 44....:e5+ 45 d4 ':e4+ 46 <it>xe4 .1xc7. 38 ... f6! 39 d6 .1xe5+ 40 .:xeS .:xd6+ 0-1 I was extremely proud of this game and Gelfand was very sporting; he said that he didn't mind losing such a game and that I would have good chances to win both the best game and best novelty prizes in Informator (in fact I won neither!). The following game was played in the penultimate round (round 12). I was on +3 and having a very good tournament by any standards, but after eight rounds I had been on +4, sharing the lead with Kasparov. However, Kasparov was on even better form and beat both myself and Karpov to take the lead; he went on to win the tournament in very convincing style. 
Game 21 V. Anand - E. Bareev Linares 1993 French Defence 1 e4 e6 2 d4 d5 3 ltJc3 ltJf6 4 e5 lbfd7 5 f4 c5 6 lbf3 a6 7 .1e3 b5 I'm not sure what the point of de- laying ...lbc6 is (the same idea was played in Kamsky-Ivanchuk, Tilburg 1992). I just responded with natural moves. 8 _d2 .1e7 The Kamsky-Ivanchuk game con- tinued 8....1b7 9 .1d3 b4 10 ltJdl lbc6 11 0-0 cxd4 12lbxd4 .1e7 13 ':f3 0-0 14 ':h3 g6 15 ltJf2 and was also won by White. 9 .1d3 g6 This is really asking a bit too much of Black's position. Of course the position is somewhat closed but Black cannot completely neglect his development. 10 0-0 .1b7 11 lbd1 Forcing Black to commit himself. If White is allowed to play c3 then, having reinforced his centre he will be free to play lbf2 followed by ei- ther g4 or ltJg4 as appropriate. 11 ... 12 lbxd4 cxd4 lbcS (D) .-  . .  -. - . - w.-*-. .... .. ..... .'_'D . .  " . d  u d . B1i.m . 1\1\illi1 .I\ oUo_ .OU  ..:  dt.z.Jd  13 b4! A similar idea to White's 18th move in Game 14. White locks the queenside pawn structure and in- creases his dark-square control on that side of the board, too. 13 ... lba4 13...lbxd3 14 cxd3lbc6 15 ':bl is clearly better for White (see Game 14 for a similar type of position), while after 13...lbe4 14 'ii'el, fol- lowed by a4, Black's queenside is undermined. Bareev chooses the best plan, which is to manoeuvre his knight to c4. 
ANAND - BAREEV, LINARES 1993 14 c3 ltJb6! 15 .tf2?! Probably not the most accurate; as we will see, it allows Black a good chance at move 17. It would have been better to play 15ltJb2ltJc4 16 ltJxc4 dxc4 17 .tc2, followed by a4, effectively gaining a tempo over the note to White's 17th move. 15 ... 'fic7 16 ltJb2 ltJc4 17 'fie2?! (D) Even here 17 ltJxc4 would be bet- ter, for example 17...dxc4 18 .te2 ltJc6 19 a4ltJxd4 20 .txd4, followed by 'fie3 and .tf3, with a slight edge for White. .- ... .  . .t.  i  .  i B  _ _  i. .i.i. .i.iD . RJIa\ R  .  U" U  . D.t. . 8 ....80   .   . : ? 17 ... ltJc6? A grave error after which Black's position declines dramatically. It is essential to hold up a4 and 17...ltJa3! seems to do the job for a few moves. After 18 ltJc2 Black just returns to c4 (not 18...'fixc3? 19 .td4), while after 18 ':ac 1 Black just continues 111 his development by 18...ltJd7 19ltJb3 ltJb6 20 ltJa5 ltJbc4, and White's queenside play has been stymied. White's problem here is the poorly placed knight on b2; in the game White manages to swap it off for the well-placed c4-knight. The blockading move .. .ltJa3! is rather unusual and I had simply missed the possibility. 18 a4 Now White has a distinct advan- tage. 18 ... ltJxd4 19 .txd4 (D) .. ... .  . .t. i.i B  _ _  i. .iBiB Bi.iD . 8JIa\   .  U" U  . B.tB .   . .. .  8 R    U R H. .: -  .  19 ... .tc6? A loss of tempo when Black can ill afford it. 19...0-0 was a better chance, although White has various lines that preserve his advantage: 1) 20 a5ltJxb2 (20...f6!? 21 exf6 .txf6 22 'fixe6+ q;g7 gives Black some counterplay for the pawn) 21 'fixb2 f5 although here it would be hard to make progress. 
112 VISHY ANAND: My BEST GAMES OF CHESS 2) 20 axb5 axb5 21ltJxc4 dxc4 22 i.e4 with an edge. 3) 20 ltJd 1 (perhaps the strongest line; White preserves his knight for the kingside attack) 20...bxa4 (oth- erwise White plays a5, and then feeds his knight to the kingside) 21 ltJf2! f6 (21.. .i.c6 22 ltJg4 is even more dangerous) 22 exf6 i.xf6 23 i.xf6 :Xf6 24ltJg4 with advantage. Note that 24....:xf4 loses to 25 1Wxe6+ h8 26 ':xf4 1Wxf4 27 ':fl 1Wd6 281Wf7, with decisive threats. 20 axb5 axb5 21 .:xa8+ The correct move-order. After 21 ltJxc4 dxc4 22 i.e4? Black is not forced to exchange as he can just play 22...0-0. 21 ... .txa8 (D) .i._ ... _ w. II _i.i . .i.i_ _i_iB .  . .  . /  ,; /  . . ii.. .  .\WJB8"  ..  . . .:= 22 ltJxc4! bxc4 Forced, as 22...dxc4 23 i.e4 0-0 (23...i.xe4 24 1Wxe4 0-0 25 ':al is similar) 24:al wins; this may seem a strong evaluation but I don't see any hope for Black. Although in terms of structure Black has the 'good' bishop, in fact his bishop has no squares at all. White can pene- trate down the a-file and Black will soon lose the pawns on b5 and c4. It's an unusual case, but here White's 'bad' bishop is far better than Black's 'good' bishop, which is se- verely restricted by White's pawn chains. 23 i.c2 i.c6 24 'ii'e3 Here White doesn't have a deci- sive entry on the a-file, but he has both a middlegame advantage (at- tack on the kingside) and an end- game advantage (protected passed b-pawn). The move played prepares a strong reply in case Black castles. 24 ... 0-0 (D) After 24...d7 White just contin- ues 25 g4, followed by f5. . . -.. .  i.i w. _ _  R.i..i.iB . BiD . i  .  _  d .  illil . . U _  _ii.B B80 . . .: - - -  25 f5! exf5 
ANAND - BAREEV, LINARES 1993 After 25....:a8 26 f6 .1f8 White can either start an attack with 27 h4 or simply seal Black's bishop in by g4-g5, continuing with both an extra piece and an extra king! 26 .1xf5 (D) B B ..B .  ' . B. _ ,,, . . . B.t.B B'B B .'Di.. ". . . U _ . . . "  . . U _ . . . .D . . .:= Now White has the possibility of e6, opening up the kingside. Once again we can see that White's bish- ops are much more effective than Black's. 26 ... 'iWd8 Or 26....:a8 (26...gxf5?? loses the queen after 27 'ii'g3+ h8 28 e6+) and now 27 e6 gxf5 28 exf7+ xf7 29 'ii'h6 ':f8! is not conclusive, so White should play 27 .1g4, followed by e6, much as in the game. 27 i.g4! To help with e6. The bishop has no more work to do on the bl-h7 di- agonal. 113 27 ... .1g5 28 'ii'e2 h5 (D) Black was almost in zugzwang: 1) 28....1d7 29 e6 wins. 2) 28...'ii'e7 29 .1c5 picks up the exchange. 3) 28....:e8 29 e6 (the simplest) 29...f6 (now that the rook has left f8 White can meet 29...f5 by 30 .1xf5 gxf5 31 ':xf5 winning) 30 g3, fol- lowed by h4, and wins. .  . B   - -  wB . ... B"'. B'. . .." . . . U _ ". .. U _ ...... . 0 . . . BiV.8D . . B:= 29 .1xh5 gxh5 30 'ii'xh5 i.e8 Or 30...'ii'e7 31 h4! winning after 31....1xh4 32 ':f4 or 31....1d2 32 ':f3. 31 M6! 1-0 The end might be 31....1d2 32 e6 .1e3+ 33 hl fxe6 34 ':g6+. A finish reminiscent of the fa- mous encounter Fischer-Benko, US Championship 1963/4. This win put me injoint second place with Karpov on +4 - it was an excep- tionally good result. 
114 VISHY ANAND: My BEST GAMES OF CHESS At the Madrid tournament in June I started with five draws. No matter what I did I would either have to defend a worse position and hang on for a draw, or not be able to break through. At this point there was an article in the tournament bulletin, which said that I was "on vacation". It infuriated me. I don't know if there was any connection, but I won my last four games and finished in joint first place. The following game was the second of the series. 
Game 22 V. Anand - F. Izeta Madrid 1993 Pirc Defence 1 e4 d6 2 d4 ltJf6 3 c3 c6 I was happy with this line as it gave me the chance to play some- thing really sharp. 4 f3!? e5 5 .1e3 .1e7 6 'iid2 0-0 7 0-0-0 b5 8 g4 exd4 8...b4 9 ce2 exd4 10 xd4 c5 11 f5! gives White a good attack, while after 8oo.'iia5!? 9 bl b4 10 ce2 .1e6 11 cl c5 12 d5 .1d7 13 ge2, followed by g3, White is also slightly better. 9 'iixd4!? The right recapture. You shouldn't take with the queen if Black still has the option of ... c6, but in all the lines with .ooc6, taking with the queen is the normal capture. The logic is the same as in the Sicilian after 1 e4 c5 2 f3 c6 3 d4 cxd4 4 xd4; Black mustn't play ...xd4 too early, as the queen occupies a domi- nating post in the centre. After 9 .1xd4 Black could con- tinue 9...b4 10 ce2 c5, and grab- bing a pawn by 11 .1xf6? .1xf6 12 'iixd6 'iia5 would give Black excel- lent compensation. 9 ... i.e6 10 g5 (D) III . _.. B ' B .._. ..t.- . - - - - ... . ' . -. . .  . .  - . ". . n u . . U . :.i.: -  -  10 ... e8 For a time I was worried about 10. ..c5, but it turns out that White can gain the advantage by an accu- rate sequence of moves: 11 'iid3! (not 11 'iid2 b4 12 gxf6 bxc3 13 'iixc3 .1xf6 and Black is a little bet- ter) 11...c4 (11....1c4 12 'iid2 favours White) 12 'iie2! (12 'iid4 c6 13 gxf6 xd4 14 fxe7 'iixe7 15 .1xd4 is risky as ...b4 and ...c3 might rip open the white king position before his minor pieces can come into play) 12...e8 13 xb5 and White stands 
116 VISHY ANAND: My BEST GAMES OF CHESS very well, since 13....1xg5 may be met by 14 f4 and 15 e5. 11 f4 'fia5 12 a3 f5? Black had two better lines: 1) 12...c5!? 13 'fid2 b4 and now: la) 14 ltJd5 .1xd5 15 exd5 (15 'ii'xd5 bxa3! is good for Black) and after 15...ltJd7 or 15....1d8 Black has good counterplay. Ib) 14 ltJbl ! ltJc6 15ltJf3 :b8 16 f5-! .1a2! (16.. ..1d7 17 .1c4 bxa3 18 ltJxa3 with advantage to White) 17 b3! .1xbl 18 xbl bxa3 (18...'fixa3 19 .1c4ltJa5 20 'fid3 is also good for White) 19 'fixa5ltJxa5 20 a2 and White has the two bishops and a structural advantage. 2) 12...f6 13 h4ltJc7 is rather un- clear. After the game !zeta explained that he did not like 12...f6 because White is not obliged to take on f6, but can play 13 h4 (as in line 2 above). He therefore preferred 12...f5, because he thought that it would force 13 gxf6; otherwise White would be left with a backward f-pawn and not much chance of an attack. However, this argument is flawed because af- ter the exchange on f5 Black cannot maintain his bishop on that square, and once it has gone, the white queen can occupy the powerful cen- tral square e4, ready for moves such as .1d3 and f5. 13 exf5! .1xf5 14 ltJge2 Now White has a definite advan- tage. 14 ... ltJc7 15 ltJg3! .1g4(D) If 15...ltJe6, then 16 'fid2 .1g4 17 :el followed by .1d3 with a prom- ising attacking position. -.. / & %;i. /  . .I . - . / . "//  . /"% W %;i   // !t::: / '/ .i' . i. .  _ . . U .  . . _ U.... D     // ;; m+ v/.  /.  ' ,/. ' /  ' , ','; /., "8. .  u d . u  /0:.i..: d /!t::: d . 16 .1d3! Although this is not bad, White misses the most efficient continua- tion, 16 'fie4!: 1) 16....1xdl 17 .1d3 g6 18 f5! (18 'fixe7 is also good) 18....1f3 19 'fixe7! ':e8 (19....:f7 20 'fid8+ ':f8 21 'fixd6 .1xhl 22 fxg6 wins) 20 'fixd6 ':xe3 21 fxg6 with a decisive attack. 2) 16...:e8 17.1d3 g618 f5 .1xdl (18....1xg5 19 'fixg4 .1xe3+ 20 bl is hopeless) 19 fxg6! (19 ltJxd 1 .1xg5 is unclear; Black will win a piece, but meanwhile White will get in fxg6) 19....1f8 (19....1xg5 20 gxh7+ h8 21 'fid4+ ':e5 22.1xg5 wins) 20 gxh7+ h8 (20...g7 21 
ANAND - IZErA, MADRID 1993 h8'ii'+ xh8 22 'ii'h7#) 21 .1d4+ i.g7 22 .1xg7 + xg7 23 'ii' g6+ f8 24 'ii'f6#. 16 ... d5(D) 16...i.xdI17 'iie4 transposes into the preceding note. .. B ..B  -  i w_ _ _ _ Bi. B B illI1 . . R .i.i. u .  .t.   u  D  / .  ,. i. 0 y;  ,.   %j  + + v/. %j  h ' //. ,  / . ; 8. .  u d d u p :. .: .  . d 17 f5! Of course White cannot play 17 ':de 1 c5 18 'iie5 ltJc6 and his queen is trapped, but I always intended to sacrifice the exchange here. 17 ... .1xdl 18 ltJxdl (D) This is again a slight inaccuracy. The alternatives are: 1) 18Jhdl? 19'ii'e5(19'ii'f4 ltJe6! 20 'ii'e5 'ii'xe3+! and Black wins) 19....1d6 20 'ii'xd6 (20 .1xb6 .1xe5 is lost for White) 20...'ii'xe3+ 21 blltJe8 22 'ii'd8 (22 'ii'b4 'iixg5 is winning) 22...ltJf6! 23 'ii'c7 'ii'xg5 24 'ii'b7 ltJbd7 25 'ii'xc6 and Black has a clear advantage. 2) 18 f6! Jhf6! (18...gxf6 19ltJf5 and 18...i.xf6 19 gxf6:xt6 20 Jhdl 117 are hopeless for Black) 19 gxf6 .1xf6 20 'ii'f4 .1xc3 21 ':xdl and White has an extremely strong attack. . . .. - . - . B ' II _ .i BiB B . i.i.8 _ d d u .  . .  .i.  u . _  "8. .  u d . u B =l2JB .: 18 ... .1d6? Black should have tried 18.. .c5 ! 19 'ii'g4 (19 'ii'h4 c4 20 f6 cxd3 21 fxe7 ':e8 22 ltJf5 'iia4! is unclear) 19...c4 and now: 1) 20 'ii'h5? cxd3 21 g6 h6 22 .1xh6 'ii'a4! 23 ltJe3 'ii'h4 24 .1xg7 'ii'xh5 (24...xg7?? 25 f6+! wins for White) 25 ltJxh5 .1g5! and Black WIns. 2) 20 .1e2 (it is not very attrac- tive to have to retreat the bishop, but White retains a very dangerous at- tack) 20....1d6 (20...h8 21 f6 gxf6 22ltJf5 ltJc6 23 g6 wins) 21 f6 g6 22 h4 and the position is still very diffi- cult for Black. 19 4 Now White's attack is decisive. 19 ... ltJe8 Alternatively, 19...ltJd7 20 f6 g6 21 'iih6 ':f7 22 i.xg6 hxg6 23 
118 VISHY ANAND: My BEST GAMES OF CHESS _xg6+ f8 24ltJf5 and White's at- tack breaks through. 20 i.d4 _c7 20...c5 loses immediately after 21 i.xg7! ltJxg7 22 f6. 21 f6 (D) .- ..-.. mmu._ B _ _  .i _iE 0 . mimim  _ _ _ U .  .  . - . - 6 mi..  u _ _  68. .  U d _ U . . .: _ *"'Z.J_ _ 21 ... gxf6 After 21...g6 the most convincing line is 22 :e I! (22 i.xg6 hxg6 23 f7+ xf7 24 'iih7+ ltJg7 is less con- vincing) and now: 1) 22... _f7 23 ltJe3 'iie6 24 ltJgf5! wins. 2) 22...c5 23 i.xg6 hxg6 (White also wins after 23...cxd4 24 :e7 'iixe7 25 fxe7 hxg6 26 exf8'ii+) 24 f7+ xf7 25 'iih7+ ltJg7 26 'iixg7#. 22 gxf6 .:f7 23 :gl h8 24 i.xh7! .:xh7 25 f7+ ltJg7 26 i.xg7+ 1-0 As 26...xg7 27 ltJf5+ f8 28 'iixh7 leads to mate. The summer of 1993 was Interzonal time again. I started with two draws, and the following game was played in the third round. 
Game 23 V. Anand - L. Ftacnik Bie//nterzona/1993 Sicilian, Najdorf 1 e4 2 ltJf3 3 d4 4 ltJxd4 5 ltJc3 6 i.e3 I had quite a few interesting ideas in the .te3 line against the Najdorf, but this is a very difficult and com- plex variation to analyse. It has taken many years for theory to con- verge on what are now considered 'main lines'. The positions are so tricky that you can never be sure that your ideas are correct; the advantage is that your opponent has the same problem! 6 ... e6 7f3 b5 8 g4 h6 9 'iid2 .tb7 This game showed for the first time that 9. ...tb 7 is just a mistake with this move-order. Black has to play 9...ltJbd7, when we reach the main line (after 10 0-0-0 .1b7). In this case White would have no time for 10 h4 because of 10...b4 11 ltJce2 d5, and White cannot push his e-pawn since the e5-square is cov- ered. c5 d6 cxd4 ltJf6 a6 10 h4 Here Ftanik thought for some time and realized that the attempt to transpose to the main line with 10...ltJbd7 is bad after 11 ':gl, with the immediate threat of g5-g6. 10 ... b4(D) After 10....te7 11 0-0-0 White is slightly better because Black has been forced to spend a tempo on ....te7, which he can normally avoid in this system. -- -.. . w..... .._ . . //..   / - - . . . .  8.8n _  . U .  8. 8"8  .   . ' U _     .  i. . : - .   11 ltJce2 d5 11.. .e5 12 'iixb4! gives White a clear advantage after 12...'iid7 13 ltJb3 d5 14 ltJc5 or 12...'iic7 13 'iia4+ ltJbd7 14ltJf5. 12 e5 ltJfd7 
120 VISHY ANAND: My BEST GAMES OF CHESS 13 f4! We have reached a kind of super- French position, but in the French Black would not weaken his pawn structure by playing ....h6. 13 ... ltJc5 14 .1g2! (D) Better than 14 ltJg3 .1e7 - the knight might go to f4 later. . " / ma .   I, _._ ..t. .. B____ .B ...  . d d _ . %j . /.,,/.  . d , u d   D A D _ /.,   0   . . . B 88l2J..t. u _ d .  .  .: _ d  d 14 ... llJbd7 After 14...ltJe4 White can start the attack against e6 with 15 .1xe4 dxe4 16 f5!, followed by llJf4. 15 0-0-0 .1e7 The alternatives are also favour- able for White: 1) 15...ltJb6 16 b3 a5 (16...ltJe4 17 .1xe4 dxe4 18 f5 .1d5 19 ltJf4 and again e6 is the weak point) 17 f5 a4 18 fxe6 axb3 19 cxb3 fxe6 20 ltJf4 iic7 21 bl is good for White. 2) 15...a5 16 g5 (16 bl is also possible, with the idea of ltJb5) 16...h5 17 f5ltJxe5 18llJf4ltJc4 19 iie2 (a typical line; the weakness of e6 proves fatal) 19.. .e5 20 ltJxd5 i.xd5 21 llJc6 'fic7 22 :xd5 win- nIng. 3) 15...ltJe4 16 .1xe4 dxe4 17 f5 ltJxe5 18 f xe6 ltJc4 ( 18.. .ltJf3 19 exf7+ xf7 20 ltJxf3 exf3 21 'fixd8 ':xd8 22 ':xd8 fxe2 23 ':d7+ wins) and now both 19 'fielltJxe3 20 'fif2 fxe6 21 'fixe3 and 19 exf7+ xf7 20 iielltJxe3 21 iif2+ g8 22 iixe3 are promising for White. 4) 15...'ifb6 16 g5 h5 17 g6! fxg6 18 .1h3 with a strong attack. 16 g5! White has the advantage, but he mustn't waste too much time! 16 ... h5 (D) Black is distinctly worse after 16...hxg5 17 hxg5 ':g8 18 g6! or 16...g6 17 h5! gxh5 18 ':xh5. .B _.. _ .J..' w _ _... _ .B B'B B . . u/./. . . " U U   D D _ z,    B B . . 88..t. U Bt.flJ_ _ . /:. .: d , d d 17 f5! The best way to soften up the e6- square. After 17 g6 fxg6 18 .1h3 ltJf8 or 17 b 1 'fib6 18 g6 fxg6 19 .1h3 ltJf8 the situation is less clear. 
ANAND - FTACNIK, BIEL INTERZONAL 1993 121 None of this was home prepara- tion. I had been working on the Eng- lish Attack with Patrick Wolff, and he mentioned that 9....1b7 is bad be- cause of 10 h4, and we left it at that. However, all White's moves appear perfectly natural, so it was not diffi- cult to play. In such messy and com- plex positions, I think it is better not to calculate too much - the tree of variations can get enormously dense. I prefer to wait to see what my oppo- nent plays, and that immediately re- moves a large percentage of possible branches. 17 ... ltJxe5 (D) Forced, as 17.. .exf5 loses to 18 ltJxf5 ltJxe5 19 ltJxg7 +. .. -.. - .J.H i w. . _ _ iB BiB B . -i-8 R i . _ _ U   . p _  . u B B m B 808_ttJ.ii.B . :. .: -  - - 18 ltJf4! 18 fxe6! ltJc4 (18...fxe6? 19ltJf4 wins) 19 ext7+ is also strong, e.g.: 1) 19...<itd7 (Ftanik) 20 .1h3+ <itc7 21 i.f4+ wins. 2) 19...xf7 20 'ifel :e8 (the line 20.. . ltJxe 3 21 'ii'f2+ <itg8 22 'ii'xe3 is also good for White) 21ltJf5 (21 .1f4 <itg8 is less clear) 21...:c8 (21.. .ltJxe3 22 ltJxe3 and d5 hangs) 22 .1d4 .1xg5+ (or else White has a very strong attack in any case) 23 hxg5 'ii'xg5+ 24 <itbl 'ii'xg2 25 :gl with a decisive attack. 3) 19...<itf8 20 'ii'el 'ii'a5 (White wins the queen after 20.. .ltJxe3 21 ltJf4ltJxg2 22ltJfe6+ ) 21 <itb Ilbxe3 22 ltJf4 (D) with the two possibili- ties: .B B . .  . J.. '8' B  .  ,Y ,, iB B B B illi1 _..  .   .. _ .... U" _ %'        0  /', , , /. , /. ,  B . . B 808. Bii.B B.:_ B: 3a) 22...ltJxg2 23ltJg6+ <itxf7 24 'ii'xe7+ <itxg6 25 :hfl! lbxh4 (the lines 25....:he8 26 'ii'f7+ <ith7 27 'ii'xh5+ <itg8 28 g6 and 25...:hf8 26 :xf8 :xf8 27 'ii'xf8 are no better) 26 'ii'f7+ <ith7 (26...<itxg5 27 :gl +) 27 'ii'xh5+ g8 28 'ii'f7+ h7 29 :h 1 mating. 3b) 22...ltJxdl 23ltJg6+ <itxf7 24 'ii'xe7+ <itxg6 (24...<itg8 25 ltJf5 :h7 26 :flltJc3+ 27 bxc3 bxc3 28 'iff8+ :Xf8 29lbfe7#) 25 'ifd6+ with a final branch: 
122 VISHY ANAND: My BEST GAMES OF CHESS 3bl) 25...<ith7 26 g6+ <itg8 (or 26...<ith6 27lbf5#) 27 .i.xd5+ .i.xd5 28 'iixd5+ <itf8 29 'iif7#. 3b2) 25...<itf7 26 g6+ <itg8 (mate is also inevitable after 26... <ite8 27 lbf5) 27 .1xd5+ .1xd5 28 'iixd5+ <itf8 29 'iif7#. 18 ... lbc4 18...exf5 19 lbxf5 is clearly very good for White. 19 'iie2 (D) .- -.- - ..t. . B____ ._ _._ B . -.B8 W . . _ . U _  0 _ _ _ B 8D8B-Bi.B - =:_ B: 19 ... 'iia5 Going for queenside counterplay. The alternative is 19...e5: 1) 20 lbxd5 and now: la) 20...lbxe3 21lbc6! (21 'iixe3 .1xd5 22lL\c6 .1xg2 23lbxd8 .1xh 1 24 'ii'xe5 transposes to the unclear line 1 b below) 21.. .lbxg2 (21.. ..1xc6 22 lbf6+ gxf6 23 .1xc6+ <itf8 24 ':xd8+ ':xd8 25 'ii'xe3 also wins) 22 lbxd8 :Xd8 23 lbxe7 winning. Ib) 20....1xd5 and now: Ibl) 21lbb3lbxe3! (21...lbxb3+ 22 axb3 'ii'a5 23 bxc4 Jlxg2 24 'ii'xg2 'iial+ 25 <itd2 ':d8+ 26 <ite2 'iixb2 27 'iic6+ <itf8 28 ':xd8+ .1xd8 29 .1c5+ <itg8 30 'iie8+ wins for White) 22 'iixe3lbxb3+ 23 axb3 .1xg2 24 :Xd8+ :Xd8 favours Black. Ib2) 21 lbc6 .1xg2 (21...lbxe3 22 .1xd5 'iic7 23 'iixe3 and White wins) 22lbxd8 (22 ':xd8+ :Xd8 23 lbxd8lbxe3 24 ':el <itxd8 25 'iixe3 .1e4 is also unclear) 22.. .lbxe3! (not 22....1xhl? 23 .1xc5) 23 'iixe3 .1xhl 24 'iixe5 (24lbxf7 <itxf7 25 :Xhl is again unclear) 24...f6! (24...:Xd8 25 ':xd8+ <itxd8 26 'iib8+ <itd7 27 'iixh8 g6 28 'iih7! should win for White) and the position remains un- clear. 2) 20 f6!? gxf6 21 lbxd5 .1xd5 (21...lbxe3 22 lbc6 lbxd5 23lbxd8 ':xd8 24 gxf6 .1xf6 25 'iic4 and White wins) 22lbf5lbxe3 23lbxe3 with advantage to White. 3) 20 lbc6! .1xc6 21 lbxd5 (21 'iixc4 dxc4 22 .1xc6+ <itf8 23 ':xd8+ ':xd8 24 lbd5 lbd7 is good for Black) 21...'iia5 22lbxe7! (after 22 .1xc5 'iixa2 23 lbc7+ f8 24 Jlxe7+ <itxe7 or 22 'iixc4 .1b5 23 'iixb5+ axb5 24lbxe7lbd7 25 .1xa8 'iixa8 26lbd5 'iixa2 Black has the advantage) and now: 3a) 22....1b5 23 .1c6+ .1xc6 (or 23...<itf8 24 .1xc5 'iixa2 25 lbc8+! and mates) 24 'iixc4 .1xhl 25.1xc5 .1f3 26 g6! ':f8 27 gxf7+ ':xf7 28 lbc6 and wins. 3b) 22....1xg2 23 'iixc4! lbb3+ (23....1xhl 24 .1xc5 transposes to 
ANAND - FTACNIK, BIEL INTERZONAL 1993 line 3a) 24 cxb3 and White is win- nin:c;) 22....xa2 23 i.xc6+ f8 2 lbg6+ <i1i>g8 2 _xc4 _xc4 26 i.xa8. · & r White WInS 10 ·  b2 ( D ) 20 <i1i>bl x The alternatives are winning for White: 2 1) 20...lLJa4 21 fxe6 lLJcxb ?  f7 23 i.d2 with a decIsIve exf7+ wx lLJ dl 24 attack for example 23... x ':xdlltJc3+ 25 i.xc3 bxc3 26 _e6+ <it>f8 27 i.xd5!. . 2) 20...e5 21lbxd5lbxe3 (White d 4 22 i.xd4 i.xd5 wins after 21.. .ex 23 i.xd5) 22 _xe3 i.xd5 23 xd5 exd4 24 ':xd4! ':c8 25 f6 fimshes Black. .. .=-..- ..i.. Ei. w f. .i'R0. :.  _ //, /. . & .8i ' " -., 0 /.::  '/.  .    - /      _ z    . ., . ' -.\Wi. . 8 8..!,- _<t>_:_ R: 21 fxe6!! White should just ignore evry- thing, except mate itself, and Just hack away. I was on e verge ;: playing the really amazIng ve I'":\. b3 '? if onl y because it elllmnates "LJ · · , lei After all threats to White's ng. 123 21 lbxb3 (21..._a3 22 i.xc5lba4 23..i.d4 lbc3+ 24 i.xc3 bxc3 25 lbd3 wins) 22 axb3lbxdl 23 ':xdl 0-0-0 24 fxe6 I would prefer to be White, but in the end 21 fxe6 seemed even stronger! . . Note that 21 <i1i>xb2? is bad In vIew of 21. ..lba4+ 22 <i1i>c 1 lbc3 23 _d3 e5. 21 ... Or: 1) 21...0-0 (Ftanik) 22 _xh5 lbxd 1 23 ':xd 1 lba4 24 i.d2 lbc3+ 25 i.xc3 bxc3 26 exf7+:Xf7 2?g6 _b4+ 28 lbb3 and Black's posItion collapses. r 2) 21...lbxdl 22 exf7+ <i1i>xf7 (0 22... <i1i>d7 23 ':xd 1) 23 ':xd 1 (D) and 0-0-0 now: .. ."..- R.i._ E.. B !_ _' '-0.:-  //. . ' i_ ,i -_   0 . . . R :-.--i. .<t>.:. . 2a) 23...lbe4 24 i.xe4 dxe4 25 _c4+ e8 26lbde6 ':c8 27lbxg7+ <i1i>f8 28lbg6+ <i1i>xg7 29 i.d4+ leads to mate. I'":\. 2 5 2b) 23...lba4 24 i.d2 c3+ i.xc3 bxc3 26 _e6+ <i1i>e8 27 lbxd5 
124 VISHY ANAND: My BEST GAMES OF CHESS .1xd5 28 .1xd5 _b4+ 29 .1b3 win- ning. 2c) 23....:af8 24lbxd5 winning. 3) 21...fxe6 22lbdxe6! (22 xb2 lba4+ 23 cllbc3 is less clear-cut) 22...lbxdl 23lbxg7+ f7 24 ':xdl wi th a decisive attack. 22 b2! lba4+ 23 cl b3 (D) 23...lbc3 24 _d3 _xa2 25 exf7 wins as Black has no real threats. ..- . . . J.  .   i- w.  . _ i. .. .   . i .  i _   U  .    ...   u BiB _ . .B.B.i.. . m:. .: -  - - 24 lbxb3! Not 24 cxb3? .1a3+ 25 c2 _c7+ 26 <itbllbc3+ and Black wins. 24 ... .1a3+ 25 bl lbc3+ Or 25..._c3 26.1d4. 26 al _a4 Black's compensation is inade- quate. 27 _d3 .1b4 28 lbcl b8 (D) Other lines are also hopeless: I) 28...lbxdl 29 ':xdl fxe6 30 lbxe6 ':d7 31 .1d4 winning. 2) 28....1c6 29 exf7 .1b5 30 .1h3+ b8 31 _d4 also wins. 3) 28....1a5 29 .1d4 (29 exf7? _b4 30 lbb3 _a3 31 lbc 1 _b4 re- peats moves) 29...'iVb4 30 .1xc3 con- solidates the extra piece. Ii //:: . '  . _ ' ,   . %4  fi  /  / wBJ.. .i. i. .. . . BiB 0' \Wr .  P -. _  u  B  -. .  - - . B. ..i..  :. .:   . . 29 .1d4 :c8 30 .1e5+ Even the calm 30 exf7 would have been sufficient. 30 ... a7 (D) 30... a8 loses to 31 lbxd5. ... . .  J. . ., w_   _ ,. .. . . .,m p,  _ _ U \Wr .  R -. .  U - ... . .. ..i.. m :. 8:   - - 
ANAND - FTAtNIK, BIEL INTERZONAL 1993 125 31 'ii'e3+ ':c5 32 ':d3 'ii'xc2 32...d4 33 xd4 'ii'xc2 34 .1xc5+ a8 35 .1xb7+xb7 36':d7+c8 37 'ii'd3 wins. 33 .1xc3 34 .:xc3 35 'ii'xc3 36 exf7 .1xc3+ _xc3+ .:xc3 .:fS 37 g6 1-0 I spent only about 30-35 minutes on this game. Afterwards, I was very proud of my cold-bloodedness dur- ing the game; I hadn't prepared it at home, but even so I played the critical and best moves with very little ef- fort. This game won me the prestig- ious Best Game Prize in Informator. The following game was played in the very next round, and was part of my best streak of the event. 
Game 24 L. 011 - V. Anand Bie//nterzona/1993 Slav Defence 1 d4 d5 2c4 c6 3 ltJf3 ltJf6 4 ltJc3 a6 I played this simply because it was fashionable. 5e3 b5 6 b3 I don't think this is the most dan- gerous system for Black as his reply is quite straightforward: he just takes his bishop outside the pawn chain to g4, swaps it off and then plays ...e6. It is difficult for White to make any- thing of his two bishops. 6 ... .1g4 7 h3 More recently, the refinement 7 _c2 has been introduced. White doesn't mind the exchange on f3, and by playing very precisely he may gain a slight advantage. 7 ... .1xf3 8 _xf3 e6 The alternative is the sharp line 8...bxc4 9 bxc4 e5, but I didn't see the need for such extreme measures. In particular, I didn't like the idea of opening the position by exchanging on c4. White has no particular threats on the queenside, so there seems no reason for Black to clarify the pawn structure. 9 .1d2 9 cxd5 cxd5 10 .1d3ltJc6 is equal. 9 ... .1b4! (D) .. ..  / - - . - wB . .i_i iBiBili . Bi.iB . 8" . . _ U . . .8 "\Wr.8   u. 8.  8.  _ u d  . i..: _ d  d 10 _dl 10 a3 Jlxc3 11 Jlxc3 bxc4 12 bxc4 ltJe4 13 ':cl ltJxc3 14 ':xc3 'ii'a5 15 d2 is certainly not better for White. 10 ... 0-0 11 .1e2 After 11 cxd5 (11 .1d3 dxc4 12 bxc4 c5! is also fine for Black) 11...cxd5 12 a4 bxa4 13 ltJxa4 (13 ':xa4 a5 is also comfortable for Black) 13...ltJc6 14 .1xb4ltJxb4 15 .1e2 'ii'd6 the position is level. 
