Eric Schiller - 639 Essential Endgame Positions

CARDOZA PUBLISHING CHESS TITLES
Our philosophy is to bring you the best quality chess books from the top
authors and authorities in the chess \Vorld, featuring words (as opposed to
hieroglyphics), clear explanations (as opposed to gibberish), quality presenta
tions (as opposed to books simply slapped together), and authoritative £nfonna
tion. And all this at reasonable prices. We hope you like the results.
ABOUT THE AUTHOR

Eric Schiller, widely considered one of the world's foremost chess ana
lysts, writers and teachers, is internationally recognized for his definitive works
on openings. He is the author of90 chess books including Cardoza Publishing's
definitive series on openings, World Champion Openings, Standard Chess Open
ings, and Unorthodox Chess Openings-an exhaustive opening library of more
.
than 1700 pages. He's also the author of Encyclopedia of Chess Wisdo1n, Ga1nbit
Opening Repertoire for White, Gambit Opening Repertoire for Black, Complete De-
fense to King Pawn Openings, Complete Defense to Queen Pawn Openings, and mul
tiple other chess titles for Cardoza Publishing. (For listings of all chess titles
published by Cardoza Publishing, go online to www.cardozapub.com.)
Eric Schiller is a National and Life Master, an International Arbiter of
F.I.D.E., and the official trainer for many of America's top young players. He
has also presided over world championship matches dating back to 1983.
OTHER CARDOZA PUBLISHING BOOKS BY ERIC SCHILLER

STANDARD CHESS OPENINGS -The new standard on opening chess play in the 20th century,
references every important opening and variation played- more than 3,000 opening strategies!
The standard reference book necessary for competitive play. A must have!!! 768 pgs, $24.95.
UNORTHODOX CHESS OPENINGS- The comprehensive guide to every unorthodox opening

contains more than 1 ,200 weird, contentious, controversial, unconventional, arrogant and out
right strange opening strategies. Great against unprepared opponents. Great Fun/528 pgs, $24.95.
WORLD CHAMPION OPENINGS - Covers the essential opening theory and moves of every
major chess opening and variation as played by a// the world champions. Learn the insights,
concepts and secrets as used by the greatest players of all time. 384 pages, $18.95
COMPLETE DEFENSE TO QUEEN PAWN OPENINGS- T his aggressive counterattacking rep

ertoire covers Black opening systems against virtually every chess opening except for 1.e4 (in
cluding most flank games), based on the powerful Tarrasch Defense. 288 pages, $16.95.
COMPLETE DEFENSE TO KING PAWN OPENINGS- Learn a complete defensive system against
1.e4 based on the mighty Caro-Kann, a favorite weapon of many great players. All White's op
tions are shown, and a plan is given for Black to combat them all. 288 pages, $16.95.
Get online now to learn more about upcoming titles! www.cardozapub.com

THE ESSENTIAL GUIDE
TO CLOSING OUT AND WINNING GAMES
Eric Schiller
CARDOZA PUBLISHING
ACKNOWLEDGEMENTS
This book is dedicated to all the teachers, players and students who
have analyzed endgames with me over the years . I'd especially like to
thank Ray Keene, Leonid Shamkovich, Vlastimiljansa,Jonathan Tisdall,
Jonathan Speelman, and Alexander Nikitin, as well as Hans Kmoch
and Billy Colias, who are no longer with us to enjoy the fascinating
chess endgames which continue to appear each year.
First Edition
Copyright © 2000 by Eric Schiller

- All Rights Reserved -
Library of Congress Catalogue Card No: 98-74281

ISBN: 1-58042-016-8
CARDOZA PUBLISHING
PO Box 1500 Cooper Station, New York, NY 1 0276
Phone (718)743-5229 • Fax(718)743-8284
Email: cardozapub@aol.com
Web Site - www.cardozapub.com
Write for your free catalogue of gaming and chess books,
equipment, software and computer games.

TABLE OF CONTENTS
I. INTRODUCTION 13
What is An Esssential Endgame? 15
Ho\v to Think in the Endgame 17
How to Usc This Book 18
2. PAWN ENDGAMES 22
King+ 1 vs. King 22
Pinter vs. Georgiev 26
King+ 1 vs. King+ 1 27
Comas Fabrego vs. Baburin 28
Analysis of Gulko vs. Yepishin 28
Study by Richard Reti 29
Adamson: "The Jailor" 31
Kurajica vs. Cifuentes Parada 35
Continuation of Topalov vs. Short 36
Instructional Position 38
Kurajica vs. Cifuentes Parada 40
King+ 3 vs. King+I 41
Analysis of Shirov vs. Almasi 41
Avrukh vs. Gurevich 43
Topalov vs. Short 44
Continuation of Nogueiras vs. Hjartarson 47
De Waal vs. Sadkowski 49
Won�vichit vs. McKay 51
King + 4 vs. King+ 4 51
Hector vs. Speelman 52
Analysis of Karpov vs. Kasparov 54
Analysis of Zeigler vs. Schiller 55
Yudasin vs. Van der Sterren 56
Letay vs. Schubert 58
Ivanchuk vs. Karpov GO
Conclusion of Burmakin vs. Z\jaginsev 62
(36) Yudasin vs. Van der Sterren 63
Sadler vs. Bologan 64
Krasilnikov vs. Popov 65
Analysis of Zeigler vs. Schiller 66
Koneru vs. Sashikiran 68
Analysis of Euwe's Study 69
Karpov vs. Kasparov 70
Burmakin vs. Z\jaginsev 73
.Gowor vs. Beckett 79
Euwe Study 81
3. PAWNLESS ENDGAMES 84
Rook vs. Bishop 84
Leitao vs. Janssen 84
Zetic vs. Kutuzovic 85
Rook vs. Knight 86
Kelecevic vs. Potterat 87
Dolmatov vs. Bologan 88
Olbrich, vs. Krupkova 90
Dunnington vs. Szemok 92
Two Rooks vs Rook and Knight 93
Kasparov vs. Bareyev 94
Queen vs. Rook 94
Neff vs. Seeman 95
Queen vs. Rook and Knight 101
Valle vs. Batista 102
Queen and Rook vs. Queen 102
Moreno vs. Aguilar 102
Queen and Rook vs. Queen and Rook 103
Ivanov vs. Grinsell 103
4. KNIGHT ENDGAMES 104

Knight vs. 2 104
Analysis of Ibragimov vs. Rublevsky 105
Knight+ 1 vs. Knight 105
Andersson vs. Hector 106
Gulko vs. Blatny 107
Study by Reti 108
Knight + 2 vs. Knight 110
Andersson vs. Hector 1 10
Knight+ 2 vs. Knight + 1 111
Bacrot vs. Agdestein 112
Knight + 4 vs. Knight+ 2 112
Larsen vs. Fries Nielsen 113
Honfi vs. Moiseyev 1 14
Knight + 5 vs. Knight + 4 116
Larsen vs. Fries Nielsen 116
Wunnink vs. Schiller 1 18
Knight + 6 vs. Knight+ 6 121
Shirov vs. Almasi 121
Knight + 7 vs. Knight + 7 125
Wiley vs. Mazi 126
5. BISHOP ENDGAMES 13 1
Bishop vs. 2 131
Sahetchian vs. Gurieli 132
Bishop + 1 vs. 1 133
Continuation of Kramnik vs. Shirov 133
Keitlinghaus vs. Atalik 134
Bishop + 1 vs. Bishop 134
From a Study by Crum ( 192 1) 135
Bishop + 1 vs. Bishop + 1 135
Kamsky vs. Shirov 136
Karpov Ljubojevic 137
Bishop + 2 vs. Bishop 139
Continuation of Speelman vs. Rublevsky 139
Kramnik vs. Shirov 140
Aseyev vs. Pigusov 142
Berelovich vs. Schlosser 145
Jussupow vs. Kotronias 147
Continuation of Shipov vs. Goldin 152
Continuation of Rublevsky vs. Spraggett 154
Shipov vs. Goldin 155
Continuation of Tuffery vs. Sinclair 156
D. Gurevich vs. Fedorowicz 158
Continuation of Rublevsky vs. Spraggett 159
Tuffery vs. Sinclair 160
Continuation of de Firmian vs. Fishbein 16 1
Barbero vs. Barthel 162
Krasenkov vs. Kosten 163
Frois vs. Teran Alvarez 164
Speelman vs. Rublevsky 165
de Firmian vs. Fishbein 170
Vaganian vs. Yermolinsky 173
Chan Chee Keong vs. Flores Jr. 175
Zontakh vs. Ilincic 177
Sherzer vs. Wojtkiewicz 180
Bishop+ 5 vs. Bishop + 5 181
Yepishin vs. Lugovoi 18 1
Bellon Lopez vs. Schneider 182
Gleizerov vs. Barua 184
Korchnoi vs. Shirov 187
6. BISHOP US. KNIGHT ENDGAMES 189

Bishop vs. Knight+ 1 190
Analysis of Leko vs. Akopian 190
Bishop vs. Knight+ 2 190
Analysis of Kasparov vs. Timman 19 1
Bishop + 1 vs. Knight+ 3 191
Conclusion of lbragimov vs. Rublevsky 192
Continuation of Sokolov vs. Timman 194
Sokolov vs. Timman 197
Leko vs. Akopian 198
Ibragimov vs. Rublevsky 201
Kasparov vs. Timman 202
Analysis of Kramnik vs. Ivanchuk 205
Analysis of Kramnik vs. Ivanchuk 206
Bishop + 4 vs. Knight + 4 208
Ftacnik vs. Gurevich 208
Bishop + 4 vs. Knight + 5 2IO
Shaked vs. Kobalija 211
Bishop + 5 vs. Knight + 4 2I2
Analysis of Kramnik vs. lvanchuk 212
Bishop + 5 vs. Knight +5 2I4
Dreyev vs. Vaganian 214
Schiller vs. Kahane 217
Schiller vs. Levin 221
Schiller vs. Bisguier 223
Zeigler vs. Schiller 225
Psakhis vs. Aleksandrov 227
Kramnik vs. Ivanchuk 229
l. ROOK ENDGAMES 233

Rook vs. I 233
Analysis of Yakovich vs. Kosten 234
Lopez vs. Alvarado 234
Rook vs. 2 234
Tarrasch vs. Janowsky 235
Analysis of Panchenko vs. Braslavsky 237
Rook + I vs. Rook 237
Lucena Position 237
Philidor Position 238
Karpov vs. Yusupov 239
Kubala vs. Gipslis 240
Nickoloff vs. Kovacevic 241
Schueller vs. Modr 241
Rook + I vs. Rook + I 244
Rook + 2 vs. Rook 245
Karpov vs. Yusupov 246
Gligoric vs. Smyslov 247
Marconnet vs. Colias 250
Somak vs. Meera 251
Rook + 2 vs. Rook + I 252
Olsen vs. Jakobsen 253
Rook + 2 vs. Rook + 2 253
De Firmian vs. Gallagher 257
Rook+ 3 vs. Rook + 2 257
Gulko vs. Yepishin 258
Andersson vs. Hubner 262
Rook + 3 vs. Rook + 3 262
Serper vs. Schw·artzman 263
Nogueiras vs. Hjartarson 264
Continuation of Andersson vs. Hubner 266
Panchenko vs. Braslavsky 267
Andersson vs. Hubner 269
Malada vs. Gara 270
Ribeiro vs. Reis 271
Rook + 5 vs. Rook+ 5 273
Rublevsky vs. Gelfand 274
Kolev vs. Marinkovic 276
Zvereva vs. Rakitskaya 279
Hrvacic vs. Malada 282
Rook+ 6 vs. Rook+ 6 288
Akhmadeyev vs. Nenashev 288
Rook+ 7 vs. Rook+ 6 294
Krasilnikov vs. Popov 294
8. ROOK AND MINOR PIECE ENDGAMES 29'l

Rook vs. Knight + 2 297
Djurhuus vs. Fyllingen 298
Rook+ Bishop + 3 vs. Rook+ Bishop + 4 298
Krasenkov vs. Kosten 298
Rook+ Knight + 2 vs. Rook+ Knight+ 2 300
Maljush vs. Timoshenko 300
Rook+ Bishop + 4 vs. Rook+ Knight + 4 303
Polgar vs. Leko 304
Rook+ 2 vs. Knight + 2 306
Psakhis vs. Smyslov 307
Rook+ 2 vs. Knight+ 5 308
Djurhuus vs. Fyllingen 308
Rook + 3 vs. Knight+ 3 310
Arnason vs. Petursson 310
Rook+ 3 vs. Bishop + 4 312
Yakovich vs. Kosten 312
Rook + 5 vs. Knight+ 6 315
Kostic vs. Glodeanu 315
Rook vs. Two Minor Pieces 319
Pavasovic vs. Ron1anishin 320
Rodrigues vs. Galego 322
Khalifman vs. Ivanchuk 325
9. QUEEN ENDGAMES 326

Queen vs. 1 326
Kavcic vs. Ivanovic 327
Riedner vs. Loebler 327
Queen vs. 2 327
Analysis of Hector vs. Speelman 328
Queen vs. 1 (more) 328
Queen and Bishop Pawn Draw 328
Queen + 1 vs. Queen 329
Van der Wiel vs. Fedorowicz 329
Queen+ 3 vs. 2 332
Babula vs. Mallahi 332
Queen + 3 vs. Queen+ 3 333
Boensch vs. Jussupov 334
Queen + 4 vs. Queen + 3 335
Babula vs. Mallahi 335
I 0. MISCELLANEOUS ENDGAMES 337

Yepishin vs. Lugovoi 337
Analysis of Korchnoi vs. Shirov 339
Ostermeyer vs. Bischoff 342
Toth vs. Muhari 343
Sahetchian vs. Gurieli 345
Rublevsky vs. Spraggett 347
Sokolov vs. Malisauskas 348
Goormachtigh vs. Van Beek 353
Kasparov vs. Kramnik 354
I I. ELEMENTARY ENDGAMES 357

Winning and Drawing: the Rules 357
Checkmate with Queen an d Rook vs. King 358
Scholtes vs. Opponent 359
Checkmate with Queen vs. King 359
Opponent vs. Crispin 359
Checkmate with Rook vs. King 360
Opponent vs. Diehle 360
Checkmate with 2 Bishop s 360
Opponent vs. Milanollo 361
CA RDOZA PU BLISHING • ERIC SCH ILLER
Checkmate with Bishop + Knight 361

Ljubojevic vs. Polgar 362
Rook vs. Pawn 365
Lamorelle vs. Bernard 366
Dunnington vs. Szemok 366
Stalemate 367
Karpov vs. Polgar 367
12. QUIZZES 368

N ogueiras vs. Hjartarson 368
Psakhis vs. Smyslov 369
Tarrasch vs. Jano\vsky 369
Boensch vs. Jussupov 370
Comas Fabrego vs. Baburin 370
Larsen vs. Fries Nielsen 37 1
Wongwichit vs. McKay 37 1
Karpov vs. J. Polgar 372
Goormachtigh vs. Van Beek 372
De Waal vs. Sadkowski 373
Hr vacic vs. Malada 374
Hr vacic vs. Malada 374
Akhmadeyev vs. Nenashev 375
Letay vs. Schubert 375
13. WORKOUTS 37 6
I. A Classic Rook and Pawn Endgame 376
Akopian vs. Georgiev (Preview) 377
Alekhine vs. Capablanca 377
Akopian vs. Georgiev (Analysis) 382
2. Evaluation of the Endgame 386
Wang vs. Mont-Reynaud 386
3. The Endgame and the Op ening 388
Tarrasch Defense- Classical Opening 389
INTRODUCTION
If you've ever had to work out a difficult endgame, you know that
every endgame is essential! There arc hundreds of critical positions,
many of 'vhich cannot be easily solved at the chessboard. This book
contains endgames drawn primarily from recent games by top players .
A few i mportant composed stu di es, and examples of elementary
endgames from amateur games have also been included. If you study
each of these endgames carefully, you will g reatly elevate your endgame
skills in no time at all !
Why should anyone devote so much time and energy to positions
with so few pieces on the board? The answer is simple: the endgame
usually determines the fate of the game if no middlegame attack has
succeeded and neither player has blundered signficiant material. A
mistake here can undo all great opening and middlegame accomplish
ments. In so many cases, a small advantage can be converted into a win
only by applying considerable endgame technique.
The prolifer ation of "sudden death" time controls makes endgame
knowledge more valuable than ever. Often, you just don't have time to
work things out, counting and then recounting and checking all lines .
Sometimes, you just need to let your instincts gu ide your hand. In these
cases, your experience and knowledge of endgame theory will have a
great impact on the result of the game.
The variety of endgames can be intimidating. There is so much to
learn, that even a book on basic chess endgames runs hundreds of
pages . Studying endgames is difficult work, and I've tried to help by
decomposing complicated endgames into simpler ones . General ly,
'vhenever a piece or pawn is captured, resulting in a different type of
ending, the game is continued in the section that covers the newly
created configuratio n. A section on rook and three pawns vs. rook and
four pawns will lead to a different section on rook and three pawns vs .
rook and two pawns if a pair of pawns are exchanged .
13
CARDOZA PU BLISH I N G • ERIC SCH I LLER
O nly a small percentage of chess endgames have been definitively

solved by chess computer programs. All endgames with a total of five
pieces (including the two kings ) were solved by Ken Thompson and
have been available for a few years . Six piece endgames are not yet
available, however, and endgames with a dozen or more pieces, very
common in tournament play, will not be solved in the near future .
If computers have so much difficulty in the endgame, how can
humans be expected to master the art? Fortunately, we humans have an
advantage. Our thought process provides a lot more flexibility in pat
tern matching than computers are presently capable of. We immedi
ately reject many candidate moves and never give them a second thought.
In the endgame, as countless composers have reminded us, this
efficient search mechanism has one big flaw. Because relatively fewer
candidate moves leave pieces in a position to be captured, a higher
percentage of our candidate moves deserve consideration. Few humans
will expand the range beyond six or seven moves after looking over the
position for a while . We choose moves that will advance our goals, but
we can miss some indirect paths .
The endgames in this book range from triv ial checkmates to mind
boggling puzzles. The positions are organized according to the mate
rial present at the start of the discussion. In complex endings a capture
may result in the analysis being continued in a different section. In this
way I have been able to avoid the distraction of sidelines which are not
necessari ly irrelevant, but which teach les sons other than the one at
hand.
The vast majority of the examples are from recent games. I 've in
cluded a couple of uniquely valuable old chestnuts as well as several
composed studies, but for reasons discussed elsewhere in this intro
duction, I have preferred actual games which show typical patterns
and errors, rather than spectacular insights and tactical combinations.
The unfamiliarity of most of the positions will make the book more
interesting to advanced players. In a few cases, I deleted the name of
the loser in an elementary checkmate. Most of those were played by
children and while the one who successfully navigated the endgame
should deserve a pat on the back there is no need to embarrass the
loser. I think you ' ll find a lot of new and fascinating positions, and if
you learn nothing else, you'll realize how easy it is for even the stron
gest Grandmaster to make a critical mistake.
Before you start lowering your opinions of Grandmaster endgame
play, make sure you consider the circums tances of the game. When
ever possible, the actual move n;umber of the game is used. Most games
14
INTRODUCTION
are played under the time control of forty moves in two hours follo,ved
by twenty moves per hour and then sudden death, and so metimes the
sudden death takes place right after move forty. Errors made bet\veen
moves 32 and 40 can often be the result of time pressure. Moves 56-60
can also be treacherous . Be kind, assume that time pressure \vas re
sponsible for many of the mistakes .
I had originally intended to discuss 365 positions, so that a dia
g ram a day \Vould lead to absorbing the entire book in a year. In the
end, I \vould up \Vith over 600 positio ns to discuss and decided that
rather than reduce the number of diagrams, which would make the
book harder to follow, I'd just keep them all. In any case, the different
endgame skills possessed by readers \Vil l vary greatly, and it is difficult
to parcel out the material for all of them. If you devote 15 minutes of
study each day, I think you can achieve the same results and after one
year you will see great improvement in your endgame play.
WH AT IS A N ESSSENTI A L EN DG A M E?
This book is intended to help you achieve g reater practical results
at the chessboard . You can't learn all of endgame theory from a refer
ence book. In fact, it is probably impossible to learn all of what \Ve
kno\v about the endgame. Databases of hundreds of thousands of posi
tions with minimal material are fine for computers, but a human being
must rely on a combination of three factors:
• specific endgame knowledge derived from experience.
• specific target positions acquired through study.
• general principles (and their exceptions! ) .
Each time you play an endgame you gain valuable experience. The
analysis that you do at the board can be very instructive, and if you
analyze each game after \vard, you can further enrich your understand
ing. But experience alone is not enough .
You need to kno\v certain positions to use as targets for your analy
sis . How many positions do you really need to know to play effectively
in tournaments? Oh, perhaps a few thousand, maybe a bit more. Clearly,
going through encyclopedic references and databases isn't going to get
the job done quickly. Which positions do you need to kno\v, and \Vhich
can you leave for later study? The learning process can be quite com
plex.
The task is made much easier by applying general principles , such
as healthy pawn structure, active pieces, safe king, etc . These consider
ations must be combined with accurate tactical analysis and kno,vl
edge of a fe,v dozen basic positions for the casual player, to perhaps a
15
CARDOZA PU BLISHING • ERIC SCH I LLER
few hundred for the tournament competitor.

There are several diffe rent approaches to presenting endgames .
Some books confine the mselves to positions i n isolation, mostly with
very few pieces on the board. This eliminates clutter, but many stu
dents find it hard to integrate this material into the decision making
process of the early endgame, when choice of piece captures and pawn
advances are critical .
In our examination, the positions are drawn from real tournament
games, mostly quite recent ones so that the positions will be relatively
unexplored in the existing literature . There are some composed stud
ies, especially of positions with minimal material, which are used to
illustrate fundamental ideas. However, I 've generally avoided using
compositions unless necessary.
Practical examples are generally considered inferior to composed
studies because they are often flawed. From an artistic standpoint, there
is no doubt that the endgame study is the highest form of chess purity.
There are no spurious elements, single solutions to each problem, and
usually an underlying tactical trick. To appreciate the chess endgame
fully, study of compositions is needed. Our concerns, however, are more
immediate .
Even the greatest Grandmasters make mistakes in the endgame,
and you'll find plenty of them in our examples. Each error, whether it
is committed by an absolute beginner or an experienced professional,
is instructive . In fact, I 've even included an index of mistakes at the
b ack of the book, because studying these errors can help you avoid
them.
While each piece in a composed study is essential, in real games
there are all sorts of extra pieces which may or may not have an influ
ence on the overall strategy. Your task is to evaluate the position, tak
ing into consideration the possibility that some features may be irrel
evant. In a study, you know there is a solution. In the game, you have
no idea whether the position can be solved or not. Seeing how real
players cope with real situations is closer to the experience of most
amateur, and even professional players .
So, what makes an endgame "essential" for the student? The most
important practical considerations are endgames with elements that
display general rules, both in pure form and as a combination of con
flicting situations. When general principles come into conflict, it pro
vides an opportunity to observe the true dynamics of the endgame . A
strong tactician will solve many positions by brute force, and a weak
tactician will have poorer results .
16
I NTRODUCTION
Someone whose tactical skills aren't quite up to the task of calculat

ing complicated endgame combinations may think that studying com
positions will help, but the real problem lies in improving tactical abil
ity in general. Learning specific tactical tricks in the endgame is not
very efficient, because the situations you encounter will only very rarely
involve those individual tactics .
The more time you spend on the endgame, the better your results
will be. Nevertheless , you want to devote your energy to the type of
study that will bring you the best results. The endgame collection in
this book places you in a great variety of situations , and the positions
often involve a number of pieces and pawns , with several different
themes . The coverage is broad.
For detailed analysis of e ach type of position, you may wish to con
sult more specialized literature . A comprehensive book on simple
endgames with a piece for each side and a single pawn would ru n well
over a thousand pages, judging from the size of individual books you
can find with hundreds of pages devote d to , for example , Rook + 1 vs .
Rook. So use this volume as a s tarting point, o r as a continuation, of
your j ourney through the jungles and deserts of endgame play. Once
you have the "essentials" down, there is plenty more to explore !
HOW TO TH I N K I N TH E EN DG A M E
O nce the game' s major battles have died down, you have to acljust
your thinking to the new circumstances presented in the endgame. Your
goal is no longer a checkmating attack against the enemy king, in most
cases . First you mus t promote a pawn and gain a new queen. Circum
stances may give you the luxury of allowing your opponent to get one,
too , but in that case you will need a material advantage or the initiative
to gain a win .
The j ourney to the promotion square is a dangerous one. Each step
up the board moves the pawn further and further from its home base,
and it becomes harder to defend. As it works its way through the en
emy camp, the pawn must avoid capture , even if an enemy piece is the
compensation. The easiest way to draw a game where your opponent
has a single remaining pawn often involves sacrificing a piece for the
pawn. No paw·n-no new queen.
Calculation is a bit different, too . Instead of aiming at a single tar
get, the enemy king, you are trying to obtain one of the many positio ns
that you know to be a win against best defense. Experience is impor
tant. The more endgames you play and analyze , the more likely you are
to avoid letting a win slip from your hands . Sometimes you just need to
17
CARDOZA PU BLISHING • ERIC SCH ILLER
follow your intuition, mov ing pieces to squares where they will be more
useful, hoping that your personal muse will remind you when the ap
propriate patterns become relevant.
In a few lucky cases , we can reduce the calculation to simple count
ing. More often, we need to rely on some " rules of thumb" . There are
some geometry tasks we must execute perfectly, knowing, for example,
when our king is in the square of the enemy pawn. If we have knights
on the edge of the board we need to watch for traps and make sure that
their mobility is not restricted by enemy pieces. In rook endgames we
need to remember which operations are successful on the "long side"
(4 files or ranks to edge) and which can be effective on the "short
side" (3 files or ranks to edge ) . When we have a bishop , we must keep
in mind whether an enemy bishop is of the same color or opposite, and
adjust strategy accordingly.
You'll find examples of all of these ideas in the col lect io n of
endgames in this book. By combining general principles with precise,
deep calculation you may be able to solve all the positions, but more
likely you'll find that you took at least one false step, at least in the
more compl icated examples .
This is not a " rate yourself" book. This is a " raise your rating" book.
Even the strongest players have weak spots in the endgame, and are
constantly working to improve . Sometimes Grandmasters miss elemen
tary wins , as you'll find out in a number of the endgames . The goal is
to take note of the positions you didn't solve correctly. Perhaps it was
just a tactical oversight. Don't worry too much about that, just don't do
it again! If, however, you failed to make a proper strategic choice, espe
cially regarding a pos sible exchange, then make sure you study the
other positions in the same chapter. Then you will avoid making a simi
lar error in tournament play.
HOW TO USE TH IS BOOK

This book is intended for used by readers with a variety of skill
levels and experience. More advanced players will find some of the
endgame examples to be triv ial. Some of the endgames are so compli
cated that even Gr andmasters fail to play them properly. I'm going to
prov ide some adv ice here fo r several different types of readers so that
each may get the most out of the endgame experiences presented here .
The endgames have been categorized according to the material on
the board. In a complicated endgame, exchanging material may lead
to an endgame which is discussed in a separate section devoted to that
type of ending. The coverage does not split each time pieces are ex-
18
I NTRODUCTION
changed, because often the resulting positions are very simple and
straightforward, or do not show typical characteristics of the reduced
e ndgam e . Whe n p a,vns are p ro m ote d to que e n s the re m ai n i n g
endgames are usually brief and not of significant ins tructional value .
A special index at the back of the book together with page references
make it possible to play through a complete endgame extrac ted from a
tournament game.
However, I recommend that all readers spend the few moments at
the position where an endgame jumps to another page and try to work
out a continuation . You may wish to use a chess computer to analyze
the positions as you study, but please remember that computer pro
grams have a lot of difficulty in complicated endgames. Often it is
necessary to look very deep into a position. This is something humans
are better at because we concentrate on what we consider the only
plausible continuations . It is true that we often overlook resources along
the way but in the vast majority of cases, we ge t to the correct final
position of at least one major variation. Still, because you have the
analysis in front of you as you work, it may be constructive to compare
the contents of the book with the opinions of your electronic counse
lor. Sometimes your compute r may be able to improve on the analysis
given here . After centuries of work, there are still mistakes being un
covered in a great deal of traditional literature .
Beginners should start by studying the chapter on elementary check
mates and then move on to the first few examples in each chapter.
C oncentrate on the endgames where the defender has just a single
pawn fo r no pawns at all. Do not attempt to solve the quiz positions
until this task has been completed. Pay special atte ntion to the final
fe,v moves of each example as the finishes must be fully understood in
order to master the basics of endgame play. Then return to the quiz
section and try to solve each one . It is best if you work without a chess
board, but if you find this difficult, use a chess set. Do not use a chess
computer for these positions.
When you have finished, and refreshed your understanding of po
sitions you misevaluated, then go to the workout section. Look for the
bes t move, but try to unders tand the actual move played. These are
positions for study, your job is to find out the truth! Using a chess
computer can be valuable here, if only to remind you how difficult
endgames are . You will find that even the strongest programs cannot
find solutions in a reasonable amount of time. I used the latest version
ofJunior(5.0) while working on this book. It is truly amazing how the
computer was still pursuing a false path even after 30 ply of analysis !
19
Those of you who already know the basics can quickly pass through
the elementary sections, but you should take a look at each starting
diagram and make sure that you know how to bring the game to a
succes sful conclusion. Try a few of the quiz positio ns right away and if
you correctly solve most of them, move on to the more complicated
endgames and follow them through all of the relevant branches which
refer to other sections . I suggest that you refrain using computer pro
grams while you study because you may get lost in details and miss the
most important conceptual points. Try to go through the boldface moves
of each example at least once without bothering to examine any of the
variations. Try to create a picture of the flow of the endgame by de
scribing the ending in three or four sentences .
If you are already a good enough player that you can solve all of the
quiz questions, start with the workout chapter. This contains positions
from the main lines and notes of the endgames included in this book.
Try to solve each o f them without the use of a chessboard . Treat each
position as an adjourned game with a lot of money at stake . Write out,
or, better, enter into a compu ter database, an exhaustive analysis of
each position and then compare it with the annotations I have pro
v ided. And please let me know if you find any mistakes I may have
made. When it comes to the endgame, neither man nor machine is
perfect. As I s aid before , you'll f ind plenty of examples of the greatest
players misplaying endgames, and modern analysts are still f inding
errors in even the most established endgame references .
All the remarks above refer to study by a single individual . In fact,
it is much easier to learn endgames by working with a coach or trainer.
The problem with working alone is that after you have come to some
conclusions regarding a position and look up the answer to f ind that
you have made an error, you are simply absorbing a reference rather
than continuing to work for yourself. A trainer can help by taking charge
of the book and tolling out only the minimal hints needed to encour
age you to f ind the correct solution. The trainer doesn't even have to
be a particularly good endgame player as all of the needed material is ,
I hope , included in the analysis . If you have a friend who is of approxi
mately equal strength, you can both improve your results by taking
turns acting as trainer and student.
I ' m going to close this section with a tip from the training camp of
Garry Kasparov which I got from his trainer Alexander Nikitin in the
early 1 980s. He explained to me that it was important not only to find
the best move in a practical endgame, but also constructive to find the
move actually played in the game , assuming that the player involved is
20
I NTRODUCTION
a famous endgame technician . He used many of the endgames of such

endgame wizards as Capablanca and Lasker. The idea is that you want
to be able to think like a great endgame player. Even if that player
sometimes makes a mistake , it is especially useful to s tudy those posi
tions and try to figure out what the player had in mind or what tactic
or strategy was overlooked.
However you use these book to improve your endgame knowledge,
I hope that you will also receive pleasure from the beautiful tactical
·combinations and the positional strategies that are found in these fas
cinating endgames . Endgame study is generally considered ted ious ,
but I hope that I have presented the material in a way that will help
keep your mind stimulated and focused, and that the experience will
be pleasurable as well.
21
PAWN ENDGAMES
Pawns are the weaklings of the opening and n1iddlegame, and re
main fragile in the endgame. As the other pieces drop off the bo ard,
however, the importance of the lowly foot soldier increases as it be
comes more and more possible for the pawn to march across the board
and attain promotio n to a more powerful piece .
You tnight think that pawn endgames are j ust a matter of counting,
and in a sense that is true, but the calculations can be complex and it is
much easier to rely on a fe,v fundamental guidelines built on the pat
terns you will see in the sections below. Pawn endgames require ex
treme accuracy, in most cases absolute perfection. A tiny error can
turn victory into defeat. You ' ll see this happen many times in your O\Vn
gan1es, and even in the games of the world's greatest players .
The following sections discuss many different con1binatio ns of
pawns, from positions with just a single pawn to endgames with seven
pawns each! Among then1 you'll find almost all of the strategic and
tactical tricks commonly found in these "simple" endgames . Study then1
carefully, because the art of endgame play often reduces to decisions
about exchanging pieces leading to king and pawn endgames . In later
chapters you will discover that mistakes are often made at these critical
junctures. Usually these errors are not the result of short-tern1 tactical
oversights but of a failure to understand the true nature of the reduced
endgame .
KING + I US. KING

When strong players compete, you don't see too many endgames
with just a single pawn. Since both players kno\v the inevitable result,
the game is usually resigned or drawn before this stage is reached, or
as soon as the king and pawn endgame is on the board. There are two
basic concepts that must be remen1bered: the square of the king and
the opposition.
22
PAWN EN DGA M ES
The square of the king determines whether a pawn can be pre

vented from promoting when it races unassisted against an enemy king.
a b c d e f g h
[ 1] Draw with either side to move.
There is no Black king in this diagram because the point is to mea

sure the power of the White king agains t a pawn. We can draw a square
around the b l -b6-g6-g 1 s quares. As long as the White king is inside the
square , the pawn can be stopped. Using the rule of the square rather
than counting can reduce errors .
The opposition is somewhat more abstract, but is essential to proper
endgame play. The following discussion is adapted from my Encyclope
dia of Chess Wisdom.
The opposition refers to the geometric relationship between the
two kings . The kings act like sumo wres tlers , trying to get the enemy to
move aside . The key position is the following:
a b c d e f g h
[2] Kings in opposition
White ' s goal is to make progress up the board. The mighty king
will try to understand the concept of pawn promotio n by reaching the
23
CARDOZA PU B LIS H I N G • ERIC SCH ILLER
8 L11 rank. The first step is to cross the meridian to the fifth rank . This
can be accomplished here because it is Black's turn to move . White has
the opposition, since any move by Black must give ground and allow
the White king to move forward . If it were White to move, Black would
have the opposition.
When the kings are close to each other, as in the previous diagram,
this is fairly obvious . The same rule holds when the kings are far apart,
but only when there are an odd number of squares between them .
a b c d e f g h
{31 Kings in opposition
With White to move , Black has the opposition. If White plays l . Kd4,
Black plays l ....Kd6 . O n l .Ke4 there is l . . . Ke6 . l . Kf4 allows l . Kf6 . In
. .
each case we get the simple opposition position, with White to move .
White can move along the third rank, but the Black king will still make
progress, for example l.Kf3 Kd6. But not l . . . Kf6??; 2 . Kf4 when White
has the opposition!
2 .Ke4 Ke6 ! Moving White's king back another rank, the opposi
tion magically returns to White! There are now an even number of
squares, four, between the two kings .
3
2
a b c d e f g h
{41 Kings in opposition
24
PAWN EN DGA M ES
If it is White ' s turn, then l .Kd3 seizes the opposition. If Black plays
l ...Kd6, then 2.Kd4 reaches a simple oppositio n positio n. The same
result follo\vs l ... Ke6 ; 2.Ke4 .
Unfortunately, mos t endgame battles d o not take place on a single
file. Long dis tance and short distance belly bumping can take place
any,vhere. Wherever the t\VO kings are , dra\v an imaginary rectangle
using the kings as the corners .
a b c d e f g h
[5] Diagonal Opposition
Imagine a line dra\vn from c3 to c7 to g7 to g3 . That \vould form a

square .
3
2
a b c d e f g h
[6] Square of opposition
When the shape is a square , whoever is not on the move has the
oppositio n. If White is to move in the diag ram above , then Black has
the opposition, and v ice versa.
There is even a more general rule: if the four corners of the rect
angle are of the same color, then the side that is not on the move has
the oppositio n. Since c3 , c7, g7 and g3 are all dark squares, the rule
-
holds. The same would be true for a position with kings at c2, c6, e2
25
CARDOZA PU B LISH ING • ERIC SCH I LLER
and e6. Try to picture that without a chess board. Are all the squares of
the same color? If so, then wherever you put the kings , if it is White to
move, Black has the opposition.
This rule is very important, especially when we get to king and
pawn endgames . Let' s look at one final example:
a b c d e f g h
[7] Rectangle of opposition
White grabs the opposition with l .Kf4! If you are not convinced,
set up the position on a chessboard and try it out. You can't give it to
your chess computer, because technically the position is a draw accord
ing to the rules, and most ches splaying programs won't touch it!
PINTER VS. GEORGIEV

Recklinghausen, Germany, 1998
3
2
a b c d e f g h
{ 8] Black to move
This is a classic winning position. The White king will move in

front of the pawn on the g-file, securing the win .
74 ... Kh5; 75 .Kf5 Kh6. 75 . . . Kh4; 76.g4 and the Black king i s forced
forward, and cannot keep up with the pawn. 76.g4 Kg7. 76 ... Kh7; 77.Kf6
26
PAWN EN DG A M ES
Kh8; 78.Kg6 Kg8 ; 79.g5 Kh8 ; 80. Kf7 Kh7 ; 81. g6+ Kh8; 8 2 .g7+ Kh7 ;
83.g8Q+. 77.Kg5.
a b c d e f g h
[9] Black to move
Once the king is in front of the pawn, the opposition is in his hands .
77 . . . Kg8; 78.Kg6. Black resigned. Whichever way the king steps , White
moves in the opposite direction, seizes the 7th rank and then the pawn
advances without interruption.
KING + I US. KING + I

When each side has one pawn, the opposition also comes into play.
If the pawns are up against each other on the same file, then the oppo
sition applies just as it did in the previous section. A fter all, your goal is
to get at the enemy pawn, which is the square in front of your own
pawn. That is the target when you have a pawn against a bare king, so
the technique should be familiar. Once you win the pawn, you may or
may not be in a winning endgame, depending on whether you can
keep the opposition .
27
CARDOZA PU BLIS H I N G • ERIC SCH ILLER
COMAS FABREGO VS. BABURIN

Capablanca Memorial, Havana, 1 999
a b c d e f g h
[ 101 Black to move
After
49 . . . Ke4; 50.Kg3 Ke3 ; 5 l .Kg2 Kxf4; 52.Kf2 and the game
was agreed drawn.
ANALYSIS OF GULKO VS. YEPISHIN

Chicago International, 1998
a b c d e f g h
[ 111 White to 1nove
This is a win for White regardless of who is on the move. 1 4.Kd5

K£3; 1 5 .Kd4 would put Black in zugzwang immediately, but even when
it is Black's turn, defeat cannot be avo ided. This maneuver is known as
triangulation. Instead of moving to a square directly, the king makes a
triangular journey (here Kc4-d4 v ia d5 ). Triangulation is only possible
for the king and queen, but the mighty queen rarely needs to make use
of it. In pawn endings , it is one of the principal weapons of combat.
1 3 . . . Kg3. 1 3 ...Kf3; 14.Kd4 is the familiar zugzwang. 14.Kd5 Kf3;
15.Kd4.
28
PAW N EN DG A M ES
a b c d e f g h
[ 12] Black to move
The goal is reached. Black loses the pawn and the game. 15 . . . Kg4;
16.Kxe4 Kg5; 1 7 .Ke5 with a simple win .
STUDY BY RICHARD RETI

1922
a b c d e f g h
[ 13] White to move
White cannot catch the advanced h-pawn. Therefore the only pos
sible defense must involve the c-pawn. Black threatens to gobble it up
after . . . Kb6, and the king is too far away to help . So all seems lost. The
solutio n involves bringing the king to the d-file to protect the pawn.
That takes four moves, which seems impossible .
1 .Kg7. White's first move is obvious enough since l . c7?? loses to
Kb7 . 1 h4. l . Kb6; 2 . Kf6 transposes to the main line . 2.Kf6 Kb6.
... ..
29
CA RDOZA PU B LISH ING • ERIC SCH I LLER
2 . . . h3; 3.Ke 7 h2 (3 . . . Kb6 ; 4 . Kd7 h2; 5.c7 h l Q; 6 . c8Q is the same as

the main line ) 4 . c 7 Kb7; 5.Kd7 hlQ; 6 . c8Q+ with a draw.
a b c d e f g h
[ 14] Black to move
3.Ke5 ! h3. 3 . . . Kxc6; 4 . Kf4 h3; 5. Kg3 h2 ; 6 . Kxh2 leaves the land
scape bare .
a b c d e f g h
[ 15] White to move
4.Kd6!
a b c d e f g h
[ 16] Black to move
30
PAWN EN DGA M ES
4...h2 ; 5.c7 h 1 Q; 6.c8Q and each side has a queen, with the result
that the game should be drawn.
ADAMSON: "THE JAILOR"

Study, 19 1 5
This study takes 2 0 moves to reach checkmate, with a lot o f varia
tions . Obv iously trying to work this out by brute calculation at the board
is very difficult. If you know which kingside configurations are a win
and 'vhich draw, you can reduce the amount of effort considerable ,
since you only need to work out enough o f the solution to reach one of
the target positions you already know.
a b c d e f g h
[ 17] White to move
A deceptively simple position . In fact, it is a very complex position .

If the Black king becomes active , a draw is likely. White cannot just go
running after the h-pawn , but must carefully keep the enemy monarch
at the edge of the board.
1.Kc7! l . Kd7? Kb 7 ; 2 . Ke6 Kc6 ; 3. Kf6 Kd6 ; 4 . Kg6 Ke5! and now
White has to worr y about . . . Kf4-g3, but if 5.g4 then 5 . . . Kf4 draws .
1 ... Ka6. There is very little point in going to a8 , which just loses
time . l . . . Ka8 ; 2 . g4f Now the Black h-pawn cannot advance, and Black
has only a single move . 2 . . . Ka7 . Here Black does not have the possibil
ity in the note on l . Kd7? and therefore White can win straightforwardly:
3.Kd6 Kb6 ; 4.Ke6 Kc7 ; 5. Kf6 Kd8 ; 6 . Kg6 Ke8 ; 7 . Kxh6 Kf8; 8 . Kh7 and
the g-pawn marches up the board . 1 . . . h5? is never a good idea in such
positions, because it brings the pawn closer to the White g-pawn. This
theme is echoed below in a different form.
Black's hope of defense in the l . Kd7 line relied on being able to
bring the king close enough to the g-pawn so that when the White king
approaches the pawn at h5, it can slip in and go after the remaining
pawn. White now wins with the direct 2 . Kd6 Kb6; 3.Ke6 Kc6; 4 . Kf5 .
31
CARDOZA PU B LISH ING • ERIC SCH I LLER
a b c d e f g h
[ 18] Black to move
. 4 . . . Kd5 (4 . . . Kd7 ; 5 . Kg5 Ke6; 6 . Kxh5 Kf6; 7 . g4 transposes . ) 5 . Kg5

Ke6 ; 6 . Kxh5 Kf6; 7 . g4 Kg7 ; 8 . Kg5 . White has the opposition and wins
as follows: S . . . K£7; 9 . Kh6 Kf8; 1 0 . Kh 7 K£7; l l .g5 and the pawn cannot
be s topped. 2.Kc6!
a b c d e f g h
[ 19] Black to move
This is a very important move . Black's king is not allowed to slip on

to the b-file . The king cannot return to a7, because then it is too far
from the g-pawn, so it mus t move to a5 . Advancing the h-pawn loses by
the same method as l . . . h5 ?
2 . . . Ka5. 2 . . . Ka7 ; 3.g4 ! and the White winds up with the opposition
after 3 . . . Kb8 ; 4 . Kd6 Kc8 ; 5 . Ke6 Kd8; 6 . Kf6 Ke8 ; 7 . Kg6 Kf8; 8 . K.x h6
Kg8; 9 . g5 . The rest is forced. 9 . . . Kh8 ; (9 . . Kf7; IO.Kh 7 ) IO.g6 Kg8; l l .g7
.
3.Kc5 . White escorts the Black king up the board. If Black retreats,
the win is achieved by the same method as in the prev ious note . We
mus t think in terms of strategy here , as the winning position still lies
far in the future.
32
PAW N E N DG A M ES
The plan now is to bring the Black king to the second rank, after
which an eventual g4 will put the pawn far enough forward so that the
enemy king cannot catch it. 3 . . . Ka4; 4.Kc4 Ka3; 5.Kc3 Ka2; 6.Kc2.
a b c d e f g h
[20] Black to move
Black is now on the horns of a dilemma. Retreating runs into a

familiar scenario but going into the corner loses because the king is
too far away.
6 . . . Ka3. 6 . . . Ka l ; 7 . g4 Ka2 ; 8 . Kd3 Kb3 ; 9 . Ke4 Kc4 ; 1 0 . Kf5 Kd5 ;
l l . Kg6 Ke6; 1 2 . Kxh6 K£7; 1 3 . g5 Kg8 ; 1 4 . Kg6 ! ( 1 4.g6 ?? Kh8; 1 5 .g7+
Kg8; 1 6. Kg6 is s talemate . ) 14 . . . Kh8 ; 1 5 . Kf7 Kh7 ; 1 6 .g6+ and wins . 7.g3 !
Precision is required! 7 .g4? allows the king to get back and maintain
the opposition . 7 . . . Kb4 ; 8.Kd3 Kc5 ; 9 . Ke4 Kd6; l O .K£5.
3
2
a b c d e f g h
[21] Black to move
Things seem to be under control , but Black has a surprising re

source ! 1 0 . . . h5 ! White can either push the pawn or capture . l l . g5 h4;
1 2 . g6 ( 1 2 . Kg4 Ke6 ; 1 3 . Kxh4 Kf7 ; 1 4 .Kh5 Kg7; 1 5.g6 Kg8; 1 6 . Kh6 Kh8;
1 7 . g7+ Kg8 ; 1 8 . Kg6 is our stalemate again . ) 12 . . . h3 ; 1 3 . g7 h2; 1 4 .g8Q
h l Q is a draw. l l .gxh5 also fails to win: l l . . . Ke 7; 1 2 . Kg6 Kf8; 1 3 . Kh7
33
CARDOZA PU BLISH I N G • ERIC SCH I LLER
( 1 3 .h6 Kg8 ; 1 4 .h7+ Kh 8 ; 1 5 . Kh6 with another s talemate . ) 1 3 . . . Kf7;

1 4 . Kh8 Kf8 and the White king cannot get out of the corner.
7 . . . Kb4. 7 . .. Ka2 lets White win by advancing the g-pawn. 8.g4 ! Ka l
( 8 . . . Ka3 ; 9 . Kd3 Kb3; 1 0 . Ke4 Kc4; l l .K£5 Kd5 ; 1 2 . Kg6 Ke6 ; 1 3 . Kxh6
K£7; 1 4 .g5 Kf8; 1 5 .Kh7 should be familiar by now. ) 9 . Kd3 Kb2 ; 10.Ke4
Kc3 ; l l . Kf5 Kd4; 1 2 . Kg6 Ke5 ; 1 3 . Kxh6 Kf6 ; 1 4 .g5+ K£7; 1 5 .Kh7 with
the usual inevitable result.
8.Kd3. Both sides move closer to the kingside . 8 . . . Kc5; 9.Ke4 Kd6;
IO.Kf5 Kd5.
a b c d e f g h
[22] White to move
White is winning, but must not be hasty. Going directly after the h
pawn allows Black to slip away with a draw.
l l .g4! l l . Kg6 Ke4 ! ; 1 2 .g4 ( 1 2 . Kxh6 K£3) 1 2 . . . Kf4 ; 1 3 . Kh5 Kg3 with
a draw. l l . . . Kd6; 1 2.Kg6 Ke6; 13.Kxh6 K£7; 14.g5 .
a b c d e f g h
[23] Black to move
White has achieved the goal, and the pawn cannot be stopped.
14 . . . Kg8; 1 5 .Kg6 Kh8; 1 6 .Kf7 Kh7 ; 1 7.g6+ Kh8 ; 1 8.g7+ Kh7;
19.g8Q+ Kh6 ; 20.Qg6#.
34
PAWN EN DGA M ES
KING + 2 US. KING + I

It is hard to generalize about endgames with two pawns against
one, because much depends on where the pawn are . The side with the
extra pawn would like to have the pawns as far apart as possible, mak
ing it harder for the enemy king to advance. Since pawns cannot move
horizontally, a player is pretty much restricted to playing the cards that
are on the table. The main skill required is knowing which king and
pawn vs . bare king endgames are winning, so that a pair of pawns can
be exchanged appropriately.
KURAJI CA VS. CIFUENTES PARADA

Malaga Open, Spain, 1 998
a b c d e f g h
{24] Black to move
73 . . . Ka8; 74.b5 Kb8; 75.a3! It is important for White to insure that

the Black king is on a8 when the breakthrough at b6 takes place . 75.a4
Ka8 ; 7 6 . a5 Kb8 ; 7 7 .b 6 Ka8 ! ; 7 8 . Kb5 ( 78 .b7+ is a draw. ) 78 . . . Kb7 ; 7 9 . a6+
Kb8 ; 80.b7 Kc7 ; 8 1 . Kc5 Kb8 is a draw. 82 . Kc6 is stalemate . 75 . . . Ka8;
76.a4 Kb8; 77.a5 Ka8; 78.b6 axb6. 78 . . . Kb8 ; 7 9 . b 7 Kc7; 8 0 . Kxa7 is also
hopeless. 79.axb6 Kb8. Black resigned. 8 0 .b 7 would have ended mat
ters . The significance of White ' s 75th move is now clear. Stalemate was
avoided and the win was preserved.
35
CA RDOZA PU B LISHING • ERIC SCH I LLER
CONTINUATION OF TOPALOV VS. SHORT

Madrid International, Spain, 1 997
a b c d e f g h
[25] White to move
White has doubled pawns , which usually increase in value as other

pieces and pawns leave the board. Although a doubled pawn is weak, it
still performs useful functions, for example defending squares. In ad
dition, the rear pawn can take over the main duties of the forward
pawn should it be necessary to sacrifice it.
67 .Ke5 ! h4; 68.£6 Kg6; 69.Ke6.
a b c d e f g h
[26] Black to move
Black resigned. 69 . . . h3; 70.f7 h2 ; 7 l .f8Q wins because of 7 l . . . h l Q;

7 2 . Qg8+ Kh5 ; 7 3 . Qh7+.
36
PAWN EN DGA M ES

a b c d e f g h
[27] White to move
1 2 .Kc3 is the obvious move, and it is correct. Black has two replies .
We'll examine 1 2 . . . Kf5 here , since 1 2 . . . Kg4; 1 3 .Kxc4 leads to an impor
tant king and p awn endgame (# 1 1 ) .
1 3.Kxc4 Ke5; 14.Kc5 Ke6; 15.Kd4 Kf5; 1 6.Kd5.
3
2
a b c d e f g h
[28] Black to move
White wins the pawn. 1 6 ... Kf6; 1 7.Kxe4 Ke6 .
37
CARDOZA PU BLISHING • ERIC SCH ILLER
a b c d e f g h
[291 White to 1nove
_A crucial king and pawn endgame position . White cannot take the
opposition and cannot make progres s .
18.Kd4. 1 8 . Kf4 i s a mirror of 1 8 . Kd4, while 1 8 . Kf3 K£5 ; 1 9 .e4+
Ke6; 2 0 . Kf4 Kf6 is similar.
1 8 . . . Kd6; 1 9 .e4 Ke6; 20.e5 Kd7; 2 l .Kd5 Ke7; 22.e6 Ke8; 23.Kd6
Kd8; 24.e 7+ Ke8; 25.Ke6 with stalemate .
INSTRUCTIONAL POSITION
3
2
a b c d e f g h
[301 White to move
l .Ke5! l . c6+ Kc8 ! ; 2 . Kd6 Kb8; 3 . c7+ Kc8 is a draw. l . . . Kc6. l . . . Ke7 ;
2 . c6! bxc6; 3 .b 7 gains a new queen. 2.Kd4 Kd7; 3.Kd5 .
38
PAWN EN DG A M ES
a b c d e f g h
[31] Black to move
3 ... Kc8; 4.Ke6 Kd8. Necessary, as the corner is not safe . 4 . . Kb8;
.
5 . Kd7 transposes . 5.Kd6 Kc8; 6.Ke7 Kb8; 7.Kd7 Ka8.
a b c d e f g h
[32] White to move
8.c6 bxc6; 9.Kc7! White steps in front of the Black pawn in order to
get to b 7 .
a b c d e f g h
[33] Black to move
9 ... c5; IO.b7+ Ka7; l l .b8Q+ Ka6; 12.Qb6#.

39
CARDOZ A PU BLISHING • ERIC SCH I LLER
KING + 2 US. KING + 2

There are no general principles that pertain specifically to posi
tions with two pawns (or more ) each . When pawns block each other,
you can sometimes think of the endgames in terms of simpler endgames
without the blocked pawns, but the pawns usually do have a role to
play, if only as occupiers of territory that must be crossed. Concentrate
on exchanges , and evaluate the resulting single-pawn endings carefully.
Just try to promote a pawn, that's what pawn endgames are all about!
KURAJICA VS. CIFUENTES PARADA

Malaga Open, Spain, 1 998
a b c d e f g h
[34] Wh ite to move
In king and pawn endgames with all pawns on one side, the closer
king usually wins . It is important to watch out for potential rook pawn
draws, however. White wins cleanly in the present example.
64.Kd4 Kg2; 65.Kd5 K£3; 66.Kd6 Ke4. Black has been keeping pace,
but the White king takes the initiative by attacking Black pawns . 67 .Kc 7
b5; 68.Kc6 b4; 69.Kc5 Ke5 . 69 . . . a5 ; 7 0 . Kb5 Kd4; 7 l . Kxa5 Kc3 ; 7 2 . Ka4
forces the Black king to abandon the remaining pawn.
70.Kxb4 Kd6; 7 l .Kb5 Kc7; 72.Ka6 Kb8; 73.b4.
40
PAW N E N DG A M ES
a b c d e f g h
[35] Black to move
We now have the classic winning method for rim pawns vs . a single
defender. (#24 ) .
K I N G + 3 US. K I N G + I
Usually three pawns against one is enough to win, prov ided that
the weaker side doesn't promote first. In that case, the stronger side
requires shelter from enemy checks . While two pawns are insufficient,
three connected pawns , even if two are doubled, do provide the neces
sary refuge .
ANALYSIS OF SHIROV VS. ALMASI

Tilburg, 1 996
a b c d e f g h
[36] White to move
59.Kh5 ! As usual, accuracy is essential. 5 9 . Kf5 b4; 60 .g6 b3; 6 l .g7

b2; 62.g8Q b l Q+ won't work. 59. Kf6 b4; 60 .g6 b3; 6 l .g7 b2; 62 .g8Q
b l Q; 6 3 . Qg3+ Qd3; 64 .Qe l + Kb3 . No more checks ! 59 . . . Kb3. 59 . . . b4;
6 0 .g6 b3; 6 l .g7 b2; 62.g8Q b l Q; 6 3 . Qg7+ Kb3 ; 64 .Qb 7+ Kc2; 65 .Qe4+
41
CARDOZA PUB LIS H I N G • ERIC SCH I LLER
Kb2 ; 66.Qb7+ Ka l ; 6 7 . Qa7+ Qa2 ; 6 8 . Qxc5 is difficult, even \Vith the

White king so far away. 60.g6 c3; 6 l.g7 c2; 62.g8Q+.
a b c d e f g h
[37] Black to move
The three pawns cannot provide enough counterplay to earn more

than a dra\v.
62 ... Kb2. O r 62 . . . c4; 63 .Qg5 ! Kc3; 64. Kg4 b4; 65 . Qe5+ Kd3 ; 66.Qe l
c3 ; 6 7 . Kf3 b3; 6 8 . Qe4+ Kd2 ; 69 . Qe3+ Kd l ; 7 0 . Kf2 ! with checkmate at
e l or e2 .
63.Qg7+ Kb l ; 64.Qg6 b4; 65.Qe4 b3; 66.Qd3 Kb2; 67.Qd2 Kb l ;
68 .Qd3 Ka2; 69.Qc3 c4; 70.Qxc4 Kb2; 7 1 .Qe2 Kb l ; 72.Qd3 Kb2;
73.Qd4+ Kb l ; 74.Qe3 b2. 74 . . . c 1 Q? allows 75 . Qxb3+.
75 .Qd3 Kc l ; 76.Qa3 Kb l ; 77.Qd3 Ka l ; 78.Qa3+ Kb l draws .