OLL - ANAND, BIEL INTERZONAL 1993 127 11 ... bxc4 12 bxc4 c5! At this point I felt that I had equal- ized completely. 13 dxc5? After this error White is strug- gling. The alternatives are: 1) 13 0-0 cxd4 14lbxd5 lbxd5 15 cxd5.1xd2 16 'iixd2 dxe3 17 'iixe3 exd5 18 .1f3 and now 18...lbc6 19 ':adl lbe7 20 ':fel ':e8 21 .1xd5 lbxd5 22 'iixe8+ 'iixe8 23 ':xe8+ :Xe8 24 ':xd5 is a dead draw, while 18...d4!, followed by ....:a7-d7, is a possible way for Black to play for the advantage. 2) 13 cxd5 (the soundest line) 13...cxd4 14 exd4 .1xc3 (14...lbxd5 15lbxd5 .1xd2+ 16 'iixd2 exd5 17 0-0 is a little better for White) 15 .1xc3 lbxd5 with equality. Even though White has an isolated pawn, the two bishops and the potentially weak pawn on a6 will ensure that he is not worse. 13 ... d4! (D) .11 Ii _..  . _  . WA ... W  _ _ .. .'11 . . 0 . . 8 . . - - . . .    . .8 .  U  8. 8 . - . ....u   ..  . .:  .   14 exd4 Forced, as 14 lba4 'iia5 favours Black, while 14 lbe4? lbxe4 15 .1xb4 dxe3 is just lost for White. 14 ... 'iixd4 15 'iic2 15 :cl :d8 prevents castling and is very awkward for White. 15 ... lbc6 16 0-0 (D) .. . -.. B. . .._. '...'11 . .  . . _ u _ _ 8 .  . - - .  .   . -  . 8 -   - 8  . \Wri. 8  .  ... U    .  ..:   -  16 ... 'iie5 16....:ad8 was also good, for ex- ample 17 ':ad 1 'iie5 18 .1d3 .1xc5 or 17 ':fdl 'iie5 18 .1d3 (18 .1f3 :xd2!) 18...lbd4 19 'iiD2 .1xc5 with a clear advantage to Black in either case. 17 'iia4 Alternatively: 1) 17 ':fe 1 ':ad8 18 .1f3 ( 18 ':adl lbd4 is also good for Black) 18...lbd4! (not 18....:xd2 19 'iixd2 .1xc3 20 ':xe5) 19 ':xe5 lbxc2 20 ':dl ':xd2 21 ':xd2 .1xc3 and Black wIns. 
128 VISHY ANAND: My BEST GAMES OF CHESS 2) 17 :ael :ad8 (the simple 17..._xc5 is also good) 18 .1f3lbd4 19 ':xe5 lL\xc2 20 lbbl .1xd2 21 :e2 lbd4 22 :xd2 lbxf3+ 23 gxf3 ':c8 with a clear endgame advantage. 3) 17 .1d 1 ! is relatively best, but Black retains the advantage after 17...:ad8 or 17...lbd4. 17 ... :ad 8 18 .1el lbd4 (D) B . _.B wB B B... .B B.. _ . R  . . U _  .8- n. _ - - - - _  B B8 8B Bii.D8B  . : M . _  19 'iixb4 19 .1d3 .1xc3 20 .1xc3lbf3+ 21 gxf3 'iixc3 is winning for Black, while after 19 .1dl .1xc5, intending ...lbf5-g3, White has a miserable po- sition. 19 ... lbxe2+ 20 lbxe2?! Losing quickly. The last chance was 20 <iii>h 1 ':d3 (20.. .lbe4 21lbxe4 'iixal 22lbd6) 21 :dl :Xc3 22.1xc3 lbxc3, when Black is much better although the passed c-pawns offer some hope of counterplay. 20 ... 'iixal 21 lbc3 'iicl! 22 'iia5 'iif4 23 'iixa6 :a8 24 'iWd6 'iixc4 0-1 As White will soon lose the c5- pawn. In the fifth round I won against Khalifman, and was in joint first place. Then I lost recklessly against Gelfand; had I defended well I could have held the position, but I was careless and suddenly it was allover. This put me on +2, below the level necessary for qualifying. I tried very hard to improve my score but despite reaching a number of winning positions I failed to score the full point in any of them. I had given up hope when I met Korchnoi in round 12. In this game Korchnoi completely outplayed me and I felt a huge weight off my shoulders. It was suddenly clear that I wasn't going to qualify and I felt enormously relieved at the lifting of the burden. Then Korchnoi started to have difficulty winning and I began to regain my interest in drawing the game. Korchnoi gradually went completely awry and even lost the game. Suddenly I had moved from absolutely nowhere to having a theoretical chance of qualifying. A win in the last round would have guaranteed qualification, but I could only draw with Epishin. Then it required five different games to have the 
BIEL INTERZONAL 1993 129 right result for me to qualify. In the end everything worked out perfectly: all the people with worse tie-breaks than me finished level with me on +3 and all the people with better tie-breaks ended up on +2. It wasn't very convincing, but at least I had reached the Candidates again. The following game is from the final of the European Clubs Cup, in which I was playing for Lyons. Curiously, the only French player in the Lyons team was Lautier. Michael had a horrible event: he lost all three games (of which this was the first). He then went on to Groningen for the PCA Qualifier and almost lost the first game there. However, he salvaged a draw and, typically for Michael, then went on to win the tournament! 
Game 25 M. Adams - V. Anand European Clubs Cup Final, Hilversum 1993 Sicilian, Kan 1 e4 c5 2 lbc3 e6 3 lbf3 While White gains some advan- tages with this move-order, it allows Black to adopt a system in which ...lbf6 is delayed. 3 ... a6 4 d4 cxd4 5 lbxd4 d6 6 f4 b5 7 .1d3 .1b7 Thanks to the omission of .. .l2Jf6, Black is able to accelerate his queen- side development. 8 0-0 One advantage of Black's system is that 8 _f3?! doesn't really work when Black hasn't played ...lbf6 since g4-g5 threatens nothing! The game Anand-Wojtkiewicz, Manila Olympiad 1992 continued 8...lbd7 9 .1e3 ':c8 10 g4lbc5 II g5 lbe7 12 lbb3lba4 13lbxa4 bxa4 14lbd2 d5 and Black was slightly better. If White wants to adopt the _f3 and g4-g5 system then he shouldn't play 2l2Jc3. 8 ... lbf6 Now that White has effectively abandoned the _f3 system (because the advance of the g-pawn doesn't fit in with kingside castling) I decided to play ...lbf6. Black can still tinker with his move-order by 8...l2Jd7!?, but after 9 <iii>h I (not 9 f5? _b6!) he probably has nothing better than 9...l2Jgf6. 9 a3 Meeting Black's threat to the e- pawn, but after this loss of tempo Black has no problems. 9 ... lbbd7 10 <iii>hl .1e7 (D) 10....:c8 and 10...g6 are playable alternatives. ZB _._ . ..JIa\_'' W ........ _ .. .- . . - - . _._ _ B . 8" . .  U . o i._ B R8. .8 u .  u a fI.B:B<iit 11 b4?! This plan can be effective, but only when Black's knight has been 
ADAMS - ANAND, HILVERSUM 1993 developed to c6. Here Black has a ready counter in the form of ...:c8 and ...b6-c4. 11 ... 0-0 12 .1b2 ':c8! 13 'iie2 Preparing e5, which Black meets by providing the f6-knight with a well-supported square on d5. 13 ... lbb6! (D) ...  . . . - - . ....B .. w. d _ _ .- .- . - - - . B'B B . " 8" .     " . .  .... . 8B..8" _ d d  a . .:. 14 e5? If White tries queenside play by 14 a4 c4 15 .1cl 'iib6 16 f3 :fd8, then Black has a promising position; he has played normal moves but White has done disgusting things on the queenside. The continuation might be 17 axb5 axb5 18 e5 and now: 1) 18...dxe5?! 19 fxe5: la) 19....1xb4 20 a2 .1xf3 21 .:xt3 d5 22 .fu.b4 xb4 23 .1xh7+ xh7 24 :h3+ g8 25 1Wh5 wins for White. Ib) 19...d5 20 g5 .1xg5 21 .1xg5 xc3 22 .1xh7+ xh7 23 131 'iih5+ g8 24 'iixf7+ h8 25 :f6 e4 (25....1xg2+ 26 xg2 'iic6+ 27 h3 e4 28 .1h6! and White wins) 26 .1h6 f2+ 27 :xf2 gxh6 (27....:g8 281Wh5! g6 291Wh4 wins) with a likely draw. lc) 19...d7 20 g5 .1xg5 21 .1xg5 :f8 22 .1f4 is unclear. 2) 18...d5! 19 xd5 .1xd5 is fine for Black. 14 ... dxeS 15 fxe5 fd5! (D) Not 15...'iixd4?? 16 exf6 .1xf6 17 a4! (17 d 1 ? 'iih4! 18 .i.xf6 gxf6 favours Black) 17... 'iid8 18 .1xf6 gxf6 19 xb6 1Wxb6 20 1Wh5 f5 21 'iig5+ h8 22 'iif6+ g8 23 :xf5 and White wins. ...  . . . - - . .... .. w ..... _ _ .11 B'B . .'BaD .  " . .   . .  i.. . u  . . _8."'.80 a . .:. 16 xd5 I spent a long time on the alterna- tives, but I eventually decided that they were not worrying for Black: 1 ) 16 xe6 fxe6 1 7 'iih5 :f5! (17...g6 18 .1xg6 hxg6 19 'iixg6+ h8 20 'iih6+ g8 21 'iixe6+ is 
132 VISHY ANAND: My BEST GAMES OF CHESS only dangerous for Black) 18 ':xf5 g6! 19 ':g5 .i.xg5 20 .i.xg6 ':c7! would be winning for Black. 2) 16 e4 and now: 2a) 16...a4?! 17 'iih5 with a further branch: 2al) 17...h6?! 18 .1cl with very dangerous threats. 2a2) 17...lbxb2?! 18 ltJg5! (18 f6+ ltJxf6 19 exf6 lbxd3 wins for Black) 18...h6 (D) and now: . .. illt1  .  . .-- - .  . .t. i w  _ _ iB BiB _  .  i  . ."      ". .    . , r#/  .  ..t .  . // /   - 8. 88" __i a B .:. 2a21) 19 ':xf7 f4 20 ':xg7+ tj;xg7 21 gxe6+ g8 22 'iixh6 .1xg2+ 23 gl 'iixd4+ 24 xd4 bxd3 25 cxd3 .1d5 is unclear. 2a22) 19 .1h7+ h8 20 xf7+ ':xf7 21 ':xf7 f4 22 ':xf4 xh7 and Black defends. 2a23) 19 xf7 ':xf7 (but not 19...1We8 20 .1g6 and now 20...'iid7 21 ltJxh6+ gxh6 22 .1f7+ ':xf7 23 ':xf7 f4 24 ':xf4 .1g5 25 'iig6+ 'iig7 26 'iixe6+ wins for White) 20 1Wxf7+ h8 21 xe6 'iig8 22 'iig6 !iJxd3 with a murky position. 2a3) 17...g6! 18 'iih6 xb2 19 ':f3 lbf4! 20 'iixf4 (20 ':xf4 .i.xe4 21 ':xe4 xd3 22 cxd3 .1g5 with a clear advantage for Black) 20...xd3 21 cxd3 .1xe4 and Black has the same type of advantage as in the game. 2b) 16...lbc4! (even stronger than line 2a3 above) 17 'iih5 (17 .1xc4 :Xc4 followed by ...'iia8 gives Black strong pressure) 17...h6! (17...g6 18 'iih6 xb2 is line 2a3) 18 .1c 1 de3! with a large advantage for Black. 16 ... .1xd5 Now Black has a safe advantage. The exchange of knights has extin- guished White's hopes of a success- ful kingside attack, and his position is structurally much worse. 17 .:ael .1c4 (D) W .. illil _ . _ /-      (!f;/   .&  . / .& W  / rf$4 0 ..  ,:::0;; . / /', / i RiB . _ _  0/ Bi. B . "I. . . ... W . D B..tB B 8.iY . 8" - .   . . D:B 18 f3 The sacrifice 18 .1xh7+? <iii>xh7 19 'iih5+ g8 is not dangerous: 1) 20 ':f3 g6 21 ':g3 g7 22 ':fl (22 f5+ exf5 23 e6+ .1f6 24 e7 
ADAMS - ANAND, HILVERSUM 1993 'iid6 25 exf8'ii+ J%xf8 consolidating the extra material) 22...'iie8 23 J%f6 J%h8 24 'iig5 d7 and Black wins. 2) 20 J%e3! forces accurate de- fence: 2a) 20...g6? 21 J%g3 'iie8 (the line 21....1xfl 22 J%xg6+ also draws) 22 1txf7 (22 J%f6 rt;g7) 22...'iixf7 23 1hg6+ 'iixg6 24 'iixg6+ rt;h8 with a draw. 2b) 20....ig5? 21 J%h3 .ih6 22 J%f6! with a dangerous attack. 2c) 20.. ..ixfl! 21 J%h3 .ixg2+! 22 rt;xg2 'iid5+ 23 rt;gl f5! (utiliz- ing the pin) 24 'iih7+ rt;f7 25 J%g3 (25 'iih5+ g6 26 'iih7+ rt;e8 27 'iixg6+ rt;d7 and the king escapes) 25... rt;e8 26 J%xg7 J%c7 is hopeless for White. 18 ... .ixd3 19 cxd3 (D) ... ... B 8 / ' . .. d__ .. ..- . ... 0 . 4' . '  :, . %g D B .8..  .".8" ,...u . . a:. 19 ... d5? A clear error. Black has too many pieces wanting to occupy d5, but whereas the queen does not have a 133 good alternative post, the knight would also be well-placed on a4. Therefore the correct arrangement is queen on d5 and knight on a4, which Black could have achieved by either 19...'iid5! or 19...ltJa4!, with a dis- tinct advantage. 20 'iid2! Now I realized my error. I consid- ered 20...a5!? 21 bxa5 b4, but after 22 J%al I decided that the complete dissolution of the queenside would not help Black's winning chances. However, this line might have been objectively best in that Black would preserve a slight edge. 20 ... J%a8 Perhaps 20...ltJb6!? is best, re- verting to the correct plan, although this would be psychologically diffi- cult to play. 21 ltJd4 .ig5 22 'iin (D) ..  .. -_. B B / . .//. - >  /,,/// .. ... . ..p."  ---   . . U  $   r8. . u . _ .  . 8 _W_U . . a:. 22 ... .ih4 23 g3 
134 VISHY ANAND: My BEST GAMES OF CHESS This may prove to be weakening in the future, but one cannot pretend that it is a serious matter. 23 ... .1g5 24 h4 .1h6 25 h2 :c8 26 .1cl?? (D) But this is a serious mistake. 26 ':e4? is also bad after 26...f5! 27 exf6 (27 ':e2 1Wd7 should also win for Black, because 28...f4 is a threat) 27...ltJxf6, but 26lbb3! is the correct move. White has a good outpost on c5 and this compensates for Black's theoretical structural advantage; the position would be roughly equal. 8" .8 . - - . B _ _ B'_' .. ... . _._n . o  . "    " 888 " u d d   . Wit  . . - * 8  :. . - - . 26 ... .:xc 1 ! 27 .:xci e3 28 h3? 28 ltJc6! was the best chance: I) 28...1Wxd3 29 1Wf3! (not 29 1Wxf7+ :xf7 30 ltJe7+ f8! nor 29 ltJe7+ h8 30 1Wxf7 ltJxfl +, as Black wins in both cases) 29...1Wd2+ (29...ltJxfl + 30 :xfl1Wxf3 31 ':xf3 is only slightly better for Black) 30 h3 lbxfl 31 :xfl and White has defensive chances. 2) 28...ltJg4+ 29 g2 (forced, as 29 gl 1Wxd3 wins) 29...1Wxd3 (not 29...ltJxf2? 30 ltJxd8ltJxd3 31 :c8! defending) 30 1Wf3 and again the win for Black is not guaranteed. The text-move plays for a trap, but if Black avoids this the game is over. 28 ... lbxn! Avoiding 28...1Wxd4?? 291Wxf7+ :xf7 30 :c8+ :f8 31 ':fxf8#. 29 .:xn _d5! (D) . . -.. .  .&& wR  .-.- .8 ..B  d _ d _ .'..0 .   8    .   .88   ; - '&I __12M _ M m:. A dream position for Black. White has several weak pawns, an exposed king and passively placed pieces. Black only needs to bring his rook into play and White will start shed- ding pawns. 30 g4 g6 31 :el :c8 32 lbf3 :c3 33 g5 .1f8 
ADAMS - ANAND, HILVERSUM 1993 135 34 ':e3 35 d4 36 _xe3 37 g3 38 4 .:xa3 .:xe3 .1xb4 as a4 39 _cl _c4 40 _bl a3 0-1 The finish might be 41 'iie4 a2 42 _a8+ .1f8. I haven't played much club chess, the total being the one year I played for Lyons. That year was quite pleasant because I just turned out for the big matches. However, I wasn't especially excited by club chess, so I wasn't too disappointed when the club just folded up - at least I didn't have to resign from the team! I might play for a club again in the future, but there is no im- mediate prospect of this. The next game is from the PCA Qualifier event held in Groningen during December 1993. 
Game 26 V. Anand - A. Beliavsky PCA Qualifier, Groningen 1993 Sicilian, Najdorf le4 c5 2 ltJf3 d6 3 d4 cxd4 4 ltJxd4 ltJf6 5 ltJc3 a6 Round about this time Beliavsky had started playing the Najdorf and it was quite clear why - he had been one of Kasparov's seconds for his match against Short. Since Short was one of the players who popularized 6 .i.e3, they must have looked at it very deeply. Unfortunately, I didn't have anything else against the Naj- dorf, so I decided to play it in any case. 6 .i.e3 e6 7 f3 b5 8 g4 h6 9 'iid2 ltJbd7 Beliavsky doesn't repeat Ftacnik's imprecise move-order - see Game 23. 10 0-0-0 .i.b7 11 h4 b4 12 ltJce2 d5 The big question was whether to play 13 .i.h3 or 13ltJg3. The former seemed rather foolhardy, as it in- volves various pawn sacrifices. How- ever, when we had the position after 12.. .d5 on the board, I just decided to gamble. 13 .i.h3!? White is committed because quiet play doesn't work, for example 13 exd5 ltJxd5 14 ltJf4 ltJxe3 15 'iWxe3 'iWb6 16 .i.c4 0-0-0 17ltJd3 'iWc7 was slightly better for Black in Sax- Anand, Philadelphia 1986. 13 ... dxe4 14 g5 (D) .}  .  . a.  _ mJ B .i...... .   '//  %/,> :; . ' . . ' . U// . / / c,> : ;;:;; ;;:j  /'/ /// 0! //" / / %:f:%   ' :yt: . ' // //// / / // . /  fij ';' ///,' . ;  . ,. ; , f!J  '.  "/  / / '/'l////" / Y %:1:: . , .  /:;; . / . . i. //0/ /  / "//" ' f::   1/:<://  ffj '   "/ l2J <:(;; f;:::;;: o /; / /. % 'i /;;i/;:' /" '/ ;; % /// ///,> . ;.:;;:;'0 ,/,    ;;.::::;:/ ;;:///;;'-h -  " %" .   % 14 ... hxg5 Beliavsky played all these moves more or less instantaneously. The text-move is the best move-order, as after 14...exf3 the reply 15 gxf6! fxe2 16 'iWxe2 is very dangerous for Black, for example 16.. ..i.xh 1 17 
ANAND - BEUAVSKY, PCA QUAUFIER, GRONINGEN 1993 137 ltJxe6 or 16..:iixf6 17 ':hfl. Prepa- ration for a World Championship match needs to be extremely thor- ough - one must not only take the existing theory a little bit further, but in fact almost reinvent the lines you expect to use since it must withstand several months of scrutiny by a team of grandmasters. I had noticed this little detail when I looked at the line, but didn't expect that players such as Kasparov and Beliavsky would have missed something like this when preparing for a world championship match! 15 hxg5 exf3 16 ltJf4 Now 16 gxf6 fxe2 17 'iixe2 'iixf6 18 ':hfl 'iie5! is fine for Black, as the bishop on h3 is hanging. 16llJg3 is a major alternative, but in this book I am avoiding getting embroiled in opening theory. 16 ... ltJe4 17 'iiel (D) .  .  /.' //' .. 0' %i'  / ,/ 0     //:/7& %  /,,; ,/. B.....4\... .. ... . . . .  _ n////// /%;  . f%  /. 4\  / . ;       /, /  /. /. / . . ;..i. D. . . . : .: .  -  This was all theory and I was wondering when his novelty was go- ing to appear. 17 ... f2 This was it, but it became clear the following year that it was not best. The alternative 17...ltJxg5? is bad: 18 ltJdxe6! fxe6 19 xe6! (19 .1xe6 is also possible: 19...ltJxe6 20 ':xh8 'iif6 21 ':d6! 0-0-0 22 ltJxe6 ':e8 23 'iixb4 f2 24 'iic4+ b8 25 'iic7+ a8 26 :Xa6+ 1-0 Romero Holmes- Tukmakov, Wijk aan Zee 1991) 19.. .ltJxe6 20 .1xe6 ':xh 1 21 'ii xh 1 ltJf6 (21...'iia5 22 .1xd7+ f7 23 'iih2 is clearly winning for White) 22 ':xd8+ ':xd8 23 'iigl and White has a clear advantage. After my success in the Beliavsky game, I continued playing the line as I felt that if I had won against Kaspa- rov's preparation, it must be good for White. However, the following year I played the variation once too often and Ljubojevic brilliantly re- futed White's play by 17....:xh3! 18 ltJxh3 e5 19 ltJb3 (19 ltJf5? 'iia5 20 b 1 .1d5 21 a3 ':b8, threatening 22...ltJc3+, and 19 ltJxf3? 'iia5 20 bl c3+ 21 bxc3 .1xf3 are both very good for Black) 19.. .a5 with a clear advantage for Black, Anand- Ljubojevic, Sicilian theme tourna- ment, Buenos Aires 1994. I won this game anyway, but only due to Ljubo's blunders in time pressure. At the present time the line remains unplayable for White. 
138 VISHY ANAND: My BEST GAMES OF CHESS Actually the Ljubo near-disaster was the result of a lapse of memory on my part. Ljubo had come to Gron- ingen and after the Beliavsky game he mentioned 17...':xh3! to me, but later I imagined that it was Van Wely who suggested it. I therefore felt it was safe to play the line against Ljubo several months later, but not surprisingly Ljubo was baffled as to why I had played the line against him when he had already told me the refutation. These days I note down not only the ideas, but also who told them to me! 18 .1xf2 'fixg5 Not 18...xf2? and White has a pleasant choice: 1) 19 .i.xe6! fxe6 (19.. .':xh 1 ? 20 .1xf7+! xf7 21 'fie6#) 20 dxe6 'fie7 21 ':xh8 xdl 22 g6 win- . mng. 2) 19 dxe6 fxe6 20 'fixe6+ 'fie7 21 'fig6+ d8 22 ':xd7+ 'fixd7 23 'ii'b6+ e7 24 g6+ is also deci- Sl ve. 19 .1e3 'iih4 After 17...f2 I was of course wor- ried about my preparation, as if this was indeed part of Kasparov's world championship preparation, then I could expect that it would be very well analysed. However, this move really surprised me, because I saw that I could reply 20 dxe6. The more I looked at it, the better it seemed, and it appears that the Kasparov team must have overlooked some- thing in their analysis. 19...'fie7 20 dxe6 ':xh3 21 xg7+ .1xg7 22 ':xh3 was a possi- ble alternative, with an unclear posi- tion. 20 dxe6! (D) Not 20 'fie2 g3!. I.. ... . 8.tBJIL\.' / B. ....  .. .lb_ _ _ . - -  .JIL\  - .. - _ _ _ Bio B_ _ _ . : .: .  - . 20 ... 'fixel ! Beliavsky played this move very quickly. The alternative is 20...fxe6 21 .1xe6 'fixhl (21...'fixel 22.1xd7+ f7 23 .1e6+! should win for White after 23... f6 24 .1d4+ <it>g5 25 ':hxe 1 <it>xf4 26 ':fl + <it>g5 27 ':f5+ or 23...<it>e8 24 ':hxel) 22 .1xd7+ f7 (22... <it>d8 23 .1b6+ <it>e7 24 'fixb4+ and 22...<it>e7 23 'fixb4+ are hopeless for Black) and now: 1) 23 'fixhl ':xhl 24 ':xhl c5 25 ':d 1 ':d8 is unclear. 2) 23 .1e6+ and now: 2a) 23...<it>f6 24 .1d4+ <it>g5 25 'fie3 'fixdl+ (25...':h2 26 d5+ <it>h5 27 .1xg7! 'fixd 1 + 28 <it>xd 1 
ANAND - BEUAVSKY, PCA QUAUFIER, GRONINGEN 1993 139 J:thl+ 29 <it>e2 and White wins) 26 xd 1 ':h 1 + 27 e2 ':h2+ 28lbg2+ g6 29 .1f5+ xf5 30 1Wf4+ fol- lowed by 31 1Wxh2, winning. 2b) 23...e8 24 1Wxhl J:txhl 25 J:txhl is slightly better for White. 3) 231We2! and now: 3a) 23...1Wh6 24 'ii'c4+ e7 25 1Wc7 wins. 3b) 23....1d6 24 1Wc4+ e7 25 ltJg6+ <it>xd7 26ltJe5+ d8 27 .1b6+ e8 28 'ii'f7#. 3c) 23..._xdl+ 241Wxdl .1d625 1Wfl! and the two rooks are not a match for White's queen, for exam- ple 25....1xf4 261Wxf4+ e7 27 .1h3 with a very strong attack. 21 ltJxg7+! Not 21 lbc7+? d8 22 ':hxel ':xh3, which favours Black after 23 lbxa8 ':xe3! 24 ':xe3 .1xa8 or 23 .1b6 ':h6. 21 ... d8 (D) 21....1xg7 22 .1xd7+ f8 23 ':hxe 1 leads to a slightly better end- ing for White. ...  . z   .  - - w..l.a_. .. . . . . . . .  .a . _ d  d . . m Bi. D. _ . . : B: /_. 22 J:thxel! After 22 .1b6+?! e7! 23lbf5+ (23 J:thxel lbxb6 wins for Black) 23...f6 (not 23...e8 24 ':hxel lbxb6 25 .1g2 and White retains some advantage) 24 .1d4+ g5 25 ':dxe 1 ':xh3 26 lbxh3+ xf5 27 J:thfl + g6 28 ':g 1 + the position re- solves to perpetual check. 22 ... Jbh3 23 lbxh3 .1xg7 24 ltJg5! (D) Now I was really happy - White is going to get Black's last kingside pawn and as a result can play for a win at absolutely no risk. Did Beliav- sky miss that 24...lbxg5 is met by 25 .1b6+ ? ... .  z . . -  - - BB.l.a_.. .. . B B - . .  . Ba. . . .  .   -  D. . . . : - -  - - 24 ... <it>e8? (D) Not the best square for the king. The alternatives are: 1) 24...lbxg5?? 25 .1b6+ c8 26 ':e8#. 2) 24...e7?? 25 lbxe4 .1xe4 26 .1g5+ wins. 
140 VISHY ANAND: My BEST GAMES OF CHESS 3) 24...ltJf6 25 ltJxf7+ c7 26 .i.f4+ b6 27 ':e6+ is clearly better for White. 4) 24... c8! 25 ltJxf7 a5 was the correct choice, with just a minimal advantage for White. Black should draw but White can still poke around for a few moves. ... ... . i..JIL\.' ' W _ _"h ,; .. . . . . . .. : "fi%; C'/  - ... . · - m · . .  . 8?jf]8. . . ;  '0 " a /'      .  %  . ;ij/// , '/ uu  25 ltJxf7! Now I felt that I was going to win this game. Black is in considerable difficulties. 25 ... ltJe5 Here the alternatives make grim reading: 1) 25....:c8? 26 ltJd6+ ltJxd6 27 ':xd6 wins. 2) 25...xf7 26 1txd7+ is also hopeless for Black. 3) 25....i.f8 26 ltJg5 ltJdf6 27 ltJxe4 ltJxe4 28 ':d4 .i.e7 29 .1f4 ltJf6 30 .1d6 ltJg8 31 1txb4 .1c6 32 ':g4 with excellent winning chances for White, much as occurs in the game. 4) 25.. .ltJdf6 26 ltJd6+ ltJxd6 27 ':xd6 and White has a clear plus. 26 ltJd6+! ltJxd6 27 Jbd6 f8?! (D) Or: 1) 27...f7? 28 .i.d4 ':e8 (White wins after 28...llJf3 29 1td7+) 29 .1xe5 .1xe5 (or 29...1txe5 30 ':xe5 .ixe5 31 ':d7+) 30 ':d7+ f6 31 1tfl+! e6 32 1txb7 and wins. 2) 27...ltJf3? 28 ':e2 is very good for White. 3) 27....:d8! (probably Black's best chance) 28 1tb6 gives White a distinct advantage. . A/0 .   . r:rfj  r . // % ijj% 0   ,/ n . . %  . )j;; -  I ' ;:' . ; f% {&J3.    0&   ' w /;/;;0'  uu/ f$'/ 4»:i/'  ;@    ' //;/; c%; ;,;/;   tI / / 'w W&0 ;,&' / + //  . fd?  . 888. . .   = :;: A /0,;, a /'   IJ /. // &   /;  /;///$  , -:. ,/,,-j,/; z. . _ uW .Aa% w..... ..  . 28 .i.h6! Very strong. If White can exchange bishops and win the b-pawn, then Black will be facing defeat, since Black's king will be too far away to defend the queenside. 28 .id4? is less accurate on account of 28....:e8. 28 ... .1xh6+ 28...1te81oses to 29 ':f6+ g8 30 1tgl ':e7 31 1tb6. 
ANAND - BEUAVSKY, PCA QUAUFIER, GRONINGEN 1993 141 29 Jbh6 ltJf7 30 J:tb6 .1d5 31 Jbb4 J:tc8 31...a5 might have made things slightly more difficult, but would not fundamentally change the position. 32 J:tb6 as 32...J:tc6 33 J:txc6 .1xc6 34 d2 is a simple win. 33 a4! (D) Not 33 J:tb5? .1xa2 34 b3 a4 35 b2 axb3 36 cxb3 .1xb3 drawing. B.B / . B B _ ... R / rf E W . . .  B.t.B 1&  ?W _ J}i . . B . . . . . . . . U . _ . .   . ._/ 33 ... .1c6 34 b3 .1d7?! Once again Black could have dragged the game out by defending more accurately with 34...ltJd8, but in view of White's material advan- tage and Black's poorly placed king, White should win in the long run. I was just going to play quietly with 35 b2, but White still has to break Black's resistance. 35 b2 J:tc5 35...ltJd8 36 J:tf6+ drives the king even further away. 36 J:tb8+ g7 37 :b7 .1c8 Or 37....1f5 (Black also loses af- ter 37...J:td5 38 J:te7 and 37...J..c6 38 J:tc7) 38 c4 f6 39 J:tb5 lbe5 40 c3lbd7 41 d4 and the queenside pawns will decide. 38 J:tb5 1-0 The importance of this game, which was played in round five, cannot be overestimated; it was the encounter which put me on the road to qualification for the PCA Candidates. Before it, Beliavsky was on +4 and I was on +2. This win propelled me towards the lead while Beliavsky, who had managed a fantastic start, subsequently collapsed and failed to qualify. The following game was played the round after the Beliavsky game. 