The basic winning technique for converting a 3-2 p awn kingside
majority is to drive back the enemy king. Often the extra p awn is used
to win a tempo. The only danger White must avoid is falling into a
drawn king and pawn endgame .
42
PAWN EN DGA M ES
AVRUKH VS. GUREVICH

Lost Boys Open, Antwerp, 1998
8
7
3
2
a b c d e f g h
[38] White to move
5 l .Ke3. The king takes up a position which guards the entry squares
e4 and f4 . Trying to circle around the b ack would not succeed. 5 l . Kd4
Kf4; 5 2 . Kd5 g4; 5 3 . h4 g3 ; 54 . f3 Ke3 or 5 3 .hxg4 hxg4; 54.Ke6 g3 ; 55.f3
Ke3 ; 5 6 . Ke5 Kf2 ; 5 7 . f4 Kxg2 ; 5 8 . f5 Kh l ; 5 9 . f6 g2 ; 60.£7 g l Q; 6 l .f8Q.
5 1 . . . Ke5. No better is 5 l . . . g4; 5 2 .hxg4+ when both recap tures lose:
52 . . . Kxg4; 5 3 . Ke4 h4; 54.Ke5 h3; 5 5 .gxh3+ Kxh3 ; 5 6 . f4 or 52 . . . hxg4;
5 3 . f4 gxf3 ; 54.Kxf3 Kg5 ; 5 5 . Kg3 Kf5 ; 56.Kh4 Kg6; 5 7 . Kg4 . 52.g4!
3
2
a b c d e f g h
[39] Black to move
52 ... h4. There is no hope in 52 . . . hxg4 ; 5 3 . hxg4 Kf6; 54.Ke4 Ke6 ;

5 5 . £3 . 53.£3.
43
CARDOZA PU BLISH ING • ERIC SCH ILLER
a b c d e f g h
[40] Black to 1nove
53 . . . Kf6; 54.Kd4!
a b c d e f g h
[41] Black to 1nove
Black resigned. Control of the opposition gives White an easy win.

For example, 54 . . . Ke6; 5 5 . K e4 Kf6; 5 6 . Kd5 K£7 ; 5 7 . Ke5 Kg6 ; 5 8 . Ke6
Kg7; 5 9 . Kf5 Kh6; 6 0 . Kf6.
TOPALOV VS. SHORT

Madrid International, Spain, 1 997
a b c d e f g h
[42] White to nzove
44
PAWN EN DGA M ES
This is not a simple endgame . Black' s king is in position to get to

f5 . White must control that square in order to advance the pawn.
59.Kg3 Ke6; 60.Kf2 Kf6. A fascinating variant is 60 . . . Kf5 ; 6 1 . Ke3
Kf6; 6 2 . Ke4 Ke6 .
8
7
3
2
1
a b c d e f g h
[42] White to move
White gives up a pawn to gain a winning position . 63 .f5+! gx£5+;

64 .Kd4 ! (64.Kf4? Kf6 is a draw. ) 64 . . . Kd6; 65.f4 ! Black must allow the
king to get around to the f-pawn. 65 . . . Ke6 ; 66.Kc5 Ke 7; 6 7 . Kd5 Kf6 ;
6 8 . Kd6 Kf7; 69.Ke5 Kg6 ; 7 0 . Ke6 etc.
6 1 .Ke2 Ke6; 62.Kd3.
3
2
a b c d e f g h
[43] Black to move
62 . . . Kd5. 62 . . . Kd6 ; 6 3 . Ke4 Ke6 ; 64 .f5+ gxf5+; 6 5 . Kd4 transposes to

the previous note . 62 . . . Kf5 ; 6 3 . Ke3 Kf6; 64 . Ke4 Ke6 ; 65 . f5+ reaches the
same position we have seen above .
63 .Ke3 Ke6. Black must give up the e4-square. 64.Ke4 Kf6; 65.f5.
45
CA RDOZA PUB LIS H I NG • ERIC SCH I LLER
8
7
3
2
a b c d e f g h
[44] Black to move
. If Black captures the pawn, White gets the king to d5 and then e5,
so there is only o ne move .
65 . . . g5; 66.hxg5+ Kxg5.
6
5
3
2
a b c d e f g h
[45] White to move
The King + 2 vs . King + 1 endgame is a win for White , as seen in

(#25 ) .
K I N G + 3 US. K I N G + 3
W h en each side has three pawns , the play is similar to previous
situations but some odd considerations can come into play. For ex
ample, if you imagine sets of blocked pawns at b4/b5 , e4/ e5 and g4/
g5 you realize that with each king on its own side of the board, the
game must be drawn as there is no way to cross the meridian. This
feature will be seen more often in the positions with large numbers of
p awns, as we 'll see later.
46
PA WN EN DGA M ES
CONTINUATION OF NOGUEIRAS VS. HJARTARSON

Biel Intezonal, 1 993
a b c d e f g h
[46] Black to move
The king and pawn endgame is winning for Black, but it takes deep
calculation.
40 Kh5 ! The only way to win. 40 . . . Kg5 ; 4 l . Kf3 h6; 42 . g4 Kh4;
. . .
43 .Kf4 Kh3 and now 44 . Kf3 ? is incorrect, because of 44 . . . Kh2 ! ; 45. Kf4
Kg2 ; 4 6 . Ke5 Kg3 ; 4 7 . Kf5 Kf3 . 44 . g5 is best, for example 44 . . . hxg5+;
45 .Kxg5 Kg3; 46.Kf5 K£3; 47.Ke5 Ke3; 48.Kd5 Kd3; 49 . Kc5 Kc2; 50.Kxb5
Kxb2; 5 l . Kxc4 with a simple win .
a b c d e f g h
[47] White to move
4 l .Kf3. The alternatives are not attractive : 4 l . Kg2 Kg4; 42 . Kh2 Kf3 ;
43 .Kh3 h5; 44 . Kh4 Kg2 ; 45.g4 hxg4; 46.Kxg4 K f2 i n contrast to the
40 . . . Kg5 line, Black is on the move and wins the race. 47. Kf4 Ke2 ;
48.Ke4 Kd2 ; 4 9 . Kd4 Kc2; 5 0 . Kc5 Kxb2 ; 5 l . Kb4 Kc2 ; 5 2 . Kxb5 Kxc3 etc.
4 1 .Ke3 Kg4; 42 .Kd4 Kxg3 ; 43.Kc5 is much too slow. 43 . . . h5 ; 44.Kxb5
h4; 45.Kxc4 h3; 46 .b4 h2 ; 47 .b5 h l Q etc .
41
CA RDOZA PU B LISH I NG • ERIC SCH I LLER
4 l . . . Kg5; 42.Kf2. 42 . g4 Kh4 ; 43 . Kf4 h6; 44 . Kf3 Kh3; 45 .Kf4 Kg2 ;

46.Kf5 (46 . Ke5 Kg3 ; 4 7 . Kf5 K£3 ! transposes . ) 46 . . . Kf3 ; 4 7 . Kg6 Kxg4;
4 8 . Kxh6 Kf5 wins quickly.
a b c d e f g h
[48] Black to move
42 . . . Kg4; 43.Kg2 h6. 43 . . . h5; 44 . Kh2 Kf3; 45 .Kh3 Kf2; 46.Kh4 Kg2 ;
4 7 . g4 hxg4; 48 .Kxg4 see the 4 1 . Kg2 line .
44.Kf2. 44.Kh2 K£3 ; 45.Kh3 h5 ; 46.Kh2 Kf2; 47.Kh3 Kgl is another
path to a familiar position. After 4 8 . Kh4 Kg2 Black wins whichever
move White chooses . 44 . . . Kh3; 45.Kf3 h5.
a b c d e f g h
{49] White to move
White resigned, realizing that transposition to one of the losing

lines given above is inevitable.

These endgames are more interesting when there are pawns on
both flanks . It takes longer to create a passed pawn when there are
more pawns to deal with, and this gives rise to more opportunities to
switch to another flank. It is important to consider the role of the pawns
48
PAW N EN DG A M ES
on the far flank, because if they are advanced too soon, they become
easier targets for the enemy king.
DE WAAL VS. SADKOWSKI

Belgian Team Championship, Antwerp, 1 999
a b c d e f g h
[50] Black to move
First the queenside p awns are advanced. 40 ...b5; 4 l .a3.
a b c d e f g h
[51] Black to move
White ' s extra p awn is the weakling at f2, but eve n such a pawn can
prove decisive by eventually consuming a tempo. Black uses a small
tactic to trade off a pair of pawns , leaving him with the better (outside)
passed p awn. This does not save the game, but if Black moved the king
then Kg5 would have been very strong.
4 l . . .g5+; 42.hxg5+ Kg6.
49
CARDOZA PU B LISHING • ERIC SCH ILLER
a b c d e f g h
[52] White to move
White is forced to return the pawn immediately.

43 .Ke5 Kxg5; 44.f4+! The only way to win is to play aggressively!
44 ... Kg6; 45.f5+ K£7; 46.Kf4 Kf6.
a b c d e f g h
[53] White to move
White wins, because Black mus t trade the h-pawn for the forward f
pawn, but its little brother is still around to finish the j ob .
47.f3 h4; 48.Kg4 h3; 49.Kxh3 Kxf5; 50.Kg3 Kg5; 5 1 .f4+ Kf5; 52.Kf3
Kf6; 53.Ke4 Ke6 ; 54.Kd4 Kf5. 54 . . . Kd6; 55 . f5 Kc6; 56.f6 Kd7 ; 5 7 . Kc5
1s even worse.
55.Kc5 Kxf4; 56.Kxb5 Ke5; 57.a4.
50
PAWN EN DGA M ES
a b c d e f g h
[54] Black to move
Black resign ed . The White king keeps Black from reaching the a
file. Sometimes a rook pawn does win !
WONGWICHIT VS . MCKAY
Australian junior Championship, 1999
a b c d e f g h
[55] Black to move
Black resigned , but why?

49 ... Kd6; 50.c5+ Ke6; 5 l.c6 Kd6; 52.c7 Kxc 7 ; 53.Kxe5 Kd7 and
Black wins ! If you arrived here from the quiz, you get a full point either
way, since after all , White did win if only because Black blundered by
resigning!

As the number of pawns increases, there are more combinations of
pawn structures to consider. Sometimes things are rather symmetrical,
as in the first example, but all sorts of configurations can arise.
51
CARDOZA PU B LIS H I NG • ERIC SCH I LLER
HECTOR VS. SPEELMAN

Roskilde Open, Denmark, 1998
a b c d e ·r g h
[56] Black to move
White is going to be able to get to the kingside . The position seems

hopeless for Black, but in fact, Black wins !
4 1 . . .b5! The b-pawn is advanced so that the White pawns can be
undermined. White must continue with the kingside plan. 42.Kf5. What
choice does White have? 42.b4+ Kc4; 43.Kf5 Kb3; 44 .Kg6 Kxa3 ; 45.Kxh6
Kxb4; 46.Kxg5 a5 ; 47 .h4 a4; 48 .h5 a3 ; 49 .h6 a2 ; 50.h7 a l Q and Black
covers the promotion square . Not better is 42 . a4 bxa4; 43.bxa4 Kb4 ;
44 . Kf5 Kxa4; 45.Kg6 a5 ; 46.Kxh6 Kb5; 4 7 . Kxg5 a4; 48 .h4 a3 ; 49.h5 a2 ;
5 0 .h6 a l Q; 5 l . Kg6 Qh8; 5 2 .g5 leads to a winning queen vs . pawns
endgame. (See #533 . )
42 .. b4.
.
a b c d e f g h
[57] White to move
Black takes the initiative and earns a passed pawn.

43.axb4+. 4 3 . a4 Kd4 ; 44 . Kg6 Kc3 ; 45 . Kxh6 Kxb3; 46.Kxg5 Kxa4;
47 .h4 b3; 48 .h5 b2; 49 .h6 b l Q wins.
52
PAW N EN DGAM ES
43 ... Kxb4; 44.Kg6 Kxb3; 45.Kxh6.
a b c d e f g h
[58] Black to move
Black wins by a tempo.

45 ... a5; 46.Kxg5 a4; 4 7.Kh6 a3 ; 48.g5 a2 ; 49.g6 a lQ; 50.g7 Qf6+.
a b c d e f g h
[59] White to move
White resigned because the pawns are doomed. 5 l . Kh7 Qh4+;

5 2 . Kg8 Qxh2 ; 5 3 . Kf7 Qh 7; 54.Kf8 Qf5+; 55 . Ke8 Qg6+; 5 6 . Kf8 Qf6+;
5 7 . Kg8 Kc4; 5 8 . Kh7 Q£7 ; 5 9 . Kh8 Qh5+; 6 0 . Kg8 Kd5 and mate in 7 .
53
CARDOZA PUB LISH I N G • ERIC SCH I LLER
ANALYSIS OF KARPOV VS. KASPAROV

Las Palmas, 1 996
a b c d e f g �
[60] White to move
Black must give way, and White quickly brings the p awn to e5 .
53.Kf3 Ke6 ; 54.Kf4 Kf6; 55.e5+ Ke6; 56.Ke4! Not 5 6 . Kg5 ? because
56 . . . Kxe5 ; 5 7 . Kxg6 Kd4 ; 5 8 . Kxh5 Kc3 ; 5 9 . Kg5 Kxb3; 60 .h5 Kxa4; 6 l .h6
b3; 62 .h7 b2 ; 63 .h8Q b l Q means prolonged suffering for White .
56 . . . Kf7; 57.Kd5 Ke7; 58.e6 would be resignable, were it not for
the g-pawn.
a b c d e f g h
[61] Black to move
58 ... g5! saves the day: 59.hxg5 h4; 60.g6 h3; 6 l .g7 h2; 62.g8Q hlQ+
With check! 63.Kc5 Qc l+; 64.Kb6 and things are grim for Black.
64 . . . Qe3+; 65 .Ka6. 65 .Kxa5 Qc5+; 66.Ka6 Qc6+ draws . 65 ... Qe2+;
66.K.xa5 Qe5+; 67 .Kxb4 Qd4+ with a miracle draw!
54
PAWN EN DGA M ES
a b c d e f g h
[62] White to move
The White king cannot hide from the checks .
ANALYSIS OF ZEIGLER VS. SCHILLER

USA Today Network, 1 990
a b c d e f g h
[63[ Black to move
The best plan is 33 . . . gxh5 ; 34 .Kxh5 Kf5; 35.Kxh6 Kf4; 3 6 . Kg6 Kxf3;
3 7 . Kf6 Ke4 and Black wins .
33 . Kf7 is the harder win, but it has some instructive value .
. .
34.Kh4 gxh5 ; 35.Kxh5 Kg7; 36.Kh4 Kg6; -3 7.Kg4 h5+; 38.Kh4.
55
CARDOZA PU B LISHING • ERIC SCH ILLER
8
7
3
2
a b c d e f g h
[64] Black to move
The position does not seem to contain hidden tricks, but it does!
There is only one way to win .
38 ... Kh6! ! (38 . . . Kf5 ; 3 9 . Kxh5 Kf4 ; 40 . Kg6 Kx£3 ; 4 1 .Kf5 Ke3; 42 . Ke5
Kd3 ; 4 3 . Kd5 etc . 38 . . . Kf6; 3 9 . Kxh5 Kf5 ; 40 .Kh6 Kf4; 4 1 . Kg6 Kx£3 ;
42 . Kf5 is the same . ) 39.f4. If 3 9 . Kg3 Kg5 ; 4 0 . Kg2 h4; 4 1 .Kh3 Kf4 ;
42 . Kxh4 Kx£3 ; 43 . Kg5 Ke4, Black gets to the queenside first.
39 . . . Kg6; 40.Kg3 Kf5; 4 l .Kf3 h4 with a win .
YUDASIN VS. VAN DER STERREN

Biel Interzonal, Switzerland, 1 993
8
7
a b c d e f g h
[65] White to move
53.f4 g4!
56
PAWN E N DG A M ES
53 . . . gxf4+? would toss away the draw. 54. Kxf4 Kd5 ; 5 5 . Kg5 Kd4 ;
56.Kxh5 Kc3 ; 5 7 . Kg6 Kb2; 58 .h4 Kxa2 ; 59 .h5 Kxb3; 60.h6 Ka2; 6 l .h7
b3; 62.h8Q b2; 63 .Qh2 is a win for White. The queen can force the
Black king to be boxed in, and then the king can \Vander over to finish
things off. 63 . . . a5 ; 64 . Qc2 Ka l ; 65 . Qa4+ Kb l ; 66 . Qxa5 Kc l ; 6 7 . Qa3
Kc2 ; 68.Qa2 Kc l ; 69 . Qc4+ Kd l ; 7 0 . Qb3+ Kc l ; 7 l . Qc3+ Kb l ; 7 2 . Kf5
etc.
a b c d e f g h
[66] Black to move
54.hxg4 hxg4; 55.Ke4 Ke6.
a b c d e f g h
[67] White to move
Agreed drawn. 56.£5+ Kf6; 5 7. Kf4 is met by g3 58.Kxg3 Kx£5 ; 59.Kf3.
57
CA RDOZA PU B LISHING • ERIC SCH ILLER
LETAY VS. SCHUBERT

15th Spring Festival, Budapest, 1 999
a b c d e f g h
[681 White to move
Black should never have considered heading for such a position.

True, it takes 1 1 moves for Black to make a queen on the a-file, while
White needs a dozen or more . This shows the danger of simply doing
arithmetic . There is geometry involved, and White can delay Black's
plans .
5 1 .Ke3 Kd6; 52 .Kd4!
a b c d e f g h
[691 Black to move
Black can move the a-pawn, f-pawn or king. Each requires examina
tion .
52 . . . a5. ( 5 2 . . . a6; 5 3 . c5+ Ke6 ; 54.c6 is pretty much the same thing. )
After 52 . . . £'5 ; 5 3 . c5+ Ke6 ; 54.c6 Kd6; 55.c7 Kxc7 ; 5 6 . Ke5 g4; 5 7 . Kxf5
gxh3 ; 58.gxh3 Black still needs 1 1 moves, but White only needs eight.
Though all pawns are on the rim, the position is still a win. 58 . . . Kd7
( 5 8 . . . Kb6; 5 9 . Kg4 Ka5 ; 6 0 . Kxh4 Ka4; 6 1 . Kg5 Kxa3 ; 62 .h4 etc . ) 5 9 . Kg5
Ke6 ; 6 0 . Kxh4 Kf5 ; 6 1 . Kg3 Kg5 ; 62 . Kf3 Kh4; 63 . Ke4 Kxh3; 64. Kd5 Kg4;
58
PAWN EN DGA M ES
65 . Kc6 K£5 ; 6 6 . Kb7 Ke6 ; 6 7 . Kxa7 Kd6 .
a b c d e f g h
[701 White to move
68 . Kb6. Only this ! ( 6 8 . a4 Kc7 draws . ) 68 . . . Kd7; 6 9 . Kb7 Kd6 ; 7 0 . a4

Kc5 ; 7 l . a5 Kb5 ; 72 . a6 . White wins . Once again the value of a single
tempo makes itself felt in a very long variation .
53.c5+ Ke6; 54.c6 Kd6; 55.Ke4 Kxc6; 56.Kf5 Kb5.
a b c d e f g h
[711 Black to move
Black resigned here, or lost on time . I hope it was the latter, be
cause the position leads to a Queen and h-pawn vs . Queen endgame .
5 7.Kxf6 Ka4; 58.Kxg5 Kxa3; 59.Kxh4 a4; 60.g4 Kb4; 6 1 .g5 a3;
62 .g6 a2; 63.g7 a l Q; 64.g8Q.
59
CA RDOZA PU B LISHING • ERIC SCH I LLER
a b c d e f g h
[72] Black to move
It is possible to win this endgame, but very, very difficult if Black

can · achieve the proper defense. See the discussion of Fedorowicz vs .
Van der Wiel in the Queen + 1 vs . Queen section.
( 10 1 ) IVANCHUK VS. KARPOV

Melody Amber (Blindfold), Monte Carlo, 1 999
a b c d e f g h
[73] White to move
This sort of ending is not too difficult to defend if you know the
proper way to handle the queenside. Black's main threat is to walk the
king to e3 and advance the kingside pawns . If the White king strays too
far to the left, then Black can invade and win the g-pawn or simply
push the pawns . Should White bring the king to the kingside, then
Black can invade the weak light squares. Precise defense is needed.
Ivanchuk' s first two moves show the proper formation.
45.Kd3 Kd5 ; 46.b3! The purpose of this move is two-fold. White
robs Black of the c4-square and als o prepares c4 in the event the enemy
monarch heads to the kingside. Looking deeper, we see a situation
60
PAWN EN DGA M ES
\vhere all the kingside pa,vns disappear and Black gets the king to d3 .
Even there al l is not lost, as you will see.
46 ... Kc5; 47.Ke3 g5; 48.Kd3 a5 .
a b c d e f g h
[741 White to move
The situation still looks bad for White . Black threatens to push the
a-pawn and disrupt the queenside . Ivanchuk takes a bold decision to
resolve the situation, one which is not highly regarded by computers .
49.a4! 4 9 . Kd2 a4; 5 0 .bxa4 bxa4; 5 l . Kd3 loses to 5 l . . . Kd5 (5 l . . . f4 ?
is premature . 5 2 . gxf4 gxf4; 5 3 . Ke4 Kc4; 54. Kxf4 Kxc3 ; 5 5 . Ke3 Kb3;
56.Kd3 Kxa3; 57.Kc3 with a draw. ) 52.c4+. Moving the king allows . . . Kc4 .
52 . . . Kc5 ; 53.Kc3 f4; 54.gxf4 gxf4 ; 55.Kd3 f3 ; 5 6 . Ke3 Kxc4 ; 57.Kxf3 Kb3;
5 8 . Ke3 Kxa3 ; 5 9 . Kd3 Kb2 and White comes up one move short.
49 . . . bxa4; 50.bxa4.
a b c d e f g h
[751 Black to move
At least Black must now find the right plan .

50 . . . Kd5! 50 . . . g4; 5 l . Ke3 Kc4; 5 2 . Kf4 Kxc3 ; 5 3 . Kxf5 I\.b4; 54.Kxg4
Kxa4; 5 5 . Kf5 Kb3 ; 5 6 . g4 a4; 5 7 . g5 a3 ; 5 8 . g6 a2; 5 9 . g7 a l Q; 6 0 .g8Q+
61
CARDOZA PU BLIS H I NG • ERIC SCH I LLER
will draw. 5 0 . . . f4; 5 l . gxf4 gxf4 ; 5 2 . Ke4 Kc4; 53.Kxf4 Kxc3 ; 54 . Ke3 Kb4 ;
5 5 . Kd4 Kxa4; 5 6 . Kc4 draws .
5 1 .c4+ Kc5 ; 52 .Kc3 f4; 53.gxf4 agreed drawn. A fter 53 ... gxf4;
54.Kd3 we reach the following position .
a b c d e f g h
[76] Black to move
Black would have drawn as follows . 54 . . . f3 . (54 . . . Kb4; 55 . Ke4 Kxc4;

56.Kxf4 Kb4; 5 7 . Ke3 Kxa4; 5 8 . Kd2 Kb3; 59.Kc l Ka2 ; 6 0 . Kc2 a4 ; 6 l . Kc l
a3 ; 6 2 . Kc2 Ka l ; 63 . Kc l a2 i s stalemate . )
55.Ke3 f2; 5 6.Kxf2 Kxc4; 57.Ke3 Kc3; 58.Ke2 Kb3 ; 59.Kd2 Kxa4;
60.Kc2 Ka3; 6 l .Kb l and the king gets to the corner.
CONCLUSION OF BURMAKIN VS. ZVJAGINSEV

Russian Championship, Elista, 1997
a b c d e f g h
[77] Black to move
White ' s extra pawn is useless and sooner or later will fall. Black j ust
has to distract the king, and that is easily done. S till, the line must be
calculated carefully because it is just b arely a win .
62
PAWN ENDG A M ES
42 ... e4! ; 43.fxe4 fxe4; 44.Kc3 Kxa4; 45 .Kd2 Ka3 ; 46.Ke3 Kxa2 ;
47.Kxe4 Ka3; 48.Kd3 Kb4.
a b c d e f g h
[78] Wh ite to move
The famous pawn-winning tactic is based on control of the opposi

tion, something that could hardly be foreseen ten moves ago , espe
cially if there was time trouble .
49.Kd2 Kxc4; 50.Kc2 Kd4; 5 1 .Kd2 Ke4 and White resigned.

Once there are five o r more pawns on one side, the room to ma
neuver can be limited by the pawns . There are fewer open files, and
blockading positions can be constructed.
( 36 ) YUDASIN VS. VAN DER STERREN

Biel Interzonal, 1 993
a b c d e f g h
[79] Black to move
Black can hold this game, but only with extremely accurate play.
63
CARDOZA PUB LIS H I NG • ERIC SCH I LLER
5 1 . . .b4! 5 l . . . Kxd5 loses to 52.f4 ! g4. The only try. (52 . . . gxf4+; 53 .Kxf4
Kc5 ; 54 . Kg5 Kb4 ; 5 5 . Kxh5 Ka3 ; 5 6 . Kg5 Kxa2 ; 5 7 .h4 a5 ; 5 8 .h5 a4 ;
59 .bxa4 b4; 6 0 . a5 ! b3; 6 1 . a6 b2 ; 62 . a7 b l Q; 63.a8Q+ Kb2 ; 64 . Qb7+
Kc2 ; 6 5 . Qxb l + e tc. ) 5 3 .hxg4 hxg4; 54.Kf2 ! Retreat is sometimes the
bes t way to advance in the endgame. 54 . . . Ke4 ; 5 5 . Kg3 K£5; 56.b4 Black
is in zugzwang.
52.d6! 5 2 . f4+ gxf4+; 53.Kf3 Kxd5 ; 54. Kxf4 Kd4 lets Black get the
queen first. 5 5 . Kg5 Kc3; 56.Kxh5 Kb2 with 5 7 . Kg6 Kxa2; 5 8 .h4 a5 ;
59 .h5 a4; 6 0 .bxa4 b 3 ; 6 l .h6 b 2 ; 62.h7 b l Q+ or 5 7 . Kg5 , followed by
5 7 . . . Kxa2 ; 58 .h4 Kxb3; 59 .h5 Ka2 ; 60 . h6 b3 ; 6 1 . h7 b 2 ; 62.h8Q b l Q;
63 . Qd4 , where White is groveling for a draw. 52 . . . Kxd6. The pawn
sacrifice gives rise to a drawn King + 4 vs . King + 4 endgame (#65 ) .
a b c d e f g h
[80] White to move
SADLER VS. BOLOGAN

III Chess Masters Enghien-les-Bains, France, 1999
a b c d e f g h
[81] Black to move
64
PAWN EN DGA M ES
The unfortunate situation on the kingside has White ' s game hang
ing by a thread, which is about to be snapped. All Black has to do to
gain entrance is drive the White king away from the d3-square . A bold
stroke accomplishes the j ob. 55 . . . b5+! The deflection succeeds , even
though it gives White a passed pawn.
56.Kxb5 Kd3; 57.Kc6. 5 7 . a4 Ke2 ; 5 8 . a5 Kx£2 ; 5 9 . a6 Kxg3 ; 6 0 . a7 f2 ;
6 l . a8Q f1 Q+; 6 2 . Kc6 Qf6 ; 63 . Qa3+ Kxh4 and the king can advance the
pawns. White cannot afford to grab the d-pawn. 64 . Qxd6 Qxd6+;
65 . Kxd6 g3 ; 66.Kc7 g2 ; 67.d6 g l Q; 6 8 . d7 Qg8 etc.
57 ... Ke2; 58.Kxd6 Kx£2; 59.Ke5 Kxg3; 60.d6 f2; 6 1 .d7 flQ; 62 .d8Q.
a b c d e f g h
[82] Black to move
Black is not distracted by the a-pawn and concentrates on winning

the h-pawn with a quiet but very annoying move .
62 . . . Kh3; 63 .Qd7. There was no way to save the pawn from . . . Qe l +.
63 . . . Kxh4. White resigned.
KING + 5 US. KI N G + 5
KRASILNIKOV VS. POPOV

N ovgorod Open, Russia 1 999
a b c d e f g h
{83] Black to 1nove 65
CARDOZA PU B LIS HING • ERIC SCH I LLER
Black easily establishes a kingside pawn maj ority on the rim and
can run over to the queenside and eat all the pawns . But White can
prevent that.
40 ... g5; 4 1 .£5+. 4 1 . fxg5 fxg5 ; 42.c4 h4 and White must not rush to
play 43.c3?? which would use up his last spare tempo.
4l . .. Kd6; 42.c4 h4; 43.h3.
a b c d e f g h
[84] Black to move
The position is now dravvn if the kings j ust tango . Black should be
content with this . But he stubbornly presses on.
43 ... g4??; 44.hxg4 h3 ; 45. Kf3 Ke5. O r 4 5 ... h2 ; 4 6 . Kg2 h l Q+;
47.Kxh l Ke5 ; 48.Kh2 Kf4; 49 .Kh3 Kg5 ; 5 0 . Kg3 Kh6 ; 5 1 . Kf4 and White
wins . 46.Kg3 h2; 47.Kxh2 Kf4; 48.Kh3. Black resigned.
ANALYSIS OF ZEIGLER VS . SCHILLER

USA Today Network, 1 990
a b c d e f g h
[85] White to move
Despite the 'veakness of White's pawns , this position is dra,vn thanks

to the active position of the White king.
66
PAWN EN DG A M ES
29.a3. 29.Kg4 leads to a long forcing variation: 29 . . . Ke5 ; 30.f4+ Ke4 ;

3 l . Kg3 a6; 3 2 . Kg4 b5; 3 3 . a3 b4; 34.a4 a5 ; 3 5 . Kg3 Kd3; 36.Kg4 Kc3 ;
37.Kf5 Kxb3; 38.Kg6 Kxa4 ; 39.Kxg7 b3 ; 40.£5 b2 ; 4 l .f6 b l Q; 4 2 . £7 Qg l +;
43 . Kh8 Qd4+; 44 . Kg8 Qxc4 ; 4 5 . Kh8 Qf4 ; 46.Kg8 Qx£7+.
29 . . . b6. 29 . . . a5 can also lead to a draw. 30.b4.
a b c d e f g h
[86] Black to move
30 ... a5 . 30 . . . cxb4 creates an outside passed pawn but it doesn't win.