Game 27 J. Benjamin - V. Anand PCA Qualifier, Groningen 1993 Sicilian, Richter-Rauzer le4 c5 2 lbf3 d6 3 d4 cxd4 4 lbxd4 lL\f6 5 lL\c3 lbc6 6 .1g5 e6 7 'iWd2 .1e7 8 0-0-0 0-0 9 lbb3 'ifb6 I had prepared the Classical Sicil- ian to be my main defence for this tournament, and in view of Game 4 in this book, it was ironic that we en- tered the same variation, but with colours reversed. 10 f3 :d8 11 bl (D) .-...- -.. i .  i-i B.  _ . i-  . .--. -  . . . m . .. . _tiJ .. R  .   U _  u ..:..: 11 ... 'iWc7 An interesting move which forces White to decide how he is going to prevent ...d5. 12 .1xf6! When Benjamin took on f6 I just couldn't believe it - I hadn't consid- ered this move at all, believing that nobody would voluntarily give up the dark-squared bishop! However, it is a reasonable move; in return for surrendering an important minor piece, White gains time to push his kingside pawns. There are a number of alternatives, for example 12ltJb5, 12 .1f4 or 12 h4, but we will leave the relative merits of these to a book on opening theory. 12 ... .1xf6 13 g4 g6 . This is a critical moment for Black. There is an argument for 13...g5, for example 14 h4 h6 15 hxg5 hxg5 16 :h5 'iWe7, followed by ...lL\e5-g6, dominating the dark squares on the kingside. However, this involves a certain amount of risk, as if the posi- tion opens up, the exposed state of Black's king may be more important than control of a few dark squares. Black can continue with 13...a6!? 14 g5 .1e7, but after 15 f4 White 
BENJAMIN - ANAND, PCA QUAUFIER, GRONINGEN 1993 143 probably has a slight advantage. I preferred the text-move because it seemed to me that Black's queenside attack, supported by the bishop on the long diagonal, would be very dangerous. I still hadn't taken Ben- jamin's idea seriously, but the next few moves show that it is not so easy for Black. 14 h4 a6 15 g5! Now 15 h5?! would be bad, as 15...g5! 16 h6 'ike7 seizes the dark squares without opening the h-file. 15 ... .i.g7 16 h5 b5 17 hxg6 hxg6 (D) ..-*-_ ... .  .. w _  _ .._._.. B.B B n _ _ _ U . -. . .llJ .. nmfi . . U _   ..:..: 18 f4! After 18 .i.d3 lL\e5 19 f4 lL\xd3! (not 19...ltJc4 20 .i.xc4 'ikxc4 21 'ikh2 with a dangerous attack) 20 cxd3 (20 'ir'xd3 b4! 21 lL\e2 a5 is similar) 20...b4 21lL\e2 a5 the position is un- clear, with both sides having attack- ing chances. The text-move threatens 19 f5, which would win as Black cannot reply 19.. .exf5 because of 20 lL\d5. Thus Black is forced to drive the knight away from c3. 18 ... b4 19 lL\a4! (D) A very comfortable square for the knight, blocking Black's queenside attack. 19 lL\e2 a5! would be much weaker. ..-*-_ ... .  .. B. _ . _ .._.... B B .  _ _ _ u llJ .  .  _  U  .llJ_ . . p . . U _   ..:..: 19 ... :b8 After 19...e5 20 fxe5! (20 .i.c4 exf4 21 'ikxf4lL\e5 22 .i.d5 .i.b7 is unclear) 20...lL\xe5 (both 20...dxe5 21 .i.d3 and 20.. ..i.xe5 21 .i.c4! fa- vour White) 21 'ikxb4 Black has in- sufficient compensation, but at least he doesn't have to worry about his king! In fact this drastic remedy might be necessary; I just hadn't re- alized how critical my position was. 20 '5b2 This is one way to build up on the h-file, but it was also very dangerous 
144 VISHY ANAND: My BEST GAMES OF CHESS to play 20 .1d3!?, followed by :h4, 'ifh2 and :h 1. The queen on h2 is not only effec- tive down the h-file; in some lines it can act along the h2-b8 diagonal, for example by f5 followed by ltJac5. 20 ... f8 (D) Time to get the king out! The al- ternatives are: 1) 20...e5? 21 f5 gxf5 22 'ifh7+ f8 23 exf5 and White wins. 2) 20....1d7 21 :d3 ltJe7? (the best move is 21...f8, transposing to the following note) 22 :h3 f8 23 :h 7! ltJg8 24 f5!? (24 :xg7 xg7 25 'ir'h8+ f8 26 :h7 is less clear after 26...e7) and now: 2a) 24....1e5 25 :xf7+ xn 26 'ir'h7+ .1g7 27 fxg6+ f8 28 .1d3 with a decisive attack. 2b) 24....1xct4 25 :xg7 xg7 26 'ifh8+ f8 27 :h7 .1e8 28 f6 forc- ing mate. 2c) 24.. .exf5 25 ltJac5 .1c8 26 :xg7 xg7 27 'ir'h8+ f8 28 :h7 winning for White. %'i /u _ ' ///- 11 /.1.:   . /.            % /uu/.t. U ,/ /,   .  .. w. _ . _ '.4\.'.'. . . . 0 lb/ . 8 ' . / /.   .lb. . . 8"8. .  U . . _ .<&t>.:..: 21 :d3 e5 21....1d7 was also possible, for ex- ample 22 :h3 (22 f5 exf5 23 ltJac5 is unclear) 22... e 7! 23 'ife2 (23 :h7 :h8 24 'iff2 1/ 2 _1/ 2 Landen- bergue-Georges, Swiss Team Cham- pionship 1994) 23...e5! (23...f8 24 :h7 ltJe7 25 'ifh2 and 23...ltJa5 24 f5 .1xa4 25 f6+ .1xf6 26 gxf6+ xf6 27 'ife3 .1xb3 28 :f3+! ate good for White) 24 :h7 :h8 with an unclear position. 22 f5 22 .1h3 exf4 23 'ir'xf4lbe5 is fine for Black. 22 ... gxf5 23 :h3! ltJe7?! (D) 23...f4? 24 :h7! and 23...fxe4? 24 :h7 .1e6 25 :xg7 win for White, but 23...e7! 24 :h7 :g8 was a bet- ter defence. - - '/ - ..t.. %  . .  . "/. w. _ , _ . . ,' . . . . . d_u lb' .8. . / / - - - -%'(( - - Rt.z.JB . .: A ' A  . // '/}     Ii ' ' "//. O o ..  % //  '0f%;  "  B<t>. _B: 24 :h8+ A critical moment. At first sight 24 :h7 appears very dangerous, but Black can defend: 
BENJAMIN - ANAND, PCA QUAUFIER, GRONINGEN 1993 145 1) 24...ltJg8? 25 J:txg7 q;xg7 26 'fih8+ q;f8 27 J:th7 .1e6 28 exf5 :dc8 29 ltJc3 bxc3 30 fxe6! and WIns. 2) 24...ltJg6! 25 J:txg7 q;xg7 26 'fih6+ q;g8 and now Black is better. The continuation might be 27 .1e2 'fie7 28 .1h5 'iff8 29 .1xg6 'ifxh6 30 :xh6 fxg6 31 J:txg6+ q;h7 with a winning ending. 24 ... ltJg8 Not 24....1xh8? 25 'ifxh8+ lL\g8 26 J:th7 transposing to line 1 of the previous note. 25 .:xg8+! After 25 'ifh7 .1xh8 26 1fxh8 'fie7! the attack peters out. 25 ... Q;xg8 26 7+ Q;f8 27 exf5 (D) -   / -  _... / - . B _ _ _.__ ..  . . -  - -  . 8" . . _ u lD. _ . . .lD. . . 808_ . _ _. ..i..: 27 ... .1xf5! 27...'ife7!? is possible, but White is slightly better after 28 f6 .1xf6 29 gxf6 'ifxf6 30 .1h3. 28 'ifxf5 'ifc6 29 g6! J:tb7 30 J:th7?! Later on the computer Deep Blue suggested 30 'ifg5!. Black's best ap- pears to be 30...J:tc8 31 'ifg2 J:tbc7 (31...e4? 32ltJa5 really does win) 32 'ifxc6 J:txc6 33 .1d3 d5!. White is better, but his minor pieces are not well placed and so Black might be able to generate a lot of counterplay. Coincidentally, Benjamin later ended up working for the Deep Blue team in their matches against Kaspa- rov! 30 ... 'ifxa4 (D)  - %/ . .   . w... ...: .. . .8. . . /iV. _ _ Ym _ .. . - - alD. . . 808. . . .. ..i.. 31 'ifg5?! Here is White's last chance to force a clear-cut draw, for example: 1) 31 .i.c4 d5 (31... 'ifd7 32 'ifh5 is unclear, but 31...J:tdd7 32 'iff3 is probably good for White) 32 ltJc5 'ifc6 33 ltJxb7 'ifxb7 34 .1xa6 'ifd7 leads to a drawn ending. 2) 31 J:txg7 Q;xg7 32 gxf7 and now: 
146 VISHY ANAND: My BEST GAMES OF CHESS 2a) 32...:xf7? 33 'ii'g5+ h7 34 i.d3+ wins for White. 2b) 32...:f8? 33 'ii'g4+ xf7 34 .1c4+ f6 35 'ii'f3+ is also winning. 2c) 32...f8 33 'ii'g5 'ii'd7 (not 33...:bb8 34 .1c4 d5 35 'ii'g8+ e7 36 'ii'g7 :i8 37lbc5 and Black loses) 34 .1c4 d5 35 'ii'g8+ e7 36 'ii'g5+ xf7 371Wh5+ g8 38 'ii'g5+ draws. 2d) 32...'ii'd7 33 'ii'g5+ xf7 34 lL\c5 dxc5 (34...'ii'e7? 35 'ii'h5+) 35 .1c4+ with perpetual check. 31 ... 'ii'e8 White still has an amazing number of attacking possibilities, but Black seems to be able to hold out with ac- curate defence. Here he must avoid 31...:c8 32 :xg7 xg7 33 gxf7+ xf7 34 'ii'f5+ and 31.. .:e8 32 Lg7 xg7 33 gxf7+ xf7 34 .1c4+, with a win for White in either case. 32 .1xa6 (D)  .  \\Ur_  .   .-.   . I. .   . .-: B_ .t. _ ..  .  .    - - . . . - .  . .tlJ. . . D. . . .<iii>. . . After 32 .1g2!? (32ltJa5 :c7 33 ltJc6? f6 wins for Black) Black again must be very careful: 1) 32...'ii'e7 33 'ii'h5 wins. 2) 32...:c7? 33 :xg7 xg7 34 gxf7+ xf7 35 .1d5+ also wins. 3) 32...e4? 33 :xg7 xg7 34 gxf7+ xf7 35 'ii'd5+ gaining mate- rial. 4) 32...d5! 33ltJc5 'ii'e7! defends. 32 ... :e7! (D) . --. . w. . _.E:  - / - - ....   .. . . /%a m     - " . - . .tlJ. . . D. . . .<iii>. . . 33 .1d3 Threatening to win by 34 :xg7 xg7 35 gxf7+ xf7 36 .1g6+. 33 ... e4! 34 i.b5 White has nothing better, for ex- ample: 1) 34 .1xe4 :xe4 35 gxf7 :el+ 36 ltJc 1 :xc 1 +! 37 'ii'xc 1 'ii'e5 de- fends. 2) 34 .1c4 d5 35 .1xd5 :e5 36 gxf7 :xg5 37 fxe8'ii'+ :xe8 wins for Black. 3) 34 :xg7 xg7 35 gxf7+ rl;xf7 36 .1c4+ d5 37 .i.xd5+ :xd5 38 'ii'xd5+ f8 with good winning chances for Black. 
BENJAMIN - ANAND, PCA QUAUFIER, GRONINGEN 1993 147 34 ... ':e5! Forcing the following liquidation. 35 gxf7 .:xg5 36 fxe81f+ .:xe8 37 .1xe8 ':gl+! 38 lL\cl xe8 (D) . -.. wB . _ .  .  // .    - - . .  BiB B - - - - B B _ . 88B . . U _  _ .<it> .  .  . - . .: . 39 84?? A blunder caused by time-trouble. White had to try 39 c3 (getting rid of the c-pawn to release the king) 39...bxc3 (39...b3? 40 axb3 i.f6 41 c2 i.g5 42lL\e2 :el is ingenious, but leads to less than nothing after 43 ':h2!) 40 bxc3 .1xc3 41 c2 i.f6, when Black is clearly better, but the reduced material gives White some drawing chances. 39 ... bxa3 40 bxa3 i.c3! With total paralysis. 41 ':h4 d5 0-1 This game gave me the clear lead and, unlike at Biel, I reached the neces- sary score very comfortably. At the beginning of 1994 I played my first-round match in the FIDE Can- didates cycle against Artur Yusupov, which I won 4 1 /2-2 1 /2. In the draw for the second-round matches I was paired against Gata Kamsky. Before the match took place, Gata and I participated in the Linares tournament and we met in the very first round. In view of the forthcoming match, the game had more importance than a normal tournament game. In fact we were to spend a lot of time with each other in the following years... 
Game 28 V. Anand - G. Kamsky Linares 1994 Sicilian, Najdorf 1 e4 c5 2 lbf3 d6 3 d4 cxd4 4 ltJxd4 ltJf6 5 ltJc3 a6 At the time Kamsky played many openings, but I wasn't sure whether he had really studied them or whether he gave priority to being difficult to prepare for. Later it became clear that he does study a lot and under- stands a lot of different openings, but not too well! I often encountered holes in his repertoire. 6 .1e3 e5 7 ltJb3 .1e6 8 f3 .1e7 The thematic Sicilian thrust 8...d5 does not equalize here: 9 exd5 ltJxd5 10 ltJxd5 .1xd5 11 c4 .1b4+ 12 f2 .1e6 13 'iixd8+ xd8 14 J:tdl+ gives White an advantageous ending. 9 'iid2 ltJbd7 10 g4 h6 10...b5 is less accurate as White may continue 11 a4! b4 12 ltJd5 .1xd5 13 exd5, when 13...ltJb6 fails to 14 a5! ltJbxd5 15 g5 ltJxe3 16 gxf6 ltJxfl 17 fxe7 'iixe7 18 J:txfl and White wins a piece. 11 h4 b5 (D) .. ..   . . / . ... wd . _ j,; .. .J. .  / uu _ "U/// B.. . . /%% $ /'/uu @J)} . .888" / :/;;// _ U . l\. dVLJ .0. 8"8 . . U  . .  . i..: - -  - 12 J:tgl This idea, which came to me over the board, is borrowed from the game against Ftacnik (Game 23). Although the position is completely different, the idea is the same - White saves a tempo by missing out 0-0-0 and uses it to push through g5 as quickly as possible. 12 ... b4 12...ltJb6!? 13 g5 hxg5 14 hxg5 ltJfd7 is another possible continua- tion. 13 ltJa4 Later it turned out that 12 J:tgl had been played before, only to be fol- lowed up by the weak 13 ltJe2?, when 13...a5 14 g5ltJh5 gave Black 
ANAND - KAMSKY, LINARES 1994 a fine position in Los-De Boer, Gron- ingen Open 1990. 13 ... d5 (D) 13. ..a5 14 g5 hxg5 15 hxg5 ltJh5 is unclear. _ruu__ .. .*. - . ... W. . _  .. ..t  .d_. . B'. . / .888" vz.J/  _ _  .l2J. ;5 8 . 88. . . a '  . ' /0  D    . / %"'/1  /J ,   , / / 14 g5 d4 It is hard to judge how the disap- pearance of the h-pawns affects the position. After 14.. .hxg5 15 hxg5 d4 16 .1xd4 (16 gxf6 dxe3 17 'ifxe3 ltJxf6! and now White has to play 18 i.d3 in order to castle; the resulting position is unclear) Black can try: 1) 16...i.xb3 17 gxf6 i.xf6 18 axb3 exd4 19 0-0-0 lbe5 with the same position as in the game, minus the h-pawns. Certainly the h- file is bad for Black if he castles, but in some other lines it is useful. Still, 20 f4 looks good for White anyway af- ter 20...ltJf3 21 'iig2ltJxgl 22 'iixgl! (here 22 e5 allows 22....1xe5! 23 fxe5 ltJh3 and the knight escapes) and I think White has great long- term compensation. 149 2) 16...ltJxe4 17 fxe4 .1xb3 18 axb3 exd4 (here the interpolation of ...hxg5 favours Black - compare the note to Black's 15th move in the game) 190-0-0 ltJe5 is unclear. 15 i.xd4! Not 15 i.f2? ltJh5 and Black is better. 15 ... .1xb3 (D) After 15...ltJxe4 16 fxe4 .1xb3 17 axb3 (not 17 .1b6 ltJxb6 and Black has no problems) 17...exd4 18 'ifxd4 hxg5 19 O-O-O! (19 hxg5 i.xg5 20 'ifxg7 .1h4+ 21 cjj>e2 .1f6 22 '*' g3 is unclear) White seems to have a pleasant edge, for example after 19...:xh4 both 20 'ifxg7 .1f6 21 'iig8+ cjj>e7 22 :xd7+ cjj>xd7 23 'iixf7+ and the simple 20 .1c4 are very good for White. .I. .. _  .  ... w . _  .. " -  . d _ _ " "   d . ,, u l2J _ 8.  _ ,,, / d u .J.. .8. 8"8 " " u _ . . D ' %%   ; =  '0  a / '  /  :- %...  / , / W;ffi ,, / 16 gxf6 16 axb3 ltJxe4 17 fxe4 is also promising, transposing into the note to Black's 15th move. 16 ... .1xf6 
150 VISHY ANAND: My BEST GAMES OF CHESS 16...i.xa4? 17 fxe7 'ikxe7 18 i.f2 is good for White. 17 axb3 exd4 18 0-0-0 Now White threatens simply f4 and e5, with an overwhelming posi- tion, so Black has to take counter- measures. 18 ... ltJe5 (D) The critical1ine as 18...i.xh4? 19 'ikxd4, 18...0-0 19 'ikxh6 and 18...'ikaS 19 f4 (followed by e5) are all very good for White. . .  .  .   -  - w. . ... . .  .   . d _ _ . . II .   . . n t.tJ_ _  U .. .. n  illI1  .   .  U _   . =:.ii.a 19 f4!! The most forcing continuation. The alternatives are: 1) 19 i.g2 i.xh4 20 f4 O-O! is not very dangerous for Black. 2) 19 'ikg2 ltJg6 20 h5 ltJf4 21 'ik g4 i.e5 22 ltJc5 0-0 23 'ikf5 with an edge for White. 3) 19 i.e2, intending 20 f4, is also promising as 19...d3 may be met by 20 'ike3!. 19 ... ltJf3 20 'ikg2 ltJxgl 21 e5 0-0 After 21...i.xh4 22 'ikxg7 ':f8 23 'ikxgl, followed by i.g2, ':xd4, etc., White has an enormous attack. 22 i.d3! White wants to take on gl with his rook, so as to tie Black down to the defence of g7. After 22 exf6 'ikxf6 23 'ikxgl 'ikxf4+ 24 cjj>bl 'ikxh4 White is not as well placed to attack Black's king as after 22 i.d3. 22 ... i.xe5! White wins after 22...i.xh4 23 ':xgl g6 (or 23...g5 24 'ikh3!) 24 .1xg6 cjj>h8 25 i.h7! i.g5 26 fxg5 cjj>xh7 27 'ike4+ cjj>h8 28 'ikh4 cjj>g8 29 'ikxh6, followed by ':h 1 or g6. 23 fxe5 'ikxb4 24 Jbgl (D) .. . -.. B. . .._ & . '  . ' .  .. .  - . .0. /   .   t.tJ  _  _ ..ii.. .    .  .. .  U,  .  M  .      - 24 ... 'ikf4+ The ending after 24...'ikg5+ 25 'ikxg5 hxg5 26 ':xg5 is favourable for White. 25 cjj>bl! 
ANAND - KAMSKY, LINARES 1994 25 'ir'd2 was also possible, but the text is more incisive. 25 ... 'ir'xe5 (D) .. . ... w. . ... .. .  .   _ _ ,#A _ . . . .  I / . . '-'lJ_ %  _  ..i.. . B. ... .<iti>. . a In this position Black has a rook and three pawns for a bishop and a knight, and there is no obvious way for White's attack to break through, so it might appear favourable for Black. However, it turns out that the most important factor is the initia- tive. White can keep harassing Black before he can coordinate his rooks. 26 lbc5! :a7 Black could have removed the danger to his king by jettisoning a couple of pawns: 26....:ad8 (26...11fd8 27lbd7 'ir'g5 28 'ir'hl! is also very good for White) 27ltJxa6 and now: 1) 27...:fe8 28lbxb4 g5 29 ':fl (not 29 lbc6?? _el+ mating), fol- lowed by lbc6. 2) 27....:a8 28lbxb4, followed by ltJc6 and b4-b5, when White should . WIn. 27 'ir'c6 151 Threatening 28 'ir'xh6. 27 ... 'ir'e3 Or 27...h8 (27...:e8 28 lbd7 'ir'e3 29lbf6+ wins) 28lbd7 'ir'e3 29 lbxf8! 'ir'xgl+ 30 a2 (the threat is 31 'ir'c8) 30...g6 31 'ir'c5 :a8 32 'ir'e5+ g8 33 lbxg6 fxg6 34 'ir'd5+ and wins. 28 ':g2 29 ':e2 30 a2 (D) Now the threat is 31 :e8 g8 32 :Xf8+ xf8 33 _d6+ mating. Black has still not been able to coordinate his pieces and, indeed, in order to meet White's threat he is obliged to retract his 26th move. h8 'ir'gl+ . .    . - - B. . ... .... .  - -  - .  . .   . . _ _ d  ..i.. . <&t>8.:. . . . . . 30 ... :aa8 Now White aims to transfer his knight to e5, when the weakness of f7 will tell. 31 llJd7 :ac8 Black loses after 31...:fc8 32 'ir'f3 or 31...:fd8 32lbe5. 32 'ir'f3 f5 
152 VISHY ANAND: My BEST GAMES OF CHESS After 32....:fe8 33 ltJe5 the f7- square is fatally weak. 33 ltJxf8 ':c5 (D) Or 33....:xf8 34 'ifb7 and White's initiative is too strong. The text- move is a dangerous try, but White has a good reply ready. "/  /. - ::fJ1:i  / -  /  /uu   . >fif//   fll .  .i' _ %:m %    _ _ :0 . /u/u /u/u m/ y/,,/,:/ __i..'iV. D.:_ .  _ _ ill11 ..- . .  - f/ / W. . all//, i @ffi@ '// .  34 ltJg6+! The simplest method, although 34 'ir'xf5 ':xf5 35 .i.xf5 'ir'fl (35...g5 36 ':e7 and 35...g8 36 .i.g6! also win for White) 36ltJg6+ h7 37llJh4+! would also have been effective. 34 ... 'ir'xg6 After 34...h7 White prevents the mate by 35 ':e5. 35 ':el 36 'ir'a8+ 37 .i.c4 White's attack is too strong. 37 ... ':c6 38 'ir'g8+ g6 39 ':gl+ 1-0 As 39...h5 40 .i.e2+ h4 41 'iWd5 leads to mate. 'ir'f6 h7 Although the above game was a good start, the rest of the tournament didn't go so well- I barely made 50%, which amounted to a fairly lousy re- sult. The following game was the only other high spot - but it was quite pleasant! 
Game 29 V. Anand - J. Polgar Linares 1994 Sicilian, Najdorf le4 c5 2 ltJf3 e6 3 d4 cxd4 4 ltJxd4 ltJf6 5 ltJc3 d6 Giving White the chance to playa Scheveningen, but I decided to trans- pose into the English Attack. 6 i.e3 a6 7 f3 ltJbd7 I was surprised that she chose the ...ltJbd7 system, because an earlier game of Kasparov had seemed to show that White could gain a strong initiative. However, Judit comes up with an improvement. 8 g4 h6 9 1:tgl b5! Better than 9...'ifb6?!, when the continuation 10 a3ltJe5 11 i.f2 'ikc7 12 f4ltJc4 13 i.xc4 'ikxc4 14 'ii'f3 e5 15 ltJf5 .i.xf5 16 gxf5 d5 17 fxe5 ltJxe4 18 :g4 was very good for White in Kasparov- Kamsky, Linares 1993. 10 h4 ltJb6 10...g6 has been played before, so this is the innovation. 11 g5 ltJfd7! After 11...hxg512hxg5ltJfd713 g6, with the idea 13.. .1:th2 14 gxf7 + xf7 15 i.f4!, White has some ad- vantage. After the text, however, 12 g6 is impossible because the h4- pawn is hanging with check. 12 'ii'e2!? 12 'ifd2 may be met by 12...b4 13 ltJd 1 d5 or 12.. .ltJe5 . 12 ... hxg5 After 12.. .ltJc4 13 0-0-0 ltJxe3 14 'ii'xe3 hxg5 15 hxg5 g6 16 f4 White's lead in development compensates for the two bishops. 13 hxg5 g6 (D) Not 13...b4? 14ltJc6 'ii'c7 15ltJxb4 d5 16 ltJd3 and White wins a pawn, but 13...ltJc4 140-0-0 ltJxe3 15 'ifxe3 g6 16 f4 is still playable, transposing to the preceding note. .I   . / .t. //./"///. .  /% _ /  W  f / 0    w %: %' // /;    /.  /, ,/ / / wB .... . .... _/,i__ ... . "  .  U . A. . . ;o . .  %8. /'  ";. 808... .  / /  : = 0 /0 i. a  /  ;%i  % /+:/-   .;y;4//    / /  / '///m/ /,"// / / 
154 VISHY ANAND: My BEST GAMES OF CHESS 14 0-0-0 ltJc4 15 i.f2 i.b7 16 f4 (D) ._ mii . _ _ .. -  . B........ . .. -.... ... . 0 .JI1\"8 . d.. u d .  . . 8  8 .\Wr  . - u d-.  . :.i. _  d _ 16 ... 'ifa5 After 16....:c8 White also plays 17 'ifg4, with the plan of eliminating the dangerous knight by i.xc4. 17 'ifg4! Once again White's priority is to eliminate the knight. Black's queen- side play appears dangerous, but she cannot land a ...llJxb2 blow before the knight is swapped off. Note that 17llJb3 'ifc7 is inferior, since next move 18...i.g7 really will threaten to take on b2. 17 ... O-O-O? This allows an unusual combina- tion, so 17.. .':c8 would have been better: 1) 18 llJxe6? ltJxb2 is good for Black. 2) 18 i.xc4 ':xc4 19 llJxe6! fxe6 ( 19. ..':xc3 20 bxc3 fxe6 21 'ifxe6+ d8 22 i.d4 ':h2 23 ':h I! with advantage to White) 20 'ifxe6+ d8 21 i.d4 is unclear. 3) 18llJb3!? 'ifc7 19 i.xc4 'ifxc4 (19...bxc4 20 llJd4 and Black's cen- tralized king is a ready target) 20 llJa5 'ifc7 21 llJxb7 'ifxb7 22 i.d4 leading to a double-edged position. 18 i.xc4 bxc4 (D) - . - - - . -  - w......... .. .... - - - - mii . . R _ . . u ..80.. .  . . 808. . . . :.  -  - - 19 llJxe6! While this sacrifice is perfectly normal when Black's king is still on e8, it is unusual when Black has al- ready castled queenside. However, here White is not aiming for a direct attack but for positional compensa- tion. 19 ... fxe6 20 'ir'xe6 b8 After 20...i.g7 (the continuation 20...d5 21llJxd5 is also favourable for White) 21 i.d4! i.xd4 22 ]hd4 6 23 ':gdl! Black will shed some more pawns (note that 23....:hl?? fails to 24 ':xc4+). 21 'ir'xg6! 
ANAND - J P . OWAR, LINARES 1994 Gaining a third awn Unless Blac k d p for the piece . evelo p · qUIckl y th e s counte rp l ay . passed g cIde the game. -pawn will de- 21 ... " 21 lb 5? .h3 (D) 23 ';£6!. c . loses to 22 i.xc5 .xc5 22 bl ' The most str. . nullify Black' t a h lghtforward wa y to t s reat of " s ead, for exam I ....xc3.In- 23 i.xc3 _ 2 p e, 22 i.d4? :xc 3 ' xa would I . essary com I .. a low unne c 2 p Icatlons 2 2 \WI - 3 exf5 :f3 24: f1 : _f5 .xf5 more complicate IS playable, but 25 lbe2 (25 f 6 J':-. after 24...i. g 7' . th . "lJxf6 26 ' WI Just an ed gxf6 i.xf6 26 i. d4 " ge for White ) 25  .e8 Wh" ...,*,c8' if Black retms eIS till better, bu ending would b ,pIece, then the ferred to keep the dIfficult, so I pre- 22 e queens on. . ... 1lf3 WhIte wins aft hI. er 22...:xc3 23 23 i.d4 155 23 :gfl! would h precise _ ther e ave been more was no up one of the need to give 23 pawns. ... ]bf: 24 'iVh7! (D) 'I 24 Black's -;; dS? 25 estchanc g6lbxe4 26 e was 24...lbc5 lbxc3+ 28 i.xc3 f;8 i.xg7 27 :xg7 to an unclear positi bxc3 i.e4leads c1, although I thin) 28....d5 29 a clear advantage. Hit W!rlte has safe, while Black' . king IS ultra- O pp . s IS ve ry , oSIte-coloured bis weak. gIn to exert th . hops only be- elr draw. h once the ma'or' IS tendenc changed bt h PIeces have been e: . ' ere all h pIeces are S ti 0 ll t e ma 1 0 r ,. on the b  IS g010g to come d oard so Black attack. un er a very strong After the text-m comfortably. ove White wins 25 6 N g i.cS e o 25...dxe4 26 7,t, , ffective and Wh' g, .JLxg7 IS in- Ite wIns easily by 
156 VISHY ANAND: My BEST GAMES OF CHESS continuing 27 :xg7 .i.c6 28 _h6 _c7 29 .i.b6. 26 .i.xc5 ltJxc5 Or 26..._xc5 27 g7ltJf6 28 g8_ :xg8 (28...ltJxg8 29 :Xg8) 29 :Xg8+ ltJxg8 30 _xg8+ a7 31 exd5 with a simple win.. 27 g7 a7 28 g8_ 28 :xd5 was also very good. 28 ... Jbg8 29 _xg8 ltJxe4 30 ltJxe4 :xe4 31 'fig7 'fic5 32 :gel :f4 33 :e7 _b6 34 _g5 c3 35 b3 _b4 36 :c7 :e4 36...b6 37 :xc3. 37 _xd5 b8 (D)  . _#&: w.J.  / ///// "/ " & "x;:/ .. x >:!;f:(: /'/; // " :f{ '>; . .  . 0 ... > . '/ . ' . :::/ t% 'A% ?'//% j/t;,; ;>/;;; :;;;;c,;; /  >;;3J; . ¥WJ::  _ rf&Afd  .8. . . A ; A ;;   oao. I::i  . .. ;  .'itiI.a  f} . /;%;;;; . :: f:::}/ :(@ ::1, . /.  / ;1Ji >""" 38 _d8+ a7 39 a3 _xa3 1-0 White mates in five: 40 :xb7+ xb7 41 :d7+ c6 42 _c7+ b5 43 :d5+ b4 44 _b6#. 
Game 30 V. Anand - G. Kamsky PCA Candidates (3), Las Palmas 1995 Ruy Lopez, Arkhangelsk le4 e5 2 ltJf3 ltJc6 3 i.bS a6 4 i.a4 ltJf6 5 0-0 bS 6 i.b3 i.b7 Kamsky plays a lot of different openings, and here he decides to play the Arkhangelsk. However, this was no surprise; he played it in the 1994 FIDE Candidates match in Sanghi Nagar, which I lost, and sub- sequently played it against Short in their PCA Candidates match. 7 1:tel i.c5 8 c3 d6 9 d4 i.b6 10 i.e3 0-0 Not 10...ltJxe4? 11 d5 and White . . wIns a pIece. 11 ltJbd2 h6 12 h3 I was following the game Short- Kamsky, PCA Candidates (6), Lin- ares 1994, which continued 12...ltJd7 13 a3 ltJe7 14 i.a2 h8 15 b4 and was eventually won by Black. I had prepared some new ideas in this game, but in fact Kamsky was the first to vary. 12 ... 'ii'b8!? (D) Z . .. - . - . ..t .i w.  _ ;; i . 0. ' .. /  . %j"// /  /   /  : " " // /  BiB _ . B DB B  . i." t2J .   u _  "   .  U  U    . \Wr  _ /R_  This looks an odd move, but Black is trying to batter down d4 by . ..'ii'a7. The problem is that it allows White to play d5. 13 d5! ltJe7 The tactical variation 13.. .i.xe3 14 dxc6 i.xd2 15 cxb7 i.xel 16 bxa8'ii' i.xf2+ 17 xf2 'ii'xa8 18 ltJd2ltJxe4+ 19ltJxe4'ii'xe4 20 'ii'd5! leads to an ending in which White has a slight advantage. Gata played 13...ltJe7 very quickly so he obvi- ously didn't believe this line. 14 i.xb6 cxb6 15 i.c2! After 15 a4 bxa4 16 i.xa4 'ii'd8 Black has a satisfactory position. 15 ... ltJd7 
158 VISHY ANAND: My BEST GAMES OF CHESS A critical moment. One of the points of 12...'ii'b8 (as opposed to the similar line 12....:e8 13 d5l[je7 14 -'.xb6 cxb6 15 -'.c2 l[jd7) is that by leaving the rook on f8 Black sets up the immediate threat of ...f5. It is also helpful for Black that his queen defends d6, since after ...f5 a white knight might easily come to e4. I decided that it is imperative to stop.. .f5, and to this end I was able to use the negative side of ... 'ii'b8, namely that on b8 the queen does not exert a latent influence on the d8-h4 diagonal. 16 l[jh4! 'ir'd8 (D) 16.. .f5 is bad in view of 17 l[jxf5 ltJxf5 18 exf5 -'.xd5 19 l[je4, so Gata decides simply to return the queen to d8. However, this in itself shows that White's strategy was cor- rect.  . mu_ . .. - .*. w_..._._._ .   .       _'_8_ . . .8.  . D . .8 8ni.. n8 .  u  u    . \WJ  - .-  17 l[jn White can hardly playa weaken- ing move such as g3, so the knight has to remain undefended on h4. Fortunately, after the text-move Black is not able to exploit this. 17 ... g5? After 17.. .l[jxd5 18 l[jf5, followed by 19l[jxd6, White gains the advan- tage. However, this variation helps us to find the best line for Black, namely 17...b4! 18 cxb4l[jxd5. Even in this case the continuation 19l[jg6! (19 l[jf5 l[jxb4 20 l[jxd6l[jxc2 21 'ii'xc2 is also slightly better for White) 19...fxg6 20.1b3 gives White some advantage. It is odd that in this variation the doubled b-pawns are liquidated, but Black gets doubled g-pawns instead! In this match Kamsky had a ten- dency to take really drastic measures when faced by minor positional problems, and the text-move is a good example. Pushing the g-pawn clarifies the position, but at the cost of seriously weakening Black's king- side. 18 l[jf3! (D) A difficult choice as 18l[jf5l[jxf5 19 exf5 l[jf6 was also tempting. Af- ter, for example, 20 .i.e4 ':c8 21 f3 ':c5 22 l[je3 g7, White has ce- mented everything but the position becomes rather closed and I wasn't sure that I would be able to break: through on the kingside later. The text-move is more compli- cated, but promises a larger advan- tage if everything works out. One factor in the decision was that at this 
ANAND - KAMSKY, PCA CAND. (3), LAs PAIMAS 1995 159 point of the match I was a point down and badly needed a win. .. II ... 8......111.. 1.-   .   - -  -  . 1. . 8' -   - . .8. . .   .  . 8 _ u t.tJ 8 R i.B R8 .  u  u    . %"CiJ -  -  18 ... f5 After 18...g7 19 llJe3 White again prevents ...f5. 19 exf5 llJxf5 20 l2J3h2! A key move, with the idea of re- grouping the knight via g4. White intends using all the light squares and the weakness of the bl-h7 di- agonal. Black is unable to mount a counter-attack against d5 quickly enough to deflect White from his plan. 20 ... 'ir'f6 After 20.. .llJe 7? 21 llJe3 White consolidates his grip. 21 llJg4 'ir'g7 The queen comes across to sup- port the weakened kingside. 22 llJge3! After 22llJfe3 llJe7 the knight on g4 has nowhere to go. 22 ... llJxe3 Now, however, 22...llJe7 23 llJg3 leaves both knights ideally posted. 23 llJxe3 Black is in a very bad way be- cause it is almost impossible to de- fend f5. 23 ... :14 (D) After 23....:f6 24 'ir'h5 'ir'f7 25 'ir'e2, threatening 26 llJg4, White also has a clear advantage. .. . ... w....... _ I.'/   .    -  -  . 1. . 8-    - - . . . .  .     . 8  u   8Ri. .  P8 .  u  u   .\WJ    III   24 a4! ! White would like to occupy f5, bu t neither 24 -'.f5 ':f8 25 -'.e6+ h8 nor 24 g3 ':f6 25 -'.f5 :af8 is really good for White. However, 24 llJf5 'ir'f6 25 llJg3! was a good alter- native to the text-move, with ideas of llJh5/llJe4 and 'ir'h5. 24 ... :ar8 The point of White's play is that if one pair of rooks can be exchanged, then the occupation of f5 will be per- manent, e.g. 24...bxa4 (24...b4 25 g3 ':ff8 26 cxb4 is just a clear extra pawn) 25 ':xa4 and now: 
160 VISHY. ANAND: My BEST GAMES OF CHESS 1) 25...:af8 26 :xf4 :xf4 (both 26...gxf4 27 llJf5 'iff6 28 'ifh5 and 26.. .exf4 27 ltJf5 'iff6 28 :e6 are also excellent for White) 27 .i.f5! and Black has no defence against g3. 2) 25...:xa4 26 i.xa4 b5 27 i.c2 ltJf6 28 b3 when White has control of f5 and can continue with c4. Gata decides simply to abandon a queenside pawn. 25 axb5 as Forced, because Black cannot al- low White to create a passed pawn on a6, for example 25...:xf2 26 bxa6 i.a8 27 1:.fl 'iff7 28 'ifel forcing ex- changes. 26 :n i.c8 27 g3 (D) R.i.B '''.B B. _._ Hi % . / / 1 % /   /   -  /  /"/   /", , /  8 . 8  /  / - . /'//  //  - - . - . n  '8 . /  /",  "i.. " . U . U .  ...: - . .  27 ... :4f7 After 27...ltJf6 White can win with the complicated 28 gxf4 gxf4+ 29 h2 fxe3 30 fxe3! 'ifg5 31 :a4! or the simple 28 h2!, which just leaves the rook trapped. 28 b4 I decided that it was time to stop ...llJc5, but it would have been safer to play 28 i.e4! ltJc5 29 .i.g2, de- fending the slightly weak kingside. After 29...e4 30 ltJc4! White starts to exploit Black's weak pawns. 28 ... e4! A good try, activating the knight and creating some kingside counter- play. One of Kamsky's strengths is that once he realizes his position is critical, he doesn't hesitate to make the necessary sacrifices to stir up com- plications. Here one wasted tempo would be too late: if White could play i.e4, Black's position would be hopeless. 29 i.xe4 ltJe5 (D) After 29...'ifxc3 30 bxa5 bxa5 31 ltJf5! White wins, e.g. 31.. .:xf5 32 i.xf5 :xf5 33 :cl or 31..:i6f6 32 'ifh5 :h7 33 f4. _.i._  ',._ . . . illi1 w. . .z.   .  -_._ 88"//  . . /'l  D _i.B B . "  "8 .  // U . . 8 . a ...:= 30 i.g2 axb4 After 30...ltJf3+ 31 hl g4 32 bxa5 gxh3 33 i.xf3 :xf3 34 axb6 
ANAND - KAMSKY, PCA CAND. (3), LAs PALM AS 1995 161 :xf2 35 ':xf2 ':xf2 36 ':a8! Black's counterplay comes to an end. 31 cxb4 lbf3+ 32 i.xf3 Necessary; White can't continue with this knight stuck in the middle of his position. 32 ... Jbf3 33 ':a8 .1xh3 33...Wb7 34 'ii'al !, threatening 35 ':a7, is hopeless for Black. 34 lWxf3 Jba8 Despite the inaccuracy commit- ted at move 28, White retains a large advantage. 35 ':c1 .:f8 (D) 35...Wb2 is met by 36 ':c2. . . w. . . /  .  - . .8.8. . o . . . . . \WIp d d -U.L . .  . _ _ U _ .  .  .  .  ... -  - 36 'ii'e2 37 ':c7 38 ':b7 39 n After 39 h2 'ii'f6 Black will de- velop more counterplay than in the game. 39 ... i.d7 .:f7 lWa1+ g7 The ending after 39...'ii'a8 4O:Xb6 'ii'xd5 41 'ii'd2 'ii'xd2 42 lbxd2 is winning for White. 40 Jbb6 'ii'd4 41 ':b8 'ii'xb4 Or 41...'ii'xd5 42 Wb2+ 'ii'e5 (the line 42.. .':f6 43 ':f8! 'ii'e5 44 'ii'xe5 dxe5 45 ':xf6 xf6 46 b6 i.c6 47 e3 e7 48 c4 also wins) 43 'ii'xe5+ dxe5 44 e3 and the b5- pawn advances. 42 e3 h5 43 b6 Not 43 'ii'xh5? 'ii'el+ 44 h2 ':xf2+ 45 g2 ':xg2+! 46 xg2 'ii'd2+ 47 fllWcl+ 48 e2 'ii'c4+ with perpetual check. 43 ... h4 (D) 43...i.b5 loses instantly after 44 lbf5+ !. a . . .  .  ..t . .- w  .   "   .   .  U _   . .8 .   . .  -  . .  - . . -  . .."  .  U  .  .\WIn .  .-U . . . .  .;. 44 g4 This pawn is very important since it shuts Black's bishop out of the game. It may appear weakening, but Black cannot exploit this because of 
162 VISHY ANAND: My BEST GAMES OF CHESS White's dangerous passed pawn and the fact that Black's king is also ex- posed. 44 ... i.b5 45 1Wd1 From the practical point of view, 45 f5+ h7 461Wdl would be bet- ter as it cuts out some tactical ideas by Black, but White has not gone wrong yet. 45 ... Wb2 46 f5+ .:xt5 46...h7Ioses to 471Wd41Wxd4 48 xd4 i.a6 49 e6 ':b7 50 xg5+. 47 gxf5 i.e2 48 1Wa4?! Unnecessarily complicating mat- ters. After 48 1We 1 ! "h3 49 ':e8 White would win comfortably. 48 ... i.f3! (D) a . . . w. /. . Ii p  . . u _ d d . .8.8_ .. . . - . . .....  . R . _ . u . . . .      A brilliant resource, which forces White to play very accurately. After 48...1Wbl + 49 h2 1Wxf5 50 1Wd4+ f7 51 b7 White wins far more eas- ily. 49 1Wd7+ 49 f6+ is simply met by 49..1Wxf6. 49 ... h6 Incredibly, there is no mate for White. 50 1We6+ h5 (D) a . . . . .   w. /. . .  \Wr.  u ..d d . .8.8_. . . . - . . .....  . p . _ . u . . . .  _ Wf{% :  51 1We8+! An important finesse, which forces Black to block the g4-square with his king. After 511We l1Wd4 the posi- tion would be a draw. 51 ... g4 52 1We1 Now everything is under control agaIn. 52 ... .1xd5 53 ':e8 i.f3 53...1Wxb6 fails to 541Wdl+ .1f3 55 ':e4+ while 53...xf5 opens the position up and allows White to re- sume his attack by 54 ':f8+ g6 55 1We8+ g7 56 h2!, when Black is helpless. 54 f6 There are just too many pawns. 