3 l .axb4 and now 3 l . . . a5 ? can be met by 3 2 . c5 ! a4 (32 . . . bxc5 ; 33 .bxa5 )
33.cxb6 a3 ; 34.b7 a2; 35 .b8Q a 1 Q; 3 6 . Qd6+ Kf7; 3 7 . Qd5+ Kf8 ( 3 7 . . . Kf6;
3 8 . Qf5+ Ke7; 3 9 . Qe5+ Qxe5+; 40 . Kxe5 wins . ) 3 8 . Qc5+ K£7; 39 .b5 Qb2 ;
40 . Kg3 and White ' s king i s safe, so the pawn can advance. Black can
arrange a check at e 1 , but then the king will be safe re treating to g2 ,
and bringing the queen back to f2 if necessary. So Black should play
3 l . . . Ke6 ! ; 32 . Ke4 a5 ; 3 3 . c5 ! axb4 ; 34.cxb6 Kd6; 3 5 . Kd4 Kc6; 3 6 . Kc4
Kxb6 37.Kxb4 with a draw.
3 l .b5. 3 l .bxc5 bxc5 ; 32 . a4 g6; 3 3 . Kg4 . leads to a King + 4 vs . King
+ 4 endgame and is treated in that section (#63 ) . 3 l ... a4. 3 l . . . g6 ; 32 .hx.g6
Kxg6; 3 3 . Ke5 h5 is not significantly different from the main line, with
only the a-file changed. 34.f4 h4 ; 35 . Ke6 h3; 3 6 . f5+ Kg7 ; 3 7 . Ke 7 h2 ;
3 8 . f6+ Kg6 ; 3 9 . f7 h l Q; 4 0 . f8 Q Qe4+; 4 1 . Kd7 Qxc4 ; 42 . Qd6+ Kf7 ;
43 . Qxb6 is better for White, but is not clearly won.
32.Ke4 g6. M oving the king would be a mistake . 32 . . . Kg5 ? ; 3 3 . Kd5
Kxh5; 34 . Kc6 Kg5 ; 3 5 . Kxb6 h5 ; 36.Kxc5 h4 ; 3 7.b6 h3 ; 38 .b7 h2; 3 9 .b8Q
h l Q; 40 . Qe5+ with a position that should be won for White . 32 . . . Ke6?;
33.f4 forces the Black king to give way and allow White to infiltrate one
of the flanks .
33.hxg6 Kxg6; 34.Ke5 h5; 35.f4. After 35 . Kf4 Kf6 the position can
not be held. 35 ... h4; 36 .Ke6 h3 .
67
CA RDOZA PU B LIS H I N G • ERIC SCH ILLER
a b c d e f g h
[87] White to move
. 3 7 .£5+ Kg7; 38.Ke7 h2; 39.f6+ Kg6; 40.f7 h lQ; 4 1 .f8Q Qe4+; 42.Kd7
Qxc4; 43.Qd6+. White escapes with at least a draw.
KONERU VS. SASHIKIRAN

l Oth Goodricke International, Calcutta 1 999
a b c d e f g h
[88] White to move
5 1 .Kh4. Draw agreed. Can't Black break through with an early . . . f5?
No . In fact there is some danger in 5 1 . . . g5+; 52.Kg3 Kg6; 53.Kf2 Kf7;
54.Ke3 g6; 55.Kd3 f5. After 56.gxf5 gxf5 ; 57.exf5, 57 . . . Ke 7 is correct,
hoping for 5 8 . Ke4?? Kf6 but willing to settle for 5 8 . Ke3 K£7. Black
must avoid 57 . . . Kf6??; 58.Ke4.
68
PAWN EN DGA M ES
a b c d e f g h
{891 Black to move
The position is hopelessly lost! 58 . . . Kf7; 59.Kxe5 Ke7; 60.£6+ Kf7;

6 1 .Kf5 etc.
ANALYSIS OF EUWE' S STUDY
(Improved)
a b c d e f g h
[901 Black to move
Although many moves are involved, the result of this position is

forced because there are really quite few options for either player.
l l . . . h6; 1 2 .h4 Kf6; 1 3 .g5+ hxg5 ; 14.hxg5+ Kxg5 ; 1 5 .Kxe5 Kg4;
1 6.Kd6 Kf4; 1 7 .Kxc6 Ke4; 18.Kxc5 Kd3; 19.Kxb4 Kd4; 20.Ka3 Kc5;
2 1 .Ka4 Kd4; 22.Kb4 Ke5; 23.Kc5 . There is no possible defense for
Black.
KING + 6 US. KING+ 6

We look at five starting positions here, because such rich endgames
display many characteristics of multiple-pawn endgames in general. In
addition, there are many opportunities for exchanges which require
the player (or analyst) to properly evaluate the resulting endgames .
69
CA RDOZA PU B LISH,I NG • ERIC SCH I LLER
KARPOV VS. KASPAROV

Las Palmas, 1996
a b c d e f g h
[91] White to move
A truly difficult king and pawn endgame . Two World Champions

are doing battle, and each makes a critical mistake ! White has four
candidate moves . The king can defend from c3 , advance to e4, or the e
p awn or h-pawn can advance .
38.h4? First error. 3 8 . Kc3 e5; 39 .fxe5 fxe5; 40.b4+ axb4+; 4 l . axb4+
Kd5 ; 42. e4+ seals most of the board, but the g- and h-files are available.
42 . . . Ke6 ; 43.Kd3 Kf6; 44. Ke3 Kg5 ; 45 . Ke2 Kf4; 46 . Kf2 h5 ; 4 7 . Kg2 g5 ;
48.Kf2 h4 ! secures the g3-square. Although Black is faced with zugzwang,
it is White to move, and the Black king wil l be able to get to e3 or g3 .
49 . Kg2 Ke3 and the p awns drop.
Nothing is to be gained by 3 8 . e4 e5 ; 3 9 . fxe5 fxe5 ; 40.h4 b4; 4 l . a4
Kc6 . The kings can do the mambo on the queenside forever without
result. Going to the kingside is the only option, but after 42.Ke3 Kd6;
4 3 . Kf2 Ke6; 44 . Kg3 Kf6 Black just has to exercise a little care . 45 .Kg4.
a b c d e f g h
[92] Black to move
lO
PAW N E N DG A M ES
Black causes himself problems with 45 . . . h5+? , which requires more

skill to defend. 46.Kg3 Ke6 ; 47.Kf2 Kd6; 48.Ke2 Kc5 ; 49.Kd3 Kd6 ;
5 0 .Ke3 Kc5 ; 5 l . f4 ! 5 l . . . exf4+ allows White a touch of brilliance with
5 2 . Kf3 ! !
a b c d e f g h
[931 Black to move
The point is that Black has to either retreat the king, after which
the b-pawn is not in any danger, or advance. 52 . . . Kd4; 5 3 . Kxf4 Kc3 .
a b c d e f g h
[941 White to move
The die is cast. 54.e5 Kxb3; 55.e6 Ka2 ; 56.e7 b3; 57 .e8Q b2; 5 8 . Qe2
Ka l ; 59 .Qe5 Ka2 ; 60.Qe6+ Ka l ; 6 l . Qxg6 b l Q; 62.Qxb l + Kxb l ; 63.Kg5 .
White wins the race .
So Black is stuck with 5 l . . . Kd6 ! ; 5 2 . fxe5+ Kxe5 reaching a King + 4
vs . King + 4 endgame (#60 ) . 45 . . . h6; 46 . Kg3 (46 .h5 g5 is a dead draw. )
46 . . . Ke6 and again no progress has been made . The position is similar
to the final position of the game.
38 Kd5; 39 .b4 axb4; 40.axb4.
...
71
CA RDOZA PU BLISHING • ERIC SCH I LLER
a b c d e f g h
[95] Black to 1nove
. White is tied down to the defense of c4 , and therefore the only king
move is to c3. Black still has a two time killer ( . . . h6 and . . . h5 ) while
White has only e4 as a respectable move . Kasparov fails to notice that
the path to v ictory leads around the kingside, and makes a move that
takes away a possible square for the king. The last move of time control
is always dangerous !
40 . . . h6? 40 . . . e5 ! was correct, and would have won, according to
Karpov. First we consider the capture : 4 l .fxe5 fxe5; 42.e4+ Ke6 ; 43.Ke 3
K£6 ; 44 . f4. Or else Black plays . . . g5 . 4 4 . . . h 6 ; 45 . Kf3 exf4 ; 46. Kxf4 g5+;
47.hxg5+ hxg5+; 4 8 . Kg4 Ke5 ; 49.Kxg5 Kxe4 and wins .
Next, we examine 4 l . e4+ Ke6 and find that White can no longer
get the king across the fourth rank, but neither can Black! 42 . Ke3
( 42 . fxe5 fxe5 reaches the previous note . ) 42 . . . exf4+ ( 42 . . . h6 is similar
to the game . ) 4 3 . Kxf4 h5 ; 44 . Ke3 g5 gives Black the outside passed
pawn, but 45 . f4 ! is a sav ing trick . The Black king is going to have prob
lems infiltrating. 45 . . . gxh4 ( 45 . . . g4 ; 46.Kf2 will not allow Black to make
any progress. 45 . . . gxf4+; 46. Kxf4 and White wins ! ) 46.Kf2 and Black
must sit idly as White collects the pawns on the h-file . 46 . . . Kf7; 47.Kg2
Kg6; 48.Kh3 Kh6; 4 9 . Kxh4 Kg6 ; 5 0 . Kg3 Kh6 with a draw.
4 1 .e4+ Kd6; 42.Ke3 e5; 43.fxe5+ fxe5 ; 44.Kf2 Ke6; 45.Kg2. Agreed
drawn.
72
PAW N E N DG A M ES
3
2
a b c d e f g h
[96] Black to move
45 . . . Kf6; 46.Kg3 is similar to the play from diagram #92 .
BURMAKIN VS. ZVJAGINSEV

Russian Championship, Elista, 1 997
a b c d e f g h
[97] White to move
White is faced with a clear choice. He must either capture the pawn
at h4 or advance to g4 . There are no other options . Computers have a
hard time with this position. Even looking 1 4 moves deep ( 2 8 ply) the
machines evaluate the options as hav ing equal merit, both leading to a
draw. You can understand why humans can make the same m istake !
White captured, and went down to defeat.
37.gxh4? 3 7 .g4 ! is the right move .
73
CA RDOZA PU B LISHING • ERIC SCH ILLER
a b c d e f g h
[981 Black to move
. 37 . . . e4+ ( 3 7 . . . Ke6; 38.gxf5+ Kxf5 ; 3 9 . Ke3 ! Kf6 ; 40 . Ke4 Ke6 ; 4 l . a3

Kf6; 42 . Kd5 K£5; 4 3 . Kxc5 e4! ; 44.fxe4+ Kxe4; 45.Kb5 g4; 46.c5 gxh3 ;
4 7 . c6 h2; 48.c7 h l Q; 49 .c8Q Qf3 should be drawn. ) 3 8 . Ke3 (38 . fxe4??
fxg4 ! and White is lost. ) 3 8 . . . exf3 ; 39.gxf5 (39.Kxf3 f4 ! ) 39 . . . Ke5 ; 40.Kxf3
Kxf5 is a draw as the Black king remains in the area e5-f5-f6 , and the
White king can triangulate to keep it from getting to f4 .
37 . . . gxh4; 38.Ke3. 3 8 .a3 a4 ! is similar. 38.a4 prevents . . . a4, but af-
ter 3 8 . . . f4; 3 9 . Kd2 Black wins : Not with 39 . . . e4; 4 0 . fxe4 Ke5 ; 4 l .Kd3 f3;
42 . Ke3 f2 ; 4 3 . Kxf2 Kxe4; 44 . Ke2 where the Black king cannot get in.
However, 3 9 . . . Ke6 ; 4 0 . Kc3 K£5; 4 l . Kd2 ( 4 l . Kd3 e4+; 4 2 . fxe4+ Ke5 wins
because the king cannot come to e3, and the pawn at e4 must fall . )
4 l . . . e 4 ; 42.Ke2 Ke5 !
3 8 . . . a4!
a b c d e f g h
[991 White to move
The pawn sacrifice cannot be declined, because Black threatens to

exchange at b3 and then circle the king v ia a5 to attack the queenside
74
PAWN E N DG A M ES
pawns which cannot be defended because the White king must remain
in the center to stop Black's pawns .
39.bxa4 Kc6; 40.Kd3 Kb6; 4 1.Kc3 Ka5; 42.Kb3. What else? Black
threatens to consume all of the White queenside pawns . The king can
not stay on the queenside, however.
a b c d e f g h
[ 100] Black to move
Continued in ( #77 ) .
ANALYSIS OF SHIROV VS. ALMAS!

Tilburg, 1 996
a b c d e f g h
The key is the kingside pawn maj ority. Eventually it \vill be reduced
to a single pa\vn for White, which Black may well win . However, by that
time the White king has entry into the queenside . There is no way that
Black can safely establish a passed pawn.
24 . .. Kf7; 25.Ke3 Ke6; 26.Ke4 Kf6 . If 2 6 . . . d5+; 2 7 . cxd5+ cxd5+;
lS
CA RDOZA PU B LIS H I NG • ERIC SCH ILLER
2 8 . Kd4 Kd6 ; 2 9 . a4 ! White preserves a tempo move (b3 ) , which is al

'vays a tremendous asset in a king and pawn ending. Eventually, White
'vill exchange the f-pa,vn for the d-pawn, gobble the queenside pawns
and make a new queen . 29 . . . g6; 3 0 .g4 Ke6; 3 1 .f5+ gxf5 ; 3 2 . gxf5+ Kxf5 ;
3 3 . Kxd5 Kg4 ; 34.Kd6 Kh3 ; 35 . Kc7 Kxh2 .
a b c d e f g h
[ 1 02] White to move
White must choose the co rrect pawn here . 3 6 . Kxb7 ! ( 3 6 . Kxb6 Kg3 ;
3 7 . Kxb7 h5 is just a draw. Black queens with check. ) 36 . . . Kgl ; 3 7 . Kxb 6
h5 ; 3 8 . a5 h 4 ; 3 9 . a6 h3 ; 40 . a7 h2; 4 1 . a8Q h 1 Q; 4 2 . Qxh l + Kxh l ; 43 .b4 i s
the 'vin.
2 7 .g4 Ke6; 28.f5+ Ke7; 29.g5 K£7; 30 .h4 Ke7; 3 l .h5 Kf7; 32.b4!
a b c d e f g h
[ 1 03] Black to 1nove
32 . . . Ke7. 32 . . . c5 ; 3 3 . Kd5 ! Ke7; 34.f6+ gxf6 ; 35.gxf6+ Kxf6 ; 36.Kxd6

Kg5 ; 3 7 . Kc7 Kxh5 ; 3 8 . Kxb 7 Kg5 ; ( 3 8 . . . Kg4? ; 39.Kxb6 h5 ; 40.bxc5 h4;
4 1 . c6 h3; 42 . c7 h2; 4 3 . c8Q+) 3 9 . Kxb6 h5 ; 40 .bxc5 h4 ; 4 l .c6 h3; 42.c7
h2; 43. c8Q h 1 Q; 44.Qc5+ 'vith a simple Queen + 2 vs. Queen endgame .
33 .b5 ! Kf7 ! This is better than 3 3 . . . c5; 34.Kd5 Kd7 ; 35 .f6 gxf6 ;
36.gxf6 h6 ; 3 7 . £7 Ke7 ; 38.f8Q+ Kxf8; 3 9 . Kxd6 etc.
l&
PAWN ENDG A M ES
34.h6. Not 34.g6+? hxg6; 35.fxg6+ Kf6; 3 6 . Kf4 d5 ; 3 7 . cxd5 cxd5 ;

38 .a3 d4 ! ( 3 8 . . . Ke6 ; 3 9 . Kg5 d4; 40 .h6 gxh6+; 4 l .Kxh6 d3 ; 42 .g7. White
queens with check . ) 3 9 . Ke4 Kg5 ; 4 0 . Kxd4 Kxh5 ; 4 l .Ke5 Kxg6; 42.Kd6
Kf5; 43 .Kc7 g5 ; 44 . Kxb6 g4; 45.Kxb 7 g3 ; 46.b6 g2 ; 47.Ka6 g l Q and
Black wins !
a b c d e f g h
[ 1 04] Black to move
34 . . . g6? This is the more interesting and instructive line . The best
plan is 34 . . . gxh6 ! ; 35 .gxh6 K£6 K£4 which should be drawn.
35.Kf4!
a b c d e f g h
[1 05] Black to move
The Black king must s tay at f7 to prevent the exchange at g6 when

Black would like to recapture \vith a pawn. 35.f6 Ke6; 36.Kd3 Kf7; 37.Kd4
is just a draw.
35 . . . Kf8. 35 . . . cxb5; 36.cxb5 d5 ; 3 7 . f6 Ke6; 3 8 . £7 ! Kxf7; 3 9 .Ke5 Ke 7;
40.Kxd5 Kd7; 4 l .Ke5 Ke7 ; 42.a3 ! and Black mus t give way. Again the
availability of a tempo on the a-file is critical .
77
CARDOZA PU B LISHING • ERIC SCH I LLER
36.fxg6 hxg6; 3 7 .a4 Kf7; 38.Ke3 Kg8; 39.Kd2 Kf7; 40.Kc3 Kg8;
4 l .Kb4 Kf7. Or 4 1 . . . c5+; 42 . Kc3 K£7; 43 . Kd3 Kg8; 44. Ke4 etc .
42.a5.
a b c d e f g h
Faced with the threat of a6 , Black mus t capture at a5 , and then the
king gets to b6. 42 ... c5+; 43.Ka4 bxa5; 44.Kxa5 etc.
ANALYSIS OF SHIROV VS. ALMAS!

Tilburg, 1996
a b c d e f g h
[ 107] White to move
Black's main problem is that the queenside majority can not be

converted into a passed pawn. There is little hope that the h-pawn will
be unleashed, either. White can win with aggressive kingside play.
28.Ke4 Ke6; 29.£5+ Kd6; 30.£6! Ke6 ( 3 0 . . . gxf6 ; 3 l .gxf6 Ke6; 3 2 . g5
c6; 3 3 . a3 h6; 34.gxh6 Kxf6 ; 35.h7 Kg7; 36.Ke5 Kxh7; 3 7 . Kd6 eats all
the queenside pawns . )
3 l .fxg7 Kf7; 32 .g6+ hxg6; 33.g8Q+ Kxg8; 34.Kd5 Kf7 . 3 4 . . . c4 ;
3 5 . Ke6 Kg7; 36.g5 c6 ; 3 7 . Kd7 K£7; 3 8 . Kc7 Ke6; 3 9 . Kxb7 Kf5 ; 4 0 . Kxc6
ll
PAWN EN DGA M ES
Kxg5 ; 4 1 . Kxb5 Kf4 ; 42 . a4 g5 ; 43 .a5 g4; 44.a6 g3 ; 45 . a7 g2 ; 4 6 . a8Q g l Q;

47.Qf8+ Ke4 (47 . . . Ke3 ; 48.Qc5+) ; 48.Qe7+ Kf3 ; 49 . Qf7+ Ke2 ; 5 0 . Qxc4+
Kd2 ; 5 1 .b4 wins without difficulty.
a b c d e f g h
35.g5 Ke7; 36.Kxc5 Ke6; 37.Kxb5 Kf5; 38.c4 Kxg5; 39.c5 Kf4; 40.a4
g5; 4 l .a5 g4; 42.c6 bxc6+; 43.Kc4. (43 . Kxc6 g3 ; 44 . a6 g2 ; 45 . a7 g l Q;
46.a8Q Qg2+! and the queen falls . )
4 3 ... g3; 44.a6 g2; 45.a7 glQ; 46.a8Q will lead to a split point.
GOWOR VS. BECKETT

Australian junior Championship, 1 999
We 've seen some mistakes made by Grandmasters, so it is hardly
surprising that young players make poor decisions, too . The following
game is presented as support for my adv ice that you should never re
sign, or agree to a draw, unless the weaker players in the audience
would understand why. The balance swings from side to side in this
game, thanks to numerous errors .
a b c d e f g h
[ 1 09] White to 1nove
79
CA RDOZA PU B LISH I N G • ERIC SCH I LLER
One way to win quickly, several ways to use the same plan later, and
a few ways to draw. But the conclusion to this one is truly amazing.
34.£4?! 3 4 . a4 ! is very simple . The next move is b4, and then a passed
a-pawn guarantees victory. 34 . . . £5. You can just imagine Black, sitting
at the board, making rnoves while simply praying that White doesn't
find a4 !
35 .h3?! g6; 36.g4?? Instead of creating an outside passed pawn,
White grants one to his opponent.
36 . . . fxg4; 37 .hxg4 h5; 38.£5?
a b c d e f g h
[110] Black to move
Here the young players agreed to a dr aw. But now Black is winning!
38.gxh5 would have made things tougher. 38 . . . gxh5 ; 39.Kd3 Kd5; 40 . Ke3
h4; 4 l .f5 c5 ! ( 4 l . . . h3 ; 42 . Kf3 c5; 43.dxc5 Ke5 ; 44 . Kg3 Kxf5; 45 .Kxh3
Ke5 ; 46.c6 Kd6; 4 7 . Kg4 Kxc6; 48 . Kf4 Kc5 ; 49.Ke3 Kc4; 50 .Kd2 holds a
draw. ) 42 . dxc5 Ke5 ; 4 3 . f6 ! Kxf6; 44. Kf4 Ke6 ; 4 5 . Kg4 Kd5 ; 46.c6 Kxc6;
4 7 . Kxh4 Kd5 and Black wins : 48.Kg5 Kd4; 4 9 . Kf4 Kd3 ; 5 0 . Kf3 Kc2 etc .
38 ... h4; 39.fxg6. O r 3 9 . f6 Ke6; 4 0 . g5 h3 . 39 ... Ke6. The Black king
stops the White pawns, while the White king cannot catch up with the
h-pawn.
KING + 'l US. KING + 'l

The following study is included because it has a great deal of rel
evance for anyone who plays or defends the Exchange Variation of the
Spanish Game ( l .e4 e5 ; 2 . Nf3 Nc6; 3 . Bb5 a6; 4 . Bxc6 dxc6 ) . It demon
strates the consequences of allowing the pawns to be doubled and not
keeping pieces on the board.
80
PAW N EN DGA M ES
EUWE STUDY
1 940 (Improved)
a b c d e f g h
[111] White to move
Here is a study that is certainly of practical value! The pawn struc-

ture is that of the Exchange Variation of the Spanish Game . Euwe sub
j ected it to detailed scrutiny, publishing his findings in 1 94 0 . Yet there
were more treasures to be discovered, and scholars have been adding
to the legacy for half a century. The verdict is that White wins .
Black must not allow all the pieces to come off while the p awn
structure remains unb alanced . White must take advantage of the
kingside p awn majority to try to create a p assed p awn on that wing.
That will keep the Black king busy so that the White king can satiate
his appetite on the other flank.
l .Ke2 Ke7; 2.Ke3 Ke6; 3.£4 c5; 4.c4 c6; 5.a4 b5; 6.b3 f6; 7.a5 b4.
a b c d e f g h
8.£5+. 8 . g4 will also work, and was Euwe's solution . If Black chooses
8 . . . g6 then 9 .f5+! is decisive . 9 . . . gxf5 (9 . . . Ke5; 1 0 . fxg6 hxg6; l l .h4 is
81
CA RDOZA PU B LISH I NG • ERIC SCHI LLER
also winning for White . ) 1 O . exf5+! Ke5 Black must play this , as other
wise the White king will get to f4 to help the pawn advance to g5 , but
Black quickly runs out of moves . l l .h3 ! ( l l .h4? h6 and White cannot
make progres s . ) l l . . . h6; 1 2 .h4 . Now Black has no useful moves. 1 2 . . . h5
( 1 2 . . . Kd6 ; 1 3 .Kf4 Ke7 ; 1 4 .g5 hxg5+; 1 5 .hxg5 fxg5+; 1 6. Kxg5 is more
prosaic. ) 1 3 .gxh5 Kxf5 ; 1 4 . Kf3 is another win, found by Minev. 1 4 . . . Ke6;
1 5 .Kf4 Kf7; 1 6 .Kf5 Kg7; 1 7 .h6+ Kxh6; 1 8 . Kxf6 Kh5; 1 9 . Ke6 .
8 . . . g5 loses to a surprising temporary pawn sacrifice : 9.e5 !
a b c d e f g h
9 . . . gxf4+; 1 0 . Kxf4 fxe5+; l l . Ke4 and the kingside pawn majority

wins easily. (See diagram #90 . ) Returning to the game, we see the more
efficient win.
a b c d e f g h
8 . . . Ke5 . This plan was worked out and refined by British Grand
master Tony Miles . It seems most illogical to allow Black to blockade
the e-pawn and tie down the White king, but there is a powerful counter
plan that soon gives the White king access to the queenside !
82
PAWN EN DGA M ES
9.g4 g6; 10.fxg6 hxg6; l l .h4.
a b c d e f g h
Miles has discovered a way to doubly implement the diversion of

the enemy king! l l . . . Ke6. The three alte rnatives are not promising.
1 l . . . g5 ; 1 2 . h5 Ke6 ; 1 3 .h6 K£7; 1 4 . e5 fxe5 ( 1 4 . . . Kg6 ; 1 5 . e6 and one
of the p awns promote s . ) 1 5 . Ke4 Kg6 ; 1 6 . Kxe5 Kxh 6 ; 1 7 . Kf6 Kh 7 ;
1 8 .Kxg5 Kg7; 1 9 . Kf5 and White wins without difficulty.
1 1 . . . £5; 1 2 . exf5 gx£5 ; 1 3 .g5 f4+; 1 4 . Kf3 Kf5 ; 1 5 . g6 Kxg6; 1 6 .Kxf4
Kf6; 1 7 .h5 and White wins .
1 l . . . Kd6 ; 1 2 .h5 gxh5 ; 1 3 .gxh5 Ke6; 1 4.Kf4 K£7; 1 5 . Kf5 Kg7; 1 6 .h6+
Kxh6; 1 7 .Kxf6 and the e-pawn promotes .
1 2 .Kf4 Kf7; 13.e5 fxe5+; 14.Kxe5 Ke7; 15 .h5 gxh5; 1 6.gxh5.
a b c d e f g h
{ 116] Black to move
The win is straightforward.
83
PAWNLESS ENDGAMES
Endgames without pawns tend to fall into two categories-triv ially
simple or fiendishly difficult. The triv ial type can be seen in the chap
ter on Elementary Endgames . In this chapter we look at the trickier
type, including Rook vs . Bishop. Rook vs. Knight, Two Rooks vs . Rook
and Knight and the often exasperating Queen vs . Rook.
ROOK US. BISHOP

A king and rook cannot usually win against king and bishop, even
if the weaker king is stuck in the corner. The bishop retreats to act as
guardian, and there is no way to set up a checkmate .
LEITAO VS. JANSSEN

Hoogovens Tournament, Wijk aan Zee, 1 999
a b c d e f g h
1 16 . . Ka3 ; 1 1 7.Bb l . Agreed drawn. If the rook does not pin the
.
bishop, the bishop is free to move and can always come back to b 1 . If
the rook pins the bishop to the king (Rh 1 ) the position is stalemate .
84
PAW N LESS EN DGA M ES
The next example shows one of the rare cases where the stronger
side can win, with a little help from his opponent.
ZETIC VS. KUTUZOVIC

Zagreb, 1997
a b c d e f g h
{ 118] White to move
There are only three legal moves for White, but which ones lose
and which ones draw? White makes a fatal error here .
73.Kfl? The only losing move ! The king gets trapped on the " short
side. "
The correct plan is 73 . Kd l ! Ke3 ; 74 .Rc 8 !
a b c d e f g h
This is the defensive formation to remember. 74 . . . Ra l + ( 74 . . . Rg2 ;

75 .Kc l Rgl +; 76.Kb 2 illustrates the difference between the long side
and the short side . Compare the position wi th the game in the note to
move 78, where the king runs out of room. ) 75 .Rc l Ra2 ; 76.Rc3+ Kf2 ;
77.Rh 3 and White has nothing to worry about.
CARDOZA PU B LISHING • ERIC SCHI LLER
73 . Ke l ? doesn't work either. 73 . . . Ke3 ; 74 . Rd8 is not as good a de

fense as the White king is forced to the short side . 74 . . . Rg2 ; 7 5 . Re8
( 75 . Kfl Rh2 . The threat is . . . Rh l #, so White must play 76.Rg8 Ra2 and
we are back in the main line of the game after Black' s 76th move. )
73 ... Ke3; 74.Kgl Rg2+; 75.K.fl Rh2; 76.Rg8. White resigned, without
waiting for 7 6 . Rg8 Ra2 .
a b c d e f g h
7 7 . Rg3+ ( 7 7 . Kg l Ra l +; 78.Kh2 Rh l +; 79.Kg3 Rg l + wins the White

rook, while 7 7 . Ke l runs into 7 7 . . . Ra l #) 7 7 . . . Bf3 .
a b c d e f g h
The mate threat is 78.Kg l Ra l +; 79.Kh2 Rh l # so White would have

to give up the rook.
ROOK US. KN IGHT

This is a draw, but whether it is a technical draw is open to debate.
As an International Arbiter, I have been asked on many occasions to
rule (or confirm a ruling) awarding a draw on a claim by the side with
86
PAW N LESS EN DGA M ES
a knigh t. Consider a case from the 1 9 96 World Youth Championships,

where I was a member of the appeal tribunal .
One of the girls claimed a draw when the king and kn ight were well
coordinated in the center of the board. Before making the claim, she
showed surprising maturity by playing for a while just to show she was
not going to cooperate and walk her king to the rim and allow some
potential trick. She kept the king and the knight in optimal position.
This was in the sudden death period, which is the norm these days.
Her opponent had a large time advantage and the defender was almost
out of time. The arbiter ruled against her, and she lost on time. In the
appeal, she was awarded the draw. The arbiter, a nice fello'v but not a
very experienced player, simply didn't know that the endgame is a draw
except in a very few cases which require simply ter rible play.
The rules of chess should stipulate that Rook vs . Knigh t is a techni
cal dr aw, noting relevant exceptions. But that' s not the current situa
tion . So you had better make sure that you understand what to avoid,
and you can do that by studying the following games.
KELECEVIC VS. POTTERAT

Liechtenstein, 1 994
a b c d e f g h
Things look bad, but Black is nevertheless able to draw. Black's

king is trapped on the back rank, and mus t move the knight, but with
careful play can avoid disaster. .
6 8 . . . N a 3 . 6 8 . . . Kd l ? ? would wa lk into 6 9 . Rh l # . 6 9 . Ra2 Nb l .
69 . . . Nb5??; 70.Ra6 ! Nc7 ( 70 . . . Kb l ; 7 1 .Rb6 or 70 . . . Kd l ; 7 l .Ra l #) 7 1 .Rc6+
wins the knight. 70.Rc2+ Kd l is a draw.
87
CA RDOZA PU B LIS H I N G • ERIC SCH I LLER
a b c d e f g h
[1 23] White to move
· Black can always choose betwee n . . . Kc l and . . . Na3 , and the White
king cannot come close enough to help . Even with the king on the edge
and the knight in the corner, the position is drawn, with correct play,
as we see in the next example .
DOLMATOV VS. BOLOGAN

Goodricke International, Calcutta, 1999
a b c d e f g h
The game continued 63 .Ng6+ Kf6 and then 64.Nf4?? Oh dear. The
correct plan is 64.Nh4 ! Rh l ; 65.Kh5 Rg l ; 66.Nf3 Rg3; 67.Nh4 Rg5+;
68.Kh6 and there is no win.
88
PAWN LESS E N DG A M ES
a b c d e f g h
Here the game was drawn. By agreement? I doubt it. Perhaps Black's
flag fell. I don't know, but in any case this is no way to draw!
64.Nf4, the move allegedly played in the game (perhaps a typo
graphical error? ) loses instantly to 64 . Rh l +; 65 .Nh5+ K£5 .
. .
a b c d e f g h
The key is the position for the Black king. It must be able to move
into a position in opposition to the knight, \vhile the pin is on the rim .
89
CA RDOZA PU BLIS H-NG • ERIC SCH I LLER
OLBRICH, VS. KRUPKOVA

Germany, 1 995
a b c d e f g h
[ 1 2 7] White to move
When the king is not stuck in the corner, there are additional de
fensive possib ilities . Still, you have to be careful, as this example shows .
78.Nh6 Rh7.
a b c d e f g h
79.Ng4? The knight runs away, but without a king to protect it, falls
v ictim to a v icious hunt. 79 .Ng8 ! is the correct defense.
79 . . . Rh4. The knight will inevitably be trapped by the enemy rook
and king. It has strayed too far from its master. 80.Ne3.
90
PAW N LESS E N DGA M ES
a b c d e f g h
[ 1 28} Black to move
80 . . . Re4; 8 1 .Ng2. The alternative is 8 l .Nd l Rf4+; 8 2 . Kg7 Rf3 ;

83 .Nb2 Kd5 ; 84 . Kg6 Kd4; 85.Kg5 Rfl ; 86.Na4 Rb l . The knight's days
are numbered.
8 1 ... Kf5; 82.Kf7 Kg4; 83.Kf6 Re2.
a b c d e f g h
[1 29} White to move
The knight is trapped, so White resigned.
91
CA RDOZA PU BLISH I N G • ERIC SCH ILLER
DUNNINGTON VS. SZEMOK

Cappelle Ia Grande, 1995
a b c d e f g h
[1 301 White to move
Black was force d to promote to a knight, as anything else would

have allowed 7 7 . Re l #.
77.Kf3 !
a b c d e f g h
[ 1 31 1 Black to move
77 ... Kfl . Any other move loses the knight immediately. 78.Rg2 !
92
PAW N LESS EN DG A M ES
a b c d e f g h
[ 1 32] Black to tnove
78 . . . Ke l ; 79.Rgl+.
a b c d e f g h
[ 133] Black to tnove
The knight is lost, and the game must be abandoned.
TWO ROOKS US ROOK A N D K N IGHT

Most of the time two rooks cannot defeat rook and knight, because
if a pair of rooks are exchanged the rook cannot win against the knight.
There are a few tactical tricks, of course . Sometimes they are not easy
to see, and in the following game a top player falls for one .
93
CA RDOZA PU B LISH.ING • ERIC SCH ILLER
KASPAROV VS. BAREYEV

Novgorod International, Russia, 1 997
a b c d e f g h
Black would be j ust fine if he moved the king off the second rank,
breaking the pin. But instead he managed to find the losing move .
72 ... Re3??; 73 .Re6! Black resigned because of 73 . . . Rxe6; 74 . Kxe6
and the knight is lost. Or 73 . . . Rf3+; 74 . Ke4 Rf2 ( 74 . . . Rf4+; 75 . Ke3 )
75 . Ke3.
QUEEN US. ROOK

Winning with queen against rook is not as easy as it looks . As long
as the king can protect the rook, it is difficult for White to bring both
the queen and king into a coordinated position. Like all endgames
\Vith 5 pieces or less ( including the kings ) , this has been worked out
exhaustively by computers . Unfortunately, humans are likely to make
mistakes, and one must not forget the 50-move rule! If Black can last
50 moves from the start of the Queen vs . Rook endgame then the
game is drawn by rule. Efficiency is important, but it only requires 3 1
moves to win from the most difficult of starting positions.
94
PAWN LESS EN DGA M ES
NEFF VS. SEEMAN

Estonian Championship (7), 1996
a b c d e f g h
[ 135] Black to 1nove
80 ... Rd3+; 8 l .Kb2 Rd4; 82.Qc7+ Kd5. 82 . . . Kb4?? loses immediately

to 83 . Qc3+.
83.Kc3.
a b c d e f g h
O nce the king approaches, the weaker side must be very careful. If
the rook moves on to d l , g4 or a4 then Qd7+ wins . O n Rh4, . . . Qd8+
works . The squares b4 , c4 , and f4 are unavailable, that leaves only e4.
83 . . . Re4; 84.Qf7+ Re6. 84 . . . Ke5; 85 . Qh5+ Kf6; 86.Kd3 allows White
to make a little progress, but might be better.
85.Kd3 !
95
CARDOZA PUBLISH,NG • ERIC SCHILLER
a b c d e f g h
- Black is now reduced to a choice of d5 or d6 for the king, since the

rook cannot move.
85 . . . Ke5; 86.Ke3 Rf6. 86 . . . Kd5+ is a good defense. In the end White
will still win, but it will take a lot of time . 8 7 .Qh 7 Re6. Keeping the
rook on the " third square" is considered the best form of defense. The
third "square" is the square formed by c3, f3, f6 and c6 . If the rook can
stay on this line, then it is hard for the king to approach.
88.Qh5+ Kd6+; 89.Kd4.
a b c d e f g h
This is the key position in the roundup of the Black forces . The
kings stand in opposition, the intervening rank is controlled by White .
89 . . .Ke7; 90.Qg5+ Kf7; 9 1 .Kd5.
96
PAW N LESS EN DGAM ES
a ·b c d e f g h
[ 1 391 Black to move
Another important p osition. The Black king cannot move, and the
rook has limited options .
9 1 . . . Re7; 92.Qg4. Black resigned (or possibly lost on time). The
Black king has almost been forced back to the last rank. Let' s examine
a possible conclusion. This will not be the most efficient path, because
I want to show some special positions along the way. O n 92 . . . Ra7;
9 3 . Qf4+ and White gives check on e3, g3 or d4, snaring the rook. Also
bad is: 92 . . . Re8; 93.Qf5+ Kg7; 94.Qd7+ Kf8. That leaves 92 . . . Rb7; 93.Kc6
Re7; 94.Kd6 with two lines .
First we consider 94 . . . Kf8; 9 5 . Qc8+ ReS ( 95 . . Kf7; 9 6 . Qf5+ Ke8;
.
9 7 . Qg6+ and Black loses the rook and is quickly mated. ) 9 6 . Qf5+ Kg7 ;
9 7 . Qg5+ K£7 ; 9 8 . Kd7 Rg8 ; 9 9 . Qf5+ Kg7 ; 1 0 0 . Ke7 Kh6. This is a back
rank! Remember that in endgames without pawns all four edges of the
board are equal . 1 0 l . Qh3+ Kg6 ; 1 0 2 . Qg4+ Kh7; 1 0 3 . Qh5+ Kg7 .
a b c d e f g h
[ 1401 White to move
1 04. Qd5 ! This wins quickly, but the obvious alternative doesn' t \vork.
97
CARDOZA PU B LISH I NG • ERIC SCH I LLER
( 1 04 . Qf7+ Kh8; 1 0 5 . Kf6? The rook must stay in control of g7, and must
abandon the king. But there is a trick! 1 05 . . . Rg6+ ! ! The rook cannot be
captured, because of stalemate . 1 06.Kf5 Rg5+; 1 0 7 . Kf4 Rg4+; 1 0 8 . Ke5
Rg5+; 1 0 9 . Ke 6 Re 5+; 1 1 0 . Kd7 with more wo rk ahead . ) 1 04 . . . Kh7
( 1 04 . . . Rh8; 1 05 . Qg5+ Kh7; 1 06 . Kf7 mates quickly. ) 1 05 . Kf7. Mate fol
lows on the h-file .
94 ... Ra7. The rook thre atens to get back to the
The other choice is
third r ank by checking at a6 . 95 .Qc4+ Kf8. All other replies lose the
rook (or worse ! ) to an assortment of diagonal tricks .
95 . . . Ke8; 9 6 . Qg8# is to be avoided at all costs !
95 . . . Kg6; 9 6 . Qe4+ Kh5; 9 7 . Qh 1 + Kg4; 9 8 . Qg 1 + Kh3 ; 9 9 . Qxa7.
95 . . . Kf6; 9 6 . Qd4+.
95 . . . Kg7; 9 6 . Qd4+.
· 96.Ke6 . Once again, Black has only one move .
a b c d e f g h
96 . . . Rf7. 96 . . . Rh7; 97.Qc5+! Kg8; 98.Qc8+ Kg7; 99.Qc7+ forces mate

in 1 0 ! 99 . . . Kh6; 1 00 . Qf4+ Kh5 ; 1 0 1 . Qf5+ Kh6; 1 02 . Qf6+ ( 1 02 .Kf6 Rf7+! !
is one of those little traps that is easy to miss after many hours of play. )
1 02 . . . Kh5; 1 03 . Kf5 . Mate in 4 .
97.Qc5+ Kg8. 9 7 . . . Kg7; 98.Qg5+ K£8; 9 9 . Qh5 Re 7+; 1 00 . Kf6 .
98
PAW N LESS ENDG A M ES
a b c d e f g h
This is another winning position . Black has no useful moves and is

mated in 5 .
98.Qg5+ Rg7; 99.Qd5.
a b c d e f g h
[143] Black to move
The threat of discovered check keeps the initiative in White ' s hands .
The rook is allowed to get back to the third rank, but this is only tem
porary.
99 ...Rg6+. We must also look at 99 . . Rf7; 1 0 0 . Qg2+ Rg7; 1 0 l . Qa8+
.
Kh7 ; 1 02 . Qe4+ Kh8; 1 03 . Kf6 Kg8 ; 1 04 . Qe6+ Kh7 ; 1 05 . Qe 8 .
99
CA RDOZA PU B LIS H I.NG • ERIC SCH I LLER
a b c d e f g h
{ 1441 Black to move
·This is a variation on a position worked out by Philidor in the 1 8th

century! He had the queen on e8, but it really doesn't matter. White
will win the rook in 5 moves or less . 1 05 . . . Rc7 ( 1 05 . . . Kh6 ; 1 06 . Qf8. O r
1 05 . . . Rgl ; 1 06.Qh5+ Kg8; 1 07.Qd5+ Kh7 ; 1 08.Qe4+ Kg8; 1 09.Qa8+ Kh7;
1 1 0 . Qa7+ etc. ) 1 06.Qh5+ Kg8; 1 0 7 . Qd5+ Kh7; 1 08.Qd3+ Kg8; 1 09.Qd8+
etc.
1 00.Kf5+ Kh8.
a b c d e f g h
{ 1451 White to 1nove
A nother little stalemate trick, but it can be avoided. 1 00 . . . Kh 7;

l O l . Q£7+ Rg7 ( 1 0 l . . . Kh8; 1 0 2 . Qe8+! Kh7; 1 03 . Qxg6+) 1 02 . Qh5+ Kg8;
1 0 3 . Qe8+ Kh7; 1 04 . Kf6 reaches the Philidor position again .
10 1 .Qd4+! Kh7. 1 0 1 . . . Rg7; 1 02 . Kf6 Kh7; 1 03 . Qe4+ Kh8; 1 04.Qh l +
Kg8 ( 1 04 . . . Rh7; 1 05 . Qa8#) 1 05 . Qd5+ Kh8; 1 06 . Qh l + Kg8; 1 0 7.Qh5
Rg l ; 1 0 8.Qe8+ Kh7; 1 09.Qe4+ Kg8; 1 1 0 . Qa8+ Kh7; , 1 l l .Qa7+ is a theme
we have already seen.
102.Qh4+ Rh6; 1 03.Qe 7+ Kg8; 1 04.Kg5.
1 00
PAWNLESS ENDGAMES
3
2
a b c d e f g h
Yet another pattern to remember. The rook cannot maintain its

position near the king.
1 04 . . . Rh7; 1 05.Qe8+ Kg7; 1 06.Qe4! Kg8; 1 07.Kg6 Rg7+; 1 08.Kh6.
The primary threat is Qe8#, but secondarily Black has to worry about
saving the rook. It isn't possible. If you find this endgame enormously
complicated, don't worry. It is. In fact, 6-time United States Champion
Walter Browne failed to win the endgame against a computer, and that
was back in 1 9 78 when computers weren't that strong. Just try to ab
sorb some of the winning positions displayed in the game, and try to
find a tactical way to reach them.
QUEEN US. ROO K AND KNIGHT

Unless the queen can force the win of one of Black' s pieces, the
game will end in a draw by the 50-move rule, or perhaps terminal bore
dom. This endgame is pure tactics. If the defender just keeps the pieces
close together, no progress can be made.
10 1
CARDOZA PUBLISHING • ERIC SCHILLER
VALLE VS. BATISTA

Maalox Plus Open, Brasilia, 1999
3
2
a b c d e f g h
6 1 . . . Rc5; 62.Qa4+ Ke5; 63 .Qe8+ Kd4; 64.Qh8+ Kc4; 65.Qg8 Rc6;

66.Kh4. Agreed drawn.
1.1 QUEEN AND ROO K US. QUEEN

This endgame is always a win for the side with the rook unless the
king cannot find shelter from enemy checks. It is likely to arise only in
the case of a promotion taking place while the defender already has a
queen. Here is one such case.
MORENO VS. AGUILAR

Peru Regional Championship, Lima, 1999
3
2
a b c d e f g h
[ 148] White to 1nove
This position is only a draw because the White king cannot escape
the checks. Even if the queen could come back to defend, Black can
exchange queens and then take the rook.
102
PAWNLESS ENDGAMES
74.Kc l . 74 . Kb3 Qd3+; 75 .Ka4 Qd4+; 76 . Qb4 ( 76 . Kb5 Qxg7)

76 ... Qxb4+; 77.Kxb4 Kxg7. 74 . . . Qc4+; 75 .Kb l . 75.Kd2 Qd4+; 76 .Ke2
Qxg7; 75 ... Qb3+. Agreed drawn. Black will win the rook. 75 . . .Qb3+;
76.Kal (76.Kc l Qc3+ is the same.) 76 . . . Qc3+; 77.Ka2 Qxg7.
QUEEN AND ROO K US. QUEEN AND ROO K

When each side has a queen and a rook, the side to move usually
wins. The winning side finds a way to take advantage of the inferior
position of enemy pieces.
IVANOV VS. GRINSELL

International Tournament, Benas que, 1997
3
2
a b c d e f g h
In this case, the rook on e7 is protected by the queen at e l , but that

line of sight can be broken by forcing the enemy king to step on the e
file.
63.Qf8+ Kg5; 64.Qh6+ Kg4; 65.Rg6+ Kf3; 66.Qf8+. Black resigned.
103
KNIGHT ENDGAMES
Knight endgames are primarily tactical in nature. Once in a while
it is necessary to exploit a long term weakness in the enemy positions.
In �his case, strategic thinking comes to the forefront. For the most
part, however, this is just a matter of counting and calculating.
One crucial tactical point is that the knight cannot lose a tempo.
The knight will always take an even number of moves to reach the
square of the same color as the one it presently occupies. To switch to
a square of a different color, the knight will have to use an odd number
of moves. Therefore, the king will often have to use the technique known
as triangulation to win a tempo if necessary.
Knights don't have an easy time maneuvering when there are a lot
of pawns on the board. Even one zone pawns can get in the way. This
can lead to some very lengthy knightjourneys. You will see these themes
illustrated in the positions presented below.
KNIGHT US.2
A knight has great difficulty coping with past pawns on two flanks.
Its limited reach makes each intercontinental journey a laborious task.
In the best of circumstances, a knight can travel from one rim to an
other in four moves. Considering that a pawn on its home square
needs no more than five moves to reach the promotion square, knights
have a hard time keeping up.
104
KNIGHT ENDGAMES
ANALYSIS OF IBRAGIMOV VS. RUBLEVSKY

President's Cup, Elista, 1998
3
2
a b c d e f g h
57.Ng8! Going after the a-pawn doesn't work: 57.Kb3 a2; 58.Kxa2
h5 ; 59.Ng8 h4 is too slow.
3
2
a b c d e f g h
Black wins! The knight cannot get back to stop the pa\vn, for ex
ample. 60.Nf6+ K£3; 6 l .Nh7 Kf4; 62.Nf6 h3; 63 .Nd5+ K£3 and the h
pawn marches to the promotion square.
57 ... a2; 58.Kb2 Ke3; 59.Nf6 h6; 60.Ng4+ with a draw as both Black
pawns fall.
KNIGHT + I US KNIGHT
Knight and pa,vn against knight is usually a draw, because the de
fending knight can sacrifice itself for the enemy pawn. In some cases
the enemy knight is simply too far a\vay and cannot catch up, but that
is a mere matter of calculation. There are some tricky situations, how-
105
CARDOZA PUBLIS�ING • ERIC SCHILLER
ever, where the defense proves difficult because the stronger side's
forces, combined to create a fence that the Knight can neither j ump
over nor run around.
ANDERSSON VS. HECTOR

Sigeman International, Malmo, 1998
3
2
a b c d e f g h
[152] White to move
White needs to hasten with the pawn, as the Black king is close.
6 l.g6 Ke7; 62.g7.
3
2
a b c d e f g h
This is the target position for knight and pawn vs. king. White wins
because the knight guards the promotion square, and does not worry
about the enemy king as the pawn will promote in any case. Notice that
the White king keeps the enemy horse at bay by controlling e3 and d4.
106
KNIGHT ENDGAMES
GULKO VS. BLATNY

Saitek US Masters Honolulu, 1998
3
2
a b c d e f g h
White needs to control a number of key squares to win this endgame.

The central pawn does not restrict the movement of the Black knight,
so White must use his knight and king to seize control.
67.Ng4 Ne7. Otherwise Kd6 eliminates the final flight square.
68.Nh6. This removes g8 and f5 from Black's options, allowing the
knight only one post. 68 . . . Nc8; 69.Nf5. Now e7 and d6 are covered, so
it is just a matter of corralling the knight with the king.
69 ... Ka4; 70.Kc6 Ka5; 7 l .Kc7 Nb6.
3
2
a b c d e f g h
The knight is preserved, but to stop the pawn, Black will need to
use c4 or d5.
72.Ne3! Eliminating both options, and securing the win.
101
STUDY BY RETI
Reti's Endgame Studies, 1929
3
2
a b c d e f g h
[1561 White to move
Black is close to promoting the pawn and the White king is rushing
to catch up. Black only needs to play .. . Kgl and then two more moves
to a new queen. White uses a series of tricks to keep the pawn from
advancing.
l .Kh4. The king comes closer to the pawn. l . . . Kgl . Black gets the
king out of the way, so that the pawn can advance. How can this be
prevented?
2.Ng4!
3
2
a b c d e f g h
[1571 Black to move
2 . . . Kg2 ! The threat is 3 .Nxf2? h2 ! and the knight cannot stay in

place to block the pawn. 3.Ne3+ Kh2. Black has managed to lose a
tempo, so that we have the same position as after White's first move,
but with White to move.
4.Nc2!
108
KNIGHT ENDGAMES
3
2
a b c d e f g h
[158] Black to move
White must be accurate. The inferior 4.Nfl +? loses to 4 . . . Kgl ; 5.Ng3

Kg2; 6.Nf5 h2; 7.Ng3 Ne4; 8.Nxe4 hlQ+.
4 . . . Nd3; 5.Kg4 Ne5+; 6.Kh4 Nf3+; 7.Kg4 Ng5 !
3
2
a b c d e f g h
[159] White to move
This is Black's best chance.

S.Ne l . 8.Kxg5 Kg l ; 9.Ne l Kf2 ! ( 9 . . . h2; 1 0 .Nf3+ Kg2 ; l l . Nxh2
draws . ) 1 0 .Nd3+ Kg3 . The pawn promotes. S .. . Kg l ; 9.Nf3+! Kg2.
9 . . . Nxf3; 1 0 .Kxh3 is drawn by rule.
1 0.Nh4+ Kf2. 10 . . . Kh2; l l.Nf3+ repeats the position. l l .Nf3 .
109
3
2
a b c d e f g h
[160] Black to move
The position is now a draw because the pawn will not be able to
advance. The White knight can go to h4 or h2 as needed, returning to
f3.
K NIG HT + 2 US. KNIGHT

When one side has two pawns and the other side has none, a knight
ending will result in victory unless the defender can somehow elimi
nate one of the pawns without allowing the other to promote. Most of
these endings are not very interesting, however, sometimes they can
lead to instructive endgames.
ANDERSSON VS. HECTOR

Sigeman International, Malmo, 1 998
3
2
a b c d e f g h
[161] White to move
White must play carefully to get the full point, because the pawns
are weak.
110
KNIGHT ENDGAMES
60.d6+! This sacrifice deflects the king, 'vhich is the only defender
because the knight cannot get to any useful square. At best, it takes
four moves: . . . Nb4, . . . Nc6, . . . Nd8 and . . . Ne6. Black, in check, has no
time for the knight. 60.Kd3 Nb4+; 6 l .Kc4 Nc2; ( 6 l . . . Nxd5; 62 .Kxd5
K£7; 63.Ke5 Kg6; 64.Kf4) 62.Kc5 Ne3; 63.d6+ Kd8; 64.g6 Nf5; 65.Kd5
Ng7; 66.Ke5 will get the job done, but it takes a lot longer.
60 . . . Kxd6. We have reached a knight endgame \Vhich is a simple
Wln.
3
2
a b c d e f g h
[162] White to move
If you do not see the winning plan, see the continuation in diagram
# 1 52.
KNIGHT + 2 US. KNIGHT + I

In most cases an endgame with knight plus t\vo pa,vns vs . knight
plus one will either be resolved by tactical means or will lead to a single
pa,vn ending. If there are two connected pa,vns facing a pawn 'vhich
stands in the \vay of the promotion square, there are some peculiar
possibilities as illustrated in our example.
I I I
BACROT VS. AGDESTEIN

Capablanca Memorial, Cuba, 1998
3
2
a b c d e f g h
[ 1631 Black to move
The knight will be sacrificed to advance the pawn. Note that White's
knight is so far away as to be irrelevant to the action.
64 . . .Nd3+; 65.Ke3 f4+! ; 66.Kxd3. 66.Ke2 f3+; 6 7.Kfl Kh3; 68.Nd6
Kxh2 is equally hopeless.
66 ... g3. White resigned. The pawn will reach the promotion square.
3
2
a b c d e f g h
[1641 White to move
For example: 6 7.hxg3+ fxg3; 68.Ke2 Kh3; 69.Nd6 g2; 70.Kf2 Kh2.
KNIGHT + 4 US. KNIGHT + 2

Normally a two-pawn advantage is sufficient to win, but if the pawns
are weak, some skill may be required. Even when the defensive forma
tion is quite solid, as in the position below, the defender will have a
hard time keeping control of all the important squares.
1 12
KNIGHT ENDGAMES
LARSEN VS. FRIES NIELSEN

Danish Championship, Aarhus, 1999
3
2
a b c d e f g h
[165] White to move
Things don't look too bad for Black, as the White knight is rooted
to b7 to defend the a-pawn and the pawn at h5 is a target. White cor
rectly chooses to abandon the a-pawn in favor of a more active knight.
5l .Nd6+! Ke6; 52.Kf4 Nxa5 ; 53.Nf5 Nc4.
3
2
a b c d e f g h
{166] White to move
White can grab the h-pawn, but must play accurately.

54.Nxh6! Ne5; 55.Nf5 Nd3+; 56.Kg5 Nxc5; 57.h6. Black must bring
the knight back to defend, since otherwise White promotes.
57 ... Ne4+. 5 7 . . Kf7; 58.h7 and a new queen cannot be prevented.
.
58.Kg6 Nf6; 59.Nd4+ Ke7.
1 13
3
2
a b c d e f g h
[167] White to move
One final finesse and the point is earned.

60.Nc6+! (60 .h7? Nxh7; 6 l .Kxh7 Kd6; 62.Nc2 Ke5 draws, because
of 63 .Kg6 Ke4; 64 .Kg5 Kd3 ; 65.Kf4 Kxc2; 66.e4 a5; 67.e5 a4; 68.e6 a3;
69.e7 a2; 70.e8Q a l Q etc.)
60 . . . Ke6; 6 1 .e4. Black resigned. The threat is e5, and then the knight
must move and allow the h-pawn to proceed.
HONFI VS. MOISEYEV

Uzhgorod, 1972
3
2
a b c d e f g h
[168] White to move
With a kingside pawn majority, White wants to create a passed pawn.