ANAND - KAMSKY, PCA CAND. (3), LAs PALM AS 1995 163 54 ... h5 54...1Wxf6 allows 55 1We6+ ex- changing queens. 55 f7 1Wd4 (D)  . :@v  1:   / % .  /@a :; 0 //j/%'j VIi?! ;%'} $% _ wB . .8. :.. u  . d . . .  . -   - . rua . _ . - . - B B ..t.B  . / . " . ?.u. .  WIt  . . -  56 ':e4 This interference move finishes Black's resistance. 56 ... 1Wf6 56....1xe4 57 f81W. 57 b7 .1xe4 58 1Wxe4 581Wdl+ .1f3 59 b81W xdl 60 f81W is much more elegant; White loses all his original pieces but he has produced two queens. 1-0 Gata had seen enough. 58...1Wal+ 59 h21Wf6 60 h3 would be a pos- sible finish. After this struggle, which equalized the scores, there was a series of five draws in which I gradually gained the initiative. In game seven I gained a de- cisive advantage, but although I failed to convert this into a win it was never- theless an important game because it dented his main opening as Black - the Flohr-Zaitsev line of the Ruy Lopez. The drawing run was finally broken in the following game. 
Game 31 V. Anand - G. Kamsky PCA Candidates (9), Las Pa/mas 1995 Ruy Lopez, Flohr-Zaitsev 1 e4 e5 2 ltJf3 ltJc6 3 b5 a6 4 a4 ltJf6 5 0-0 e7 6 ':e1 b5 7 b3 d6 8 c3 0-0 9 h3 ':e8 10 d4 b7 11 ltJbd2 f8 12 a4 h6 13 c2 exd4 14 cxd4 ltJb4 15 .1b1 lId7 (D) An unexpected and rather unusual move. In game seven he had played the more normal 15...c5. .a a...a .J.".i W _  _ i.   . ',  . - , - aia . a _ Da a _ a atba p " '. u  % . i..  - - - * 16 b3! I hadn't prepared anything espe- cially against 15.. :iid7, although I knew all the theory. Since Kamsky must have carefully prepared this line, I decided to try to find a con- tinuation which would take the game out of theory and after a time I noticed the move 16 b3 (16 e5 and 16 ':a3 had been tried before). While there is some risl\.. in leaving the known paths, I had spent so much time on the Flohr-Zaitsev that I felt confident in my general understand- ing of this type of position. 16 b3 is quite a difficult move to make if you are used to the standard theme of swinging the rook over to the kingside via a3, but blocking the third rank doesn't mean that I am abandoning the rook to its fate. It can later have an influence along the a-file, but for this White needs to de- lay axb5 until it is really effective. 16 ... g6 If 16...c5, then 17 .1b2 and White prevents the usual re-deployment of the bishop by ...g6 and ...g7. 17 b2! (D) The earlier game Van der Wiel- Karpov, Amsterdam 1991 continued 
ANAND - KAMSKY, PCA CAND. (9), LAs PAIMAS 1995 165 17 axb51Wxb5 18 d5 c6 19 i.b2ltJh5 20 dxc6ltJxc6 with an unclear posi- tion. However, in this example White played axb5 too soon; it is a useful threat and shouldn't be executed too early. .IB B.lfI.B B B.t '"B . B .B  -.  - - - B'B B B 8- 8B B _ u   B8_ B_8   8. _  U _ "i..ii  - - -  17 ... i.g7 18 1Wc1! This is really the new idea: White intends to play i.c3 and 'iWb2, build- ing up pressure on the long diagonal. Van der Wiel had analysed 18 d5 c6, which is satisfactory for Black, but the alternative 18 i.c3 c5 19 axb5 (not 19 d5? ltJxe4 20 i.xg7 ltJxd2 and Black wins) 19...axb5 20 :xa8 i.xa8 21 i.xb4 cxb4 22 i.d3 also merited attention. 18 ... ':ac8 White gains the advantage after either 18...d5 19 e5 ltJe4 20 ltJxe4 dxe4 21 i.xe4 i.xe4 22 ':xe4 ':ac8 or 18...c5 19 e5 dxe5 20 dxe5. 19 i.c3 c5 20 d5 The structure is now very similar to the Benoni Defence. White's dream is to achieve the push e4-e5 under fa- vourable circumstances. Obviously not 20 i.xb4?! cxb4 21 1Wb2 :c3 with very active play. 20 ... 1We7 Preparing .. .ltJd7 . 21 ltJn (D) Now 21 1Wb2 is not especially effective because Black can reply 21.. .ltJh5 . B.lB.lB.B ..... .- B____ .B  ..  - - - B._8B B 8_ B8B B B8m .B8 B . 08. ai.1I a= 21 ... ltJh7?! Black changes his plan and de- cides to retreat the knight to h7 in- stead. 21...ltJd7 would have been better, although White has a slight advantage after 22 axb5! (22 i.xg7 xg7 23ltJe3 h5 followed by ...lbe5 is safer for Black) 22...axb5 23 :a7 i.xc3 24 1Wxc3 :b8! (24...ltJb6 25 1Wcl h7 26 e5 :a8 27 exd61Wxel 28 1Wf4! i.xd5 29 :xf7+ g8 30 i.xg6 wins for White) 25 1Wd2 g7. This line again emphasizes the point 
166 VISHY ANAND: My BEST GAMES OF CHESS that White should delay axb5 until it results in a concrete gain. 22 xg7 xg7 23 e3! (D) .IBIB _ fl . m - B."'. _'.- .. / ..  u ,,   '8. . _2__ 8_ 888 . .8B tD_8 8 8 088 a ' i. /. ' .,/ a %i' = 0-, /0/  + '  ' /1  /:  /1 /' ' / u /' / / /, /' "  Threatening g4. 23 .., h5 This weakens the kingside but the alternative defences also favour White: 1) 23...lbg5 and now: la) 24 xg5 hxg5 (24...1Wxg5 25 axb5 axb5 26 f4! is very promisjng for White) 25 axb5 axb5 26 :a5! Wlc7 27 ':a7 when Black is in trou- ble: 1 a 1) 27... 'iib6 28 lL)f+ gxf5 29 1Wxg5+ ct.fs 30 'ii'h6+ e7 (30...g8 31 :e3 f4 32 e wins) 31 e5! with a winning ttack. la2) 27....:a8 28 f5+! gxf5 (White also wins after 28...f6 29 ':xb7 1Wxb7 30 xd6) 29 'ii'xg5+ f8 30 1Wh6+ e7 31 e5! d8 32 exd6! ':xel+ 33 h2 is an attractive finish. Ib) 24 1Wc3+ (also very strong) 24...1Wf6 (24...h7 25lbxg5+ 'it'xg5 26 g4 and the f6-square is a horri- ble weakness) 25 1Wxf6+ xf6 26 xg5 hxg5 27 axb5 axb5 28 ':a5 with a very promising ending for White. 2) 23...1Wf6 24 lbg4! 1Wxal 25 1Wxh6+ g8 26 e5 ! (D) and now: ..I..... B B"'B B'B. .  . . ' . &   /',  . .. - ...8" . _ 'u, U  A_ . .k o ,.. W B"Z.J. .8. .ltJB8 . _ 088 . ' ' '//.   . '/',0  /  =  /. %; %....   / "   "  2a) 26...1Wc3 27 .1xg6 (not 27 g5? 1Wxel+ 28 h2 1Wxe5+ 29 xe5 xg5 30 xg6 f6 and Black defends) 27...fxg6 28 f6+! xf6 29 1Wxg6+ <it>f8 30 1Wxf6+ g8 3  ':e4! winning for White. 2b) 26...dxe5 27 g5 ':c7 28 xh7 also wins. 2c) 26...xdS 27 i.xg6 'iixel+ 28 xel fxg6 29 1Wxg6+ h8 30 h6 ':f8 31 f7+ :xf7 32 1Wxn with a clear advantage for White. 2d) 26....:xe5 27 gxe5! (not 27 ':xe5? dxe5! 28 llJg5 'iixbl + 29 h2 'it'f5) 27...dxe5 28 .1xg61Wxel+ 29 xel fxg6 30 twxg6+ strongly 
ANAND - KAMSKY, PCA CAND. (9), LAs PAIMAS 1995 167 favours White. The queen and passed pawns are far more effective than Black's scattered forces. 24 1Wd2!? White starts to play against the knight on b4. One idea is to play llJd4 at some stage, and if .. .exd4 then 'ifxd4+ and 'ifxb4. 24 ... g8 Simply stepping off the danger- ous long diagonal. After 24...'iff6 25 ':a3 bxa4 26 ':xa4 or 24...bxa4!? 25 :xa4 White has a positional advan- tage, while 24...11Jf6 meets with the tactical refutation 25 axb5 axb5 26 llJf5+! gxf5 27 'if g5+ h8 28 e5 and now: 1) 28...11Jh7 29 'ifxh5 f6 30 .1xf5 .1xd5 31 llJh4! 'iff7 32 llJg6+ g7 33 e6 .1xe6 34 .1xe6 :xe6 35 :xe6 'ifxe6 36 ':a7+, winning. 2) 28...dxe5 29 ':xe5 'ifd6 30 'ifh6+ g8 31 ':xf5 llJe4 32 ':g5+ llJxg5 33 1Wxd6 with a decisive ma- terial advantage. 25 axb5 axb5(D) BIB.B.B wB.t.. .iB4\ _ . Bi. .i8. ri ,,0 _ /%ff' ,;?: _ .8. . .l\. .l\ o_ "lJ_O .  "8_ , _ U . iL.    .   26 d1!! I spent a long time on this move, because if White delays then Black can play. ..':a8 and relieve the pres- sure. My main problem against Kam- sky has always been messing up winning positions (as in the Sanghi Nagar match), so I put a special ef- fort into being precise. The point of the move is both to prevent .. .llJg5 and to prepare c3, striking at the weak b5-pawn. It turns out that to save the pawn Black has to retreat his knight from b4, but then his most active piece disap- pears. 26 ... a6 Or 26....:a8 (26...11Jf6 27 llJc3 is very awkward for Black) 27 ':xa8 ':xa8 28 llJc3 'ifd7 29 e5 dxe5 30 llJxe5 'ifd6 31 'iff4 'iff6 32 'ifxf6 llJxf6 33 d6 with a clear endgame advantage for White. 27 llJc3 b4 (D) BIBIB.. wBJ._ _i.4\ 4\. r BiB . .8. .i " .8.  . . _ _  W .8 .tLJB8 . 1Uii "8 . - . _ U  %:I '" iL   :W  @A a /' //.  + ;;!%{i  {I ,, /.    , 28 llJb5 
168 VISHY ANAND: My BEST GAMES OF CHESS 28 a4!?, followed by b6-c4, is also very strong. 28 ... c7 After 28....:b8 29 1Wf4 ':ed8 30 i.d3 White again consolidates his positional advantage. 29 i.d3 ltJxb5 30 i.xb5 ':ed8 31 i.c4! Most Benoni players would have fainted by this point! White has a dream position: the bishop on c4 supports d5 in preparation for e4-e5, Black's minor pieces are ineffective and his kingside is weak. 31 ... f6 (D) Trying to bring the knight back into play. ..- ...  .  . illt1 P w..t.. .'. .  -..  - -  . 8. .. . - - . _.8. . .8. .ttJ.8 .  "8.  _ U   .       32 6! This creates the tactical threat of e5, followed by d6 and 1Wxg6+, so Black is forced to retreat his queen. The effect is that White activates his queen with gain of tempo. 32 ... 1Wf8 Black has no time for 32....:a8 (32...xe4 33 ':a2! wins) owing to 33 ':xa8 ':xa8 34 e5 h7 (34...dxe5 35 d61Wf8 361Wxg6+ and 34...xd5 35 ltJg5 are dead lost for Black) 35 e6 driving a wedge into Black's po- sition. 33 1Wg5 1Wg7 33...ltJh7 341Wf4 and 33...1We7 34 ':a 7! are no better for Black. 34 :&7 (D) ..- ... R._rua B ..t.. ... . - -.- . /8. .  % - - r.8. . .8. .ttJ.8 . . B8. . .    - -  34 ... ':c7 There is no defence: 34....:d7 (if 34...i.a8, then 35 e5 h7 36 1Wg3 puts Black's position under intoler- able pressure) 35 e5ltJe8 (35...xd5 36 i.xd5 and 35...dxe5 36 xe5 ':dc7 37 d6 also win for White) 36 e6 ':dc7 37 exf7+ 1Wxf7 38 ':e6 wins for White. 35 i.a6! ':b8 Or 35....:dd7 36 i.xb7 ':xb7 37 ':a8+! h7 381Wf4 g8 39 e5 dxe5 40 ':xe5!, followed by ':ee8. 36 e5! 
ANAND - KAMSKY, PCA CAND. (9), LAs PAIMAS 1995 169 The thematic breakthrough comes just when Black's pieces are tied up and unable to meet the new threats generated by this move. 36 ... ltJe8 Forced, as 36...dxe5 37 d6:d7 38 ltJxe5 and 36...ltJxd5 37 exd6 are hopeless. 37 Jbb7 38 i.xb7 39 'iid8 40 :&1! 40 e6 is less accurate as Black can still resist by 40...fxe6 41 ':xe6 (41 dxe6 'iie7) 41...ltJg7 42 'iixf8+ xf8 43 ':xg6 c4. 40 ... ltJc7 41 'iid7 _b8 White can win more easily after 41..._e8? 42 _c6! 'iixc6 43 dxc6 ':b6 44 exd6 or 41.. .dxe5 42 d6 _e8 43 dxc7! _xd7 44 ':a8+ g7 45 c8_. 42 'iixd6 The only chance. 43 bxc4 ':cxb7 Jbb7 _f8 c4 b3(D) - B B.B wB-_ BiB _ _ BiB _ _D .i _B _ B Bi_ Bt;JB _ B D.  . .  - . .  44 ':b1 I thought for some time about this move in order to find a clear-cut win. In fact 44ltJg5 b2 45 ':bl ':a7 would also have won, but White needs to find 46 h2! (after 46 ':xb2 'iixb2 47 'iid8+ ltJe8! 48 'iixe8+ g7 Black is saved because White's king is too exposed) 46....:al 47 ':xb2 'iixb2 48 'iid8+ g7 (or 48...ltJe8 49 'iixe8+ g7 50 'iixf7+ h6 51 'iif8+! xg5 52 h4+ g4 53 'iif3+ xh4 54 'iif4#) 49 'iif6+ h6 50 ltJxf7+ h7 51 'iih8#. 44 ... b2 45 'iic5! (D) The idea is to bring the queen back to d4. Then Black's pieces will be tied down to defending the b2- pawn, and White can exploit his cen- tral pawns. - . B.B B .-_ Bi. - . _iB . R .i d _ u d .- . - . . _t;J_  . . _ , U . _:B _ = 45 ... ':b3 45....:a7 46 'iid4 ':a2leaves Black equally pinned down, when White wins by 47 ltJg5 followed by 48 e6. 
170 VISHY ANAND: My BEST GAMES OF CHESS 46 'iid4 (D) //'/;;W# 0;%;' /  j'  . / .  ;:;(1:: /@ 0;  P;0 _ .aU w _ ... ////:;' ////"/0''-; ''0 i :;W: .\  '; , , , '?lfi ,  ,  A  /./ '  ,  , ' , ' , : a % % 0   ;%'/; . all/// >/,/P /,u'  //0>, . ' "//, / 11 ' '/: %); w  l 8  l ?;;f). i //, "// ,,"// /////// /// /'// ,fa// . _ //%  /'0'0 ... .t.z.JR8 _ . 8. .". . %/0 - - , 47 ltJg5 White could also win by 47 e6 ':c3 48 exf7+ xf7 49ltJg5+ e7 50 d6+ (but not 50 ':el+ d7 51'iig7+ c8 52 d6! ':xc4! and the fight con- tinues as 53 ':e8+ loses to 53...b7!) 50...d7 (50...d8 51ltJf7+ d7 52 ltJe5+) 51'iig7+! xd6 52ltJe4+. 47 ... ':c3 48 'iif4! f6 Or 48....:cl+ (48...'iif8 49 ':xb2) 49 ':xc 1 (49 h2 is also winning) 49...bxcl'ii+ 50 'iixcl with three ex- tra pawns. 49 exf6 50 f7+ ltJxd5 1-0 
Game 32 V. Anand - G. Kamsky PCA Candidates (11), Las Palmas 1995 Sicilian, Najdorf le4 c5 2 ltJf3 d6 3 d4 cxd4 4 ltJxd4 ltJf6 5 ltJc3 a6 I had held the 10th game with Black fairly easily, so Gata was get- ting a bit desperate; hence his choice of the Najdorf. 6 i.e3 e6 7 .te2 .te7 8 f4 ltJc6 9 1*'d2 ltJxd4 10 1*'xd4 0-0 11 0-0-0 1*'a5? (D) A serious mistake, all the more surprising in that similar positions arise in the Richter-Rauzer, an open- ing with which Kamsky is very fa- miliar. 11...b5 or 11...1*'c7 would have been better. 12 'iib6! 1*'xb6 13 i.xb6 A dream position for White, espe- cially in view of the match situation (I needed one point from the last two games to win). White has a clear end- game advantage and can press hard for the win without the slightest risk of losing. . B: J. 0  .  ........ - . ... ... w. . _ /j",,. .. .- . . /u/y: _ . //.' '// . . .   ;m _ . . rou l\  / . rtJJf _ 0 / ,0 _ .   . 0"-1/  ,/ /; W& 88.i.8" // ;:; _ U   '  //;    /;-%? . ' '''lf ... - ;" - ,% 13 ltJe8 Black also has a poor position af- ter 13....td7 (13...ltJd7 drops a pawn to 14 .tc7 while 13...e5 14 f5 does not help Black) 14 e5 dxe5 (14...ltJe8 15 .tc5! with a decisive advantage) 15 fxe5 lbd5 16 ltJxd5 exd5 17 .tf3 %lac8 18 i.xd5 (18 .te3 .tf5 19 c3 d4 20 %lxcJ4 .tc5 21 i.xb7 :tc7 is less clear, for example 22 .txa6 :ta8) 18.. ..tf5 19 .tb3 i.g5+ 20 cJi>b 1. White seems to keep his pawn and though Black has a temporary initia- tive, White should be able to weather it. 14 e5! Immobilizing the knight on e8. 14 ... d5 
172 VISHY ANAND: My BEST GAMES OF CHESS After 14.. .f6 15 .tf3! (15 exf6 .txf6 16lbe4 is also good for White) White has strong pressure. 14...dxe5 15 fxe5 f6 16 .tf3 is similar. 15 f5! (D) The most forceful method. After 15 i.g4 g6!, followed by ...ltJg7, Black has better defensive chances. .......-.. ... -./. B_ _.  .. ... . _ .'D. . . . . .  . . D.i..D . :. .: -  - - 15 ... .td7 Or: 1) 15.. .exf5 16 ltJxd5 i.g5+ 17 cJi>bl .te6 18 .tf3, followed by h4, and White stands very well because of the e8-knight. 2) 15....tg5+ 16 cJi>bl .tf4! 17 .td4! (17 .tc5 .txe5 18 .txf8 cJi>xf8 is only slightly better for White) 17...b5 (17...ltJc7 18 g3 .tg5 19 f6! is similar) 18 J:dfl .tg5 19 f6 with a large advantage. 16 .tg4! (D) White must be accurate. After 16 fxe6 fxe6 17 .tg4 .tc8, followed by ...g6 and ...lbg7, Black might wrig- gle out. .. ..-.. ..... B _ _.... _ .. ... . . .'D. _ . .i.. .  . . ". ." U _  U . :. .: -  - - 16 ... .tc8? After this Black is in dire trouble. The alternatives were: 1) 16...1:c8? 17 f6! gxf6 (17....tc5 18 ltJxd5 gxf6 19 exf6 and 17.. ..td8 18 ltJxd5 are also winning) 18 ltJxd5 .td8 19 .txd8 exd5 (19...%lxd8 20 ltJb6) 20 .txd7 J:xd8 21 e6 winning either a clear pawn or the exchange for a pawn. 2) 16....tb4! 17 fxe6 .txe6 18 ltJxd5! (18 .tf3 i.xc3 19 bxc3 J:c8 20 .txd5 .txd5 21 J:xd5ltJc7 is not clear) 18....txg4 19 J:d4 winning a pawn. 3) 16...i.g5+ 17 cJi>bl J:c8 (after 17.. . exf5 18 .tf3 .tc6 19 ltJxd5 the e8-knight is again a serious handi- cap) 18 fxe6 .txe6 (or 18...fxe6 19 ltJxd5 and wins) 19 .tf3 with a clear advantage to White. 4) 16.. .exf5 is most simply met by 17 i.f3! as in line 3 above. 17 ltJxd5?! is less clear after 17.. ..tg5+ 18 .te3 f4! (18...fxg4 19 .txg5 and 18...i.xe3+ 19ltJxe3 i.e6 20 i.xf5 
ANAND - KAMSKY, PCA CAND. (11), LAs PALMAS 1995 173 should win for White) 19ltJe7+ (19 .txd7 fxe3 20 J:d3 J:d8 is unclear) 19...h8 (19...xe7 20 ':xd7 fxe3 21 J:xe7 strongly favours White) 20 J:xd7 fxe3, although 21ltJd5 retains some advantage for White. 17 J:hn! as 18 ltJa4! (D) .IB-*-B...B BBiB .i.i m BiB B  Bi R 8B _ _ U _ lDB . Bi.B B . B B 8"8. .8 u   U . :.:. -  - - From here the knight can move to c5 or b6, as appropriate. 18 ... f6 There is no good move, for exam- ple 18.. .exf5 19 .txf5 .te6 20 ltJc5 or 18...J:a6 19 e2 J:a8 20 .te3 and Black has serious problems in either case. The text-move is more com- plex, but White's advantage persists. 19 fxe6 fxeS 20 ltJc3 The knight has achieved its task and now heads for the excellent square d5. Round about here I was suddenly worried that I had allowed the position to become unnecessar- ily complicated, but this was only nerves due to the exceptional impor- tance of the game - in fact White has everything under control. 20 ... .tgS+ If 20...J:xfl 21 J:xflltJf6, White has a nice win: 22 J:xf6!! .txf6 (or 22...gxf6 23 ltJxd5 f8 24 ltJxe7!) 23 ltJxd5 a4 24 ltJc7 picking up ma- terial. 21 bl (D) .IB-*-B..._ B _iB B .i m B8B .  .i  -  - - _ B .i._ B  _ B 88. .8" u _  u B<iiiB:_:B 21 ... ltJf6 Or 21...J:xfl 22 J:xflltJf6 (after 22....:a6 White has the pleasant choice between 23 ltJxd5 and 23 e7 .tf6 24 .txc8 J:xb6 25 ltJxd5) 23 J:xf6 (23 h3 d4 24 ltJb5 ltJd5 is less effective) 23.. .i.xf6 24 ltJxd5 h5 25 .th3 e4 26lbc7 J:b8 27 a7 e3 28 c 1 and wins. 22 ltJxdS ltJxg4 22...ltJxd5 23 J:xf8+ xf8 24 J:xd5 is a simple win. 23 Jbf8+ This wins the exchange and leads to a technically winning ending. 23 
174 VISHY ANAND: My BEST GAMES OF CHESS e7 would have had a similar result: 23...:e8 (23...:xfl 24 e81i'+ :f8 25 1i'h5) 24ltJc7ltJxh2 25ltJxe8 i..xe7 26 :hl i..g4 (26...ltJg4 27ltJc7) 27 ltJc7 :c8 28 :de 1 i..d8 29 :xe5 i..xc7 30 i..xc7 :xc7 31 :xh2. 23 ... cJi>xf8 24 ltJc7 Not 24 i..c5+ cJi>g8 25 ltJc7 b6!. 24 ... :a6 On 24...:b8, 25 i..c5+ cJi>g8 26 i..a 7 traps the rook. 25 i..c5+! A necessary intermediate check. 25 ... cJi>g8 25...i..e7?? allows mate in one. 26 ltJxa6 i..xe6 26...bxa6 27 e7 cJi>f7 28 :d8 also leaves White the exchange up. 27 ltJc7 i..f5 28 h3 ltJf6 Or 28...ltJe3 29 i..xe3 i..xe3 30 g4 i..g6 31 ltJe6 and White wins an- other pawn. 29 g4! (D) . . ..8 .. . /. BB  d  .8_8  . / .  .  /  .  ' %ii   .JL:lj  , ;.,/ /  ,,/ 8 . .8 - 8 8 . B. . . ..:8 _ Of course the position is winning for White as Black doesn't even have a pawn for the exchange, but this forcing sequence of moves gives Black no chance to develop counter- play. 29 ... i..e4 29...i..g6 30 ltJe6 i..h4 31 i..b6 is no better. 30 ltJe6 i..h4 Or 30...i..f4 31 i..b6. 31 g5! ltJd5 32 :n h6 33 gxh6 gxh6 34 :'-8+ cJi>h7 (D) 8 . a _ w_.. . .. . .. / _ _t.z.J. u,   . .  ,,/ _ d . ..1.  d d d _ 8 _ _ . B. .. 8 .<it>. . . _ _ Wd _ 35 i..d6! ltJb4 The alternatives are 35...i..f6 36 ltJc5 i..g6 37ltJxb7 and 35...i..g2 36 i..xe5 i..xh3 37 ltJf4 ltJxf4 38 :xf4 i..g5 39 :f7+ cJi>g6 40 Jhb7. In both cases White gains another pawn. 36 i..xe5 i..xc2+ White wins easily after 36...ltJxc2 37 :f4 i..d3 38 J:xh4ltJd4+ 39 <it>cl ltJxe6, simplifying the position. 
ANAND - KAMSKY, PCA CAND. (11), LAs PALMAS 1995 175 37 cl i.e4 38 ltJf4 ltJd3+ Otherwise Black is mated on h8. 39 ltJxd3 i.xd3 40 ':h8+ White could have taken the pawn by 40 ':f7+ g6 41 ':xb7, but forc- ing the exchange of bishops leaves White with a trivial technical task. 40 g6 41 i.f4! i.g5 42 i.xg5 xg5 43 d2 .tb5 44 .:a8 i.a6 45 ':c8 h4 46 ':c5 a4 47 .:as h5 (D) 48 e3 Of course it doesn't really matter, but 48 el! was a little more accurate $I . . . ///--;/ Z;0 '//ffi:  wR'. _ B - gr . . .IL. . . . a _ _ .. .. . - ' . . . .   . . U  . . . . . . in that 48.. ..td3? 49 ':xa4+ cJi>xh3 loses the bishop after 50 ':a3. 48 i.n 49 Jba4+ 3 50  1-0 . ..and I had won the right to chal- lenge Kasparov. My success in this match was mainly due to not underestimating Kam- sky's fighting qualities. Already in Sanghi Nagar I had him beaten, but Ijust didn't finish the job. This time I was more careful in winning positions and didn't relax until the match was actually over. In the Riga tournament during May I finished second. My play was very convincing, apart from the loss to Kasparov in the Evans Gambit, and I felt on form. This was the first time that I encountered Kasparov after becoming the challenger and although I lost the individual game, I was happy with my performance. Both the tournament and the city were very pleasant, so I had every reason to be in a good mood as I started my preparations for the World Championship match. 