This should take place on the h-file, which is further from the queenside
than the g-file. The Black king is in an excellent position to defend, or
to attack weak White pawns . The White knight is temporarily tied to
114
KNIGHT ENDGAMES
the defense of the b-pawn, but only as long as the Black knight sits at
c5.
43.h5! Ke6 . 43 . . . Ne6 is a better defense. From e6, the knight helps
to control g5. 44 . Nf3+ Kf6 (44 . . . Kd5; 45 .g5 hxg5; 46.h6 Nf8; 47.Nxg5 is
a simple win.) 45 .Kd3 and now:
45 . . . Ng5 ; 46 . Nd2 Nh3 ; 47.Kc4 Nf2; 48.Kb5 Nxg4; 49. Kxb6 Ke6;
50.Kxa5 Nf6; 5 l .Kb6 Nxh5; 52.a5 Nf4; 53.a6 Nd5+; 54 .Kb7 Kd7; 55.a7
Nc7; 56.b4 h5 ; 57.b5 h4; 58.b6 wins.
45 . . . Nc7; 46.Kc4 Ke6; 47.g5 hxg5; 48.Nxg5+ Kf6; 49 . Ne4+ Kg7;
50.Nc3 Kh6; 5 1 . Nd5 Ne6; 5 2 . Nxb6 Kxh5; 53 . Nd7 Kg5; 54 .Nc5 Nc7;
55.Nb7 K£5; 56.Nxa5 and the pawn will advance without difficulty. So
even with best play, Black was already lost.
44.Kd4 Nd7; 45.Nc4.
3
2
a b c d e f g h
[169] White to move
45 . . . Nf6. This offers more resistance than the alternative: 45 . . . Nc5;

46.Nxa5! Nxa4; 47.bxa4 bxa5; 48.Kc5 and White wins. 48 . . . Ke5; 49 .Kb5
Kf4; 50.Kxa5 Kxg4; 5 l .Kb6 Kxh5; 52.a5 Kg4; 53.a6 h5; 54.a7 h4; 55.a8Q
h3; 56.Qh l etc.
46.Ne3 K£7; 47.Kc4 Nd7.
liS
3
2
a b c d e f g h
[170} White to move
48.Nf5 ! White starts by sacrificing a pawn, but has a bigger sur

prise in store. The key is the destruction of the pawn at h6, and the
remaining queenside pawns quickly follow. Black is left with a knight
to defend against three pawns-an impossible task.
48 . . . Ne5+; 49.Kb5 Nxg4; 50.Kxb6 Ke6; 5 1 .Nxh6! The knight is
sacrificed, but the pawns win. 5 1 . . . Nxh6; 5 2.Kxa5.
3
2
a b c d e f g h
[171} Black to move
52 ... Kd7; 53.Kb6 Kc8; 54.Ka7. Black resigned. The b-pawn marches
up the board, while the knight must remain on the kingside to guard
against the advance of the h-pawn.

The more pawns you have, the easier it is to give one away and live,
perhaps prosper as well. In the example here, Larsen shows us a nice
positional sacrifice.
116
KNIGHT ENDGAMES

Danish Championship, Aarhus, 1 999
3
2
a b c d e f g h
[172] White to move
White has an extra pawn, but the double isolated pawns and over
all weak pawn structure doesn't inspire a lot of confidence in his win
ning chances. Black may be content just to shuttle the knight back and
forth between c6 and b4. If the White knight moves, not only is the a
pawn vulnerable, but the enemy king can then come to e5. The posi
tion requires inspired play!
44.e5 ! ? White sacrifices the extra pawn but before it goes it will
anchor the knight at d6, so that it can attack b7. 44 ... Nc6; 45.Kf3 Kf5.
45 . . . Kd5; 46.Kf4 Ke6; 4 7 .Ke4 followed by Nd6. 46.Nd6+ Kxe5; 4 7 .Nxb7
c4. 47 . . . Kd5 allows 48.Kf4 threatening Kf5-g6xh6. 48 . . . Ne7; 49.Nd8 !
and N£7 lets the king and knight team up to win the pawn at h6.
48.bxc4. The passed pawn is at least blockaded, freeing the king to
deal with the kingside problems . 48 ... Kf5; 49.c5 Ne5+; 50.Kg3 Nc6.
White has two extra pawns, but all the pawns are weak. See diagram
# 1 65 for the continuation, but first try to find the best move for White.

Knight endgames are tricky because there are not only a lot of tac
tics, but also some amazing journeys with the horses galloping all over
the board, especially when one knight is pinned down.
I I 'l
CARDOZA PUBLISfiiNG • ERIC SCHILLER
WUNNINK VS. SCHILLER

Grandmaster Open, Groningen, 1997
3
2
a b c d e f g h
In this example, from one of my games, Black's task is to get the

king to one of the flanks . It isn't easy!
42.Nd3.
3
2
a b c d e f g h
{ 174] Black to 1nove
White had counted on this move to keep the Black king out. The
endgame strategy is not simple. If the Black king tries to infiltrate the
queenside prematurely, White will be able to advance the central pawns,
or somehow break down the defensive pawn barrier. Strategically there
are several stages.
First, Black makes a passed pa,vn on the queenside. Next, the pawn
is advanced to tie down the White knight. Third, the kingside pawn
formation is clarified. Finally, the Black knight embarks on a remark
able journey to take up a defensive post, after \Vhich Black can safely
'vin the White knight.
I ll
KNIGHT ENDGAMES
42 ... Ne6; 43.Kf3 b5. This is the most direct way of creating the
passed pawn. 44.axb5 Nd4+; 45 .Ke3 Nxb5. Stage one accomplished.
46.f5. I didn't think this move was as good as retreating the knight
so that the Black pawn cannot advance so quickly. 46.Nb2 g6; 4 7 .f3 is
correct. Note that the Black a-pawn cannot go to a5 because of Nc4+.
46 . . . a5! ; 4 7 .f4 a4; 48.Kd2 a3; 49.Ncl.
3
2
a b c d e f g h
[175] Black to move
Stage two achieved, the pawn is advanced and the White knight
must be very careful not to allow it to reach a2 . Black cannot maneuver
a knight to b4 to assist.
49 . . . Kc5. The king must play an active role, threatening to advance
toward the White pawns . 50.Kd3 Nd6; 5 1 .Na2 h5 . This pawn wants to
advance to h4, so that it will eventually be closer to the promotion
square after Black maneuvers the knight to capture the pawn at h3 .
52.h4 f6; 53.Ncl. White just moves the knight back and forth. Black
must somehow try to entice White to advance the pawn from e4 to e5,
where it will be weaker. Then the knight will need to move to e7 to
make sure the pawns cannot advance, even with the help of the White
king. Under those circumstances, Black will be able to use the king
effectively on the queenside.
51 . .. Nc4; 54.Na2 Nb2+; 55.Kc3 Ndl+; 56.Kd3 N£2+; 57.Ke3 Ng4+.
This is a useful post for the knight. Black threatens to play . . . Kc4, so
White's king must stay somewhere near the center.
58.Kd3 Nh2; 59.Ke3 Kc4; 60.e5 .
119
3
2
a b c d e f g h
[1761 Black to move
This move is forced, but another goal has been achieved. Any other
plan by White is too slow, as the Black king is ready to take out the
enemy knight.
60 . . . Ng4+; 6 l .Ke4 Nh6 ! This move will make sense when you see
the continuation.
62 .Ncl Ng8; 63 .Na2 Ne7.
a b c d e f g h
[1771 White to move
White is now lost. There are no White threats on the kingside, so

the White knight can be hunted down.
64.Nc l Kc3 ; 65.Na2+ Kb2 ; 66.Nb4 a2; 67.Nxa2 Kxa2; 68.Kd4 Kb3.
68 . . . Nxf5+; 69.Kd5 Nxh4; 70.exf6 gxf6; 7 l .Ke6 Ng2 would have been a
more efficient end to the game.
69.Kc5 Nxf5; 70.exf6 gxf6; 7 l .Kd5 Nxh4; 72.Ke6 Ng2. White re
signed, because the king cannot catch the h-pawn. 73.f5 Ne3; 74.Kxf6
h4; 75 .Kg5 Nxf5; 76.Kg4 Kc2 is a simple win.
120
KNIGHT ENDGAMES

In order to make progress in endgames \Vith many pa,vns it is im
portant to create a passed pa,vn or a threat of a passed pa,vn on one
flank, so that you can operate on the other side of the endgame.
SHIROV VS. ALMASI

Tilburg, 1996
a b c d e f g h
[178] Black to move
White has a kingside pawn majority, but that's about it. Ho\v can
White be expected to win? These are t\vo strong Grandmasters, and
one does not expect a blunder. This early in the game there is unlikely
to be time pressure. It is instructive to see ho\v Shirov imposes his \Vill
on the position, with a little help from his opponent. Just goes to show
that you shouldn't abandon an endgame as drawn, especially a knight
endgame!
2l . . . Nf5; 22.Ne4 c5. 22 . . . Nd6; 23.Nxd6+ cxd6; 24.c4 is a winning
endgame, though it requires great depth. See diagram #1 0 1 . 23.g4.
3
2
a b c d e f g h
12 1
There are five plausible retreats for the knight. Almasi does not
choose wisely. 23 Nh6?! The alternatives are:.
...
23 . . . Ne7; 24.Ng5 Nd5; 25.Kf3 h6; 26.Ne6 Kfi; 27.f5 Nb4; 28.Nxc7
Nxc2; 29.a4 Nd4+; 30.Ke4 is going to be very difficult for Black be
cause White can win the pawn at b6 by playing Nd5. 23 . . . Nd6?; 24.Nxd6+
cxd6; 25.c4 reaches the king and pawn endgame of diagram # 1 07.
23 . . . Nh4; 24.f5 ! forces 24 ... h6; 25.Kg3 g5 but the knight is still
trapped and White can force an exchange by maneuvering the knight
to f3.
23 ... Nd4 is the most active move and offers the best defensive pros
pects. 24.c3 Ne6; 25.Ke3 (25.f5 Nf4; 26.Ke3 Nd5+ should hold. ) 25 . . . Ke7;
26.f5 N£8; 27.Kf4 Nd7; 28.Ng5 N£8 (28 . . . h6; 29.Ne6) 29.a4 Kd6; 30.Ne4+
Ke7; 3 l .g5 Nd7; 32 .h4.
3
2
a b c d e f g h
[180] Black to move
White has a lot of space, and that is important in a knight endgame.

Black's position is passive, but how to break through? Meanwhile, Black
has the move. Black does not want to move the knight, when f6+ may
be good. There are still a lot of options.
32 . . . c6 weakens the kingside, but does prepare . . . b5 . 33.h5 b5; 34.a5
c4 leaves too many holes on the dark squares. 35.Ke3! c5 (Or else Kd4
and Nc5 . ) 36.h6! and now a rather forcing line is 36 . . . g6; 37.f6+ Ke6;
38.f7 Kxf7; 39 .Nd6+ Ke6; 40.Nxb7 Kd5; 4 1 .Nd8 Nb8; 42 .Kf4 Na6;
43.Nf7! Nc7; 44.Nh8 Ne6+; 45.Kg4 Nd8; 46.a6 Kc6 (46 . . . Nc6 allows the
tactic; 47.Nxg6 ! hxg6; 48.a7 Nxa7; 49.h7) 47.Kf4.
122
KNIGHT ENDGAMES
3
2
a b c d e f g h
[181] Black to move
Black cannot afford to go after the a-pawn with 4 7 . . . Kb6 because of

48.Nxg6 and after 48 . . . Ne6+; 49.Kf5 Nd4+! (49 . . . 1'\dS; 5 0 . 1'\eS the
kingside pa\vns break through. ) 50.cxd4 cxd4; 5 l . Nf4 d3; 52 .g6 hxg6+;
53.Ke4 d2; 54.Nd5+ Kxa6; 55. Nc3 b4; 56. Nd l and the h-pa,vn promotes.
There is no salvation in 48 . . . hxg6; 49. Ke5 N£7+; 50.Kf6 Nh8; 5 1 .Kg7
either.
Therefore, Black uses the defense 47 . . . Ne6+! ; 48.Kg4 Nd8! No\v
the sac doesn't work. 49.Nxg6 hxg6; 50.h7 N£7; 5 1 .Kf4 Kb6; 52.h8Q
Nxh8; 53.Ke5 with one line following 53 . . . Kxa6; 54.Kf6 Ka5; 55.Kg7
Ka4; 56.Kxh8 Kb3; 57.Kg7 Kxb2; 58.Kxg6 Kxc3. White can make things
a bit difficult. See the continuation in diagram #36. So, Black should
play 53 . . .Nf7+ so the knight will not trapped at h8, as Black's queenside
play is too slo\v. 54.Kf6 Nxg5 ; 55.Kxg5 Kxa6; 56.Kxg6 Ka5 ; 57.Kf5 Ka4;
58.Ke4 Kb3; 59.Kd5 Kxb2; 60 .Kxc5 Kxc3 \vith a vvin for Black.
24.Kf3 N£7; 25.h4.
a b c d e f g h
123
Black has three options, ignoring the queenside, where each pawn
advance creates a weakness. 24 . . . h6 is possible, but it seems that too
creates a hole, and after 25.h5 Black is being asphyxiated. That leaves a
king move, since the knight is not going to go to d6 for reasons we have
examined in the king and pawn endgames. (#1 0 1 , 1 07).
25 ... Ke7. 25 ... Kd7; 26.Ng5 is similar to the game. 26.Ng5 . If Black
exchanges knights we have a complex endgame which was examined
in position # 1 0 1 . 2 6 . . . Nd6. 26 . . . Nxg5+; 2 7 .hxg5 b5 is diagram #36.
27.Nxh7 Nc4.
a b c d e f g h
[1831 White to move
The knight is curiously ineffective. It can grab the b-pawn, but Black
still can't achieve a pawn breakthrough on the queenside.
28.f5 Nxb2; 29.g5 ! Nc4. 29 . . . Kf7; 30.g6+ Ke7.
a b c d e f g h
[1841 White to 1nove
White uses a small sacrifice to get the f-pawn off of g7. 31.f6+!
gxf6; 32.h5 Nc4; 33.g7 Kf7; 34.h6 Ne5+; 35.Kf4 Ng6+; 36.Kg4 Ne5+;
37.Kffi Nd7; 38.g8Q+ Kxg8; 39.Nxf6+ Nxf6; 40.Kxf6 Kh7; 4 l .Ke7 and
White wins.
124
KNIGHT ENDGAMES
30.h5 b5. 30 . . . Ne5+; 3 l .Ke4 Nf7; 32 .h6! Nd6+; 33.Ke5 gxh6; 34.g6
and it is over. 3l.Kf4. Eliminating any knight check. 3l . . . Na3 .
3
2
a b c d e f g h
32.h6! gxh6; 33.g6 Nc4; 34.£6+ Ke6; 35 .£7 Ne5; 36.f8N+! 36.f8Q?
Nxg6+; 3 7.Ke4 Nxf8; 38.Nxf8+ Kd6, which gives Black three pawns for
the knight.
a b c d e f g h
[ 1861 Black to move
39.Kf5 Kd5 ; 40 .Ne6 Kc4; 4 l .Nxc7 Kc3; 42 .Nxb5+ Kxc2; 43 .Nd6

with a draw.
After the underpromotion, Black resigned. The result was a fore
gone conclusion. 36 ... Ke7 loses to 3 7.Kxe5 and 36 . . . Kd6 is met by 37.g7.
KNIGHT + l US. KNIGHT + l

Positions 'vith this many pawns tend to involved blockaded and
balanced pawn structure. In many cases defense is possible simply by
moving the king back and forth. There has to be a real \veakness, which
also must be accessible to attack by king or knight, in order to make
125
progress . Still, there are times when one side is simply too passive, and
this can lead to defeat, as we see in the next example.
WILEY VS. MAZI

Bled Open, Slovakia, 1 999
3
2
a b c d e f g h
[187] White to move
With even pawns in a balanced configuration, you might think that

White has no real chances. In fact, White has two major advantages.
The White Knight will take up a powerful position at d6, after which
the Black knight is immobilized by the need to defend the pawn at b7.
The king can infiltrate the queenside via a4-a5-b6. Black cannot create
any openings for his own king to get across the meridian. Pawns at e4
and h4 can be maintained, allowing no entrance on the kingside.
40.Nd6 ! Kf8; 4 l .f3. This is not needed, and only weakens the
kingside pawns. If White sends the knight on other duties, Black can
play . . . Ne6-d4 and attack the f-pawn while threatening . . . Nc2. Never
theless, it does not eliminate winning chances.
4 l . f4 can, and should be played immediately. White does get to it
in a few moves. It is more effective right away, for example 41 ... Ke7;
42 .f5 where 42 . . . gxf5; 43.Nxf5+ K£7; 24 .Nd6+ followed by Kh3 and g4,
establishing a passed pawn. Black makes White work a bit more after
42 . . . g5, where 43 .g4 ! must be found.
126
KNIGHT ENDGAMES
3
2
a b c d e f g h
[1881 Black to 1nove
T his surprising move fo rces a \Vinning e ndgame. 4 3 ... hxg4

(43 . . . gxh4; 44.gxh5 K£8; 45 .Kh3 Kg7; 46 .Kxh4 Kh6; 47.Ne8 ! \vins the f
pawn. ) 44.hxg5 fxg5; 45 .Kg3 Kf6; 46.Kxg4 puts Black in zugz,vang.
4 1 . . . Ke7.
3
2
a b c d e f g h
42 .Nc4. The knight retreats to guard the a-pawn, \Vhich \vill be

relevant if the Black knight gets to d4 and threatens . . . Nc2xa3 .
42 . . . Kd7. The king relieves the knight of its defensive obligations.
43.K.f2 Ne6; 44.Ke3 Nd4; 45 .f4!
127
CARDOZA PUBLIS"ING • ERIC SCHILLER
3
2
a b c d e f g h
[190] Black to move
· 45 . . . Ke6; 46.fxe5 ! fxe5; 47 .Kd3 . With c2 covered, the White knight

is free to go after the pawn at b 7. 4 7 . . . Kf6? 4 7 . . . Nb3 is correct. What
can White accomplish if Black just shuttles the knight between d4 and
b3? 48.Nd6 b6! ; 49.Nb7 b5 and the knight cannot maneuver to attack
the a-pawn.
48.Nd6. Black cannot defend the pawn at b7. 48 . . . g5; 49.Nxb7!
gxh4; 50.gxh4 Ke6.
a b c d e f g h
[191] White to move
Perhaps Black counted on the dual threat of . . . Nf3 and . . . Nc2 to get
the pawn back. Establishing material equality is not sufficient, how-
ever.
5 1 .Nd6.
128
KNIGHT ENDGAMES
3
2
a b c d e f g h
[192] Black to move
Black can chase the h-pawn or the a-pawn. The decision is crucial,
since one draws and one loses !
5l . . . Nb5? Black goes after the wrong pawn. After 5 l . . . Nf3; 52 .Ne8
there are two options. 52 . . . Kd7; 53.Nf6+ Ke6; 54.Nxh5 Nxh4; 55.Ng7+
Kd7; (55 . . . Kf6; 56.Ne8+ Ke7; 57.Nc7 and the a-pawn goes. ) 56.a4 N£3;
57.Nf5 Ng5; 58.Nd6 keeps everything defended and White can oper
ate with the king on either flank.
However, 52 . . . Nxh4; 53.Nc7+ Kd7; 54.Nxa6 Kc8; 55 .a4 Kb7; 56.b5
N£3; 57.Nb4 Nd4; 58.Kc4! h4; 59.Nd3 h3; 60 .Nf2 h2 is only a draw,
because the knight is tied down by the h-pawn.
52.Nf5 Nxa3; 53.Ng7+ Kf7; 54.Nxh5 .
a b c d e f g h
[193] Black to move
Black has three weak pawns and either the king or knight will be
occupied on the kingside.
54 . . . Nb5; 55.Ng3 Nd4; 56.Nf5 Ne6; 57.Kc4 Kf6; 58.Kb3 !
129
3
2
a b c d e f g h
[194] Black to nzove
·The idea is simply Ka4-a5xa6.

58 . . . a5. Exchanging a pair of pawns is not enough. 59.bxa5 Nxc5+;
60.Kc4 Na6. (60 . . . Nxe4; 6 l .a6 Kxf5 ; 62.a7 and the knight can't get
back in time. ) 6 l .Ne3. Black resigned. Because the Black king must
deal with the h-pawn, White can use the knight to 'vin both remaining
Black pawns with a trivial win afterwards.
130
BISHOP ENDGAMES
Bishop endings come in all shapes and sizes . Bishops of opposite
color, often effective attackers in the middlegame, lead to more drawish
situations in the endgames. Amateurs seem to both overestimate and
underestimate the drawing chances provided by bishops which are stuck
patrolling different diagonals, and a number of the examples presented
in this section will show surprising results.
Unlike knight endgames, bishop endgames tend to be far more
positional. In some respects, they are easier to evaluate because it is
not difficult to determine whether your pawn structure will help or
hinder your bishop, and make the same judgement on the enemy posi
tion. There are clear general goals regarding the use of a bishop: stop
enemy pawns and protect your own.
In order to achieve these goals you usually need to keep your bishop
happy with a flexible position and plenty of mobility. If your pawns
need defense, you can place them on squares of the same color as your
bishop. If there are bishops of opposite color, the pawns will be safe,
though the bishop will be less mobile, with one less square to occupy.
The bishop can find the pawn blocking a key diagonal, or may be forced
to remain at its defense rather than assist in any attack. When the bish
ops occupy the same color squares, the pawn on that color may be
attacked.
BISHO P US. 2
A bishop cannot win against two pawns, and is generally used to
save the game by sacrificing itself for one of the pawns when the other
can be captured or contained by the king.
13 1
SAHETCHIAN VS. GURIEU

Cappelle Ia Grande, France, 1995
63 . . . d3; 64.Kd4 d2; 65.Ke3 .
3
2
a b c d e f g h
[195] Black to move
Black did not promote to a queen, as that leads to a quick draw.

65 . . . d 1N+. Underpromotion to a knight is all that can be achieved,
because it comes with check. 65 . . . d l Q; 66.Bf3+ Kh4; 67.Bxdl g2; 68.Kf2
Kh3; 69.Bf3.
66.Kd2 Nf2; 67.Ke2 Kg4; 68.Be6+ Kf4; 69.Kfl.
3
2
a b c d e f g h
[196] Black to move
Black could not make any progress and the game was eventually
drawn.
132
BISHO P ENDGAMES
BISHO P + I US. I
In most cases, a bishop and a pawn defeat a lone pawn. The major
exceptions are when the defending king can force the win of the en
emy pawn or when the pawn is a rim pawn and the bishop cannot
control the promotion squares. An example of each exception is pre
sented below.
CONTINUATION OF KRAMNIK VS. SHIROV

WCC Candidates Final, Cazorla, 1 998
3
2
a b c d e f g h
[197] Black to move
65 . . . Kf5! Agreed drawn. White can prevent the Black king from
reaching the critical f4-square, but only temporarily.
a b c d e f g h
[198] White to 1nove
After 66.Bcl b3; 67.Kh5 b2 ; 68.Bxb2 Kf4, the pawn falls.
133
KEITLINGHAUS VS . ATALIK
First Saturday GM, Budapest, 1998
3
2
a b c d e f g h
The players agreed to a draw here. The technique is very simple.

White retreats the king to al and sits there.
38.Kc3 Ke5; 39.Kb2 Be6; 40.Kal Kd5; 4l.Ka2 Kxc5+; 42.Kal Kb4;
43.Kb2 a5; 44.Kal a4; 45.Kh2 a3+; 46.Kal. Black cannot make progress,
for example 46.Kal Kb3 ; 47.Kb l Bf5+; 48.Kal a2 is stalemate.
BISHOP + I US. BISHO P

With bishops of opposite color, the endgame with one pawn is al
most always drawn. As long as the defending bishop can cover the
promotion square, the half-point is secure. In some cases, promotion
can even be allowed, if there are stalemate tricks . This is a theme com
mon in studies but almost unheard of in practical play.
With bishops of the same color, a win can be achieved if the control
of the promotion square is secured and the enemy bishop cannot keep
an eye on it. The basic winning procedure is as follows .
134
BISHO P ENDGAMES
FROM A STUDY BY CRUM ( 192 1)

British Chess Magazine
3
2
a b c d e f g h
[2001 White to move
14.Bh7 forces the enemy bishop to stay on the h3-c8 diagonal.

14 . . . Bh3; 15.Bc8! The occupation of the promotion square is crucial
to victory. Now Black is chased from both diagonals.
15 . . . Bf1 ( 1 5 . . . Bg2; 1 6.Bd7 Bb7; 1 7.Bb5 and White wins. ) 16.Bg4.
White concedes the shorter diagonal, which is harder for Black to main
tain. 16 . . . Ba6; 1 7.Bf3 Kc5; 18.Be2 .
3
2
a b c d e f g h
[2011 Black to move
Black must allow the pawn to promote in a move or t\vo.
BISHO P + I US. BISHO P + I

In most cases endgames \Vith a single pawn each will end in a draw
as one side gives up the bishop for an enemy pawn. The exceptions
arise only when there are bishops of the same color and a pawn very
close to the promotion square, especially near the rim where the de-
135
fending bishop has less room to maneuver. Then we see some real
winning chances.
KAMSKY VS. SHIROV

Buenos Aires, 1 994
3
2
a b c d e f g h
[202] Black to move
66 ... Kd4. The threat now is . . . Be4, and it cannot be answered.

67 .h4. 67 .Bb 1 Be4 is a quick end. 67 ... Be4; 68.Be6 b2; 69.Ba2.
a b c d e f g h
[203] Black to move
69 . . Kc3 ! Winning the White bishop is meaningless if Black has no

.
pawns left! 70.Kf4 Bh7. All Black has to do now is bring the king to a3.
7 1.h5 Kb4; 72.h6 Ka3.
136
BISHO P ENDGAMES
a b c d e f g h
[2041 White to move
White resigned.
(33) KARPOV VS. LJUBOJEVIC

Melody Amber (Blindfold), Monte Carlo, 1996
3
2
a b c d e f g h
[2051 White to move
The FIDE World Champion, in a blindfold game against a top

Grandmaster, shows how technique kicks in, even when you are play
ing in the dark. To win with a rook pawn against bishop you n1ust oc
cupy the heart of the corner, in this case the b7-square.
87.Kc5 Kd3; 88.Kc6 Kc4; 89.Kb7. Black cannot keep the bishop at
b8, but abandoning the short diagonal means that the a-pawn will ad
vance. There is a single, albeit temporary, defense, and l�ubojevic sees
it.
89 . . . Kb5 !
137
3
2
a b c d e f g h
[206] White to move
90.Bf2. 90.Kxb8?? Kxa6 draws. 90 . . . Ka5 . Black must continue to sit

on the pawn, watching from either a5 or b5 . Karpov eliminates one of
those options.
9 1 .Bh4 Kb5; 92.Bel!
3
2
a b c d e f g h
[207] Black to move
Black has no way out.

92 . . . Ba7. 92 . . . g3 ; 93.Bxg3 Bxg3; 94.a7 and a new queen next move.
92 . . . Ka4; 93.Bg3 wins. 93.Kxa7 Kc6; 94.Kb8.
138
BISHO P ENDGAMES
3
2
a b c d e f g h
[208] Black to move
Black resigned.
BISHO P + 2 US. BISHO P

The stronger side should win the vast majority of these endgames.
There are some drawing chances with bishops of opposite color, if the
king can safely blockade one of the pawns. Defense is difficult, how
ever, and if the attacking pieces are well coordinated the game can be
won.
CONTINUATION OF SPEELMAN VS. RUBLEVSKY

Moscow Olympiad, 1994
3
2
a b c d e f g h
[209] Black to move
White wins because a bishop at e4 can always defend both a pawn

at c6 and a pawn at g6, if necessary. This frees the king to infiltrate and
eventually win the Black bishop for a pawn.
139
6l . . . Be5. Or 6 l ... Kc7; 62 .Kc5 Bd6+; 63.Kd5 Bf8; 64.g6 Bg7; 65.Bf3!
followed by Ke6-f7.
62.Kb6 Kc8; 63.Bg4+ Kd8; 64.g6 Ke7.
a b c d e f g h
[210] White to move
65.c7 Bxc7+; 66.Kxc7 Kf6; 67.Bf5 etc.
BISHOP + 2 US. BISHO P + I

Whether you have bishops of the same color or bishops of opposite
color, the winning technique 'vith an extra pawn is to force the oppo
nent to give up the bishop for one pawn. This is easiest when the pawns
are far apart. If one pawn is on the rim, however, you need to worry
about drawing when your bishop is of the opposite color of the promo
tion square.
KRAMNIK VS. SHIROV

WCC Candidates Final, Cazorla, 1998
3
2
a b c d e f g h
[211] White to move
140
BISHO P ENDGAMES
White can win Black's bishop but the outcome of the game is still
inevitable. The game will be drawn because winning the bishop takes
the White king too far from the weak f-pawn. The pawn itself defends
against invasion via d4 or g4, but the f4-square can only be guarded by
the bishop, which has other things to worry about, namely the promo
tion of the Black pawn.
64.h6! Bxh6.
3
2
a b c d e f g h
[2121 White to move
65.Kxh6.
3
2
a b c d e f g h
We now have a bishop and pawn vs. pawn endgame, but it is just a
draw! See the continuation at diagram # 1 97.
14 1
ASEYEV VS. PIGUSOV

Russian Championship, Elista, 1997
3
2
a b c d e f g h
[214] Black to move
The essential drawing concept in endgames like this is to target the

enemy pawn which cannot be protected by the bishop. Your bishop is
then free to capture White's other pawn, even as a sacrifice. In this
interesting example White tries to take control of the e5-square, to en
able him to advance the pawn in safety. Black has a fine defense that
manages to hold the draw.
62 .. Be5 ! Black grabs the critical square and is in position to give
.
the bishop up for the pawn in a clear example of the defensive strategy.
62 . . . Bd6 is inferior. 63.f6 Ke5; 64.£7 Bf8; 65.Bc7+ Ke6; (65 . . . Ke4; 66.Kf6
Kd3; 67.Bd8 Kc2; 68.Be7 transposes. ) 66.Kg6 Kd5; 67.Kf6 when the
Black king can move to c6 or e4.
One line runs 67 . . . Kc6; 68.Be5 Kd7; 69.Bf4 Be7+; 70.Kg7 Kc6;
7 1 .Kg8 Kd5; 72 .Bh6 Kd4; 73.Bf8.
3
2
a b c d e f g h
[215] Black to 1nove
142
BISHO P ENDGAMES
This is a typical winning position for White. 73 . . . Bxf8 (73 . . . Bg5;

74.Bg7+) 74.Kxf8 Kc3 leads to a classic example of queen vs . knight
pawn technique. 75.Ke7 Kxb3; 76.f8Q Q/1 ; 76 . . . Kb2; 77.Qf6+ Ka2;
78.Kd6 b3; 79.Qc3 b2; 80.Qc2 Ka l ; 8 l .Qa4+ Kb l . The White king ap
proaches, and each time the enemy king clears the promotion square,
the queen chases him back, freeing another tempo for the invasion by
the king. 82.Kd5 Kc l ; 83 .Qc4+ Kd2 ; 84.Qb3 Kc l ; 85 .Qc3+ Kb l ; 86.Kc4
Ka2; 87.Qa5+ Kb l ; 88.Kb3 Kc l ; 89.Qe l #.
67 . . . Ke4 is somewhat trickier. 68.Bd8 Kd3; 69 .Be7 Bh6; 70.Kg6 !
3
2
a b c d e f g h
[216] Black to move
This is how a bishop gets dominated when it is confined to a 3-

square diagonal.
Let's get back to the game after 62 . . . Be5.
3
2
a b c d e f g h
[217] White to move
63.Bc5. White forces the bishop to abandon the post. 63 . . . Bc3;

64.Bd6. The bishop is now tied to the defense of the b-pawn. Black
must now go after the White pawn at b3. 64 . . . Kd3; 65.Kf4 Kc2; 66.Be5.
143
3
2
a b c d e f g h
[218] Black to move
This position is drawn.

66 . . . Bd2+. 66 . . . Kxb3; 67.f6 Ka2; 68.£7 ( 68.Bxc3 bxc3; 69.£7 c2;
70.f8Q c l Q+ is a draw. ) 68 . . . Bd2+; 69.Kf5 Bh6 draws.
67.Kg4 Bc3. Agreed drawn.
KAMSKY VS. SHIROV

Buenos Aires, 1 994
3
2
a b c d e f g h
[219] White to move
Bishops of the same color tend to lead to fewer draws, because the
stronger side can use the bishop to block or control key squares. Bish
ops of opposite color cannot fight over the same square. In general,
the t'vo major considerations are how far a pawn is advanced and espe
cially how useful the kings are. White's king is far away from the action,
and has the single threat of marching the king to g4 and winning the g
pawn. That's fine, if he can get rid of the b-pawn, but that is not so
simple.
144
BISHO P ENDGAMES
64. Kg4 Bd5!; 65.Bf5 b3. White must beware the trick . . . b2 and . . . Be6,
but that is taken care of by capturing the enemy pawn.
66.Kxg5.
3
2
a b c d e f g h
[220] Black to move
You may be able to work out the rest for yourself, but if you need
help, review diagram #202.
BERELOVICH VS. SCHLOSSER

Goodricke International, Calcutta, 1 998
a b c d e f g h
[221] White to move
This position is a little deceptive. It would seem that White has no

winning chances because the bishop is the wrong one to assist the a
pawn. If Black can sacrifice the bishop at f2 for the f-pawn, then the
game will be drawn even if the two pawns fall. So how can White win,
or at least try to maximize winning chances? The key lies in the diffi
culty Black has in defending against the f-pawn . The king must assist,
and this keeps it far from the a-file.
145
42.Kf3 h2; 43 .Kg2 Bg3; 44.a4!
3
2
a b c d e f g h
[222] Black to tnove
44 . . . Bd6; 45.a5 Bc5; 46.a6 Bgl ; 4 7.Be5 ! The threat is Bxh2,

44 ... Kd7.
but Bb8 is also useful. 47 . . . Kf7; 48.Bxh2 Ba7; 49.Be5 . The king cannot
approach the pawn. 49 . . . Ke7; 50.Kf3 Kf7; 5 l .Kg4 Kg8; 52.Kh5 Kh7;
53.f6 Bb8; 54.f7 and White wins.
45 .a5 Kc6. The king heads for the queenside, leaving the bishop to
cope with the f-pawn. 46.f6! Bd6; 47.f7 Kb5; 48.Be5 ! ! 48.Bg7?? Kxa5;
49.f8Q Bxf8; 50.Bxf8 draws. 48 . . . Bf8; 49.Bc7 Kc6; 50.Bxh2.
3
2
a b c d e f g h
Black resigned. The White king will bring sufficient force to bear.
5 0 . . . Kb5 ; 5 1 .Bc7 h5 ; 52 .Kh3 Be7; 53 .Bd8! Bf8; 54.Kh4 Kc6;
55.Kxh5 Kd7; 56.a6! Kc6; 57.a7 Kb7; 58.Bb6 Ka8; 59.Kg6 Kb7; 60.Kh7
Ka8; 6 1 .Kg8 Be7; 62.Bd4 Kb7; 63.Bg7.
146
BISHO P ENDGAMES
a b c d e f g h
[2241 Black to move
White carries out the standard winning maneuver in bishop plus

bishop pawn endgames.
6 3 . . . Kxa7; 64.Bf8 Bg5; 65 .Bc5+.
(99) JUSSUPOW VS. KOTRONIAS

Germany, 1 996
3
2
a b c d e f g h
[2251 White to move
With bishops of the same color, and Black's remaining pawn per
manently planted on the same color square as White's bishop, there is
no reason for players to agree to a draw in this position. The position
should be drawn, because the Black king cannot be forced to abandon
f7. In this game, White tried to circle around to the queenside.
50.Ba3. First, the bishop will maneuver to c7. 50 . . . Be7; 5 1 .Bb4 Bf8;
52.Ba5 Be7; 53.Bc7 Bf8. Black has been forced into a more passive
defense. 54.Bd8 Bg7; 55.Kg5. White hopes to establish a position with
a pawn at f6, after which the Black bishop will run out of room.
147
55 . . . Bb2 ; 56.Bc7 Ba3 . Black kno,vs that the White pawn is not dan
gerous, so defends from the side of the board where there is more
room. 57.Kf5 Bb4; 58.Bb6 Bd2; 59.Bd4 Bb4; 60.Bf2 Ba3 ; 6 1.Ke4. All
other plans having proved useless, White tries the circle journey. It is
important to keep Black's king away from the f-pawn, however.
6 1 . . . Bb4; 62 .Kd3 K£6; 63.Ke4 K£7; 64.Be3 K£6; 65.Bcl Kg6; 66.Bb2
Bc5; 67.Bc3 Ba3 .
3
2
a b c d e f g h
[226] White to move
68.Ba5. White cannot go to the queenside while the Black king can
get to f5. 68...Kf7; 69.Bb6 Ke7; 70.Kd3 K£6; 7 1 .Ke4 Ke7; 72 .Bf2 Bb4;
73 .Bh4+ K£7; 74.Bg5 .
3
2
a b c d e f g h
[227] Black to move
Finally, a new plan, with the f-pawn defended. The king is now free
to roam.
74 . . . Ba3; 75.Kd3 Kg6?! A small mistake. 75 . . . Ke8 was correct. 76.Kc4
Kd7; 77.f5 Bb2 is evaluated by Kotronias as equal. Indeed, the draw is
secure because the king cannot be driven from d7. Even if White gets
148
BISHOP ENDGAMES
the pawn to f7, Black could defend with . . . Bg7. 78.f6 Ke8; 79 .Kb5 K£7;
80.Kc6 Be5! and the bishop shuttles on the d5-h2 diagonal.
76.Kc4.
3
2
a b c d e f g h
[2281 Black to move
This sets some problems. Black must hasten to bring the king back
to the e-file.
76 ...Kf7. 76 . . . Bc5; 77.Kb5 K£7; 78.f5 Ke8; 79.Kc6 Ba3; 80 .Bf4 Ke7
allows the trick ( SO . . . K£7 transposes to the game after the capture at
d6. ) 8 l .Bxd6+! Bxd6; 82.f6+ Kxf6 83 .Kxd6, winning.
77.Kb5 Ke8; 78.Kc6 Bc5; 79.f5 Bb4; 80.Bh4. 80.Bf4! almost works.
SO . . . K£7; 8 l .Bxd6 Bc3; 82. Bg3 Kf6; 83 .d6 Kxf5; 84.d7 Bf6 holds.
80 . . . Ba3; 8 1 .Bg3.
3
2
a b c d e f g h
[2291 Black to move
This returns the game to the correct plan. We have the same key
position but with the bishop at g3 .
8 1 ...Kf7; 82.Bxd6 Bb2 ; 83.Bg3 Kf6; 84.d6 Kxf5.
149
3
2
a b c d e f g h
[2301 White to move
The pawn is gone, but White's king can get to c8, and then the
bishop will get to the promotion square. This is the standard technique
for winning bishop and pawn vs. bishop of the same color. Unfortu
nately, it does not work when the pawn is a center pawn, because Black
has just enough room to keep the bishop guarding from a distance.
85.Bh4 Bc3 ! The best defense, bringing the bishop to aS.
86.d7 Ba5.
3
2
a b c d e f g h
{2311 White to tnove
87 .Kd6. White tries to keep the king out. 87 .Bg3 is the only alterna
tive suggested by Kotronias. 87 . . . Bd8; 88.Bc7 (88.Kb7 Ke6; 89.Kc8 BgS;
90 .Bc7 KdS ! ; 9 l .Bd8 Bd2 is the same as the next note on 87.Kb4 . . )
88 . . . BgS; 89.Kb7 Ke6; 90.Kc8 KdS ! (Not 90 ... Bh4?? which lets White
get away with 9 l .Bd8 Be l ; 92.BgS BaS; 93 .Bf4 wins. ) 9 l .Bd8 Bd2; 92.Bh4
BaS; 93.Bg3 Kc6.
ISO
BISHO P ENDGAMES
3
2
a b c d e f g h
[2321 White to move
This is the key drawing position. Black controls c7.

Note that 87.Kb7 no longer works, because the Black king gets to
c6. 87 . . . Ke6; 88.Kc8 Kd5; 89. Bg3 Kc6 ! White cannot win, because c7 is
controlled by Black.
3
2
a b c d e f g h
[2331 Black to move
87 . . . Kg6. 87 . . . Bb4+?? was one last chance to go wrong. After 88.Kc7;

Ba5+ 89.Kc8 Ke6; 90.Bg3 Black cannot prevent Bc7. The difference
between this and the drawing lines is that from d5, Black can get the
king to c6.
88.Ke7 Kf5; 89.Kd6. 89.Ke8 Ke6; 90.Bd8 Bd2; 9 l .Ba5 Bg5; 92.Bb4
Bf6 and in a mirror of lines presented above, White cannot make use
of e7. 89 . . . Kg6; 90.Ke6 Bb6; 9 1 .Bf2 Bd8; 92.Bd4 Bh4; 93.Bc5 Bd8.
White properly abandoned his efforts at this point and the game ended
in a draw.
15 1
CONTINUATION OF SHIPOV VS. GOLDIN

Russian Championship, Sankt Petersburg, 1 998
3
2
a b c d e f g h
[2341 Black to move
White wins as long as the Black king cannot prevent the a-pawn
from queening. While the bishop is in charge of defense of a8, the
White king can eventually get in position to win the bishop by escort
ing the pawn. Note that Black does not have the resources seen in the
notes to move 54, because the White bishop guards the square in front
of the f-pawn and it cannot be deflected from the diagonal by the Black
pawn, which can be captured at b2 while defending f6.
56 . . . Kc8. Black must shuffle the king in such a way that the White
king cannot gain access to c7, where it would be able to support the
advance of the pawn. 57.Bd4 Kd7; 58.Bg7 Kd6; 59.Bc3. These moves
were likely the result of time pressure. At move 60, White executes the
winning plan.
59 ... Kd7.
3
2
a b c d e f g h
[2351 White to tnove
152
BISHO P ENDGAMES
After 59 . . . Ba8; 60.f6 B£3; 6 l .Kb5 . White is heading not for the pa,vn
but for the kingside! 6 l . . . Ba8; 62.Kc4 Bd5+; 63.Kd4 Ba8; 64.Kc3 Bc6;
65.Kf4 Ke6; 66.Kg5 .
3
2
a b c d e f g h
Black loses on 66 . . . Kf7; 67.Kf5 Bf3; 68.Ke5 ! and the king gets back
to the queenside to assist the a-pawn.
66 . . . Be4 is met by 67.Bb2 . Neither the bishop nor pa,vn can move
safely, so Black is forced to use the king, after which the queenside
opens up. 67 . . . Kf7; 68.Kf4 Ba8; 69.Ke5 etc.
60.Kc5 ! Kc7; 6l.f6 Kd7; 62.Kc4!
3
2
a b c d e f g h
White gobbles the pawns, returns up field to assist one of the pa,vns
so that Black must sacrifice the bishop, then promotes the remaining
pawns without difficulty. The important point is that the Black king
cannot keep close enough to the t\VO pawns to be an effective defender,
even with bishops of opposite colors.
153
CONTINUATION OF RUBLEVSKY VS. SPRAGGETf

FIDE World Championship Knockout, Groningen, 1997
3
2
a b c d e f g h
This position looks a lot like a composed study. White must play
very carefully and avoid minefields that can lead to draws.
6 1 . . . Bd2; 62.Bd4 Bb4. Black's bishop must be able to capture at c5
if the pawn advances, as this is the basic survival tactic in the position.
3
2
a b c d e f g h
[2391 White to move
63.Be5 . 63.c5 would have been a tragic mistake, since there is a

swindle. 63 . . . Bxc5 ; 64. Kxc5 Kh8; 65.Kd6 Kg8; 66.Ke7 Kh8 and no
progress can be made. 67.Be3 (67.Kf7?? is stalemate ! ) 67 . . . Kg8; 68.Bg5
Kh8; 69.Bf6 Kg8! White cannot force the enemy king out of the corner.
Black had only to avoid (69 . . . gxf6; 70.Kf7 f5; 7 l .g7+ and White still
wins. )
63 . . . Kh8; 64.Bd6.
154
BISHO P ENDGAMES
3
2
a b c d e f g h
[2401 Black to move
Black resigned. The bishop is forced from the diagonal, and the
pawn advances, for example 64 . . . Ba5; 65 .c5 Bd8; 66.Kc6 Kg8; 67.Kd7
Ba5; 68.c6 Bb6; 69.Bc7 Be3; 70.Ba5 Bf4; 7 l.Bb4! Threatening Bd6,
and the pawn marches.
71 ... Bb8; 72.Bd6. No further resistance is possible.
BISHO P + 2 US. BISHO P + 2

As we add more pawns to the bishop endings, more targets arise.
The more accessible the pawns are to attack, the more tactical the game
becomes . With fewer pawns, you want to try to place your own pawns
in positions where they cannot be attacked, freeing your pieces to go
on the offensive.
SHIPOV VS. GOLDIN

Russian Championship, Sankt Petersburg, 1998
3
2
a b c d e f g h
[2411 White to move
ISS
This endgame requires a great deal of precision. Neither advanc

ing the a-pawn nor capturing the b-pawn will lead to a win.
54.Bb2! 54.Kxb5 . Allows the Black king to get to a8, after which no
win is possible because the bishop can be sacrificed for the f-pawn,
leading to a bishop of the wrong color rook pawn endgame. 54 . . . Kc7;
55.f6 Bh5; 56.Bb2 Kb8; 57.Bxe5+ Ka8. There is no possible win. 54.a7
BaS; 55.Kxb5 Ke8; 56.Bb2 e4; 57.Bd4 Kf7 and Black will be able to get
to the pawn at f5, by sacrificing the e-pawn at e3 or e l and moving the
king to f6.
54 . . . b4. Black has a number of alternatives. 54 . . . Kd6; 55.f6 Bd5;
56.Bxe5+! Kxe5; 57 .Kc5 ! This clever move secures the win, since Black
cannot capture at f6 and one of the White pawns must advance. 57 . . . b4;
58.a7 b3; 59.£7 Bxf7; 60.a8Q b2; 6 l . Qb8+.
· 55.Bxe5 b3; 56.a7. Even though there are bishops of opposite color
and White cannot control a8, White will win. The continuation is pre
sented in position #234.
3
2
a b c d e f g h
[242] Black to move
CONTINUATION OF TUFFERY VS. SINCLAIR

New Zealand Championship, Wellington, 1995
3
2
a b c d e f g h
[243] Black to move
156
BISHOP ENDGAMES
Black has the advantage, because his pawns are further up the board,
are better protected, and the center is under control. White's bishop is
marooned on the queenside, for the moment, and the king has the
burden of having to fight against both pawns.
43 ... g3; 44.Kfl Kf3 ; 45.Ke l .
3
2
a b c d e f g h
[244] Black to move
45 .h5?? is suicide: 45 . . . g2+; 46.Ke l g l Q#.

45 ... g2; 46.Kd l glQ+; 47.Be l . This is only worth playing out for
the cute finish.
3
2
a b c d e f g h
[245] Black to move
47 . . . Ke4! and mate by . . . Qg4 next move. A dozen steps shorter than
the plan actually seen in the game! The pawn break is often the key to
rapid endgame victory.
157
(46) GUREVICH, D VS. FEDOROWICZ

New York Open, 1997
3
2
a b c d e f g h
[246] White to move
White wins because the Black bishop cannot comfortably defend

b8. It isn't necessary to march the king to the queenside. In fact, that
would just lead to a draw. Instead, the whole business can be carried
out from long range.
After 50.Kg2 Be5; 5 1 .Bgl . Black resigned. The bishop will be lost.
5 1 . . . g5; 52.b7 Bb8; 53.Bh2.
3
2
a b c d e f g h
[247] Black to move
The bishop must abandon b8, and is lost.

53 . . . Ba7; 54.hxg5! 54.b8Q+ Bxb8; 55.Bxb8 gxh4 would of course
be drawn. 54 . . . Kf7; 55.Kh3. The king must be used to protect the g
pawn. 55 . . . Kg6; 56.Kh4 Bf2+; 57.Bg3 Ba7; 58.b8Q Bxb8; 59.Bxb8 etc.
158
BISHOP ENDGAMES
CONTINUATION OF RUBLEVSKY VS. SPRAGGETT

3
2
a b c d e f g h
In this pure bishop endgame, Black has two vulnerable pawns, and
the situation is already hopeless.
58 ... Kg8. 58 . . . Ke8; 59.Bd6 Bd4; 60.Kd5 Kd7; 6 l .Bxc5 Bf6; 62.Bd4
is simple enough. 59.Bd6. The bishop is now tied to the c-pawn, and
White can bring the king to finish it off. Meanwhile, the Black king is
trapped in the corner. 59 . . . Bf2; 60.Kd5 Be3; 6 l .Bxc5.
3
2
a b c d e f g h
This is a winning B + 2 vs . B + 1 endgame. See diagram #238 for the

continuation.
BISHO P + 3 US. BISHOP + 2

With the advantage of an extra pawn, the strategy is to build enough
pressure so that your opponent must sacrifice the bishop when you
still have a pawn remaining. Keeping your bishop active and working
159
from a distance is even more important as the number of pawns in

crease.
TUFFERY VS. SINCLAIR

New Zealand Championship, Wellington, 1995
3
2
a b c d e f g h
[250] Black to move
This position is a simple win for Black, since the White b-pawn
cannot cross the sixth rank. Nevertheless, there is the short way and
then there are the long ways. Black missed the efficient winning move.
42 . . . h4! Black did manage to '\vin after 42 . . . Bb6; 43.Bf6 Bc7; 44.Bh4
Kd4; 45.Bg5 Bxg3+; 46 . Kd2 Kc5; 47.Kxd3 Kxb5; 48.Ke2 Kc4; 49.Kfl
Kd3; 50.Kg2 h4; 5 l . Be7 Ke2; 52.Bf6 Be l ; 53.Be5 h3+; 54.Kh l K£3;
5 5 . Bd6 Bd2; 5 6.Be7 g3 ; 57.Bh4 g2+; 58.Kh2 Bf4+; 5 9 . Bg3 Bxg3+;
60.Kxh3 giN#. Such a long journey was not necessary.
43.gxh4. 43.Kfl h3; 44.Bb4 K£3; 45.Ke l h2 and mate follows from
a new queen. 46.Kd l . For continuation of diagram #25 1 , see diagram
#243.
3
2
a b c d e f g h
[251] Black to move
160
BISHO P ENDGAMES
CONTINUATION OF DE FIRMIAN VS. FISHBEIN

World Open, Philadelphia, 1 997
3
2
a b c d e f g h
[252] Black to move
The king can now infiltrate the queenside and pick off the a-pawn.
What transpires on the kings ide is of little importance.
44 ... Ke5; 45.Kd3 Bb7; 46.Kc3 Ke4; 47.Bd l . White keeps control of
the d l -h5 diagonal. 4 7 . . . Ke3; 48.Kb4 K£2; 49.Bxa4 Bc8; 50.Be8 Kxg3.
Black has a passed pawn, but it cannot get passed the h5-square .
5 1.Kc5 Bb7; 5 2.b4 Bf3; 53.b5 h5 .
a b c d e f g h
[253] White to move
White correctly evaluates the endgame after the bishop sacrifice at

h5. This is the cleanest win.
54.Bxh5 ! Bxh5. There is no point in playing on if Black lets the
bishop live. 55.b6 Bf3. Forced, as otherwise the pawn gets to b7. 56.a4
Kf4. The king tries to get back in time, but fails. 57.Kd6. Black re
signed. 57 Kf5; 58.Kc7 Ke5; 59.b7 Bxb7; 60.Kxb7 Kd5; 6 l .a5. The
. . .
pawn cannot be stopped.
16 1

With an equal number of pawns, it is configuration that counts.
Doubled pawns are a liability, and connected pawns are an asset. Con
nected passed pawns are invaluable !
BARBERO VS. BARTHEL

Ambassador Open, 1 996
a b c d e f g h
[254] White to move
The power of the connected passed pawns overcomes even bishops

of opposite colors.
42.g5 d3; 43.Ba4 d2; 44.Kg4 Bc5; 45.h6+ K£7; 46.Bb3+ Kg6; 47.Bc2+
Kf7; 48.g6+.

Two extra pawns may be meaningless, especially if one is doubled.
Bishops of opposite colors are necessary for defense, in most cases.
Our example exploits the creation of a permanent outpost for the bishop
in a square blocking the path of an enemy pawn. If you can manage
that, and your king can get in the way of the remaining pawn, all that
you need is a safe square for your bishop and the pawns will be para
lyzed.
162
BISHO P ENDGAMES
KRASENKOV VS. KOSTEN

Asti, 1 996
3
2
a b c d e f g h
[2551 Black to move
There is nothing to fear in this endgame. After a few more moves it

was agreed drawn.
36 . . . g5; 37.Bc8 g4; 38.Kg2 Kc7. There is no way to make progress.
39.Ba6 Be5; 40.Bd3 f4; 4 l .Bf5 g3 42.fxg3 fxg3; 43.Bd3 and White sim
ply moves the Bishop back and forth.
3
2
a b c d e f g h
[2561 Black to move

It often happens that an endgame involves an extra pawn on a flank.
With bishops of opposite color a defense is possible, but only if the
defenders pawns are 1 00% safe. Such situations are rare, and winning
possibilities exploit the possibility of forcing the enemy king to aban
don defense of the pawn.
163
FROIS VS. TERAN ALVAREZ

Santo Antonio International, Portugal, 1 999
3
2
a b c d e f g h
[257] White to move
If you are contemplating entering into an endgame like this, you

need to keep in mind that bishops of opposite color cannot cure all ills.
White's pawns are safe on the light squares, but the Black pawns are
invulnerable to the bishop. Black may be able to infiltrate the queenside
and escort the a-pawn.
If the White king strays too far from the center, Black may be able
to take the path Kd6-c5-d4-e3-f2, and the bishop can only defend f3
from e2 or d l . In this case you have to consider the use of the a-pawn
as a decoy, hoping to exchange it for two of White's pawns, after which
the connected passed pa\vns will win. If you can visualize all this in
advance, you have to recognize that Black has winning chances. Are
they sufficient against perfect play? That requires empirical analysis.
43 .Ke2 Kd6; 44.Kd3 Kc5 ; 45.Bb3 Kb4; 46.Kc2.
a b c d e f g h
[258] Black to move
164
BISHO P ENDGAMES
This much is pretty easy to figure out. If the a-pawn moves for
ward, White plays Ba2 and shuttles the king between c2 and d3 as
needed to keep the enemy king out. Since retreating to c5 is not going
to achieve anything against Kd3, what is Black to try? It turns out that
there is a tactic to activate the bishop.
46 ... Bc l ! ; 47.Be6 Be3; 48.Bf7. White has nothing better to do than
move the bishop back and forth on the a2-g8 diagonal, keeping the
Black king from c4. If Black plays . . . Kc5, to try to get to d4, White
maintains man-on-man coverage with Kd3.
4 8 . . . Bb6; 49.Be6 a4; 50.Bf7 Ba5; 5 l .Be6 Kc5; 52.Kd3 a3; 53.Ba2
Bd8.
3
2
a b c d e f g h
[2591 White to move
Black conceded the draw here. The superb coordination of king

and bishop held the position together.
SPEELMAN VS. RUBLEVSKY

Moscow Olympiad, 1 994
3
2
a b c d e f g h
[2601 White to move
165
CARDOZA PUBLIS HING • ERIC SCHILLER
By all rights this game ought to be drawable. After all, there are
bishops of opposite colors and in addition White cannot control h8 so
that cannot be used as a promotion square. In an important team com
petition, Speelman had to rely on all of his endgame ability to squeeze
a full point out of this position.
44.c5. The pawn advances so that it can get past c5 quickly before
the enemy bishop can shut the square down. Now only one dark square,
c7, remains on the path to the promotion square. 44 . . . Kf6; 45.c6 Ba5.
The pawn is stuck at c6, but (long range) either the Black bishop or
king must stand guard at c7. 45 . . . Ke7?; 46.Bxf7 Kd6; 47.Be8 and White
wins . 46.£3. White keeps the Black king from infiltrating. 46.f4! was a
try. Play would then reach a position similar to 48.f4 below.
46 . . . Ke7.
3
2
a b c d e f g h
[261] White to move
46 . . . g5? ! ; 47.hxg5+ Kxg5; 48.Bxf7 wins a pawn, and 48 . . . h4 is met

by 49.g4, according to Speelman, but this is not the end of the story.
With 49 . . . h3 ! Black plans to play . . . Bc7, . . . h2, . . . Kh4 and White will
have no way of advancing the pawns since the king must remain at g2
or h l to prevent the pawn from queening. White will not be able to
break the bind without giving up the f-pawn to clear the diagonal for
the bishop. 50.Kg l Bc7 ! ; 5 1 .Bc4 Kh4; 52.Bfl h2+ (52 . . . Kg3??; 53 .g5
and White gets a new queen. ) 53.Kg2 (53.Kh l Bf4 is no different. )
53 . . . Kg5; 54.Bd3 Kf6; 55 .Be4 Kg5 with a draw.
4 7 .Kg2 Bd8; 48.f4.
166
BISHO P ENDGAMES
3
2
a b c d e f g h
[262] Black to move
White is making a little progress. Sooner or later Black should play

. . . f5, even though that leaves the g6-square weak. It is hardly surprising
that Black was reluctant to advance the pawn, but this hesitation leads
to trouble.
48 . . . Bb6. Or 48 . . . f6; 49.Be4 (49.Bb3 Kd6; SO.Bc2 f5; S 1 .Ba4 BaS;
S2.Bb3 Ke7! is similar. ) 49 ... f5; SO.BdS Bc7 (SO ... Kd6??; S 1 .Bf7 Kxc6;
S2.Bxg6 Kd6; S3.Bxf5 is a simple win, with K£3 , Bg6, BxhS. ) S 1 .Kf3
Ba5 is a good defense, because if the White king comes to the queenside,
. . . Be 1 picks off the pawns. Black will just move the bishop on the aS-e 1
diagonal, keeping the king at e7 to guard the c7 -square.
49.Kf3 Bc7. 49 . . . fS ! see the 48 . . . f6 line. 50.Bb3 Ba5. 50 . . . f5; S 1 .Ke2
Bb6 leads to a split point. On S2.Ba4 BaS; S3.Kd3 Kd6; S4.c7, S4 . . . Kxc7;
SS.Ke2 Kd8 maintains the opposition. S2 .Bc4 can be tried, but then
S2 . . . BaS; S3.Ke3 Bb6+; S4.Ke2 BaS; SS.Bg8 Kf8; S6.Bh7 Kg7; 57.Kd3
Kxh7; 58.Kc4 Kg7; 59.Kd5 Kfi; 60.Kd6 Kf6 keeps the opposition, where
6 1 .c7 Bxc7+; 62.Kxc7 Ke7 is a drawing line pointed out by Speelman.
5 l .Ke4.
3
2
a b c d e f g h
[263] Black to move
167
5 I . . . Bb6 ; 5 2 .Kd5 Bf2 . S 2 . . . fS j ust doesn't seem to come into

Rublevsky's calculations. 53.Ba4 Bb6. 53 . . . Kd8; S4 .Kd6 Bb6; SS.Bb3
forces S5 . . . Bc7+; S6.KcS Ke8 and White circles around the queenside.
S 7.KbS Ke7; S8.Ka6 f6; S9.Kb7 and Black cannot guard both c7 and £7.
54.Bb5.
3
2
a b c d e f g h
[264] Black to move
Black has several defensive plans. The game saw 54 . . . Ba5 .