Game 33 V. Anand - J. Timman Tal Memorial, Riga 1995 Ruy Lopez, Worrall Attack 1 e4 e5 2 lbf3 lbc6 3 .tb5 a6 4 .ta4 lbf6 5 0-0 .te7 6 1i'e2 It's perhaps a bit flippant to say that I played 1Ve2 because I was sick of %leI, but it is useful to vary your openings a bit. Even chess profes- sionals are human beings, and if you have seen a position hundreds of times then it is possible to become stale. A bit of variety helps to keep one's interest alive. Another point is that I had just played 33 games against top-level grandmasters (11 in Las Palmas and 22 in Monaco) and no one has an in- exhaustible fund of new ideas. I still had a few left in this 1i'e2 variation, which was another reason to try it in this game. 6 ... b5 7 .tb3 0-0 8 c3 d6 9 %ld1 .tg4 10 h3 It's best to force Black to decide right away whether he is going to take on f3. For this reason I regard 9.. ..tg4 as being premature. Tivia- kov, who is an expert on this line, is of the same opinion. 10 ... .th5?! (D) 10....txf3 11 1i'xf3ltJa5 12 .tc2 c5 is probably the lesser evil. -B . ... .  ".. w _ _ _ ...- - . ... _ ..t . .. . ..i.D .12J.  "\WJ. U UWU . 12J:.  r_. 11 d3 ,White could already play 11 g4, but given that Black didn't take on f3 last move, he is hardly going to do so now. 11 ... lba5 12 .tc2 c5 13 ltJbd2 ltJd7 Now 14lbfl can be answered by 14....txf3 15 1i'xi3 .tg5, so it is time to break the pin. 14 g4! i.g6 
ANAND - TIMMAN, TAL MEMORIAL, RIGA 1995 177 15 ltJn White has opted for a very solid formation in the centre and will con- duct all his play on the kingside. 15 ... lbb6 Black has many possibilities, but White retains a slight advantage in any case, for example after 15.. .h6 16ltJg3 i.g5 17 ltJxg5 hxg5 18lbf5. In this line the pawn on g5 doesn't really block White's kingside play because he can often continue with h4, meeting ...gxh4 by g5 and re- gaining the pawn on h4 at his leisure. Black's best idea may be 15...ltJc6 16lbg3 J:e8 17 lbf5 lbf8, heading for e6, when the slight weakness of f4 might be relevant later. 16 ltJg3 lbc6 (D) :I // :  // % . / / .   /.   w ;/  /;I / ; /';$/< _ ;;;!:';' . . / /;; & / . :0' & W /;; ::},'i/ 0 .. %:;i .. 'l///// '0'#/ / " '/ / ,/""  & "/ 1&\ W&% ;;;  mr .. /" ...w w;f.IL //;:/ , <'#; i // . //,,-:: '/ / 1 / ,, //;::/ /;, ///; /  0 1/1  ;;;/'/   0 0/ '(0; % :0,,,,'/ /. ,,// /0> ;tf:i;;:i: A /'0C A % ;* _O%iJor . llJ    u  z "i..jy  ;; / U _   ".  /   / (j 17 lbf5! A very awkward knight. If Black ever plays .. .i.xf5, White will take back with the g-pawn, opening the g-file for his attack. 17 ... ':e8?! 18 h4! This threatens 19 lbxe7+ li'xe7 20 h5, so now Black is forced to take. Black can, it is true, win a pawn, but White's attacking chances more than compensate. 18 ... i.xf5 After 18...f6 19 i.b3+! i.f7 (not 19...d5? 20 exd5lbxd5 21 c4! win- ning for White, and 19...cJi>h8? 20 h5 gives White a clear advantage) 20 i.xf7+ cJi>xf7 21 g5 White has excel- lent attacking prospects. 19 gxf5 (D) .- -.-.- . . "ii B  )ff;,; ,_  i . . .. .i . /:,,0   .    . /;t   ///// / f/] ;;j/',f; p  I    %Z /.  @ x/ 1' / /..1 % /;///" /,,-::'/-/// '//'/// , / / . f!R/. A iY T:'\ . . ,Lt 0 ///_ A ,  _\Wr . o ru}... _ iii ?:Qft   g "   ':: :/. '0  /-   % ? / /  ; / ;/ /  19 ... d5?! Black could have transposed into the game by 19...i.xh4! 20cJi>hl (20 ltJxh41i'xh4 21 cJi>g21i'e7 22 J:hl is also feasible) 20...d5. This move- order would have been more accu- rate, as it denies White the opportu- nity mentioned in the next note. 20 <iithl?! Not bad in itself, but Black's pre- vious move gave White an additional 
178 VISHY ANAND: My BEST GAMES OF CHESS option which he didn't exploit: 20 lbg5! (not 20 h5 h6 21 h2 .tg5 22 :g 1 f6 and Black sets up a blockade) 20...h6 (otherwise White proceeds with his attack without sacrificing a pawn) 21 _h5! hxg5 22 hxg5 .td6 (22..._d6 23 g2 g6 24 4 gxf5 25 :hl winning) 23 g6 fxg6 (23...f8 24 .tg5 fxg6 25 fxg6 and Black is lost) 24 fxg6lbe7 (24..._f6 25 i.g5 _e6 26 exd5! lbxd5 27 .tb3 with a decisive attack) 25 _h7+ f8 26 .th6 lbf5 27 .txg7+! followed by mate. 20 ... .txh4 21 :gl (D) 21 lbxh4 _xh4+ 22 g2 d4 is less clear. ._ ....B . . . - B. . .1..1. I._B _ B .1.1.8. - - - . . .8.  d d . _ _ D8BliJB 8i..\WIP . u .wu .   .  - - . - 21 ... .tf6 White's attacking chances along the open kingside files offer very good compensation for the pawn. In addition, White's bishops are very effective; the one on cl is ready for a sacrifice on h6, while the c2-bishop can go to b3 and in the long run will play an important role in exploiting Black's weak light squares. 22 lbh2 g6 23 _f3 dxe4 If 23.. .d4, White plays simply 24 lbg4. 24 dxe4 ltJd7 2S 3! _e7 After 25...c4 26 .te3, followed by :ad 1, the combined pressure on d7 from dl and h3 would be very awk- ward. 26 .te3 .tg7 27 ltJg4 ltJf6 27...gxf5?? loses immediately af- ter 28lbh6+ h8 29ltJxf5. 28 lbxf6+! If 28 .tg5?, then Black escapes from the pin by 28...ltJxg4 29 :xg4 f6. 28 ... _xf6 (D) After 28....txf6 29 :g2, followed by :agl and :h2, White has an enormous attack. .B ....B wB . _1._, '.B .'B .. 8. . - - . B _8B B .   .\WI d u _ .W 8"i.B " B u _ U M a '   . '   a ' <it>  :/jff%.  / :!z-;%  " / 
ANAND - TIMMAN, TAL MEMORIAL, RIGA 1995 179 29 ':g3? The obvious 29 i.xc5! would have been the logical culmination of White's play. This not only regains the pawn, but also prevents .. .c4, which imprisons the c2-bishop. Af- ter 29...g5 30 i.e3 h6 31 ':g2 'ile7 32 i.b3 Black's position would come under steadily increasing pressure. After the careless text-move Black gains a new lease of life. 29 ... ltJe7! 30 h2 Now 30 i.xc5? is impossible be- cause of 30...ltJxf5 31 exf5 'ilc6+. 30 ... c4 31 ':h1 (D) .. ..... B _ _ lIi_' ._ _ _i. ... r/_ .._- . .   "\WI . U _ .- "i..   U . U * - - - .: 31 ... h6 More or less forced to prevent .tg5. If Black tries to keep the pawn then 31...:ad8 32 gl h5 33 i.g5 'ilc6 34 'ilh4 f6 (34...c;t>f8 35 f6 wins) 35 .te3 g5 36 'ilxh5 and White is very much better. 32 .txh6 i.xh6 33 'iixh6 :ad 8 Not 33...'ilh8? 34 'ilxh8+ xh8 35 f6lbc6 36 g2+ g8 37 ':gh3, winning. 34 g2 'ilg7 3S 'ile3 :d6 White still has a clear advantage, but he cannot win by playing solely on the kingside. The next move starts the opening of the queenside with the aim of activating the c2-bishop. 36 b3! ':c8 (D) ... ... w. . _i_ i. _ .i. ... -. .... - ."" mu  .OU _ _ _i._ <iit_ . B _ .: 37 bxc4 bxc4 After this the bishop can become active at a4, denying critical squares to Black's pieces. The alternative was 37....:xc4, but then the bishop becomes active on b3 instead and White can gradually step up the pressure, much as in the game: 1) 38 i.b3 :c7 39 'ilg5 'ilf640 'ilh6 'ilg7 is not entirely clear. 2) 38 'ila7! 'ilf6 (38...ltJc8 39 7 :c5 40 i.b3, etc.) 39 i.b3 ':c8 40 fxg6 ltJxg6 41 i.xf7+ 'ilxf7 42 
180 VISHY ANAND: My BEST GAMES OF CHESS :h8+ cJi>xh8 43 'iixf7 ltJf4+ 44 cJi>f3 and White will win. 38 :hh3 White doesn't have a straightfor- ward winning plan. He is just probing, altering his plan of attack according to how Black arranges his defensive pieces. Winning is often as much a matter of putting pressure on the op- ponent and making life unpleasant for him as it is about specific moves. 38 ... cJi>f8 39 'iicl Now the threat is 'iihl, so Black's king has to return. 39 ... cJi>g8 (D) BE. ... W . / / / / . .  . ,.  .. - ... . . . - -   .... . . R .  : . u  ou {  / - .... u'&' .. ,-;,// R d '<0{/ 40 :f3 Just at this moment White can switch his rooks to f3 and g3, since 40...gxf5+ loses to 41 :fg3! :g6 (or 41...ltJg6 42 exf5 ltJf4+ 43 cJi>h2) 42 exf5. This will eventually cause f7 to come under strong pressure. 40 ... :cd8 41 ':hg3 Now White threatens a gradual penetration on the queenside by 'iia3, i.a4 and 'iic5. Then c4 will be attacked and White will also be threatening 'iic7. Timman decides that there is no real defence against this plan, and so goes for desperate counterplay. The result is only to ac- celerate his loss. 41 ... f6?! 42 fxg6 ':d2 42...ltJxg6? loses after 43 ':xg6 'iixg6+ 44 :g3. 43 'iibl 'iif8 (D) 43...:e2 loses to 44 'iib6, but Black might have lasted longer with 43...cJi>f8. However, after 44 ':g4 ltJxg6 45 'iib4+ ':8d6 46 'iic5 the win is just a matter of time.  . . / 0 .  iI ; / /.' //, .  .    p  / / / , W}} . - -  w.  ,. _ .. . r. . fgr  . ?   .;:j:t:{: %.  f '/ '/;:/ /// ;  tj /h/; "  m}& / 'l:;:-i:;    ///-;/  ',-',-///': // / A //.  fWJ: /  £ o ... d'&'. .\WJ .. /j.. Y . ;@ . ' . ;: . . . / ;:: ;:'> 44 'iih I! 'iig7 45 .:xf6 1-0 In view of 45...'iixf6 (45....:xc2 46 :f7 and 45...:f8 46 'iih7+ are also catastrophic) 46 'iih7+ cJi>f8 47 g7+. 
MATCH WITH KASPAROV 1995 181 The months leading up to the Kasparov match were a very exciting time, with winning the World Championship a real possibility. I had been compet- ing in the FIDE and PCA cycles for roughly two years and there had been many tense and exciting moments. When I finally reached the Kasparov match there was a feeling of anticlimax, as if I was already spent from the earlier efforts. I had the feeling that, having played so many matches, I was fairly exposed because I had already shown most of my best ideas trying to reach the world championship itself. The champion can be much better pre- pared, as he only has to play when and where he chooses and can just wait to see how the cycle develops. Of course, every challenger says much the same thing! Certainly, I would have preferred seven or eight months to prepare for Kasparov instead of just under six. Moreover, I had agreed to play tourna- ments in Monaco, Riga and Moscow (these were arranged before I knew I would play Kasparov) which ate into the possible preparation time. However, it was perhaps no bad thing that I played in these events, as six months is a long time to sit analysing without any tournament activity, although it did mean that my preparation only really began in May, and the first game even- tually started on September 11 tho I had to assemble a team in a hurry, and it was a unique experience sud- denly going from having at most two seconds to having four. These were Ubilava, who had been with me since the beginning of the cycle; Yusupov, who had already helped me earlier in the Kamsky match; Wolff, who had been my second in the Ivanchuk match (see Game 17), and Speelman, who was completely new. I had worked with Ubilava and Yusupov before, and was very happy with them. I felt that Patrick Wolff would be a help as he is very well organized and has great experience against the Sicilian. Speelman had been Short's second in his match against Kasparov and I felt that he might be able to offer me some insight into world championship chess. It was amazing how much more you could accomplish with such help, but it was also much more confusing comparing the results of one person's analysis with another's. Trying to get everybody to work together in the most efficient manner was a major task in itself. You can't have all five people ana- lysing on one board - it's just too many heads and hands. On the other hand, splitting into groups analysing the same position often leads to the groups heading off in entirely different directions and then it can be hard to decide which line you are actually going to play. Thanks to the information explo- sion, the amount of material you have to deal with is gigantic. Facing 
182 VISHY ANAND: My BEST GAMES OF CHESS Kasparov, analysis of the Sicilian was a priority, but this is one of the largest of all opening complexes. I had decided that my main defence to 1 e4 would be 1...e5, and this was also an enormous amount of work. Kasparov had shown that he was prepared to play openings other than the Ruy Lopez, so we had to spend some time on the Evans Gambit, Scotch and Irregular Open Games. He also plays 1 d4 regularly, so one can imagine the amount of work we had to do. With hindsight, and given that we were new to the job, I think that we did a reasonably good job with our preparation. I am sure that we would do better next time, based on our experience in 1995. Still, this was an area where Kas- parov had an advantage due to his vast experience preparing for world cham- pionship matches. His preparation was able to survive the close scrutiny of a world championship match while mine took some heavy blows. It was exciting finally to be given the chance to play against Kasparov for the world championship, but I look back on it now with a fair amount of disil- lusionment. The organization of the match verged on the ludicrous. First the venue for the match was changed from Cologne to New York without con- sulting or even informing me! I was still preparing for Cologne and making hotel reservations there when it was known within the PCA that New York would be the venue. I should perhaps explain that under the PCA system you are left largely to make all your own arrangements for the match, so a switch of venue is quite an inconvenience. Then, in late July or August, Bob Rice suddenly called to announce that the prize fund had been reduced from $1.5 million to $1.35 million. And so on. The only response from the PCA to the various problems regarding the organization of the match was "We are doing our best, but...". I didn't find this argument particularly convincing. I had to try to shut myself away from all these other problems to concentrate on the chess, but I didn't have total success. At some level, it kept bothering me. They tended to take the most optimistic interpretation of any good news. On the other hand, bad news would be parcelled out bit by bit in small doses, so that you would not realize the full import straight away. If it hadn't been for Frederic Friedel, who was in charge of player relations for the PCA, I probably wouldn't have been kept informed at all. A few days before the match we were suddenly told that we had to write a daily column for USA To- day. Towards the end I completely lost interest in this and produced just the barest minimum - deep notes along the lines of "He played the Sicilian." The dealings with the PCA leading up to the match were thoroughly de- pressing and, by the time I got to New York, I was just sick of the whole 
MATCH WITH KASPAROV 1995 183 thing. This was not a factor in my defeat but my inability to deal with it and take it in my stride was! Kasparov put up with a lot during his 1984 match against Karpov. I have a much better idea now of how one should just develop a thick hide in these matters. Nevertheless, a positive consequence of the whole affair was that when it was over, I was quite happy to forget about both it and the events leading up to it. 
Game 34 V. Anand - G. Kasparov PCA World Championship, New York (9) 1995 Sicilian, Scheveningen 1 e4 The World Championship had be- gun with eight draws. The absolute record, 17 consecutive draws, still belongs to Karpov vs Kasparov from Moscow 1984/5. However, in 1995 the next six games had five decisive results! 1 ... 2 ltJf3 3 d4 4 It)xd4 5 It)c3 6 ..te2 This had become one of the main battlegrounds of the match. Neither player wanted to blink first, so an- other Classical Scheveningen ap- peared on the board. 6 ... 7 0-0 8 a4 9 i.e3 10 f4 11 <it>hl 12 ..tf3 (D) Our discussion of the Schevenin- gen continues into my fifth White. I had tried 12 'iid2 in the first game and then 12 i.d3 three times. In games five and seven Garry had c5 d6 cxd4 It)f6 a6 e6 ..te7 ltJc6 0-0 'iic7 :te8 managed to find a solution to 12 i.d3, so it was time to switch varia- tions. ..1. .. . 1@ A J});.JL J A %  :;;;; i . ' '//. 0 .  i . / i B ;0  /.     /;;f/; %,  ; /'/ / / i ;f:::::; . :% . ;; i PI //. /'i/:; , /. , :; 0 f/ / . /?, /, ;'l:i%;? :-;:-:,; /'  /.: /'/ >; / 1;/;/;; ifi  ?2J '// A {fj ;;A >;;/ //).::% . / Q /  ?; ; / /' V ;;;;/} :fI i. , /// , /  ;/0 //; /";/// . ;::; +/; + i;;/; h / /. %;X0/, , ; 'l/@/; // , :://!;}: , '  fj / /0 .;-;;:i;) / /  >;;;::i 1/ / i " ; ,    ;{ '8' /:$ , :// , / , ' rD ,/i ,/ / / ';:;. /, /. . /. / '{fj  ';/:' ,  /'    / / / .;- a /    /. / // 12 ... ..td7 Kasparov had played 12...:tb8 consistently in the matches against Karpov, but then switched to this move against Van der Wiel in Am- sterdam 1987. I can't really say that I was surprised by .....td7, because I had studied the Van der Wiel game, but it wasn't uppermost in my mind. I was waiting for 12...:tb8 and was ready to whip out 13 g4, when this move was played. I had to search my memory to remember what we had found in the Van der Wiel game. 13 ltJb3 ltJa5 
ANAND - KASPAROV, PCA WORW CH., NEW YORK (9) 1995 185 14 ltJxa5 15 'ii'd3 16 :rd1! Almost all the interesting games in this line were played by Van der Wiel, against Polugaevsky and Kas- parov himself. Van der Wiel's notes were a good starting point for my own analysis. We had prepared both 16 g4 (Van der Wiel-Polugaevsky, Haninge 1989) and the text-move, which was a recommendation of Van der Wiel (a third move, 16 'fid2, was played in Van der Wiel-Kasparov). My decision to prefer 16 :fdl was made at the board. 16 ... i.c6 After 16...e5 White simply plays 17 f5. 17 b4 'fic7 18 b5 .i.d7 (D) Not 18...axb5? 19 axb5 i.d7 20 ltJa4 with advantage to White. 'ii'xa5 :ad8 j  ...ffi .;:: ? ,,/ / Wg #;/  . :  u/ .t. :« / .  . / . W { %   (2;,0 . /:; / ; /uu, .. . . _ /,: _ ffZ 8. . . A. {A . Q '0/i% ;:/;; Q r ;p;}% w m ii . i. ;f .. : %  + + %:{// / //' ::; ;  . ' . 3 A.  .  A  / /;/;: Q w j/.% Q;A v D ' /;0' : %, :; <it> %  ;:..;// . ;;//. /')/ . '/%' . -;//;://  /  ;;{ :??;:; fd% I was surprised that he was pre- pared to go down this line so blithely but later it turned out that there was a good reason: an earlier game Cuij- pers-De Boer, Dutch Championship 1988 had continued 19 ltJe2 :c8 20 bxa6 bxa6 21 'ii'xa6 :a8 22 'iid3 :xa4 23 :xa4 i.xa4 with equality. It was lucky I didn't know about this game, or I might have abandoned the whole line! 19 :abl! Clearly stronger than 19 ltJe2. 19 ... axb5 (D) After 19...:c8 20 e5 dxe5 21 fxe5 'ii'xe5 (21.. .ltJd5 22 i.xd5 exd5 23 ltJxd5 'fixe5 24 i.f4 is also good for White) 22 i.d4 'ii'c7 23 i.xf6 (after 23 b6 'ii'b8 24 .i.xf6 i.xf6 25 'fixd7 :e7 Black will regain the piece) 23...i.xf6 24 'fixd7 White will cre- ate dangerous passed pawns as Black hasn't exchanged the a-pawns. i' . · . .  ,;/;X' {o /u/'" f ffg W ...t,/'.'   .:// % . % /% & . nl' % .  .  . /  / ,i // / /z  %% ..- . - A fj( .. . . :  . ft:7 A :ti  / :'@  % Q ;f :?;- // Q   i% /// //% // u  .;:.;///'& 0?  <3;  ii % '  i. if::;--,f . / -j +  + -0(//;;/-  ;,  ,,;  :;/<'/ A f:%d? _ A  Q. Qd 8:8:8 . 20 ltJxb5! 20 axb5 :c8 offers White less than the previous note, now that the a-pawns have gone: 21 ltJa4 'ii'xc2 
186 VISHY ANAND: My BEST GAMES OF CHESS 22 lL\b6 (after 22 Wixc2 :xc2 23 lL\b6 White's compensation for the pawn is nebulous) 22...Wixd3 23 :xd3 :c7 24 e5 dxe5 25 fxe5 lL\d5 26 .i.xd5 exd5 27 lL\xd5 .i.f5! 28 lL\xc7 .i.xd3 29 lL\xe8 (or 29 :dl :d8) 29.. ..i.xb 1 30 lL\d6 .i.xd6 (or 30....i.d3) 31 exd6.i.f5 with a draw. 20 ... i.xbS This surprised me since I was ex- pecting 20...Wia5 21lL\xd6 (anything else allows Black to play .. .i.c6 with a fine position) 21....i.xa4 22 .i.b6 (22 e5 .i.xd6 23 exd6lL\d5 24 .i.xd5 :xd6 25 Wia3 :xd5 26 :xd5 exd5 is, if anything, slightly better for Black) 22...:xd6 and now: 1) 23 Wixd6 .i.xd6 24 .i.xa5 .i.xf4 (24....i.xc2? loses to 25 e5) 25 :Xb7 .i.xc2 26 :d8 :Xd8 27 .i.xd8 .i.xe4! (27...lL\xe4 28 i.c7 gives White more chances) 28:b4 .i.xf3 29 :xf4 .i.d5 30 .i.xf6 gxf6 31 :xf6 is a draw. 2) 23 .i.xa5! :xd3 24 cxd3 .i.xdl and in the resulting ending White keeps a slight advantage due to his two bishops and Black's weak b- pawn. 21 _xb5 21 :xb5!? may be even stronger than the game continuation, e.g. 21...lL\d7 (21...:c8 22 :dbl Wixc2 23 Wixc2 :xc2 24 :xb7 definitely favours White; note that 24...d5? loses to 25 e5) 22 :dbllL\c5 23 Wic4 with a clear edge for White. 21 ... :a8 22 c4 e5 (D) aB BaB.B ..ma _.. w. _ _ _ B _ II B B..-8 _ B 888B80 B B . mi.B B B B80 B.:B.:B B 23 .i.b6! A fine move, forcing the queen to go to c8. The exchange 23 fxe5? dxe5 is premature since then Black can meet 24 .i.b6 by 24...Wic6!. 23 ... Wic8 Now 23...Wic6 is bad owing to 24 Wixc6 bxc6 25 c5!, when 25...dxc5 loses a piece after 26 fxe5. Therefore the queen has to retreat to a more passive square. 24 fxe5 25 as 26 h3 dxe5 .i.f8 Wie6 (D) a. .a-.. - - - - wB'_ B'_' m 8"_ _ "\WI. // . u-.   . _8888 . _ _ .i.B8 . . 88. .':.':B B 
ANAND - KAsPAROV, PCA WORW CH., NEW YORK (9) 1995 187 27 :d5! lbxd5?? An inexplicable mistake. Black should just wait and make a useful move on the kingside, e.g. 27...h5. Then if White plays 28 c5, Black can reply 28...Wic6 or 28...:ec8. If White retreats the bishop from b6, then Black can safely take the ex- change since White needs the b6- bishop to support the advance of the pawns. White remains with the bet- ter position but has to find a way to break through. 28 exd5 White is clearly better and, more- over, his position is very easy to play: he just has to push his pawns. 28 ... Wig6 (D) .. ._ .  .... d _  wBiB Bi.i m B... n\WJ8- . U.. _ d .8B . . . . .J.B8 _ . 88. B:B B B<iit 29 c5 e4 30 i..e2 :e5 After 30...i..e7 31 d6 i..f6 32 d7 :f8 Black threatens . ..i..e5 with some counterplay, but 33 i..c7! kills any potential black activity. 31 Wid7! :g5 (D) -B B .._ wBiB_Bi_i it B .-B n P8.  u u . _ B BiB _ . . _ 88 8 BJ.B8B B:. B B<iit 31...Wig3 32 Wixb7 :g5 33 :gl is hopeless for Black. After the text-move I just had to calculate a little to be sure of victory. 32 :gl! e3 33 d6 :g3 34 Wixb7 The simplest, although 34 <it>h2 is also good. 34 ... Wie6 (D) -B B... wB-. .i.i _ D-. . "  . . UU/ _ B . . . . " 8 . d _ _ . BJ._88 _ _ B a<iit 35 <it>h2! At first I intended to continue 35 :fl, but then I saw Kasparov's trap: 
188 VISHY ANAND: My BEST GAMES OF CHESS 35..J%b8! 36 'iixb8 :xh3+ 37 <it>gl 'iie5 and, amazingly, Black forces a draw. I then saw that 35 :dl :b8 36 'iid5 would lead to a win, but the text-move is even easier, because 35...:e8 may be answered by the simple 36 d7. 1-0 This gave me my first win in a world championship match and my first win against Kasparov (in a tournament game) since Reggio Emilia 1991/2. I was ecstatic. Needless to say, the rest of the week was a cold shower. I think one of my main problems was that I had no idea how tense you could become in such a match. Although I had gained a lot of match experi- ence during the cycles, I had no idea what it was like to playa world champi- onship match; it is genuinely different to lesser matches. Looking back at the 11th and 13th games, I didn't need four seconds to tell me what I did wrong in these games - I just blundered. Kasparov's play was far from exemplary during the match, but he didn't make any real blunders. This shows that he was able to keep his nerves under much better control, which could, of course, be a function of his much greater world championship experience. If there is a next time, I believe I would be much better prepared to cope with the pressure of the match. I had noticed that a number of players had been badly affected by match defeats. Andrei Sokolov's loss to Karpov sent his career into a tailspin. Like- wise Hjartarson against Karpov. After my defeat by Kasparov, I gave some thought as to how to get my career back on track. The memory of the chess world can be very short-lived - you can become a nobody within a year if you don't back your reputation up by good results. I was still strongly motivated to stay at the top, but I understood that waiting for the next Candidates wouldn't be enough - I would have to keep making good results in the inter- vening period. After the match, my first event was the tournament at Wijk aan Zee, which was Heaven compared to the New York match. You could just go to Wijk aan Zee, everybody understands chess and is enthusiastic about chess; you can just play chess and need not be distracted by changes of venues and prize- funds and 101 other things. I felt happy in this event, despite the bitterly cold weather, and I played reasonably well. One of my targets was to regain my appetite for chess and to this end I decided to vary my openings. In some games I played 1 d4, and in others I adopted very sharp lines - all to keep my interest and motivation alive. The following game shows one of these open- ing experiments. 
Game 35 V. Anand - B. Gelfand Wijk aan lee 1996 Sicilian, Grand Prix Attack le4 c5 Gelfand is a very straightforward player who doesn't vary his open- ings much - with Black against 1 e4 he likes to play the Najdorf. I had noticed that he makes no effort to avoid the Grand Prix Attack and had shown some vulnerability against this line. For example, in his Candi- dates match against Short at Brus- sels 1991, he lost a game against the Grand Prix Attack and for the rest of the match abandoned the Sicilian. Later, however, Gelfand beat Sax when the Hungarian Grandmaster tried to repeat Short's success. In view of these games, I knew that I couldn't really catch Gelfand by sur- prise with the Grand Prix Attack, but I did have one new idea to try out... 2 ltJc3 d6 3 f4 g6 4 ltJf3 i.g7 5 .i.c4 ltJc6 6 d3 e6 7 0-0 ltJge7 8 1Vel! h6 Not 8...d5? 9 exd5 exd5 10 ltJxd5!, while after 8...0-0 9 f5!? exf5 10 1Vh4! White has an automatic king- side attack. 9 .i.b3 a6 (D) After 9...ltJd4 (9...0-0 10 _h4 is slightly better for White) 10 ltJxd4 cxd4 I1ltJe2 0-0 12 <it>h 1 f5 13 ltJg 1 <it>h8 14 ltJf3 .i.d7 15 .i.d2 :c8 16 'if g3 fxe4 17 dxe4 d5 18 exd5 exd5 White had the advantage in the game Topalov- Van Wely, Wijk aan Zee 1996. ., w£ .... MUi . .  // .. d.ILB I.   Ni., i . w ,;;;; w;&/ _  , , ,; .&   1% 1&\  . ' .& W . p/, .& . / A" A_A  all/'-''// '/uu !/,: Y:0//j ,/, // !/,: . . . . /f  u, u!/,: :/W;? ;';;;:://J, . ." . /; $; g /%/W A.P ..5iLQ_t.z..J. Yffi. ." Ui    : . ' '"   + / /" '      %/ % / u/ ",/ / y;; /, y;; 10 e5!? d h . . ., ... an t IS IS It. After 10 a4 :b8 II1Vg3ltJd4?! (11...b5!? may be better) 12ltJxd4 cxd4 13 ltJe2 b5 14 axb5 axb5 15 1Vf2! 1Vb6 16 f5 exf5 17 exf5 gxf5, which was played in Anand-Gel- fand, Reggio Emilia 1991/2, White could have gained the advantage 
190 VISHY ANAND: My BEST GAMES OF CHESS by continuing 18 lbf4, but clearly Gelfand was not going to repeat the whole variation. The idea of this type of pawn sac- rifice, which occurs relatively often in the Closed Sicilian and Grand Prix Attack, is simply to fight for the dark squares. 10 ... lbf5 If Black plays 10...d5, then 11 _f2 b6 12 i..d2lba5 (12...0-0?! 13 lbe2 i..b7 14 a4 is good for White; a5 is a threat as c5 is weak) 13lbe2 lbxb3 14 axb3 favours White, who will play b4 or d4 and exploit the ab- sence of Black's dark-squared bishop from the queenside. If 10...dxe5 11 fxe5lbxe5 (11...g5 12lbe4 g4 13lbd6+ f8 14lbg5! with a winning attack for White) 12 lbxe5 _d4+ 13 <it>h 1 (not 13 i..e3 _xe5 14 _f2? _xe3! 15 _xe3 i..d4) 13..._xe5, then 14lbe4 (14 Wif2 is also possible) 14...0-0 (14...f5 15 i..f4 _xb2 16 lbd6+ and 14..._c7 15 i..f4 e5 16 i..xf7+! are also very good for White) 15 i..xh6 f5 16 i..f4 (even stronger than 16 i..xg7 <it>xg7 17lbg5) 16..._xb2 17 :bl _d418 lbg5 and White wins. The move Gelfand played is the best. 11 <it>h1! 11 lbe4 would be premature, as Black could take on e5. 11 ... lbfd4 (D) After 11. ..0-0 White could play 12 lbe4, since if Black takes the pawn on e5 then White has g4 fol- lowed by i..xh6 at the end. Also 11. . .lbcd4 12 i..d2 (since the pres- sure on e5 has been lifted, White doesn't have to play lbe4 immedi- ately) and 11...d5 12 lbe2 (or 12 Wif2) would give White a slight ad- vantage. z...._.. _ w... .._ ..JIt\... .--. . . .   . . . U . .-p. . _ U _ .J._llJ. . . U . . U   :.<iit - - -  12 lbe4 I spent a long time thinking about 12lbxd4 cxd4 13lbe4 dxe5 14 Wig3 (14 fxe5 is met by 14...i..xe5 and not 14...lbxe5? 15 _g3! g5 16 i..f4!, winning) with the point that 14...0-0 allows 15 f5! exf5 16 i..xh6! (16 _xg6 <it>h8 17 i..xf7lbe7!) 16...lbe7! 17 i..g5! fxe4 18 _h4 with a clear advantage for White, e.g. 18...lbf5 19 :xf5 _b6 20 :f6!. However, I just couldn't find anything against 14... _e7! 15 fxe5 i..xe5 16 i..f4 i..xf4 17 :xf4 f5! (17...0-0 18 :f6! is too dangerous) 18 _xg6+ <it>d8 19 lbg3 (19lbc5 :e8!) 19..._g5!. Al- though Black needs to playa whole 
ANAND - GELFAND, WIJK AAN ZEE 1996 191 string of 'only' moves, I didn't doubt that Boris would find the correct path. In the end I settled for the more modest text-move, but this has the defect that Black can gain time against White's rook. 12 ... lbxf3 12.. .dxe5 13 lbxe5! is promising for White. 13 .:xt3 (D) Not 13lbxd6+? Wxd6 and Black wins. ..J...  . - . - ... .. B_ _ _ _ ........ .   . _ _ U _ . .ttJD . .i...:. . . U   U    . - - - .'*' 13 ... dxe5 14 fxe5 lbxe5 After 14...i.xe5 White can choose between two favourable lines: 15 Wf2 and 15lbxc5 i.xh2 16 .1xe6!. 15 .:n g5! Boris immediately found this forced move. 15...0-0 loses to 16 .1xh6! and now: 1) 16...lbxd3 17 We3 lbxb2 18 i.xg7 <i'xg7 19 Wxc5 f5 20 We5+ <i'h7 21lbg3 Wf6 22 Wxf6 :xf6 23 :abl trapping the knight. 2) 16....1xh6 17 lL\f6+ <i'g7 18 Wxe5 Wd4 19 Wg3! with a very dan- gerous attack. After other 15th moves, White just plays .1f4 and Wg3, with a very strong attack. 16 Wg3 16lbxc5 0-0 17 'ir'e4 We7 regains the pawn, but White's attack has gone. 16 ... 0-0 (D) Once again the correct move. In- stead, for example, 16...f5 (16...b6 17 .1f4! Wc7 18 :ael and 16...:g8 17lbxc5 also favour White) 17lbxc5 (17 .1f4 Wc7!) 17...We7 18 d4lbc6 19 c3 is good for White. ..J.  .  .  . - - , ...  . . w_ _  _ .. ...  - - - -  . - ' - '  , - " . BttJ. . .i... Hi . . U _ . u a m B:.<it> 17 .1xg5!? I didn't think much about this piece sacrifice, since it seemed to be the natural follow-up to White's ear- lier plan. I just checked that Black had no obvious defence, and then played it. While this may appear reckless, it fitted in with my ambition to play 
192 VISHY ANAND: My BEST GAMES OF CHESS interesting chess during the tourna- ment. With best play Black can survive, so it is probable that White should look for an improvement earlier. Still, Black has to walk a tightrope for several moves, no easy task in such a complex position. 17 lbxc5 is not dangerous for Black and he can equalize comforta- bly by 17...b6 18lbe4 ii.b7. 17 ... hxg5 18 lbxg5 (D) Threatening 19 'ii'h4. ..I.  . . .......- - - ... ..' B._,; .. ... .   /;j /. . %" . "// m /. 11  2  /j  _ %;jil z; / ' W /j///,// I I Y ::-' . . . .  -;f  )/;@ .. . IDi1 ...... - ". ." U   u a _ _:. 18 ... lbg6 Best. After 18...'ii'd4 (18...b5 19 :ae 1 ltJg6 20 ltJxf7! is good for White) 19 'ii'h3! :e8 20 :ael :e7 21 :e4 White's attack is dangerous, for example 21...'ii'xb2 (21...'ii'd6 22 :e3, intending :g3, also poses problems for Black) 22 d4! c4 23 i.a4 (so that Black doesn't get a pawn on b3, though even 23 dxe5 cxb3 24 'ii'h7+ <i'f8 25 :ef4! looks good) 23...'ii'xa2 24 dxe5 'ii'xa4 25 'ii'h7+ <i'f8 26 :ef4 and White wins. 19 .:ael! (D) Not 19 'ii'g4? 'ii'd4 20 'ii'h5 'ii'h4 and the attack collapses. K  . ?fi} .t. :  '// .  'U .   < _   gffi ;;:; , /.I,   & ?:{/p f & _ B ...P ... .. .... . . ;; f%  . . . . .i._. . ". ."   . U . &0%  rlh . . :.'8' I had seen up to here when I sacri- ficed the piece and thought that Black would have a tough job de- fending the position, but in fact he can hold on. 19 ... 'ii'e7 A critical moment. The alterna- tives are: 1) 19...c4? 20 i.xc4 b5 21 i.b3 is pointless as 21...i.b7? loses to 22 :xe6! . 2) 19...i.h6? 20 ltJxf7 :xf7 21 'ii'xg6+ wins. 3) 19.. .i.xb2 (Black can just about hang on after this move) and now: 3a) 20 ltJxe6 i.xe6 21 i.xe6 <i'g7! defends. 3b) 20 i.xe6!? i.e5 (20...fxe6 21 lbxe6 :xfl + 22 :xfl i.xe6 23 'ii'xg6+ and White may have no 
ANAND - GELFAND, WIJK AAN ZEE 1996 193 more than perpetual check) 21 :xe5 fxe6 22 :ee 1 q;g7 and Black is slightly better. 3c) 20 ltJxf7! :xf7 21 'ii'xg6+ :g7 22 i.xe6+ i.xe6 23 'ii'xe6+ q;h8 24 :e3 :h7 (24...'ii'g810ses af- ter 25 :h3+ :h7 26 1Ixh7+ 'ii'xh7 27 :f3) 25 :f7 (D) leading to a final branch: .B  .  w{:? _ ;2::ffi _ BB'. .:.. .. _if. . #;:  :ffi. . /' ;:0   r   . , . 1Y $; :0  ,,>/ ?;;/; .'///c;/ (?' ///-j,// /////// 0:///, //////// 'i'/ (( 8 a  ", %  '/j /?;'/' 1;:W;  ;  ; //-j:,; $;/  1 ;/ ////.., / ////// /.. /  8  . / ' 8 - }{ 8 rJ %,:r     ' /  '////// /, // . . . .rlh //: }%& w %j 'tiiiI 3cl) 25...:h4 26 :h3 'ii'g5 27 :xh4+ 'ii'xh4 28 :f3 wins. 3c2) 25. ..i.g7 26 :xg7! :xg7 27 :h3+ :h7 28 'ii'e5+ q;g8 29 :g3+ q;f8 30 'ii'f4+ :f7 31 'ii'h6+ q;e8 (31...q;e7 32 :e3+ q;d7 33 'ii'e6+) 32 :g8+ q;e7 33 'ii'e3+ q;d7 34 :xd8+ with a winning ending. 3c3) 25...:h5! 26 :xb7 (26 :h3 jig5 27 :xh5+ 'ii'xh5 28 :f3 :e8 wins for Black) 26...i.g7 is unclear. White has three pawns for the piece but Black has enough pieces in play to defend his king. 4) 19...i.f6!? (D) (this may also enable Black to draw) and now: EB.*.. ... w.'. B'. a_ . .a_1&\_ .-. .-.--.     £J . / :f$i;  '// :// %  /";;//;/. , //;//. // // !/,: ;::;;'1; // . . . . ..8. . 808. .80 . . a:B<t> 4a) 20 ltJxf7? q;xf7 is unsound. 4b) 20 ltJe4? i.h4 solves all Black's problems. 4c) 20 :f5?! q;g7 repulses White, but not 20...exf5? 21ltJxf7 :xf7 22 'ii'xg6+ q;h8 23 i.xf7 'ii'f8 24 :e8 'ii'xe8 25 'ii'h6#. 4d) 20 1he6!? i.xe6 (20...i.xg5? 21 1Ixg6+ wins) 21 ltJxe6 'ii'e 7 ! (not 21. ..fxe6? 22 'ii'xg6+, nor 21... 'iib8?! 22 ltJxf8 'ii'xg3 23 hxg3 with excel- lent winning chances for White) and now: 4dl) 22 ltJf4?! q;h7 (22...q;g7? 23 ltJh5+ wins for White) and Black has some advantage after 23 ltJd5 i.h4 or 23 'ii'h3+ i.h4. 4d2) 22ltJxf8 q;xf8 23 'ii'f3 q;g7 24 i.d5lbe5 25 'ii'e4 is unclear. White will gain three pawns for the knight, but Black's king is not in danger. 4d3) 22ltJc7 (with the threat 23 'ii'xg6+) 22...q;g7 23 ltJd5 'ii'e2! (23...'ii'e5 24ltJxf6 'ii'xg3 25 ltJh5+ wins while 23...i.h4 24ltJxe7 i.xg3 25 ltJf5+ q;h7 26 ltJxg3 is a very 
194 VISHY ANAND: My BEST GAMES OF CHESS favourable endgame for White) 24 :e 1 J.h4 25 :xe2 J.xg3 26 hxg3 and in view of White's broken king- side pawns, Black has no problems. 4e) 20 lbxe6 fxe6 and now: 4el) 21 he6 g7! (certainly not 21...i.xe6 22 Wxg6+ h8 23 Wh6+ g8 24 .ixe6+ winning for White) 22 :d6 We7 23 h4 We5 and Black wins. 4e2) 21 Wxg6+ .ig7 22 .ixe6+ J.xe6 23 Wxe6+ h8 and White has a perpetual check although possibly not more. 20 :IS!! (D) a.i.. ... ... .- B..__ .. ..... . . 0  .: . .  . . . - _.t.. _ ". . U .  U . . a R<it Defending the knight on g5 so that White can play Wh3. 20 ... .if6 The main alternative is 20....ih6 (20...exf5? loses to 21 :xe7 lbxe7 22 Wh4 :d8 23 Wh5) and now: 1) 21lbxe6?! i.xe6 22 .ixe6 (22 :xe6 fxe6 23 Wxg6+ Wg7 defends) 22...fxe6 23 Wxg6+ Wg7 24 Wxe6+ h7 is unclear. 2) 21lbxf7?! and now: 2a) 21...g7 22lbe5! and Black is helpless: 2al) 22...exf5 23 Wxg6+ h8 24 Wixh6+ mates. 2a2) 22...:xf5?! 23 Wxg6+ wins after 23...h8 24 Wxf5 or 23...f8 24 Wxf5+! exf5 25lbg6+. 2a3) 22...:f6 23 :f3! (threaten- ing 24 :efl) wins. 2b) 21.. .:xf7 22 Wxg6+ .1g7 (or 22...:g7 23 Wxh6) 23 :xe6 i.xe6 24 .ixe6 :af8 25 :g5! (threatening mate in one) 25...h8 (25...Wf6 26 Wxf6 - isn't that a nice pair of pins?) 26 :h5+ .ih6 27 :xh6+ ':h7 28 gl Wg7 29 ':xh7+ Wxh7 30 Wg5 should be winning for White. 2c) 21...h7! (Gelfand's sugges- tion) is unclear after 22lbxh6 xh6 or 22 lbg5+ i.xg5 23 :xg5 Wf7 24 h4 b6. 3) 21 h4 (D) and Black seems to have no adequate defence against the threat of h5: a R"'_  .. ..mu. BE.. ... .B B.... .  .: d _ d  . . . n _ _ _ U ..illU ..t.. _ BB ..  . " .Ah d 0  '&I 
ANAND - GELFAND, WIJK AAN ZEE 1996 195 3a) 21...q;g7 22 h5 i.xg5 23 Jhg5 :'h8 (23...Wf6 24 hxg6 :'h8+ 25 q;gl wins a pawn with a good posi- tion) 24 :'xg6+ q;f8 (24...fxg6? 25 'ii'xg6+ wins) 25 ':g5 and again White has a clear pawn more. 3b) 21...q;h8 22 h5 exf5 (the line 22...i.xg5 23 ':xg5 is very good for White) 23 :'xe7 lbxe7 24 lbxf7+ Jhf7 25 i.xf7 i.g7 26 Wg5 and White wins. 3c) 21...i.xg5 22 :'xg5 q;g7 23 h5 is line 3a. 3d) 21...c4!? 22 i.xc4 b5 23 i.b3 (D) (23 i.d5 ':a7 24 h5 'ii'c7! is less clear-cut) and now: ...t. ... .illU B   _' .. ....- ... .: . 8 . D ..I\.  .....0. _ 8D88 .88 . . a 8 3dl) 23...q;h8 24 h5 exf5 25 Jhe7 lbxe7 26lbxf7+ :'xf7 27 i.xf7 and White wins. 3d2) 23...i.b7 24 h5 Wb4 (White also wins after 24...exf5 25 :'xe7 lbxe7 26 lbxf7+ q;h7 27lbxh6) 25 hxg6 exf5 26 i.xf7+ ':xf7 27 gxf7+ q;f8 (or 27...q;g7 28 :'e8) 28 :'e6 is winning for White. 3d3) 23...Wb4 24lbxf7! i.g7 25 :'e4 We7 26lbd6 Jhf5 27lbxf5 and again White wins. 21 lbxe6 (D) ...t. ... B.'. _.. .. .CD_a. . - .:. . - . - .i..8m _ 88. .8P u   U . . a . 21 ... fxe6?? Up to here Black has found the correct defence time after time, but now he goes down without a fight. 21...:'e8! was the right defence: 1) 22 :'efl was my original in- tention. During the post-mortem we agreed that after 22...i.xe6 23 i.xe6 Wxe6 24 :'xf6 White has enough compensation. Later Gelfand sent me an e-mail pointing out that 22...i.h4! is very good for Black. 2) 22 Wf3 i.h4 (22...lbh4 23 Wg4+ lbg6 24 Wf3 is a likely draw, but not 22...i.e5? 23 :'xf7! Wh424 :'xe5 i.xe6 25 :'xe6 :'xe6 26 Wfl :'ae8 27 :'f6 and White wins) 23 :'xf7 Wxf7 24 Wxf7+ q;xf7 25 ltJc7+ i.e6 26 :'xe6 Jhe6 27 i.xe6+ q;e7 28lbxa8 q;xe6 29lbb6 and the ending slightly favours Black. 
196 VISHY ANAND: My BEST GAMES OF CHESS 3) 22 J:tffl i.xe6 (22...fxe6 23 'ii'xg6+ i.g7 24 :f6! 'ii'xf6 25 'ii'xe8+ 1Wf8 26 i.xe6+ i.xe6 27 1Wxe6+ 1Wf7 28 1Wb6 is unclear) 23 i.xe6 fxe6 24 1Wxg6+ i.g7 25 J:tf3 J:tf8 with advantage for Black. 4) 22 J:te4! i.xe6 23 i.xe6 fxe6 24 1Wxg6+ i.g7 and after 25 J:tf3 or 25 J:th5 1Wf6 26 1Wxf6 i.xf6 27 J:txc5 White still should be able to make a draw owing to the reduced material. However, whatever winning chances there are lie with Black. 22 .:xe6! (D) A deadly blow. Not 22 i.xe6+? i.xe6 23 ':xe6 1W g7! and Black is better, nor 22 1Wxg6+? 1Wg7 (how- ever, 22...i.g7? 23 J:txe6 wins for White) 23 1Wh5 leading only to a draw. 22 ... 1i;g7 Or 22...i.xe6 23 1Wxg6+ 1Wg7 (if 23...i.g7, then 24 i.xe6+) 24 i.xe6+ J:tf7 (24...<iti>h8 25 J:th5+ mates) 25 ...t.. ... B BiB _ B iB .:.. . ,W _ _ .  .: ... . . . B W}j  %/ i. /@, 0 /  ' '//  + fJj% A/ ///j , /  _0_ , ". ./  d  /m   i70  _ (;;; // '/////'// /;::0); 'itiiI i.xf7+ f8 26 'ii'xf6 and White . WIns. 23 .:xe7 + i.xe7 24 .:xf8 i.xf8 White's large excess of pawns would be enough to win in any case, but in fact Black doesn't even get a chance to bring his queenside pieces into play. 25 h4! 1-0 In view of 25...1i;h7 26 h5ltJe7 27 1Wf3 i.f5 28 'ii'xb7. The Wijk aan Zee tournament went well in general, although the failed brilliancy against Sokolov was an unnecessary loss. At first I thought that my loss to Topalov was also unnecessary, but it turned out later that he had seen much more than I had. In the last round I beat Tiviakov in a game lasting 107 moves, gaining revenge for a loss I had suffered against him in 1989. In fact I have only lost once to him at a normal time-limit, but he has written about it so often that most people think I have a huge negative score against him! My score of 8/13 was sufficient for second place behind Ivanchuk. In April I participated in the annual Amber tournament in Monaco, which consists of a mixture of blindfold and rapid games. At the start I played a number of difficult opponents, but in these early rounds managed 1 1 /2- 1 12 against Karpov and Lautier and 2-0 against Nikolic. However, Kramnik had 
5TH AMBER RAPID, MONTE CARLO 1996 197 raced into the lead and I was never able to catch up with him. Half-way through I had a bad patch with six draws and two losses from eight games (I was even a bit lucky to achieve this meagre total). Finally I came to the end of the bad patch, beat Xie Jun 2-0 and then met Judit in the following round. In the blindfold game she had the advantage but then played some strange moves and after a hard fight I won. Then we met in the rapid game. 
Game 36 V. Anand - J. Polgar 5th Amber Rapid, Monte Carlo 1996 Pirc Defence 1 e4 My win in the earlier blindfold game had been tiring - after getting a position where she could barely move a piece, I allowed myself to be swindled and had to win the game several times. Surely the rapid game (with my eyes open!) would be more relaxing? 1 ... g6 I didn't expect ...g6 and was quite surprised that she played it. 2 d4 i.g7 3 lbc3 d6 4 i.e3 c6 5 'iid2 b5 6 f4 I couldn't really remember what to play here, but in a rapid game you shouldn't worry too much about mi- nor details. By now the exertions of the first game had faded away, at the prospect of a good hackfest! Some- how, I couldn't bring myself to be solid, especially after three wins in a row. 6 ... lbf6 7 i.d3 Not 7 e5? b4!. 7 ... 8 lbf3 e5 exd4 9 .i.xd4 10 0-0 At first I was quite happy with my position as all White's pieces are de- veloped to reasonable squares. How- ever, after her next two moves I real- ized that my opening had not been very successful. 10 ... 11 lbe2 12 hl 13 i.gl 14 lbg3 15 b3 16 .:ael 0-0 b4 lbbd7 c5 i.b7 'iic7 .:ae8 :'e7 (D) . . ... '  _ " .,/, ... .  & _  0 & W  I'/"' I  .  '. % / . ' "' '.'.' / j', -- - :/ . - '/ . . ,  U  _..lb . .  _ _ U _ . a:m<itt Now 17 :'e2 ':fe8 18 ':felleads to a horribly passive position, and 18.. .lbg4 intending 19.. .i.c3 would 
ANAND - J. POLGAR, MONTE CARLO RAPID 1996 199 be very unpleasant. Therefore, I de- cided to throw caution to the winds and swing my queen over. 17 f5 :re8 18 _f4 lbe5 (D) 18...h6!? is also possible: after 19 fxg6 fxg6 20 :e2! (direct attempts like 20 i.c4+ don't work) the posi- tion is quite messy. Understandably, Judit tries some- thing else. _ B.B.8  _ /. //  - '''/. - ' &  % & W .../. /...    ,.,,, . / -.. .  M¥    8. . - - .  8 . - . -  888ii._lD 8888 _8U 8 8 a:m 19 _h4 20 i.e3 21 gxf3 22 i.f4 To meet ...d5 with e5. Neverthe- less, Black must play.. .d5 as other- wise White would have time for ':e2-g2. 22 ... 23 e5 24 fxg6 (D) Not 24 bxc4? dxc4 25 fxg6 hxg6! (by including the exchange on c4, White has forced Black to recapture _d8 lbxf3 'iib6 d5 c4 with ...hxg6, since 25...fxg6? loses to 26 i.e3 !, but he has outfoxed him- self, since that mighty b7-bishop is now staring at f3!) 26 lbf5 gxf5 27 i.xf5 lbd7! (now this works since Black can meet :gl with ...i.xf3+) and I suspect that Black is already . . wlnmng. 24 exf6? is also bad as Black can reply 24...:xel 25 fxg6 (25 fxg7 cxd3) 25.. .fxg6 !. 8 8.B.8 //.t.. . . B% . _ _ illi1 _ _1\_ _ . M¥O. 8 8'U B ..   - . - - 88_ii.B8 8888 . B B . a:8 24 ... hxg6 (D) 24...fxg6! is better: 1) 25 bxc4 lbd7 is clearly good for Black. 2) 25 exf6 i.xf6 26 :xe7 i.xe7! (26...:xe7? 27 i.g5) is slightly bet- ter for Black. 3) 25 lbf5!. We thought in the post-mortem that Black was better after 24...fxg6, but later I found this move, which, while not as promising as in the game, still offers White good chances. Besides, White doesn't have much of a choice at this point. I am 
200 VISHY ANAND: My BEST GAMES OF CHESS not sure how to assess the position aside from the usual cop-out of un- clear! One possible continuation is 25. ..gxf5 26 i.xf5! h6 (26.. .i.c8 27 exf6!) 27 :gl with messy complica- tions. . .Z... .t.. i W.h . _ _  . i. - . - . _ .iD . / &  .  %  %; . /. /  ..    % /   ",- ,, B.i..  /1w  'h / ..  ,, _ . n:R<iit 25 ltJf5! Now this is extremely strong and I think White is already winning. 25 ... gxf5 Or 25. ..cxd3 26 ltJxg7! and wins. 26 i.xf5 d4 26...d7 27 iih7+ f8 28 :gl is decisive. 27 :e2 d3 27...ltJd5 28 iih7+ cjj>f8 29 J:g2 f6 30 i.h6 and White mates in a few moves. 28 J:g2 :xe5 29 Jbg7+ Better than the unclear continua- tion 29 'iih6ltJg4 30 'iih7+ f8 31 :xg4 'ii'f6 32 i.xe5 (32 :Xg7 'iixg7! 33 i.xe5 :xe5 is good for BIck) 32...:xe5. 29 ... 8 Or 29...xg7 30 i.h6+! h8 31 e3+ g8 32 'ii'g5+ f8 33 i.xb6 picking up the queen. 30 'iih6 :e2(D) . ../0; . ji; /;ffi;j ...   M /      /!;2    & n W   ...  /;;;;I . /.  /u/U '/////,//, }f/ /.i , ,    ,  _ illU _ ,>f _ _ . . i.. - :  /i. . . /j"  . ' , ",- . ..i.. AA... " o !&% 0 /'/;//; ... _ U . ff ". ::;::::,;;% '/,///0 / //;/ ....  '8' 31 bxc4!! I am very proud of this move, which was the main reason why I se- lected this game. Basically, I quickly rejected 31 :g6+ followed by 32 :xf6 because of 32....txf3+. Judit had gone much further in this line, but I couldn't be bothered - I wanted something cleaner. Black's pieces are perfectly placed; indeed the only piece which can be better placed is the e8-rook (aside from the black king, of course!). I noticed :g2+ fol- lowed by :xe2 and I also saw bxc4 in connection with the move c5 dis- turbing the black queen. Suddenly I realized that the d3-pawn could not move! Bingo! After 31 :g6+ e7 32 :Xf6 Black may play: 
ANAND - J. POLGAR, MONTE CARLO RAPID 1996 201 1) 32....i.xf3+ (I had only seen this) 33 :xf3 :el+ 34 <it>g2 :gl+ 35 <it>h3 'ii'xf6. I stopped here - I was convinced I was winning and didn't want to waste my time making this work. 2) 32...:g8! (full marks to Judit for noticing this move!) and now 1 am indebted to Fritz4 for the reply 33 i.d6+! 'ii'xd6 34 i.xd3! (the only way - White needed to jettison his f4-bishop in order to cover f3 and to lure the enemy queen to d6, from where it no longer eyes the gl- square) 34...'ii'xf6 35 'ii'xf6+ <it>xf6 36 i.xe2 c3! and Black is certainly not worse here. White can also try 31 cxd3, with the same idea as in the game, but then Black can limp on with 31...<it>e7. 31 ... d2? After 31... <it>e 7 32 i.xd3! :f2 Whi te has: 1) 33 :xf2'ii'xf2 34 :xf7+ <it>xf7 35'ii'g6+ (35 i.g6+ <it>g8!) 35...<it>e7 (35...<it>e6 36 i.f5+ <it>e7 37 'ii'g7+ <it>d8 38 i.c7#) 36 'ii'g7+ <it>e6. 2) 33 :xf7+! xf7 34 'ii'g6+ <it>e7 (34...<it>e6 35 :xf2 is winning for White) 35'ii'g7+ and now either 35...<it>d8 36 :xf2 :el + (36...:g8 37 .i.c7+! 'ii'xc7 38 'ii'xf6+ is deci- sive) 37 <it>g2, or 35...<it>e6 36 :el + ltJe4 (36...i.e4 37 i.xe4 is winning) 37 i.xe4, winning for White in both cases. 32 :g2+ She had missed this one. Now 1 win the house. 1-0 Topalov was easily the most successful tournament player in 1996, finish- ing either first or joint first in Amsterdam, Novgorod, Leon, Madrid, Vienna and Dos Hermanas. I had already lost two games against him in 1996, so I was looking forward to stopping the sequence. 
Game 37 V. Anand - V. Topalov Dortmund 1996 Sicilian, Scheveningen 1 e4 c5 2 lL\f3 e6 3 lL\c3 a6 A slightly unusual move-order, but both of us seemed to have de- cided that the game was going to be a i.e2 Scheveningen regardless of the move-order! The comment in Game 25 about this being an inflexi- ble move-order for White only ap- plies if White is going to play the aggressive f4 and 'ii'f3 system. You can play the solid i.e2 line against almost anything. 4 d4 cxd4 5 lL\xd4 lL\c6 6 i.e2 d6 7 0-0 lL\f6 8 i.e3 i.e7 9 f4 0-0 10 a4 'ii'c7 11 <it>hl :e8 A position on which I had done a lot of work for the Kasparov match. Although I wasn't so successful in the match itself, the analysis paid off over the succeeding months. At the time this game was played, I was still far ahead of other grandmasters in my understanding of this line. Later on they caught up with me, but not before I had notched up several . WIns. 12 i.f3 fiJaS (D) IB.lBIB.B .. .-. w. _ _ _ .. .- . - - - - . . - . . mD B B m m.tB ". ." U _ _ u a ."B:B<it 13 g4! Kasparov played 13 i.gl against Topalov a month earlier at Dos Her- manas, but I decided to be less sub- tle. That game continued 13.. .i.f8 14 'ii'e 1 :b8 15 h3 fiJd7 16 i.h2 fiJc6 17 :d 1 lL\xd4 18 :xd4 b5 19 axb5 axb5 20 e5 with an edge for White. 13 ... fiJd7 13... fiJc4 14 i.c 1 e5 15 fiJf5 exf4 16 g5 is good for White. 14 i.g2 This move reflects one of my discoveries: that when Black plays ...ltJd7 voluntarily, it may not be 
ANAND - TOPALOV, DORTMUND 1996 203 necessary to play g5. Of course, you may want to play g5 in the end for attacking purposes, but White can time it much better. One point be- hind leaving the pawn on g4 is that an early g5 can be met by ...g6 fol- lowed by .. .e5. If the pawn is still on g4 it is much easier for White to meet this manoeuvre by the piece sacrifice f5, opening the g-file af- ter ...gxf5 gxf5. 14 ... i.f8 Black has a range of possible plans; for example he could continue 14...b6, but again White plays 'ii'el, :d 1 and any other useful moves he can find before pushing the g-pawn. 15 'ii'el b6 After 15...c4,. White plays 16 i.c 1 intending b3 and i.b2. 16 :dl i.b7 (D) .- .. . . - . - . I&\.ii w ........ _ i i. . - - . - - . . - 8- 8"8. .  U . .   . .  - . "8. .i. u . . U . .:.:. 17 'Wh4 White's plan is to move the e3- bishop out of the way (possibly to gl) and then to play %ld3-h3. After Black defends the h7-square by ...g6, ...i.g7 and ...f8 White again makes use of the position of the pawn on g4 by playing f5, since then .. .exf5 can be met by gxf5. 17 ... c6 After 17...c418i.cl g619%1d3 White proceeds with his kingside at- tack. 18 de2! White must make time for this move as 18 %lf3 lets Black free his position by 18...xd4 19 i.xd4 e5 20 %lh3 h6. 18 ... b4 19 %ld2 (D) .. ..-.. .J...ii B.... i; i. . ; - . . . . . . _ .  illi1 8. .8u8. -   . .  -  "8l2J.i. u _ . U . - .:. 19 ... 'ii'd8 If 19...d5, then 20 e5 f6 21 exf6 xf6 22 i.d4 gives White a posi- tional advantage. The text-move aims to provoke g5. This looks odd unless you have read the earlier dis- cussion about the merits of with- holding g5 ! 20 g5 f6 (D) 
204 VISHY ANAND: My BEST GAMES OF CHESS After 20...g6 21 :f3 :c8 22 J:h3 h5 23 ltJg3 a deadly sacrifice on h5 is looming, while 20.. .h6!? 21 'ii'f2 (21 ltJd4!? hxg5 22 fxg5 ltJe5 23 :df2 is also possible) 21...hxg5 22 fxg5 ltJe5 23 g6! fxg6 24 i.xb6 fa- vours White. If Black continues quietly by 20.. .:c8, White plays 21 i.g 1 fol- lowed by :f3-h3. .. -.-.. ..t... .. w _ _ _ /uu, . ... .  /u", /u"' . . . . 0 8- .8"  _ . U _ .   . .  - .  ...tn u8t.z..J. u . . .:. 21 ltJd4! Returning to exert pressure on the new weakness at e6. 21 ... fxg5 After 21.. .ltJc6 22 gxf6! Black runs into problems: 1) 22... 'ii'xf6 23 'ii'xf6ltJxf6 24 e5 ltJg4 25 i.gl ltJa5 26 ltJb3 i.xg2+ 27 :xg2 lL\h6 28 ltJe4 with a clear endgame advantage. 2) 22...ltJxf6 23 e5 (23 ltJxc6 i.xc6 24 e5 i.xg2+ 25 :xg2ltJd7 is only equal) 23...ltJd5 24 'ii'xd8ltJxd8 25 ltJxd5 i.xd5 (25...exd5 26 e6! ltJxe6 27 ltJxe6 :xe6 28 i.xb6 is very pleasant for White) 26 i.xd5 exd5 and now both 27ltJf3!? and 27 e6 ltJxe6 28 ltJxe6 :xe6 29 i.xb6 give White some advantage. 22 fxg5 ltJc6 23 :df2! :c8 After 23...ltJxd4 24 i.xd4 ltJe5 Black has finally occupied the e5- square, but it's a bit late for this to matter: White continues 25 i.xe5 dxe5 26 :17 with an excellent posi- tion. 24 ltJce2! In order to make sure that White can maintain a knight on d4. 24 ... ltJc5 After 24.. .ltJde5? White can choose between the quiet 25 b3 and the sharp 25 ltJxe6 :xe6 26 :xf8+ 'ii'xf8 27 :xf8+ :xf8 28 ltJf4 :ee8 29 ltJd5, with some advantage for White in either case. 25 ltJxc6 i.xc6 26 ltJd4 (D) .. . . . - -   /!% ; W &  %//, /?i W . B. _ % /mu . f.t.. . . . , //,," . @f .  . '// .  . /' W&/  A. A  0_ o. _ .    %: / :J,>' /::   /'///// /, //,;, /' h//:/,/ 8ri#!%  "  :'///;/;  ,JiL  . % f;: : _ rlh 'W:fJ£ ;,;,,;/:;: )//;j(  '8' 26 ... i.d7 
ANAND - TOPALOV, DORTMUND 1996 I don't think either of us looked at 26...i.xa4 for more than one sec- ond. In such a position you just know that Black cannot afford to go pawn-grabbing. White would con- tinue 27 1:tf3 (threatening 28 1:th3; 27 b4 e5 is less clear) and now: 1) 27...e5 28 1:th3 h6 29 'ii'h5! and White's attack is very strong. 2) 27...1:tc7 28 b4! and now: 2a) 28...ltJb7 29ltJxe6! 1:txe6 30 g6 and White is winning. 2b) 28...ltJd7 29 i.h3 ltJe5 30 g6!! ltJxg6 (30... 'ii' xh4 31 i.xe6+) 31 'ii'xd8 1:txd8 32 ltJxe6 and wins. 2c) 28...e5 29 bxc5 exd4 30 cxb6 is the critical line. White has the ad- vantage but the game is far from over. 27 e5! The point of Black's previous move is to set up latent threats along the c8-h3 diagonal; for example 27 1:tf3 e5 28 ltJf5 g6 29 1:th3 h5! re- futes the attack. The text-move ef- fectively counters this plan. 27 ... dxe5 After 27...d5 28 1:tf3, followed by 1:th3, White has an immense attack. 28 ltJf3 i.c6 After 28..:ii'c7 (28...i.d6 29 1:td2 'ii'c7 30 1:tfdl and wins) White con- tinues his attack by 29 g6 h6 30 ltJg5. It looks slow, but White will gradually use his greater firepower on the kingside: 30...i.e7 (30...i.c6 31 ltJf7) 31 1:tf7! i.c6 32 i.xc6 'ii'xc6+ 33 <it>gl i.f8 34 ltJh7 'ii'xa4 205 35 b4! 'ii'a3 (35... 'ii'xc2 36 ltJxf8 wins) 36 i.xh6! and the lbng-awaited sacrifice on h6 finishes Black. 29 ltJxe5 .txg2+ 30 Jbg2 (D) . . / Ii /." '//. ·  . '  .  . ' % , ... /. %...   - »/;// %"  " ,; @ " . . . B_._ . ... . - - - -  . . "//.  / .     /  AB . . Wi1 o. . . _ . .  " . . - . 0 / / A     I /. /  O- :.,.  /, / !/,: w// // //;//; /// , !/,: . . .... . gffj ; .Ii6. . 'i!i¥ The threat is 31 ltJg4 followed by 32ltJf6+. 30 ... 1:tc7 White's preponderance on the kingside is simply too great: 1) 30...i.e7 31ltJg4! and nbw: la) 31...<it>h8 32 ltJf6 i.xf6 (if 32...gxf6, then 33 g6) 33 gxf6 gxf6 (33...g6 34 i.g5) 34 1:txf6 and the white attack is decisive. 1 b) 31.. . 'ii'd5 32 ltJf6+ i.xf6 33 gxf6 %lc7 34 a5! and White is win- ning, for example 34...bxa5 35 f7+ 1:txf7 36 1:txf7 <it>xf7 37 'ii'f2+ picks up a pIece. 2) 30...'ifd5 31 g6 h6 32 i.xh6! gxh6 (32...'ii'xe5 33 i.c 1 mates) 33 ltJg4 ltJd7 34 g7 'ii'xg2+ 35 <it>xg2 %lxc2+ 36 <it>h 1 i.xg7 37 ltJxh6+ wIns. 
206 VISHY ANAND: My BEST GAMES OF CHESS 31 lL\g4 h8 White also wins after 31.. .e5 32 lbf6+ gxf6 33 gxf6+ h8 34 'ii'g3 i.d6 35 i.h6 or 31...'ii'd5 32lbf6+ gxf6 33 gxf6+ i.g7 34 i.h6. 32 'ii'h3! Threatening 33 g6 followed by a sacrifice on h6. 32 ... 'ii'd5 The continuation 32...g6 33 i.f4 i.d6 34 'ii'c3+ :g7 35 :d 1 lbe4 36 'ii'c6 is decisive. 33 g6 h6 (D) . .. - d d _ _ .  .  wg _ g _ i .i.8 - g g - . _... B 8. . Bl2J. - . - ... P8. .: u . d u . . .:. 34 i.xh6! White has more than one good continuation, but this is the most convincing. After 34lbxh6 gxh6 35 i.xh6! (not 35 g7+? i.xg7 36 i.xh6 g8, but 35 :xf8+ :xf8 36 g7+ g8 37 gxf8'ii'++ xf8 38 i.xh6+ e8 39 'ii'g4 is also very good for White) 35...i.xh6 36 'ii'xh6+ g8 37 g7 'ii'xg2+ 38 xg2 :xg7+ 39 h 1 lbxa4 White should win, but the text-move is instantly deadly. 34 ... gxh6 35 g7+! 35lbxh6 (35 :xf8+?:Xf8 36 g7+ xg7 37 'ii'xh6+ 17) 35...i.xh6 36 'ii'xh6+ g8 transposes to the last note. The order of moves in the game rules out the possibility of ...'ii'xg2+. 35 ... i.xg7 Or 35...:xg7 36 :xf8+ :xf8 37 'ii'xh6+ g8 38 lbf6+ 17 39 'ii'xg7#. 36 lbxh6 Threatening 37 lbf7++ g8 38 'ii'h8#. 36 ... 'ii'xg2+ 37 'ii'xg2 i.xh6 Or 37...h7 38 'ii'h3 i.xh6 39 :f6 g8 40 :xh6 with further ma- terial gains to follow. 38 'ii'g6 1-0 38...:g8 39 'ii'xh6+ :h7 40 'it'f6+ :hg7 41 'ii'h4+ :h7 42 'ii'd4+ :hg7 43 :17 wins at least another piece. This was almost a model game, but it is not easy to appreciate unless you are familiar with all the intrica- cies of the Scheveningen. Towards the end of the year I competed in a very strong double-round event in Las Palmas. The other players were Kasparov, Karpov, Topalov, Kramnik and Ivanchuk. I started with two draws, but in the third round my tournament 'came alive with the following game. 
Game 38 V. Anand - V. Ivanchuk Las Palmas 1996 Ruy Lopez This game was played on my birth- day. The tournament had begun slowly, with only one decisive result in the first two rounds and there had been a lot of whining amongst the public and press about it. This seems a bit unjustified, given the fighting spirit that prevails these days, but perhaps chess fans are right to worry that chess will revert back to the 17- move draws prevalent in the 1980s. 1 e4 e5 If anyone doesn't know already, I vanchuk plays everything. In fact I expected this, but of course couldn't be sure! 2 f3 c6 3 i.b5 a6 4 i.a4 f6 5 0-0 i.c5 Ivanchuk had already played this move twice during 1996. 5...b5 6 i.b3 i.c5 was all the rage in 1995, but by this time it had been super- seded by the text-move. 6 xe5 6 c3 b5 7 d4 bxa4 8 dxc5 xe4 9 ltJxe5 xe5 10 'iWd5 i.b7 was played in Short-Ivanchuk, Novgorod 1996, which ended in a draw. Somehow I didn't feel like repeating this line. 6 ... xe5 7 d4 xe4 8 :tel i.e7 9 .:xe4 g6 10 c4 0-0 11 c3 d6 If 11...c6, intending 12...b5, then 12 d5 is slightly better for White. 12 d5 i.h4 For some reason, this had escaped my attention; the point is that 13 g3 is met by 13...c6 (this actually hap- pened in Smirin- Izeta, Las Palmas 1993). After the text-move I thought for a long time, because if Black drives the knight back with ...c6 then he has solved all his opening prob- lems. Then I saw the exchange sacri- fice, but I consumed a lot of time before taking the plunge. I didn't want to concede equality too easily on my birthday, but the possibility of just being an exchange down was so- bering. Finally I decided that Black wouldn't be able to break White's bind. 13 'WIts! An innovation. 13 ... c6 (D) 13.. .i.e6 !? is an al temati ve since 14 f4? i.xc4 15 xg6 fails to 
208 VISHY ANAND: My BEST GAMES OF CHESS 15...i.xf2+!. Therefore White is bet- ter advised to play 14 i.c2. .B.t._ _.B wB'B B'_' ... .Jk\. -  .... B BtiJB .. i.-8":.  . u d _ B B _ B A D /  .   D /. / A / O  .  O  /  I: /",  /    . «; '/ '"" . , 14 Jbh4! Consistent. Otherwise Black is just better. 14 ... 'ii'xh4 Not 14...ltJxh4? 15 i.g5 with the lines: 1) 15...'fIa5 16 ltJe7+ h8 17 i.c2 h6 (17.. .ltJg6 18 i.xg6 fxg6 19 ltJxg6+ and 17...ltJf5 18ltJxf5 i.xf5 19 i.f6! are also hopeless for Black) 18 'fIxh6+! gxh6 19 i.f6#. 2) 15...f6 16 i.xh4! cxd5 17 'fIxd5+ h8 18 i.g3 with excellent compensation for the exchange. 15 'ii'xh4 ltJxh4 16 ltJb6 :b8 Trying to keep the material. Black can bailout by playing 16...i.f5, but the two bishops give White a slight edge. 17 J..f4 ltJf5 Not 17...:d8? 18 i.g5. 18 d5 Establishing the bind. 18 :el? ltJxd4 19 i.xd6 i.e6 20 i.xb8 :xb8 is even slightly better for Black. 18 ... :e8 (D) After 18...cxd5 19 cxd5 :d8 20 :e 1 <it>f8 21 h3 ltJe 7 22 g4! Black still has to find a way to untangle. _.tBI... ..R. ... w. _ .  ..  .  - . . B _8.. i.B8B ?; . B . . . 8" . "8" u . u u n   ,  .  =   //&/i _ / % " /. /. "//m,  " 19 n White would like to play 19 g4, but this is impossible because of the reply 19...:e4. The text-move, how- ever, threatens 20 g4 because White can meet 19...:e4 by 20 :el!. Other methods of preparing g4 are less ef- fective: 1) 19 h3 h5! (19...:e5 20 :dl and 19...:e2 20 g4 :xb2 21 i.b3 ltJd4 22 i.xd6 ltJxb3 23 axb3 :xb3 24 c5! favour White) 20 fl :e4! and White no longer has the move :el. 2) 19 f3 ? (this stops 19. ..:e4 and threatens g4, but there is another problem) 19...:e2! 20 g4 ltJd4 21 fl J%xb2 22 i.xd6 ltJxf3 23 i.xb8 
ANAND - IVANCHUK, LAs PALMAS 1996 209 .i.xg4 and there is no defence to mate! It is hardly necessary to mention that the greedy 19 dxc6 bxc6 20 .i.xc6 releases the bind and gives Black the advantage after 20...:e2!. 19 ... h6? Up to here Black has defended well, but after this move he gets in serious trouble. The idea of relieving the back rank in order to prevent :e 1 in response to .. .:e4 is correct, but he has chosen the wrong pawn move. The alternatives are: 1) 19...:e4? 20 :el! is good for White. 2) 19...f6 20 h3 :e5 21 :dl c5 22 .i.xe5 fxe5 23 g4ltJd4 (23...ltJe7 24 J:td3 !, followed by :b3, and White will win at least a pawn) 24 f4! is very good for White, since Black's queens ide pieces are still immobi- lized. If Black continues 24.. .exf4, then 25 :el followed by :e8 wins a pIece. 3) 19...h5!? and now: 3a) 20 dxc6 (White can always bailout this way) 20...bxc6 21 .i.xc6 .:xb6! (better than 21.. .:e 7 22 ltJd5 or 21...J:td8 22 c5!) 22 .i.xe8 :xb2 and Black should be fine. 3b) 20:d 1 :e4! is very unpleas- ant for White. 3c) 20 :el :xel+ 21 <it>xel f6 is a risky line for White. 20 h3 Now Black is in difficulties. 20 ... :e4 20...:d8 21 g4llJh4 22 c5 is good for White. 21 .i.h2 Black's problem is that ...h6 has not prevented White's g4, as 19.. .h5 would have done. 21 ... cxd5 (D) -.... ... w BiB _i] i   . . ,   //,,// . .i_.. .JtB8..B . . . . .8 8" . "8 u d u _ a _ .. 22 g4 White's threat is not so much 23 gxf5 as 23 .i.c2. Black is obliged to surrender material in the hope of ob- taining three pawns for a piece. 22 .:xc4 23 ltJxc4 dxc4 24 :e1 This intermezzo does the trick - White succeeds in exchanging rooks. After 24 :dl i.e6! (but not 24...b5 25 .i.c2 llJh4 26 .i.xd6 :b7 27 .i.g3 threatening both mate and the h4- knight) 25 gxf5 .i.xf5 Black avoids the rook-swap. 24 ... .i.e6 Or 24...ltJd4 25 :e8+ <it>h7 26 .i.xd6 :a8 27 .i.e5 b5 (27...ltJc6 28 
210 VISHY ANAND: My BEST GAMES OF CHESS i.c2+ f5 29 f4! is hopeless; the threat is 30 :xc8, and if Black exchanges on e5 the passed pawn will decide) 28 i.d 1 lbc6 29 i.c2+ f5 30 ':xc8 ':xc8 31 i.xf5+ g6 32 i.xc8 lbxe5 33 i.xa6 b4 34 <i>e2 and the minor- piece ending is won for White. 25 gxf5 i.xf5 26 i.xd6 i.xh3+ 27 <i>gl ':d8 28 ':e8+ .:xe8 . 29 i.xe8 i.e6 Black has three pawns for the bishop, but the pawns are far back and Black has only one passed pawn, so White should win, although care is required. 30 a4! g5 31 as Now there is only the kingside to worry about. 31 ... <i>g7 32 i.a4 <i>g6 33 i.dl After 33 i.c2+ i.f5 34 i.dl i.e4 Black's bishop takes up its optimum square. 33 ... i.d5 34 i.c2+ f6 35 i.c7 Stopping 35...h5 because of the reply 36 i.d8+. 35 ... <i>e6 36 i.h7 Now White prevents ...f5. Black is gradually running out of active moves, when it will be time for the white king to advance. 36 ... i.f3 37 <i>h2 (D) . B . B .. ..Bi- B_ _ _ _ .B B.B  - - - -  . .  U  . _ B.B B . B B B.t.B " . R  U . U  . . . . 37 ... <i>d5 38 i.c2! Not 38 g3? i.e4. It is still too early to allow the exchange of bish- ops. 38 ... i.e4 39 i.dl <i>d4 40 i.e2 i.d3 41 i.b6+ The two bishops form an effective team. 41 ... <i>d5 42 i.dl f5 43 g3 e5 44 i.c5 <i>f6 45 i.h5 f4+ Or 45...<i>g7 46 i.d4+ <i>h7 47 i.n, followed by i.d5, and one of Black's queenside pawns faIls. 46 <i>h2 1-0 The pawns are blockaded and White will soon win one by either i.f3 or i.f8. 
Game 39 V. Anand - A. Karpov Las Palmas 1996 Queen's Gambit Accepted In round 6 I slid back to 50% after a horrible loss to Kramnik in which I failed to put up any resistance. I spent the evening disgusted with my play and decided to adopt an uncom- promising style the next day against Karpov. The result was my best game of the tournament. 1 lbf3 There was no way I could face a boring Caro- Kann and trying to deal with an improvement on move 45 leading to a difficult ending, etc. I felt that it would be better to go down in style than to do something like that. Now, how does a move like 1 l£1f3 allow me to get interesting positions? Well, to be honest, ll£1f3 can lead to positions even more bor- ing than after 1 e4 c6, but at least they would be unfamiliar boring po- sitions! 1 ... 2 d4 3 c4 4 e4 I played this without hesitation as I wanted to liven things up immedi- ately. d5 e6 dxc4 4 ... b5 5a4 c6 6 axb5 7 b3 He had already started thinking a lot and I knew that he wasn't famil- iar with this opening - not that there is a great deal of theory on it. 7 ... b7 8 bxc4 xe4 9 cxb5 l£1f6 10 .1e2 1 0 d3 has been played more fre- quently. 10 ... 11 0-0 cxb5 .1e7 0-0 (D) . . uz,  .. .     . .. w_ . _ _ . M'. _ M8M M M M D-*-M M _ M Mt[)M . B!L"8 R . . U U \Wr.: _e.z.J__.  12 l£1c3 The game Lutz-P.Schlosser, Ber- lin 1989 continued 12l£1bd2 b7 13 l£1c4 a6 14 b6 l£1c6 15 d2 l£1d5 
212 VISHY ANAND: My BEST GAMES OF CHESS with an obscure position - it isn't clear whether the b6-pawn will be weak or strong. Developing the bl- knight to c3 appears more natural as it exerts some influence over the im- portant d5-square. 12 ... i.b7 13 ltJe5 a6 After 13...i.b4 14 i.b2 i.xc3 (or 14...a615 i.f3) 15 i.xc3 a6 (15...'ii'd5 16 ltJf3 is also a little better for Whi te) 16 i.f3! (better than 16 i.a5 'ii'd5! 17 i.f3 'ii'xb5 and White can- not profit from the b-file line-up) White has an edge. 14 i.f3 (D) a-  .... - - - . ..t.. .. B_ .. ... . ..  . .  . . .  ..t. . . "" /UU  1f.:  - .  14 ... ltJd5 Karpov finds the safest solution, leaving himself with only a slight disadvantage. Black can also try 14...i.xf3 15 'ii'xf3 'ii'xd4 16 'ii'xa8 'ii'xc3 17 i.f4, when again White has a slight edge. 15 ltJxd5 exd5 16 ':b1! After 161Wb3 axb5 17 ]has i.xa8 18 'ii'xb5 White has an edge, but fac- ing Karpov's defensive skills I pre- ferred to aim for a large advantage! 16 ... 'ii'b6 17 i.e2!! I decided that the bishop had noth- ing more to do on f3 and the best plan was to relocate it to d3. 17 ... axb5 There is no choice; both 17.. .a5 and 17...f6 18 i.e3! a5 19ltJd3ltJd7 20 i.f3 would leave White with a very strong passed b-pawn. 18 ':xb5 'ii'c7 19 i.f4 i.d6 20 i.d3 i.a6 (D) After 20...i.c6 21 ':b3, the white pieces are ominously aimed at the black kingside. .. . -.. .  ... w _ _ _ .t..  . . / ._- .:.. . . /  .  z _ . . ..t. . ' . "" . , U U . .1f.: - - -  21 i.xh7+! Here, I spent a few seconds look- ing at 21 ':xd5, which leaves White with a clear extra pawn, but as I mentioned earlier I couldn't face a 
ANAND - KARPOV, LAs P AlMAS 1996 long, technical game. In many lines Black can exchange on e5, leaving a position with 4 vs 3 on one side. De- pending on which pieces are left, this might or might not be a win, but the game would certainly continue for a long time. Then I saw i.xh7+ and didn't waste any more time on ':xd5. I spent some time analysing the sacrifice, and didn't see a de- fence for Black. By now  was too excited to analyse and decided that I would simply play it. Karpov had hardly any time left and I was sure he wouldn't find a defence. Perhaps this decision was some- what reckless but I wasn't punished for it - indeed, I was rewarded with a . . nIce wIn. 21 ... 22 'Wh5+ 23 ':b3 (D) xh7 g8 .. . ..B .  ../ B. _  n .t ' . . /  /  ,/  W B B' ..  "    U _ . B:. . B . . "8" .  U U . . .: . . .  23 ... i.xe5? After this error there is no saving Black's position. There were two 213 alternatives that would have offered Black more defensive chances, al- though White retains a very danger- ous attack in every line. Certainly it would be a monumental task to de- fend this position over the board, es- pecially taking into account Black's time shortage. The alternatives are: 1) 23...i.c8 24 ':g3 and now: la) 24....:a3 (this move was sug- gested by a New in Chess reader, Maarten de Zeeuw) (D) with a fur- ther branch: .  .  ,-..... - ,% . illU ... w. _ . '/n .  . . - - - . B .. B. . "  . . U _ . _..Md - . .  . . "8"   U U . . .: -,- lal) 25 ':xg7+ xg7 26 i.h6+ <iti>f6 27 ':e 1 ':g8 28 'iVh4+ (28 ltJg6 ':xg6 29 'iVh4+ f5 30 'iVh5+ is per- petual check, while 28 f4 is unclear, but could also be a draw) 28...<it>e6 29 ltJg4+ d7 30 ltJf6+ c6 31 ':c 1 + <it>b7 32 ':xc7+ i.xc7looks unclear. la2) 25 f3 'iVe7 and White has various attacking ideas: la21) 26 i.h6 i.xe5 27 dxe5 g6 28 i.xf8 'iVa7+ 29 hl ':al 30 
214 VISHY ANAND: My BEST GAMES OF CHESS ':xg6+ fxg6 31 'iWxg6+ leads to a draw. la22) 26 g5 f6 (26...'iWe6 27 ':h3) 27lL\g6 'iWe8 and White has no effective way to proceed. la23) 26 J:cl! 'iWf6 27 .1h6 (27 g5 xe5 28 xf6 xf6 29 'iWxd5 ':d3 is a likely draw) 27...xe5 (if Black allows White to take on g7 then White should have the advan- tage) 28 dxe5 'iWb6+ 29 hl g630 xf8 ':c3 31 'iWh6 :Xcl+ 32 'iWxcl and White wins. Ib) 24...'iWe7 (D) and now: .III-*-. _.. w. . _._ .11. . . .. .. . n  . _ U _ _ . _ _ a . . 8   U U - . .:= Ibl) 25 g5 and now Black should play 25.. .f6 26 lL\g6 'iWe8, with an unclear position, rather than 25... 'iWe6, when 26 ':h3 'iWxh3 27 gxh3 f6 28 lL\g6 fxg5 29 'iWh8+ f7 30 lL\xf8 xf8 31 f4 g4 32 'iWh5+ is very promising for White. Ib2) 25 h6! xe5 26 dxe5 g6 27 e6! .1xe6 (27...'iWxe6 28 xf8 wins material) 28 'iWe5 f6 29 ':xg6+ and White wins. 2) 23...f6! 24 ':h3 fxe5 (24...xe5 25 dxe5 transposes to the game) 25 dxe5 'iWc4! (25.. .':xf4 26 e6 f8 27 'iWh8+ e7 28 'iWxg7+ xe6 29 ':el+ ':e4 30 ':h6+ forces mate) and now: 2a) 26 'iWh7+ f7 27 e6+ f6 (27...xe6 28 ':el +! wins for White and 27...e8 28 'iWg6+ d8 29 g5+ c8 30 ':cl is clearly better for White) with another fork: 2al) 28 ':h6+? gxh6 29 'iWxh6+ f5 (29...e7 30 g5+ mates) 30 g4+ e4 31 ':el+ 'iWe2! (31...d3 32 'iWg6+ d4 33 .1xd6 probably fa- vours White, but is extremely messy; I didn't bother to analyse this line deeply as the strength of 31..._e2! made it irrelevant) 32 ':xe2+ xe2 33 .i.xd6 ':al+ 34 g2 fl+ 35 g3 ':f3+ 36 h4 ':h3+ and Black wins. 2a2) 28 g5+ xe6 29 ':el+ d7 (not 29...e5 30 ':xe5+ d6 31 ':el lL\c6 32 'iWg6+ c7 33 'iWxg7+ b6 34 ':bl + b5 35 ':hb3 winning for White) 30 'iWxg7+ c6 31 ':c3 lbd7 32 ':xc4+ xc4 with an unclear position. White has a lot of pawns on the kingside, but since Black's king is now safe he can acti- vate his pieces. 2b) 26 ':el! (cutting off the en- emy king's escape route) 26..._xf4 27 'iWh7+ f7 28 exd6 and now: 2bl) 28....:e8 29 'iWh5+ g6 30 ':e7+!! ':xe7 (30...f6 31 'iWf3! is an unexpected win) 31 'iWh7+ f6 
ANAND - KARPOV, LAs PAIMAS 1996 215 (31...f8 32 dxe7+ e8 3311i'xg6+ xe7 34 :h7+ mates) 32 1Ii'xe7+ f5 33 1Ii'f8+ e5 34 :e3+ with a decisive advantage for White. 2b2) 28...ltJc6 29 :f3 1Ii'xf3 30 gxf3 .1c4 31 h 1. White has the ad- vantage because of his d-pawn and persisting attack, but this was Black's best chance. 24 :h3 f6 25 dxe5 1Ii'e7 25...1Ii'c4 26 :el 1Ii'xf4 2711i'h7+ f7 28 e6+ e8 2911i'g6+ is also no help for Black. 26 7+ 7 27 :g3 (D) 1.- . _ . B _ . _._- .t.8 8  8 - - - - . .iD . . . m . . . . a . . R8 R .  U U . . .:= 27 ... e8 27...:g8 2811i'g6+ f8 29 exf6 is devastating. 28 Jbg7 From now on, the game pretty much plays itself - there are several ways to win on every move! One al- ternative is 28 exf6! gxf6 (the lines 28.. .:xf6 29 :xg7 1Ii'e6 30 1Ii'h5+ :17 31 :xf7 jfxf7 32 :el+ and 28...1Ii'xf6 29 :el + are also termi- nal) 29 :e3 jfxe3 30 fxe3 .1xfl 31 .1d6 and White will have a decisive material advantage. 28 ... 1Ii'e6 29 exf6 lL1c6 30 .:&t d8 31 h4 To clear the back rank - it's always nice to have time for such details. 31 ... b7 31...ltJd4 loses to 32 c7+ c8 33.1a5. 32 :c1 a6 33 .:&1 Again, there are other routes to victory, e.g. 33 .1c7+ c8 34 .1b6. 33 ... b7 34 :d1 a6 35 1fb1! 1hf6 36 g5 c8 1-0 Karpov lost on time while in the act of playing 36...c8. The reply 37 1Ii'b6 wins on the spot. Winning this game gave me a 6-3 score against Karpov in decisive games. My other games ended in draws and my score of + 1 was sufficient for out- right second place behind Kasparov. Finally, here is an effort of which I am very proud, from the Credit Suisse chess festival in Biel (July-August 1997). 
Game 40 V. Anand - J. Lautier Bie/1997 Scandinavian Defence At the opening ceremony the previ- ous day, the chess players had to play a match against representatives from the Swiss Skiing Federation. Each participant from the A and B tournaments had to play two moves and if they didn't mate the skiers within 24 moves, then the skiers won. The chess players chose the Scandi- navian. To my great surprise, this is exactly what happened in my first round game! 1 e4 d5 I don't recall Joel ever having played this before, but he hadn't com- peted at all since Monaco in April, so I assumed this was an opening he had prepared during the intervening three months. 2 exd5 3 ltJc3 4 d4 5 ltJf3 6 i.c4 Nowadays 6ltJe5 is more popular, but during my preparations for the World Championship, I noticed that the lines with 6 i.c4 were very dan- gerous for Black to navigate. Know- ing that Joel didn't have a great deal of experience with the Scandinavian, 1Ii'xd5 1Ii'a5 ltJf6 c6 I decided to test him in this critical variation. 6 ... i.f5 7 ltJe5 The variations with 7 i.d2 leave White with a slight edge, but I re- membered that 7 ltJe5 and 8 g4 gave Black more problems. 7 ... e6 8 g4 i.g6 9 h4 ltJbd7! (D) Joel chooses the best line. After 9...i.b4 10 i.d2ltJe4 11 f3! White gained some advantage in Campora- Curt Hansen, Palma de Mallorca 1989 and it was after this game that people started to look at 9...ltJbd7 more seriously. .. ... . ..... w_ . . _ .'.'rI.t.  .  . -   . Bi." .8' -    .  . . 88B  . U _ U _  .. B:  _ * d 
ANAND - LAUTIER, BIEL 1997 10 ltJxd7 ltJxd7 11 h5 i.e4 12 ':h3 i.g2 A nice finesse - if White plays ':g3, then Black will gain a tempo with a later.. .i.d6. However, if the game continuation is correct, then Black will have to abandon his fi- nesse and play 12...i.d5. Then 13 i.d3 i.d6 14 i.d2 1/ic7 15 ltJxd5 cxd5 16 1/ie2 i.f4? 17 0-0-0 0-0-0 18 i.xh7 won a pawn for White in Ochoa de Echaguen-Denker, New York Open 1989, but of course this was not forced. 13 ':e3! I had wanted to play 13 ':g3, which gives White a slight edge de- spite the fact that Black can play .. .i.d6, but a fresh look at a position during a game can often turn up bet- ter moves than those found during home preparation! 13 ... ltJb6 After 13... b5 14 i.d3 b4 15 ltJe4 Black will have to play .. .i.xe4 sooner or later, when White will be better due to his two bishops. Under- standably, Joel didn't want to resign himself to an inferior position with few prospects of counterplay. 14 i.d3! This move, which I found at the board, was the reason I decided to go for 13 ':e3 instead of 13 ':g3. 14 i.b3?! is inferior after 14...c5!, when Black has good counterplay. 14 ... ltJd5 (D) 217 The obvious reply, attacking c3 and e3.  ..  /  n /.(/2;:;;  /   ,//;0/ ,y /' ",,/  & ;;:;< %g & - & W W. 4;) :'/ . - . ;/,n , ;/,////': /y ////// _i.i. . ill ... .8 . D B8.  / / . . .  . _ .iL  ({f:W 8n8B .t.. K.U. g %l  % _ /;;\WJ = /0  . .  // 0 .y % . /  "" '/ ,  //0j z 15 f3! This remarkably calm move is the point behind White's play. He is willing to jettison a few pawns and/or the exchange in order to snare the bishop on g2. When the bishop is fi- nally trapped, Black will probably end up with a rook and two pawns for two minor pieces. Owing to the lack of open files, the minor pieces will be very much more at home in the resulting position than the rooks and this, coupled with White's lead in development, will almost inevita- bly give him a clear advantage. In C.Bauer-Prie, French Champi- onship 1996, the continuation 15 ':g3 ltJxc3 16 bxc3 i.d5 17 i.d2 1/ia4 18 1/ie2 b5 19 h6 0-0-0 was fine for Black. 15 ... i.b4 (D) After 15...ltJxc3 16 bxc3 1/ixc3+ 1 7 i.d2 1/ixd4 18 f2 i.xf3 19 
218 VISHY ANAND: My BEST GAMES OF CHESS xf3 or 15...xe3 16 xe3 'iWb6 17 f2 h3 18 :bl a position of the type mentioned above is reached; White holds the advantage in both cases. I.B B.B . //i. .ii w  _ _ _ BiBiB _ - .. .fj  R .8. _ u   B mi.a8_ 8 R 8. R.t.. u _ _ _  1f* . - - *  16 f2! xc3 The above comment also applies to the line 16.. .lL\xc3 17 bxc3 i.xc3 18 :bl xd4 19 xg2 xe3 20 xe3. 17 bxc3 'iWxc3 18 :b1 'iWxd4 White wins after 18...xf3 19 'iWxf3 'iWxd4 20 :Xb7 0-0 21 'iWe4. 19 Jbb7 :d8 (D) The other critical variation runs 19...h3 20 :xf7! (I stopped here, but Joel saw two moves further!) 20.. .c5 (Black simply protects his queen and threatens 21... xf7; if in- stead 20...xe3, then 21 xe3 'iWd6 22 :f4 is very good for White) 21 :f5!! xe3 22 xe3 'iWb2 23 :xc5 0-0 24 g3! winning, as Black will be lucky to get a single extra pawn. If 19...f4, then 20 g3 'iWd6 21 a3! xh5+ (21...'iWxa3 22 e4! also wins) 22 xg2 'iWg3+ 23 fl is winning for White. - ..- . :. .i-i w_ _ . _ Bi.i. _ B _. _fj . _ B8_ _ _i.a8. fj.fj. =.t._ . -... . After the text-move, I saw the possibility of g6 but then I realized that it didn't work immediately be- cause Black could run with his king, e.g. 20 g6 'iWxdl 21 :Xe6+ f8 22 a3+ (or 22 :xf7+ g8) 22...e7 23 i.xe7+ g8, and the attack fails. Then I saw the possibility of in- serting h6 at the start of the combi- nation, when a later ...g8 could be met by :g7+!. Suddenly, all that re- mained was to check the details... 20 h6!! gxh6? This gave me a chance for a really beautiful finish. Black could still fight on with 20...xe3, but Joel hadn't seen the idea behind h6!. The analysis runs: 1) 20...g6 21 i.xg6! 'iWxdl 22 :xe6+ f8 23 :Xf7+ g8 24 :g7+ f8 25 i.a3+ followed by mate. 
ANAND - LAUTIER, BIEL 1997 219 2) 20...ltJxe3 (absolutely the only move that doesn't lose by force) 21 xe3 'iWe5 22 hxg7 :gS and now 23 'iWc 1 ! threatening 'iWa3 and xg2 is strong (not 23 h6 'iWh2!). 21 .1g6!! (D) - ... . B _.:. .... ......t_ a... . _ II .8. . . a88 8.88 =.1. . \WJ. . . .-. . 21 ... lbe7 There is no way out: 1) 21...'iWxe3+ 22 xe3 fxg6 23 c5 wins. 2) 21...'iWf6 22 xf7+ 'iWxf7 23 :xt7 ltJxe3 24 'iWxdS+! (24 'iWe2 ltJd 1 + 25 xg2 xt7 is less clear, although White remains much better after 26 'iWe4) 24...xdS 25 xe3 .1h3 26 :xa7 :eS 27 :xh7 and Black loses several pawns, followed by his bishop! 3) The key variation is 21... 'iWxd 1 22 :xe6+ fS 23 .1xh6+ gS 24 xt7#. 22 'iWxd4 Jbd4 23 :d3! There is no need to bother with 23 :Xe6 :d7 when the text-move wins effortlessly. 23 ... :d8 24 Jbd8+ xd8 25 d3! 1-0 Because after 25....1hl 26 .1b2 :eS 27 f6 Black will soon be in zugzwang and have to surrender at least a piece. 1996 and 1997 have been extremely successful years for me. I've not only had good and consistent results, but have also produced a lot of beautiful games. I am enjoying myself immensely playing chess and hope that my cur- rent form augurs well for a renewed attempt at gaining the title of World Champion. 
Combinations - ; { . - - . . . 1 . . .&g& - :}!$} % . - . B tw >/,; & g & - ; ....    D '/. '   i:   :;j  /1,0 /:////; /"o// /,. 0  /m $A. !  w 0 _ //jZi &Q a %i ' /. . /:J" / A   //  D /'    0 i'  «'&' /.i, , '..  /,  .'if. ..tM ... . "': . . . u A. Hamed -  Anand Thessaloniki Olympiad 1984 Although White's king is in an awkward situation, for the moment the bishop on g2 is providing an ade- quate defence. How did Black press home his attack? %'///',.'(<< :%;; . %,/ ;; ,;,/jf i;:;j . / / $:3 )}$'h ,/m a ¥ J11 1  "'/'i'% ;:f;/,z: & 2 i , ] /1£1 . . . ' '% // i?I /Aiii ' . '" i :li  //'M' B    _ /0 ,  '/ ,    3 % . '%  . % 8 // . ) .. :*f::'f!:: %;j  ; i;;/'>/"'/-'l /"o// /'/" -://;/;;/ :*''< .t   /.: 8 ;? . : . iJd  . >i' . /"  ;; -;;,//;;)..{0 ///::; ';:1; /o/;., ij///: ///,/,. 0'///;;'-;   /,. /; A {; ,, '  .  ¥lj '  .o %4     t ,//  ;"  A   / ..  r/j ' D / o Wf ;W% / '/  /; w/// //;; /, y/  //  - - .,//   . .   Thkmakov -  Anand Delhi 1986 White's bishop is horribly placed and Black is effectively a piece up on the rest of the board, but does he have anything better than simply ex- changing on f2? 
COMBINATIONS 221 {/ . "dW . ;: ;://.   ./% % <;;;;% :- / -:; '/;,0;0; $: ". ii 33 wiM . . . .i/  . ifi nn , _ z . .. .  "::0" 'm//% // '/'d + ; X .  ,;  Wd .?t .:. ""     . .\WI.  ; 7M.'  , Ja. Gil - V. Anand World Junior Championship, Gausdal1986 White chose to meet the attack on his queen by 31 1Id2 (31 'iie2 would have been a better chance). How did Black finish the game at a stroke? .  /  /   / . / % '0 a ,;m  / , , / %,  4 _ ... Ri & - - . BA _ ,. . .i.i.  _ z .  " "iB . u  G . D 0 'J":\. 'f!ffi R 0 0 % '  /j'  "l..J.  % /j  /"   /"  //   -i."  _ W U . .  . /0   .  J. Levitt - V. Anand Match (3), London 1987 White is threatening to make fur- ther progress on the queenside with 29 1Icc6. How did Black pre-empt this with a kingside strike? ;!PJK / .;. / ':. 1  { ;:!>x/ jp}f _ _ . !!i (jW3 . & 5 _ ?//{ :;/7 .. A BI.B ... _ ... @ t;/;) '. ;fff$j . .A.  '0JP  0 _; . i D rih. J":\. ?f  0,  'i!iiI «,,1/,::: "l..J ; & r!1   : .    /  . / . id A   0}, , / V %//, /'>,  'W///"/ a %i 0 : . . . ..{'?i .  ;% ;%  W:0 f ;f / , u/. , , ,W);/j /.0/j/ J. Gdanski -  Anand World Junior Championship, Baguio City 1987 Although Black's protected passed pawn on a2 is very dangerous, White has an extra pawn and threatens to start eliminating the queenside pawns by lbd4+ and lbxb3. How did Black make use of his advanced pawns? Note that it is not Black's first move but his fourth which is the difficult one! 
222 VISHY ANAND: My BEST GAMES OF CHESS ..Wt1 _ . . . - - .   .  .. 6_  _ _ W.t...... _ . ... - "". .  u u d     . ttJ . m u    i. . 8 _  u u   . \Wr  M  .   V. Anand - M. Adams Lloyds Bank, London 1987 Black has just played 14...lbf6- d5, preparing to answer 15 b5 by 15...lbxc3 followed by 16....1xb5. What is the flaw in this plan which allowed White to score a quick win? . . . . 7. . . . B... . . . . . . ..ttJ. . . .   . _ U _ /;;:0';; . . . . . . . . v. Anand - P. Thipsay Coimbatore 1987 Anand had gone into this ending deliberately, even though his king is now far away from the action. Is the position a win?  .  . . "  . . .  .  . .  . \Wr..t. 8 _ _ .. w..  . . - - . - o W: , : , ? 8 . ' '/ //.  ,0£ w /.4 I /;'//-< ?:0 "( / '// /,///// /. //,-/ '// .8. .8. W .  . :;;/   %  + W1:% _ - ,; % .  . .  U . . U . . a . R. Tomczak - V. Anand Lugano Open 1988 This was a lucky escape for An- and. White has sacrificed a piece for an enormous attack, and now he could have won by 35 'iih6+ g8 36 Zle6, when there is no answer to the threat of 37 'iig5+ f8 38 Zlf6. In- stead, he rushed in with 35 Zle6, but Black replied ... what? 
COMBINATIONS 223 . . '  .   . . 9 _ _ Bi. B __ . . . . . _'-Z.J8 , _  . . . _ . . d .B BiB " . . . u _ _ d . .  R _1/ v. Anand - R. Gerber Biel1988 After 44.. ..1e3, confining White's king, Black has good drawing pros- pects, for example 45 lbd6 c7 46 xf7 (46 f5 .1c1) 46...xc6 47 h6 .1d4 4S h7 d5 49 hS_ .1xhS 50 lbxhS d4 51 f7 d3 52 e5+ c2. However, Black was impatient to attack the b2-pawn and played 44....1g7. Why was this a mistake? . . .. .  /_,/', 10 . / ; ! .  @: _ ,ili;, _ B    . ,/.  -   i.i _ . _ d . __ D . D__ . ...ml.. R . . R - . . M'*' D. Campora - V. Anand Thessaloniki Olympiad 1988 Black has a formidable line-up along the g-file, but White is threat- ening to force exchanges by _fl. How did Black make use of the ag- gressively posted rook on g2? ...1.. . ll i . i - - - - W .. ... . BiBi. . R.  _ U d _ t2JD...t. \Wr. . uwd d u   -:- M _ .  V. Ivanchuk - V. Anand Reggio Emilia 1988/9 Ivanchuk now played the serious elTor 17 hl ? (instead 17 .1f4 would have been very good for White). How did Black then win? 
224 VISHY ANAND: My BEST GAMES OF CHESS BZB . ' 12.'& . . B'. ttJ.   - . - " .88 u . _'t..J_ Di.& & & .   " . _ U _ & . D . .. . - - -   v. Ruban -  Anand Palma de Mallorca GMA 1989 Material is equal, but White has several threats. IT Black does not have something special, then he will be in big trouble. What did Anand play? . & . & 13. .if& . w .:. . . .8& . .. & ... 0 . .   . . _ U  .  .."   . U  . . . .  Anand - Ye Rongguang Manila Interzonal 1990 White would like to advance his pawn, but at the moment this would leave his rook undefended. Is 51 :C2+ followed by 52 b6 the right way to proceed? 0'//0' :,;, /ij  /,;::;%  I /% :',/$ 0% 1::'   ////,: / . :f{ J% %f% /,/", /;:/ ,/ ,/,// /,////// 0 14 ri" fj .  :;:fJ(/ ;W$' %;Z W . . ...  . / .  .  - - . 8. . .8. . . . .. . . _ 0 . Bi . . v. Anand - A. Dreev Candidates match (2), Madras 1991 The position seems to be rather awkward for White as his king is in serious danger. Not only does the possibility of a back-rank mate mean that White's queen has restricted mobility, but Black is also threaten- ing to win the h2-pawn, after which his own h-pawn will become very dangerous. If White defends it by 42 1i'gl, then 42....:g2 will drive the queen away. Can you see a clear-cut way for White to save the game? 
COMBINATIONS \Wr} i% ' _ . ,;; »'A / //<-%';  & )}, :/;;;: 1&\ ;//// & 15../ ;j}J.... B . ;'0  it . '  . ' : :do ll ' ,,/0.-;::  :0:1f. '/ m):/ '   ,;  ,  0 1  / 81 ° /  . '/ . ' . % 8 /,      //;;/://:c '/ 'u/  , %;::% (;;;}0 8 ?/:; 8 . " / . /;::i . /; / c.' ;};,/ %  //// :>4/;' ,,;;:;,   ",;::c.: u0;/,i //// /u u :'////,'/ ?1)< : : Ii "'::;{{'f) 8 :/;;*; '///// / / / /";;//'} ;;?;:././:' g% %/  0;jj  A  .,. o {Q ,,; %;% A %   '0 " . . " . , .... - - ....  Anand - J. Timman Linares 1991 White has an extra pawn, but he will have to work hard to exploit it. However, the Dutch grandmaster made the awful blunder 27...1Vd7?? Why was this so bad? SB"'.  .. 16_'_'.'_' B'. _ a.m  .  . -  - . .  . . ; ./,:  /W; '//. .  8 /. W$; M' / A ¥'    / A W o   if;;'   0 /;1d //  ,Y'/'0/ u  ,-://,//. n a\WJ n =   1i!Ir  /  / , /  ,/, / , v. Anand -  Salov Immopar Rapid, Paris 1991 16....:dS 17 'iif3 and 16....:eS 17 f4 are winning for White, so Salov decided to give up the exchange by 16...d6. What happened next? 225 k' .,   . ;(/} A w:' _ .-7:% 0/1 . / : . '/., '//, ?,{ . . /.' /%d 17 ?J  i%i:  // ///   ),}  / , B 2:0r B.R . ;-;:0 ,%$ ,,&[$ /u/u ff/;f; / . rlJ  if %1 /;;;%   ;%n w// ,/ ,// /" // //////'/, ;,,; . ''' ? ?i:/;:  0 / /:;  :::0:;/ ,:;;y-:- /  'f / V '00;:/;: -:;;:;'>' , /,  . :;?f .  //;/xi ;O'  ;% A D   g  / A  O  /O:;% //  ;., ,/ /.  :/$,,1; /$  %/ / 0;:'  /, '/; ! :  , .;.. A. Beliavsky - V. Anand Reggio Emilia 1991/2 Black has an advantage, but it ap- pears that considerable work will be needed to win. Anand made it look easy by winning a vital pawn. How? %r; . W ' ff%f :i ... 18.'. . .. B M _ R.a . ' D 0/%0 & . u/.,    . 'A / //// !.% /"  /. / . R8 Wtl .  . -  . / / fj  / . /., "/,   /./% A .    h o ,,%; //  , /w;« 8. . . . . :;ff"'  " a XjW;%;"', .... L. Ljubojevic - V. Anand Roquebrune Rapid 1992 A truly chaotic position. Black can reach a favourable ending by playing 29...'iixg5 30 'iixg5 ':xg5 31 <iti>xfl d3, but how did Anand win far more convincingly? 
226 VISHY ANAND: My BEST GAMES OF CHESS ...- - ... ....-.. 19d _ _ . W. .- H. . _ _ d _ . . . n88H. H. e.z.JU _ d d Dlt). _8. H. H. 8R   _ u . a:B.tB<iit  Anand - M. Illescas Linares 1992 In this position one might expect protracted strategic manoeuvring by White, but Anand was able to utilize his queenside pressure to launch an explosive combination. How? ........ - .H. ... 20_ . . . W . .. H. .  - . . II . _ . a.i.. . . . mlt). AmA.\WrAR    1IrU  U . .:.  --  Anand - K. Robatsch Manila Olympiad 1992 In return for the sacrificed pawn, White has a large lead in develop- ment. How did he use it to launch a deadly attack? The detailed varia- tions are quite complex, but the main point is to get the basic idea right! ... ... p\WJ. . _ 21... ... B..... _ H.... H. .'-Z.Jd u d  . 8n e.z.J_  u u .8. _ . 8. ....  - - - * B B:. . c. Garcia Palermo -  Anand Manila Olympiad 1992 White's king is very exposed, but he has a knight ready to hop into d6. How did Black win material? 
COMBINATIONS . . ... 228 _ . _ Bi. . .iB B _ II .i  .8. . . - - - 88B_ a8 .  .8 x__ 8 . B.. L. van Wely - V. Anand 1ilburg 1992 White is a pawn up and he is at- tacking Black's rook, but even so An- and found a combination leaving him with a clear advantage. What was it? .. . ... . . R i 23d .  _ W . . . .  . 8. - - . - m .. .... d..d . B _ 88 8 8 088 .wa. . - .  V. Anand - J. Piket Amsterdam VSB 1993 White has very strong threats on the kingside in return for the piece, but first of all he has to deal with the attack on his rook. Which move is correct: 42 f6 or 42 .1f6? 227 II 8.... 24_ 8 B Bi w BiB IIi Ii  .   . . .  88 88. .   .8. .   . .t.a _ 8B . . : d .  R  Anand - A. Khalifman Las Palmas 1993 White's h-file attack is certainly dangerous, but Black's queen has just leapt into White's position and is attacking the b2-rook and c3-pawn. How did White proceed with his at- tack? 
228 VISHY ANAND: My BEST GAMES OF CHESS ..- ... /% & ' . '..,//, /1 ,'Y'1:J .  & 25  A . ;/: A  ".   $/; B B'B'. . D . B'B D :  " . ,  /' U . B . . D . . i.Q  .  u .:. a .<it;> v. Anand - M. Adams PCA Candidates match (1), Linares 1994 White has a slight advantage as Black's e-pawn is slightly weak and his bishop is more active than the enemy knight. On the other hand, Black's doubled rooks on the d-file could be useful. Adams now contin- ued with 31...1if7. Why was this a mistake? W   %} /;:;r 0 . / " , ' . W  :Wa  // % / g ";;:'/; ;; // . ,@; :{<  ,,"  @  0/': . ./;//..' ..  . / 26   . :r:;: %j  /y/ ;:;///,/;/, '''' B . %J; .. <;;/ . // /j;;;;:;  . '@ %;? < ;;;) {. : / /! /  ./  . ' . /: .. :. . ': }$ /. ;//'" //;g . [:.J.j. :0- "/ '/X  /"",1,;/ W'l' · "    ttJ   ;:  l:  @,dh   (//';;ij/,/, '0)// '/':;;-,: /"", _ A ,'iY ((/ A Wf A .0. :.0.0 D.ii.. . .:B .<it;> V. Anand - B. Gelfand Dos Hermanas 1996 White has a decisive positional advantage in that he is a pawn up and his knights dominate the centre. Gel- fand therefore decided to stake eve- rything on a last gamble by 4O...i.xh3 41 gxh3 .:xf3. How did Anand re- fute this last desperate attack? 
COMBINATIONS  .. . a 0:f0/ t.z.J W // 27_ _ _.18. w  " .' I /   . //u/ . _ a.. . ... B .. B . . ."8" / uu a . . . v. Anand - J. Polgar Dos Hermanas 1997 Material is equal, but White has a dangerous passed pawn. How did Anand utilize this pawn to win the game? B _.I... 28 .iV..' _. w 'li..'.'. . . B . 8B8.811 . B8 . .:  . 8" _ d U U R.i.R .:. V. Anand - V. Ivanchuk 6th Amber Rapid, Monaco 1997 White is already a pawn up, but Black will gain a little counterplay if he has time for .. .lL\c5, attacking b3 and opening the d-file. How did An- and finish the game vigorously? 229 . B.... 29B a B'B. w'B _.IB'B . . B . B . B'B  .   %/ .    ,'h     , , ,/ /  /   .   u. _ . .:. . V. Anand -  Kramnik 6th Amber Rapid, Monaco 1997 Despite material equality, White is winning due to the imprisoned bishop on h2. What is the most effi- cient method of finishing the game? .I. . _.. /.t. . /  . . 30 /   _ _ W  . - . _ d ,. .  .  // - . , /m _ . 8/ .  . ' U /m d  R .8." U d t.z.JU m 8.i.  , , ," _ U /    .   . Wd :  / - ,  /  V. Anand - U. Andersson 6th Amber Rapid, Monaco 1997 In this innocuous-looking posi- tion, White found a way to win a pawn. How? 
Solutions 1) 42...lbdl ! Threatening to win the queen with 43...lbf2+ and thus forcing White's reply. 43 :a2 :al! A neat way to eliminate the defen- sive rook and so gain access to 12. 44 ':b2 44 ':xal lb12+ 45 'iix12 'iix12 is hopeless as White will lose the e- pawn immediately. 44...lbxb245 'iixb2 1 46 _e2 :a3 47 .1n:a2 0-1 2) 26...lbdJ ! This is much stronger than taking on f2. Black ends up in the same type of position, but with an extra pawn. 27 'iie2 This is forced because 27 fxe3 'iial+ 28 <iti>g2lbel+ and 27 'iixd3 'iix12+ 28 <iti>h 1 e2 are dead lost for White. 27... 'iixf2+ 281Vxf2 exf2+ 29  g7 Not only has Black won an impor- tant pawn, but his king has a clear run into the heart of White's posi- tion. 30 .1b5 <iti>f6 31 c5lbxc5 32 <iti>xf2 <iti>e5 0-1 White will soon lose the d5-pawn, when he will be two pawns down for nothing. 3) 31...lbf3+ 0-1 A typical combination. After 32 gxf3 _g5+ Black wins material. 4) 28....1xg3 A surprising but strong combina- tion based on the immobility of the e2-bishop. 29 fxg3 'iih3 30 .1dl White is unable to defend because his bishop gets in the way wherever it moves. The alternatives are: 1) 30 ':bc6 1Vxg3+ 31 <iti>h 1 g8 32 ':fl g4 33 'iid2 _h3+ 34 <iti>gl g3 and wins. 2) 30 ':fl 'iixg3+ 31 <iti>h 1 g4 32 ':xf8+ ':xf8, again winning. 3) 30 .1fl _xg3+ 31 _g2 (31 .1g2 and 31 hllose to 31...g4) 31..._xe3+ 32 <iti>h2 1Vxb3. 30...lbd7 White no longer defends fl and so Black can win a whole rook. 31 'iig2 _xg2+ 32 <iti>xg2 xb6 33 lbaS :ac8 34 ':c5 Jbc5 35 bxc5 lbc4 36 lbxc4 bxc4 37 c6 ':c8 38 .ta4 c3 0-1 
SOLUTIONS 231 5) 48....1a3! The only possibility, but the game is not over yet. 49ltJd4+ Or 49 ':h 1 ':d6+ 50 ltJd4+ ':xd4+ 51 cxd4 .1xb2 and the pawns are too dangerous, e.g. 52 c4 .i.cl! 53 ':xcl b2 or 52 ':h6+ f7 53 ':xh7+ g6 54 ':a7 .1cl. 49... e5 50 lbxb3 .1xb2 51':hl White takes aim at Black's last pawn. 51....:&3 ! Of course 51...al'ir' 52ltJxal, fol- lowed by ':xh7, is a draw. 52 ltJal After 52 c2 ':xb3 53 xb3 al'ir' 54 ':xal .1xal the h-pawn decides as the white king cannot reach hI (55 c2 xe4 56 d2 f3 57 el g2). 52...Jbc3+ 53 d2 .1xal The simplest solution. Now there is no question of a rook's pawn and wrong bishop draw. 54 Jbal ':a3 55 c2 h5 56 b2 .:a8 57 f4+ xe4 58 ':el + xf4 59 al h4 60 ':e7 ':h8 61 81 0-1 6) After 15 b5 ltJxc3 White played 16 'ir'd3!, threatening mate on h7. This interpolation defends the b5- pawn with gain of tempo, and so wins a piece. The finish was 16...g6 17 Jlxc3 1Wc7 18 bxa6 ltJa5 19 .i.xa5 1-0. 7) Finding the winning line (without moving the pieces!) is a good test of your powers of visualization. It is long but absolutely forced. 53... c5 After 53...d5 54ltJb6+ c5 55 ltJxa4+ c4 56 g5 b3 57 f5 xa4 58 e5, White's king is just in time to win the game. 54 ltJa3 d5 The best chance. White can no longer win the a-pawn, so Black can ultimately play his king round to b3. 55 g4 e4 56 c4 d4 57 f3 c5 58 e3 b4 59 d3! This knight sacrifice leads to a winning ending of'ir' vs 'ir'. 59... xa3 60 c5 b2 Or 60...b3 (60...b4 61 c6 a3 62 c2) 61 c6 a3 62 c7 a2 63 c8'ir' al'ir' 64 'ir'c4+ a3 65 'ir'a6+ b2 66 5+ c 1, winning as in the game. 61 c6 a3 62 c7 8163 c8'ir' a1'ir'64 _c2+ a3 65 _c5+ 81 66 _c4+ a3 67 _a6+ b2 68 'iib6+ c1 69 'ir'c5+ b2 70 'ir'b4+ 1-0 After 70...a2 71 c2 Black is mated. 8) After 35... 'ii'xe6 White resigned immediately, due to 36 fxe6 .1e4#. 9) Black had overlooked a tactical point and White won by 45 c7+ d7 46 ltJd6 xc7 47 ltJe8+ d7 48 
232 VISHY ANAND: My BEST GAMES OF CHESS ltJxg7 e7 49 f2 f6 50 h6 g6 51llJf5 1-0. 10) 33...'ir'g3! Sometimes the best move is not a spectacular combination but a sim- ple liquidation to an ending. At first sight Black has not obtained very much from his position, but a closer inspection shows that the penetra- tion of a rook into the heart of White's position has fatal consequences. 34 .1xg3 ':xe2 The pawns on a2, c3, d3 and f3 are all vulnerable to attack and in this blocked position White's bishop is unable to develop any activity. 35 .1h2 ':e3 36 ':xg8+ xg8 37 d4 A desperate attempt to bring the bishop to life. 37...f4 Black shuts the bishop out com- pletely. 38 dxe5 dxe5 39 ':gl+ h7 40 .:n Or 40 ':g5 ltJf7 41 ':xh5+ g7 42 ':f5 ':xc3 and the pawns will fall one by one. 40....:xc3 41':f2ltJf7 42 g2 b5 43 n b4 0-1 11) 17...ltJf2+! 18 ':xf2 (the alterna- tive 18 'ir'xf2 .1xf2 19 gxh3 .1e3 is hopeless) 18....1xg2+ 0-1 leads to a forced mate. 12) 48...llJf3+ 49 g2ltJh4+! An unusual case in which a knight forks another knight. 50 xg3 Or 50 ltJxh4 .1xd5+ 51 f3 (51 xg3 gxh4+ wins a whole rook) 51...:Xc3 52ltJg6+ h7 and Black's material advantage is decisive. Thus White has to surrender a piece. 50...ltJxf5+ 51 h2 ltJe7 Defending the rook and thus un- pinning the bishop. 52 e4 .1e2 The rest is easy. 53 Jbc8+ ltJxc8 54 g3 .1xh5 55 ltJd8 b6 56 f4 bxa5 57 bxa5 g7 58 e5 .1f7 59 fxg5 hxg5 60 exd6ltJxd6 0-1 13) Definitely not! 51 g2! Avoiding Black's cunning trap, 51 ':c2+ f3 52 b6? 'ir'h3+!! forc- ing a draw by stalemate. White would have to return by 52 ':c6 or 52 ':c5, but of course this loses time. 51...'ir'f5 52 ':c2+ dl 53 ':cl+ Again White must take care. 53 ':d2+ el 54 b6? again runs into a stalemate after 54...'ir'f3+ 55 gl 'ir'g2+. 53... e2 54 b6 Now the pawn can safely advance. Black's one check on f3 is harmless. 54...'iib5 55 ':c7 'ir'e5 56 ':c2+ dl 57 ':cl+ e2 58 ':bll-0 
SOLUTIONS 14) White continued: 42 'ii'gl':g2 43 as! Not 43 'ii'xg2? hxg2+ 44 xg2 g5 45 h3 f4 46 a5 .1b8 47 a6 .1a7 48 fl g3 49 e2 xh3 50 f3 .1b8! 51 a7 .1xa7 52 f4 and one way to win is the attractive ma- noeuvre 52....1e3+! 53 f5 .1g5. Paradoxically, White needs to pre- serve the black pawn on h3 to set up a stalemate. 43 1Wbl + rJ;g7 44 a5 ':xh2+ 45 gl may also draw, but it is cer- tainly a less forcing line than the text-move and offers more chances to go wrong, for example 45....:g2+ 46 hl.:xg4 47 a6? (47 'ii'b7+ g6 48 'ii'c8 is correct) 47...h2 48 a7 ':gl+ 49 'ir'xgl+ hxgl'ii'+ 50 xgl .1d4+ and wins. 43...Jbg1 + 44 xg1 Now White only needs to give up his g-pawn with g5 and his a-pawn with a7 to force a draw, but he has to make sure that Black cannot win by meeting g5 with ...f5. 44... g7 After 44...g5 45 a6 .1d4+ 46 hl f4 47 g5, to be followed by 48 a 7, Black is forced to stalemate White. 45 a6 .1b8 46 hl g8 47 gl 47 g5 f5 48 g6 f4 49 g7 f3 50 a7 would also draw. 47...848hl e849g5f550 g6 f4 51 g7 7 52 g8'ii'+ g8 53 a7 i.xa7 1/ 2 _1/ 2 233 15) The game finished 28 .1xf4+! 1-0 since 28...exf4 29 'ii'f6+ forces mate and 28...g7 29 ':dglleaves Black two pawns down with a bad posi- tion. 16) 17 .1d2 A surprise. White ignores the ex- change and, making use of the fact that d6 is now blocked, traps the black queen instead! 17... 'ii'c5 18 ':cl 'ii'xcl 19 .1xc1 .1d7 20 .1h6 .:re8 21 'ii'f3 1-0 17) 33...'ir'c5! This creates an awkward double threat: 34...ltJf6, trapping the queen, and simply 34...'ii'xe3+. 34 ltJd4 This rescues the queen, but at the cost of giving up the e5-pawn. 34...ltJf6 35 'ii'f3 'ii'xe5 Not only has Black won a pawn, but he also retains a large positional advantage. 36 ':d1 g8 37 'ii'f4 'ii'd5 38.:&1 e5 39 'ii'f5 ':c4 Further material loss is now inevi- table. 40 b3 exd4! 0-1 As 41 'ir'xd5 ':c2+ 42 f3ltJxd5 wins a piece. 18) 29....1xh3! 
234 VISHY ANAND: My BEST GAMES OF CHESS A very surprising move. Ignoring the attacked rook, Black takes a pawn which is defended three times! 30 'ji'xh3 The other lines lead to crushing material loss, e.g. 30 gxf5 'ii g4+ 31 f2'ji'g2+32el 'iixhl+,30':xh3 'iixg4+ 31 ':g3 'ii'xdl + or 30 ltJxh3 'iixg4+ 31 h2 ltJf3+. 30...'ji'e1+ 31 g2 'iie2+ 32 g3 :f3+ 32....:xg5 33 'ji'xh7+ f7 would have been even more devastating, but the move played is more than suffi- cient. 33 ltJxf3 'iixf3+ 34 h2 34 h4 g5+ 35 h5 'iif7+ 36 xg5 'iif6+ 37 h5 'iih6#. 34...ltJxg4+ 35 gl 'iixdl+ 36 g2 'ji'e2+ 37 g3 'iif2+ 38 xg4 h5+ 39 g5 'ji'f6# (0-1) 19) 32ltJxb6! Hedgehogs are not normally run over like thi s ! 32...ltJxb6 33 ltJa5 'ir'a7 34 e5 dxc5 35 bxe5 ltJe8 White also wins after 35....1xc5 36 .1xc5 ltJfd7 37 ':xd7 ':xd7 38 .i.xb6 or 35...ltJfd7 36 cxb6 ltJxb6 37 'iig3. 36 e6 ':b6 36...ltJb6 37 ':bl is similar. 37 ':b1]bb138 ]bb11-0 20) 14 ':d8+! This unusual sacrifice traps the black king in the centre long enough for the active white pieces to mount a lethal attack. 14....1xd8 15 .1xe5 .1d7 Or 15....1e7 16 'iid3 f5 (16...'ii'f6 17 'ji'b5+ leads to mate) 17 .1xe7 xe7 18 .1xf5! 'iib6 (18...exf5 19 'iid6+ mates, while 18...'iib4 19 .1xh7 gives White an enormous at- tack for almost no sacrifice) 19'ii'c3! f8 20 .1e4 e5 (or else 21ltJe5) 21 ltJxe5 .1e6 22 'iia3+ g8 23 'iie7 ':f8 24 .1f3 (threatening 25 ltJd7) with a winning attack. 16 'iid3 5 After 16....1a4 White has a very attractive win by 17ltJd4! (17'ii'd6 .1f6!) 17....1e7 (17....:c8 18 lbc6! mates and 17...f5 also loses to 18 ltJc6) 18ltJc6!! .i.xc6 (18....i.xc5 19 'ji'd8+ ':xd8 20 ':xd8#) 19 .1xc6+ f8 20 'ji'd8+ ':xd8 21 ':xd8#. If Black tries 16.. ..i.b5, then 17 'iid6 wins. 17 e4 'iixe5 17... 'iia4 18 ltJe5 .1f6 19 ltJxd7 0-0-020 .1xb7+ c7 (20...xb7 21 ':bl+ mates) allows a beautiful mat- ing continuation: 21 'ji'd6+ xb7 22 ':bl+ a8 23 ':b5! ':c8 24 'ir'd3! ':c6 25 'ji'f3 'ji'a6 (25....:hc8 26 'ii'xc6+ ':xc6 27 ':b8#) 26':a5 ':b8 27 'ji'xc6+ 'ir'xc6 28 ':xa7#. 18 'ir'xd7+ 8 19 'ji'xb7 Not 19 .1xb7 ':b8 20 .1c6 g6 and Black escapes. 19...g6 
SOLUTIONS After 19....:c8, White finishes by 20 .1c6! 'ii'xc6 21 'ii'xc6 (21':xd8+? ':xd8) 21....:xc6 22 ':xd8+ e7 23 ':xh8 with a clear extra piece. 20 'ii'xaS g7 1-0 Black is too much material down after 21 ':xd8 'ii'xc4 22 ltJd2. 21) 29...ltJxe5! 30 fxe5 After 30 ltJd6 'ii'e7 31 ':xd5 (or 31 fxe5 'ii'xh4+ 32 gl ':c2! 33 'ii'xc2 'iihl+ 34 f2 'iig2+ winning the queen) 31...'iixh4+ 32 g2lbg6 33 ':d4 ':cl Black has a winning at- tack. 30....1f3 31 'ii'd3 31 'ii'f2 .1xdl 32 ltJd6 'ii'c7 33 ltJxc8 'ii'xc8 is similar to the game: Black has an extra pawn and a large positional advantage. 31....1xdl 32ltJd6 .1e2! The key point. 33 'ii'e2 'ii'e7 34 ltJxe8 'ii'xe8 White's king is exposed and his pawns are weak, making Black's task straightforward. 35ltJb6 'ii'c5 36lL\c4 .1bl 37 2 .1d3 38ltJd6 'ii'c3 0-1 White is losing more material. 22) 44..J:f3! Threatening both 45...ltJg4+ and 45...h4. Note that the alternatives 44...ltJf3+? 45 gxf3 ':xf3 46 'ii'g5 and 44...h4 45 .1xfl hxg3+ 46 hl 'ii'xaS 47 'ii'el are unsound. 235 45 .1e2 45 gxf3 ltJxf3+ costs White his queen. 45...ltJg4+ Once again, there are various al- ternatives which fail, for example 45...ltJc4? 46 .1xc4+, 45....:xb3 46 'ii'd5+ and 45....:xg3 46 xg3lbc4+ 47 'ii'f4 h4+ 48 g4ltJe5+ 49 xh4. Black can play 45...h4, but it only draws after 46 gxf3ltJxf3+ 47 .1xf3 'ii'xg3+ 48 hl 'ii'xh3+ 49 gl 'iixf3 50 'iid8+. 46 hxg4 ]bg3 47 hl ? 47 gl would have been a better chance, but 47...'ii'c5+ 48 hl':xb3 gives Black excellent winning pros- pects, for example 49 'ii'a2 'ii'c 1 + 50 h2 'ii'c7+ 51 h 1 'ii'f7. 47...hxg4 Thanks to the position of the king on hI, Black has greater attacking possibilities down the h-file. 48 'ii'xb4 h7! 49 'ii'e4 'ii'e7 50 'ii'd5 ':e3 51 .1xg4 'ii'h4+ 52 .1h3 'ii'el + 53 h2 'ii'g3+ 0-1 23) 42 .1f6! 42 f6 is certainly tempting, but Black can reply 42...ltJce3! 43 fxe3 (43 hxg4 is also met by 43... 'ii'c3) 43... 'ii'c3 44 'ii'xc3 (44 'ii'fl 'ii' xe3+ ) 44...dxc3 45 hxg4 c2 (the bishop on g7 may not be well placed if White cannot actually give mate) 46 ':c7 ':al+ 47 f2 cl'ii' 48 ':xcl ':xcl with a drawn ending. 
236 VISHY ANAND: My BEST GAMES OF CHESS 42... <iti>f8 42...lbd2 43 :g7+ <iti>f8 44 .ie7+ wins White's queen. 43 hxg4 1Wd2 44 1Wxd2 Not 44 1Wxc4? :al+ 45 <iti>h2 1Wh6+ 46 <iti>g31Wf4+ 47 <iti>h3 :h 1 #. 44...lbxd2 45 :d7:as 46 f3ltJc4 47 g5 :al + 48 <iti>h2 :a6 49 :c7 49 g4 would have been simpler as Black's pawns cannot move at the moment. 49...lbe3 50 .ig7+ <iti>g8 51 .ixe5 d3 52 .ic3 White must not try to be too clever: 52 f6 d2 53 g6 hxg6 54 f7+ <iti>f8 55 .if4 g5! 56 .ixe3 :h6+ 57 <iti>g3 dl1W 58 .ic5+ :d6 would be embarrassing for White. 52...d2 53 .ixd2 lbn + 54 <iti>h3 lbxd2 55 f6 White wins this ending fairly easily thanks to the poor position of Black's knight. 55...:al 56 <iti>g4 :gl 57 <iti>g3 :hl 58 f4 :h4+ 59 <iti>f5 :a4 60 :g7+ <iti>f8 61 ]bh7 1-0 24) 23 :xa2! The first sacrifice defuses Black's counterplay... 23...1Wxa2 24lbxh7! ...and the second breaks through on the kingside. 24...lbxh7 25 1Wxg6+ <iti>f8 26 .:xh7 ':xh7 27 1Wxh7 The liquidation has left White with a large advantage. Black's king is permanently exposed while White has sufficient pawns nearby to be relatively safe. White must try to bring his knight up to support his queen. 27...1Wg8 After 27...'iixa4 28lbg4 'iial+ 29 f2 1Wa2+ 30 gl 1Wf7 31 1Wh8+ e7 321Wxe5+ <iti>f8 33 1Wxa5 White has enough material to win, while 27...1Wal+ 28 ltJdl 'iixa4 29 lbe3 1Wal + 30 <iti>f2 gives White a decisive attack. 28 1Wh4 ltJd7 29 ltJg4 1Wb3 30 1Wh6+ e7 31 'iixc6 'iibl+ 32 f2 1Wb2+ The alternative was 32...'iib6+ 33 1Wxb6 ltJxb6 34 .ib5 :c8 35 ltJxe5 d6 36 f4 :xc3 37 g4, although this ending should be a win, for example 37...c5 38 g5 <iti>d4 39 g6 xe4 40 g7 :c8 41 .id3+! <iti>d4 42 .ih7 lbxa4 43lbd7, followed by lbf8. 33 .ie2 :b8 34 ltJxe5! ltJxe5 35 1Wc5+ <iti>e6 36 1Wd5+ 6 37 1Wd6+ 7 381Wxe5 'iib6+ 391Wd41Wxd4+ 40 cxd4 :b4 41 .ib5! ':xd4 Or 41...:xb5 42 axb5 <iti>e7 43 e3. 42 e3 :b4 43 g4 e6 44 g5 :b3+ 45 <iti>f4 :b4 45...:xb5 46 axb5 a4 47 b6 d7 48 g6 a3 49 b 7 <iti>c7 50 g7 will pro- mote with check. 46 .ic6 <iti>d6 47 .idS :bl 48 f5 e7 49 f4 ':al 50 <iti>g6 :xa4 51 <iti>g7 :al 52 f5 :n 53 f6+ <iti>d6 54 g6 1-0 
SOLUTIONS 237 25) 32 .idS! Cutting off Black's defence of the d4-rook. Adams should have recon- ciled himself to the loss of the ex- change and played 32....:4xd5 33 cxdS exdS 34 ':eS, although even here White would retain a large ad- vantage. However, Black decided to defend his rook diagonally... 32...1Wg7? 33 ':e5 ...only to have the defence cut off again! There was an equally good al- ternative solution by 33 ..txe6+ h8 34 ..txfS. 33....:4xd5 34 cxd5 cxd5 35 g4! Opening up more lines for the rooks makes the win easy. 35...ltJe7 36 ':xe6 fxg4 37 'iih4 ':d7 38 ':be1 f8 39 f5 1-0 After 39.. .ltJxfS 40 ':f6+ ':f7 41 ':xfS ':xf5 42 1Wd8+ f7 43 ':e7+ Black loses his queen. 26) By far the simplest solution is 42 ltJef6+! gxf6 43 ':g4+ 1-0, as White will be a clear rook up. 27) 56 ':d1ltJc5 57 d7 Winning a piece, after which the remaining task is purely technical. 57...ltJe6 58 d81W ltJxd8 59 ]bd8 g5 60 ':d7 A neat echo of the earlier combi- nation. The exchange of rooks makes the task simpler. 60...Jbd7 61ltJf6+ g6 62ltJxd7 f7 63 d2 e6 64lbb6 h5 65 h4 f4 66 hxg5 f5 67 ltJc4 xg5 68 ltJd6 e3+ 69 fxe3 h4 70 ltJe4+ g4 71 e2 f5 72ltJf2 1-0 28) 22 ltJb5! axb5 22.. .e5 (22.. ...txb2 23 ltJd6 costs Black his queen) 23 1Wh4 ltJf6 24 ltJc3 is relatively best, but now the win is much simpler for White as he can just drop his knight in on dS. 23 ..txg7 <it>xg7 24 1Wh6+ Cleaning up the kingside before trapping the 1?ishop by taking back on bS. 24...g8 251Wxh7+ <iti>f8 26 'iih8+ e7 27 'iih4+ g5 27...f8 28 axbS leaves White well ahead on material, position and attack! 281Wxg5+ f6 28...ltJf6 29 ':f3 is deadly. 29 ':h7+ d6 30 1Wd2+ c7 31 cxb5 1-0 31.....txe4 32 ..txe41Wxe4 33 ':cl+ b8 34 ':xd7 is devastating. 29) 451Wxe6 1-0, since 4S....:xe6 al- lows mate, while 4S...fxe6 46 ':xh7 xh7 47 ':hl wins the bishop in perfect safety. 30) 14ltJxd4! ..txg2 15 ltJf5 1We6 16 1Wg5 
238 VISHY ANAND: My BEST GAMES OF CHESS Defending the knight on f5 with gain of tempo. 16...lL\e8 17 xg2 h6 18 'ir'g4 ltJdf6 19 'ir'f3 The extra pawn has been consoli- dated and, as a bonus, the bishop on b2 has been activated. The rest is straightforward. 19...e4 20 dxe4 lL\xe4 21 ':fdl h7 22 bxc5 bxc5 23 ':d5ltJ8f6 24 i.xf6 ltJxf6 25 ':xc5 ':ab8 26 ':dl ':b2 27ltJd4 'ir'e7 28 ':b51-0 
I ndex of Opponents Numbers refer to pages. A bold number indicates that Anand was White. Adams 130,222,228 Agdestein 16 Andersson 229 Bareev 81, 110 Beliavsky 54, 136, 225 Benjamin 23, 142 Campora 223 Dreev 224 Ftanik 119 Garcia Palermo 226 Gdanski 221 Gelfand 106, 189, 228 Gerber 223 Gil 221 Hamed 220 Hubner 87 Illescas 226 Inkiov 7 Ivanchuk 92, 102, 207, 223, 229 Izeta 115 Kamsky 98, 148, 157, 164, 171 Karpov 59, 211 Kasparov 67, 74,184 Khalifman 227 Kramnik 229 Kuijf, M. 40 Lautier 216 Levitt 221 Ljubojevic 225 Morovic Fernandez 49 Ninov 12 011 126 Petursson 43 Piket 227 Polgar, J. 153, 198, 229 Robatsch 226 Ruban 224 Salov 225 Sokolov, I. 90 Spassky 34 Tal 29 Thipsay 222 Timman 176,225 Tomczak 222 Topalov 202 Thkmakov 220 Van Wely 227 Ye Rongguang 224 
Index of Openings Numbers refer to pages. A bold number indicates that Anand was White. English 29 French Defence 74, 81, 110 Petroff Defence 87, 102 Pirc Defence 54, 115, 198 Ponziani Opening 40 Queen's Gambit Queen's Gambit Accepted 106, 211 Semi-Slav 59 Slav Defence 126 Queen's Pawn 43 Ruy Lopez 16, 207 Symbols Arkhangelsk 157 Breyer 34 Flohr-Zaitsev 164 Worrall Attack 98, 176 Scandinavian Defence 216 Sicilian Grand Prix Attack 189 Kan 12, 130 Maroczy Bind 49 Najdorf 119,136,148,153,171 Richter-Rauzer 7, 23, 92, 142 Scheveningen 67, 90, 184, 202 + Check !? Interesting move ++ Double check ?! Dubious move # Mate 1-0 Whi te wins , Good move 0-1 Black wins " Excellent move 1/ 2 _1/ 2 Draw ? Bad move (n) nth match game ?? Blunder (D) Diagram follows 
 J - . T Vishy Anand is the strongest chess player ever to come from India, and is currently rated number two in the world, with Garry Kasparov in his sights. His friendly, unpretentious personality makes him extremely popular with chess players around the world. In this, his first book, he presents detailed and entertaining commentaries to forty of his best games. Anand's ability to get to the heart of complex positions, which makes him a deadly opponent over the board, comes over supremely well in his notes, which will amply repay careful study. . Flowing attacking masterpieces - explained logically step-by-step . Anecdotes from top-level events . Practical hints from a world-championship challenger . Game introductions set each game in its practical context . Thirty puzzle positions to test your skill Viswanathan (Vishy) Anand comes from Madras in southern India. Having shot to fame as a child, playing master-standard chess at lightning speed, he became a grandmaster in 1987 after winning the World Junior Championship. By the start of 1991, at the age of 21, he had established himself among the world's elite by winning the Reggio Emilia tournament ahead of Kasparov and Karpov, defeating both in their individual games. In 1995 he challenged Garry Kasparov for the world championship in New York, and although Kasparov won in the end, it was a tough fight, with Anand leading after nine games. Since then, Anand has remained a major threat to Kasparov. Other titles from Gambit Publications: Secrets of Practical Chess John Nunn 101 Chess Opening Surprises Graham Burgess The GAMBIT Guide to the Bogo-Indian Michael Rohde S.T.A.R. Chess Paul Motwani How to Beat Your Dad at Chess Murray Chandler Gambit Publications Ltd is: Managing Director: Murray Chandler GM Chess Director: Dr John Nunn GM Editorial Director: Graham Burgess FM £15.99 $24.95 ISBN 1-901983-00-5 For further information about Gambit Publications, write to us at: Gambit Publications Ltd, 69 Masbro Road Kensington, London W14 OLS, England. Or send an e-mail to:100617.2702@compuserve.com