S4 . . . f5 may be too late now. SS.Ba4 can turn out well: SS . . . Kd8;
S6.Kd6 BaS; S7.BbS Bb4+; S8.Ke6 Be l ; 59.Kf6 Bxg3; 60.Kxg6 Bxf4;
6 l .KxhS Bg3; 62 .KgS Ke7; 63.Kg6 Kf8; 64.Bc4 Ke7; 6S .hS Kd6; 66.BbS
f4; 67.h6 Be l ; 68.Kf5 f3; 69.Ke4 £2; 70.Kd4 Bd2; 7 l .h7 Be l ; 72 .Kc3 is a
wonderful piece of analysis by Speelman. SS.Ba6 is also good. SS . . . BaS;
S6 .Bc8 Kd8; S7 .Be6 Ke7; S8.Bg8 Kf8 ; S9.Bh7! Kg7; 60.Bxg6 Kxg6;
6 l .Kd6 Kf7; 62 .c7 Bxc7+; 63.Kxc7 Ke7; 64.Kc6 Ke6; 6S .KcS ! )
S4 . . . Bc7; S5.KcS f6 and White must avoid S6.Bd3, which lets Black
escape via S6 . . . gS; S7.fxgS fxgS; SS.hxgS Bxg3 when S9.Kb6 h4; 60 .Bc4
Kf8 ; 6 l .c7 Bxc7+; 62 .Kxc7 Kg7; 63.Bd3 h3; 64.Be4 h2; 6S.Kd6 h l Q;
66 .Bxh l Kg6 dra\vs . The correct choice i s S6.Bc4, which wins. S6 . . . Kf8
( Or S6 . . . gS; S7.hxgS fxgS; SS.fxgS Bxg3 ; S9.Kb6 Kf8; 60.c7 Bxc7+;
6 l .Kxc7 Kg7; 62 .Bd3 h4; 63.Kd6 h3; 64.KeS h2; 6S .Be4.) 57.KbS gS;
S8.f5 Ke7; S9.g4 hxg4; 60.hS g3; 6 l .Bd5 Kf8; 62.h6 g2; 63 .Bxg2 Kg8;
64.Ka6 Kh7; 6S .Kb7 BeS; 66.c7 Bxc7; 67.Kxc7 Kxh6; 68.Kd6 . White
prevails in the end, as demonstrated by Speelman.
S4 . . . Kd8 ! was a strong defense, using the king to create threats
against the pa,vn at c6 which will keep the enemy bishop in defensive
mode. 5S.Kd6 BaS ; S6.Bc4 Bb4+; 57.Ke5 . Black cannot prevent the
White bishop from getting to g6 . It doesn't matter, however, since Black
can save himself with S7 . . . Bel ; SS.Bx£7 Kc7; S9.Be8! Bxg3; 60.Kf6 Bxh4+;
6 l .Kxg6 Bg3 !
168
BIS H O P ENDGAMES
55.Kc5 .
3
2
a b c d e f g h
55 . . . Kd8. 55 . . . f6 56.Bc4 Kd8 (56 . . . Bc7 transposes above to 55 . . . Bc7.)

57.Kd6 Bb4+; 58.Ke6 and the kingside pawns fall. 55 ... f5; 56.Bc4 Bd8;
57 .Kb5 Bc7; 58.Ka6 followed by Kb7.
55 ... Bc7 is logical. Speelman gives 56.Bc4 where 56 ... f5; 57.Kb5 Ke8;
58.Ka6 Ke7; 59.Kb7 Kd6; 60.Bf7 wins, so Black must try 56 . . . f6, in
which case 57.Kb5 g5 . Then 57 . . . Kd8; 58.Bf7 g5; 59.hxg5 fxg5; 60 .fxg5
Bxg3; 6 l .Bxh5 leads to a lost Bishop + 2 vs. Bishop endgame, shown in
diagram #209. On the other hand, 58.fxg5 fxg5; 59.hxg5 Bxg3; 60.Kb6
is also a win. White will play c7 and win the enemy bishop for the
pawn. Then the bishop will stop the h-pawn and the king will come to
the rescue of the g-pawn. For example:
60 . . . Kf8; 6 l .c7 Bxc7+; 62 .Kxc7 Kg7; 63.Bd3 h4; 64.Kd6 h3; 65.Ke5 .
60 . . . Kd8; 6 l .Kb7 h4; 62.g6.
60 ... h4; 6 l .c7 Bxc7+; 62 .Kxc7 h3; 63.Bd5 h2; 64.Kc6 and the king
must give ground. It cannot get to the pawn via f8, as seen in the 60 . . . Kf8
line.
56.Bc4.
169
3
2
a b c d e f g h
[266] Black to move
Black is lost.
56 . . . Be 1 . This plan, while effective in some of the previous varia
tions, comes much too late now. The alternatives are hardly attractive:
56 . . . f6; 57.Kb5 Bc7; 58.Bf7 g5; 59.hxg5 fxg5; 60.fxg5 Bxg3; 6 l .Bxh5
with a winning theme we have already seen, using the bishop to protect
pawns at g6 and c6 and the king to help advance one of the pawns.
5 6 . . . Kc7; 57.Bxf7 Be l ; 58.Bxg6 Bxg3; 59.f5 Bxh4; 60.Bxh5 is similar.
56 . . . f5; 57.Kb5 Bc7; 58.Bf7 picks off the pawns.
56 ... Ke7; 57.Kb5 Bc7; 58.Ka6 Kd6; 59.Kb7 f6; 60.Bf7 g5; 6 l .fxg5
fxg5; 62 .hxg5 Bd8; 63 .Bxh5 and the bishop cannot stop the g-pawn
because White will play c7.
57 .Kb6 Black resigned. White will get the pawn to c7 with check
and promote.
DE FIRMIAN VS. FISHBEIN

World Open, Philadelphia, 1997
3
2
a b c d e f g h
[267] Black to move
170
BISHO P ENDGAMES
This is an instructive endgame in a number of ways. Black has a

bad bishop, unable to attack any enemy pawns. White has a good bishop
with clear targets on the queenside. Ultimately, it is the seemingly safe
pa,vn at a4 that proves Black's undoing. White accurately plays the
position into a winning king and two pawns vs . king and bishop
endgame.
37 . . .Bd5; 38.Bh7. Why not 38.Bxd5 Kxd5; 39.g4 with an extra pawn
in a king and pawn endgame? Fishbein would know, he is the author of
a fine book on the subject!
3
2
a b c d e f g h
[268] Black to move
The problem is that Black controls the frontier, and White will not
be able to break through, or organize the advance of the b-pawn to b3 .
Whenever the White king reaches c2, Black will plant a king at c4.
(39.Kd3 Kc5; 40 .Kc2 Kc4; 4 l .g4 reaches the same position as the next
note.) Black' s next move will either draw or lose!
The bad move is 39 . . . Kc4?; 40.Kd2 Kb3; 4 l .Kc l Kc4; 42.Kc2 Kd5;
43.Kd3 Ke5; 44.c4 bxc4+; 45.Kxc4 K£4; 46.b4 axb3; 4 7 .Kxb3 Kxg4; 48.a4
h5; 49.a5 h4; 50 .a6 h3; 5 l .a7 h2. Then 52 .a8Q Kg3; 53.Qh l demon
strates a basic queen vs. rim pawn win.
A saving plan is available, however: 39 . . . Ke5 ! ; 40 .Kf3 K£6; 4 l .Kg3
Kg5; 42 .Kh3 h5 ; 43.gxh5 Kxh5 ; 44.Kg3 Kg5; 45 .Kf3 Kf5; 46.Ke3 Ke5;
47 .Kd3 Kd5 is only a draw! White's inability to penetrate the fourth
rank leaves no way to exploit the extra pawn.
38 . . . Bc4; 39.Kf3 Be6. 39 . . . Kd5 could have been played. 40.Kg4 Be2+;
4 l .Kf4 Kc4; 42 .Bc2 Bd3 presents two options.
43.Bxd3+?? Kxd3; 44.Kf5 Kc2 leads to a win for Black. 45.Kg6 Kxb2;
46.Kxh6 Kxc3; 47.g4 b4; 48.axb4 a3 ! ; 49 .g5 a2; 50.g6 a l Q; 5 l .g7 Qh l+;
52.Kg6 Qe4+; 53.Kh6 Qh4+; 54.Kg6 Qg4+; 55.Kf7 Qf5+; 56.Ke8 Qg6+;
57.Kf8 Qf6+; 58.Kg8 Kxb4 is the basic queen vs. knight pawn win.
Ill
43.Bdl ! wins, for example 43 . . . Bg6; 44.Ke5 Kd3; 45.Kf6 Be4!; 46.Kg7
Kd2 ! ; 47.Bh5 Kc l ; 48.Kxh6 Kxb2; 49.Be2 Bc6 ! ; 50.g4 Kxa3; 5 l .g5 Kb2 .
3
2
a b c d e f g h
[269] White to move
52.c4 would be the best chance. At least it threatens to advance the

g-pawn to g8, since Black cannot defend from d5. 52 . . . a3; 53.g6 a2;
54.g7 al Q; 55.g8Q Qh l+ holds the draw.
a b c d e f g h
[270] White to move
40.Bd3 Bd7; 4 l .Ke3 Bc6. 4 l . . . Kd5 is countered by 42.Kf4 Kc5;

43.Ke5 Bc6; 44.Be2 . From this station the bishop keeps the enemy king
from invading on the light squares, and also prevents . . . h5. The h-pawn
can now be hunted down. 44 . . . Bd7; 45.Kf6 Kd5; 46.Kg6 Ke4; 47.Kxh6
Ke3; 48.Bh5 Kd2; 49.g4 Kc l ; 50.g5 Kxb2; 5 l .g6 Kxa3; 52.g7 Be6; 53.Kg6
Kb2; 54.Kf6 Bg8; 55 .Bf7 Bh7; 56.Kg5 a3; 57.Kh6 Bb l ; 58.g8Q a2;
59.Bxa2 Bxa2; 60.Qg5 Bc4; 6 l .Qd2+. Or 55 . . . Bxf7; 56.Kxf7 a3; 57.g8Q
a2 when; 5 8 . Qg2+ Ka3 ; 5 9 . Qfl Kb2; 60.Qxb5+ Kal ; 6 l . Qa4 Kb2;
62 .Qxa2+ Kxa2; 63.c4 wins. Black has no more counterplay and the
White king will stroll oyer and help deliver checkmate.
112
BISHOP ENDGAMES
42.c4. White finally liberates the queenside. 42 . . . bxc4; 43 .Bxc4 Kf5;

44.Be2!
3
2
a b c d e f g h
White was aiming for this position. The remainder of this game is
in position #252.

In this section we present a lesson on the famous bad bishop. A
bishop is considered bad if it is on the same color as its own pa,vns .
Beginners find that strange but it makes sense when you realize that
every one of the pawn takes a\vay at least a square, and perhaps an
entire diagonal, from the bishop. The reduced mobility even makes it
hard to defend the friendly pawns !
VAGANIAN VS. YERMOLINSKY

World Team Championship, Lucerne, 1993
3
2
a b c d e f g h
173
CA RDOZA PUBLISHING • E RIC SCHILLE R
White's bishop is not just bad, it's disgusting! Having to defend a4

and another pawn is more than the poor thing can bear.
4 7 . . . g4! ; 48.fxg4 Bxg4.
3
2
a b c d e f g h
[273] White to move
Three pawns each, a blockade in the center, and only one real tar
get: the pawn at e4. The awkward position of the White pieces makes it
difficult to defend it. All Black has to do is insure that he is not some
how left with an a-pawn, since the bishop is of the wrong color to pro
mote the pawn at al .
49.Kc4 Be2+; 50.Bd3 Bf3 ! 50 . . . Bxd3+??; 5 l .Kxd3 and Black is in
zugzwang! 5 1.Kb4 Bxe4; 52 .Bfl Kf4. The king escorts the pawn. 53.a5 .
It doesn't really matter what White does, the game is already decided.
53 . . . Ke3 ; 54.axb6 axb6; 55.Kc4 d3 .
3
2
a b c d e f g h
[274] White to move
White resigned. 56.Bxd3 Bxd3+; 57.Kd5 Bxb5; 58.Kd6 Bc4; 59 .Kc6

b5 is simple enough.
1 14
BISHO P ENDGAMES
BISHO P + 5 US. BIS HO P + 4

When there are many pa,vns on the board, the availability of en
trances into the enemy position becomes critical. The chessboard be
comes a different kind of puzzle 'vhere the only goal is to someho'v get
your king across the meridian. Since kings cannot occupy adjacent
squares, they cannot directly oppose eac� other. Techniques such as
triangulation can sometimes seize the opposition, but only if the en
emy bishop is unable to move.
CHAN CHEE KEONG VS . FLORES JR.

Cairnhill Open, Singapore, 1998
3
2
a b c d e f g h
[275] White to move
Black's extra pawn is of no real significance. The key to the posi

tion is the ability of Black's king to get to e5, after 'vhich White can
easily run out of moves . It only took Black ten moves to win.
38.Bd l Ba2; 39.Be2 Ke5; 40.Bfl . This bishop must obviously re
main on the fl -c4 diagonal. It must also be available to defend the g
pawn as needed. 40 . . . Bb3; 4 l .Be2 . 4 l .Bd3 Bd l ; 42.Bxg6 Bxg4 is hope
less for White. 41 . . . Ba4.
175
C ARDOZ A PUBLISHING • ERIC SCHILLER
a b c d e f g h
[276] White to move
42.Bf3. Otherwise Black plays . . . Bdl . 42 . . .Bc2; 43.Be2 Be4? 43 . . . Bb3

would have been more efficient, but Black saw a clear winning plan
and went for it. The only problem is that it doesn't work!
44.Bd l Bg2; 45.Be2 Bh3.
3
2
a b c d e f g h
[277] White to move
46.Bd l ?? White misses his chance. 46.a4 ! works now, since Black
cannot easily attack the pawn. 46 . . . Bfl ; 47 .Bb3 a4!
176
BISHOP ENDGAMES
a b c d e f g h
[278] White to move
White resigned. 48.Bxa4 Bxc4 presents no problems, and 48.Ba2

Bh3; 49.Kf3 Kd4 is a trivial win.
ZONTAKH VS. IUNCIC

Belgrade, 1996
a b c d e f g h
[279] Black to move
White has an extra pawn in this bishops-of-the-same-color endgame.

It is an outside passed pa\vn and should be able to advance. Black has a
passed e-pawn, and a possible target at f6.
45 . . . e3; 46.Kg2. This is the only way to insure that the king will be
able to help stop the pawn, since if it goes to g4, Black can cut off the
path with ... Ke4.
46 ... Ke4; 4 7 .h4!
I ll
a b c d e f g h
[2801 White to move
White must play carefully, as there are dangers. 4 7 .g4? Kd3; 48.Be 1
Bg5; 49 .h4 Bxf6 (49 . . . Ke2; 50.hxg5 Kxe l ; 5 l .g6 ! ) 50.Kf3 b6; 5 l .a4 a5;
52.h5 Bg5 is nearly zugzwang. The bishop must stay on the e l -a5 di
agonal to keep the enemy king off of d2, and the king must remain at
£3 to guard against . . . Ke2. The next move is pretty much forced. 53 .b4
axb4; 54.Bxb4 e2 ! ; 55.Bel Bd2 ! ; 56.Kf2 f6 !
3
2
a b c d e f g h
A wonderful finesse which puts White in zugzwang. 57.h6 Bxh6;

58.Kf3 Bd2; 59.Kf2 Bxe l +; 60.Kxe l Ke3 and mate in 6. 6 l .a5 bxa5;
62.g5 a4; 63 .gxf6 a3; 64.f7 a2; 65.f8Q al Q#.
47 Kd3.
. . .
17 8
BISHO P ENDGAMES
a b c d e f g h
[282] White to move
The active Black king can't force a win, but it can salvage a draw.
48.Be l . 48.Bb4? e2; 49.Be l Bd2; 50.Kf2 Be3+; 5 l .Kf3 Bd4 will lead
to the win of the f-pawn and play similar to the note to move 47. 52.h5
(52 .g4 Bxf6; 53 .g5 Be7 and the bishop gets back to e3 via c5 . 54.Kf2
Bc5+; 55.Kf3 Be3 with . . . Bd2 to follow. ) 52 . . . Be3 ! ; 53.g4 Bg5; 54.a4
(54.Kg2 Kc2; 55.Kf3 Kdl ) 54 . . . Bd2; 55 .Kf2 aS; 56.h6 Bxh6 where;
57.Bxa5 loses to 57 . . . Bd2. 48 . . . Ke2; 49.Bc3 ! Bf8. Black must transfer
the bishop to a more useful diagonal, even though it allows White to
advance the pawns.
50.g4 Kdl ; 5 1 .g5 e2. White is willing to give up the bishop now, as
the kingside pawns can work by themselves. 52 .h5 b5; 53.h6.
a b c d e f g h
[283] Black to move
53 . . . b4. Black seems to be on the verge of winning, but White has a

surprising resource to save the game.
54.Be l ! ! 54.h7 bxc3; 55 .h8Q e l Q; 56.Qxf8 c2 and Black gets a sec
ond queen. 54 . . . Kxe l ; 55.h7 Kd2 . Black had to waste time moving the
179
king, and White uses the extra tempo to regain the bishop.
56.h8Q e lQ; 57.Qxf8 Qe2+. Agreed drawn.
SHERZER VS. WOJTKIEWICZ

World Open, Philadelphian 1 997
3
2
a b c d e f g h
[284] White to move
White has a classic case of the bad bishop blues . The kingside pawns
are fixed on dark squares. Even without the extra pawn at f7, Black
would win. The plan is to infiltrate with the king. The target square is
e4.
42.Kc2 Ba7; 43 .Be l Kd7; 44.Kd3 Kc6; 45.Kc4. White tries to keep
the king off of d5, but now gets a passive bishop. 45 ... Be3! ; 46.Bg3.
The only way to protect the f-pawn. 46 ... Bgl ; 47.Kb4. 47.Be l Bh2;
48 .Bd2 Bg3 wins the h-pawn.
4 7 ... Kd5; 48.Ka5 Ke4. Mission accomplished. The queenside is
sufficiently defended by the bishop at g 1 . 49.b4 Kf3; 50.Be 1 Kxf4;
5 1 .Kxa6 Kf3 ; 52.b5 f4; 53.Bd2 Ke4.
3
2
a b c d e f g h
[285] White to move
180
BISHO P ENDGAMES
White resigned. The f-pawn advances faster than the b-pawn.

54.b6 f3; 55.Be l f2; 56.b7. 56.Bxf2 Bxf2 ; 57.b7 Bg3 stops the pawn,
and the remaining kingside pawns fall to the Black king.
56 . . . fxe 1Q; 57.b8Q Qa l+; 58.Kb7 Qb2+; 59.Kc8 Qxb8+; 60.Kxb8
Bf2 . Black wins.

The following endgame shows how a single square, in this case d6,
can determine the fate of a game, or at least its length. Bishops can
usually shed tempi easily, and zugzwang is likely to present itself if the
kings are in close contact near a pawn.
YEPISHIN VS. LUGOVOI

St. Peterburg, Russia, 1 996
3
2
a b c d e f g h
[286] White to move
42.Kc6. The king comes forward to take control of d6. The pawn
can wait. 42.Be6+ Ke7; 43. Bxd5 Kd6 will take much longer at least, and
may not be winnable at all. 44.Bf3 Kc7; 45.Kc4 Kd6; 46. Kd3 Bc7; 47.Ke4
Kd7; 48.Kd5 Bd6 and the game can eventually be won by advancing
the pawn to d7, protecting it with the bishop, and then the king can
roam destructively. 49 .Be2 Bb4; 50.Bb5+ Ke7; 5 l .Kc6 Bc3 (5 l . . . Ba5;
52 .d5 Kd8; 53.d6 Kc8; 54.d7+ Kd8; 55.Kb7 ! The b-pawn will be re
moved, since 55 . . . Ke7 allows 56.Kc8.
42 . . . Kf8; 43 .Be6! Black loses the d-pawn and gets none of the
counterplay. 43 . . . £3. This pawn 'vasn't worth anything. 44.Bxd5 ! fxg2 ;
45.Bxg2 Ke8; 46.d5. Black is helpless. 46 . . . Ke7; 47.d6+ Ke8; 48.Bf3 !
18 1
3
2
a b c d e f g h
[287] Black to move
Black resigned. If the White king gets to d7 it is over, and Bh5+ will
in any case force the Black king to abandon that square.
BISHO P + 6 US . BISHO P + 4
Given multiple pawn islands containing a lot of weak pawns, the
possession of a passed pawn, especially an extra pawn, is of great value.
When the enemy bishop is offside and the king is distant, things should
be simple. They aren't. In the next diagram Black, already down a pawn,
is forced to immediately sacrifice the a-pawn. Down two outside passed
pawns, the win should be simple, but it still requires technique.
BELLON LOPEZ VS. SCHNEIDER

Rilton Cup, Stockholm, 1 998
3
2
a b c d e f g h
[288] Black to move
This is a fascinating position. Despite the bishops of opposite col

ors, Black is hard pressed to stop the White c-pawn from promoting.
182
BISHOP ENDGAMES
The pawn at a5 is under attack, but there is no time to capture it be

cause the bishop at f3 cannot get back to stop the c-pawn.
30 . . . Kf7. 30 . . . a4 loses to 3 1 .c6 3 l .Bxa5 Ke8; 32.c6 Bd l ; 33.c7 Kd7.
White advances the c-pawn to tie down the enemy kin g . Black needs to
keep the bishop active to be able to defend the two weak e-pawns. Since
all of White's pawns are on dark squares, Black has no targets, and
therefore no counterplay.
34.Ke l Ba4; 35.Kd2 Bc6; 36.Kc3 Bb7; 37.Kd4 Bc6; 38.Kc5 Bb7;
39.Kb6. The point of this king journey is simply to free the bishop and
provide support for the advance of the a-pawn. 39 . . . Kc8; 40.Bb4 Bd5;
4 l .g4. This insures that Black cannot try to defend with . . . h5, . . . g6 and
. . . Bf5 . 4 1 .a4 Kd7; 42.a5 Kc8; 43 .a6 Kd7; 44.a7 Kc8 doesn't accomplish
much. The key to the win lies on the kin gside. White patiently prepares
the way.
4 l . . . Kd7; 42 .h4 Kc8; 43 .a4 Kd7; 44.a5 Kc8; 45.a6 Kd7.
3
2
a b c d e f g h
[289] White to move
The queenside pawns cannot be promoted because Black has all

the light squares under control. However, should the Black kin g be
forced to c8, White may be able to get the king to d6 and from there
the kingside can be infiltrated.
46.Bf8 g6; 4 7 .g5 ! The pawn at h7 is permanently crippled. 4 7 ... Kc8;
48.Bb4 Kd7; 49.Ba5. The point of the bishop maneuver is to defend
the pawn at c7 while keeping d6 open for the king.
49 . . . Kc8; 50.Kc5 Kd7; 5 1 .Bb6 BaS.
183
3
2
a b c d e f g h
[290] White to move
Black has been reduced to shuttling the bishop between d5 and a8 .

The a-pawn \vas kept on the sixth rank to control b7. White now wins
by deflecting the king from d7.
52.c8Q+! Kxc8; 53.Kd6 Bd5; 54.Ke7. A highly instructive endgame,
showing the importance of playing both flanks in a bishops of oppo
site color endgame.

King centralization is the theme of the next game. Many players
will automatically centralize the king as soon as possible. While I will
forgive a student for rushing the king to the center and overlooking a
more important maneuver, failing to improve king position is gener
ally unforgivable. It is rare that this error is made by a Grandmaster,
but it does happen.
GLEIZEROV VS. BARUA

1 Oth Goodricke International, Calcutta, 1999
7.
a b c d e f g h
[291] Black to move
184
BISHOP ENDGAMES
White's extra pawn is not enough to win, given the bishops of op
posite colors. But Black has four pawn islands, and White has two.
That is a significant structural advantage. In order to hold the game,
Black must activate the king. For some reason, Barua waits much too
long, and his king pays with his life.
29 . .. Bc3. 29 . . . Kg 7 is correct. The pawn at c4 cannot be attacked
and should not advance until it has to. 30.Kg2 K£8! Sometimes retreat
is the best way forward! (30 . . . Kg6 runs into 3 l .Bc2+) 3 l .Kf3 Ke7; 32.Ke4
Kd6.
a b c d e f g h
{292] White to move
The position seems promising for White, who can expand his ad
vantage on the kingside. 33.f4 gxf4; 34.Kxf4 establishes connected
passed pawns, but all is not lost yet. 34 . . . Kc5 and now White is actually
in g reater danger! 35 .h4 (35 .a4 Kd4! ; 36.h4 Be5+; 37.Kf3 Kc3 ! threat
ens to trap the bishop at d l ! ) 35 . . . Kxb5; 36.h5 Kb4; 37. g4 Bg7; 38. g5
Kc3; 39.h6 Bf8.
a b c d e f g h
{293] White to 1nove
185
This is the best defensive formation against two passed pawns. When
the h-pawn advances, the bishop returns to g7. Unfortunately for Black,
White's king can get to f7 in three moves. 40.Ke5 Kd2; 4 1 .Kf6! c3;
42.h7 Kxd l ; 43.h8Q c2; 44.Kf7! Bc5; 45 .g6 c l Q; 46.Qhl+ Kd2; 47.Qxcl+
Kxc l ; 48.g7 and wins. Nevertheless, this gives White many ways to go
wrong, and takes a lot longer.
30.e3.
a b c d e f g h
[294] Black to move
White has the luxury of time, and opens up a route for the bishop
or king, while also preparing to recapture with the e-pawn after an
eventual f4.
30 . . . Bd2; 3 1 .Kfl e5. 3 1 . . . Kg7; 32.Ke2 Bb4; 33.f4 Kf6; 34.fxg5+ Kxg5;
35.h4+ Kf5; 36.Kf3 is hopeless for Black. 32.Ke2 c3; 33.Kf3 Kg7; 34.Kg4
Kf6; 35.Bc2 Be l ; 36.f4 gxf4; 37.exf4 Bd2. Or 37 . . . exf4; 38.Kxf4 and
the two connected passed pawn are only half of the problem. White
can also walk to b7, capture the a-pawn at a7, then, thanks to the block
by the pawn at c3, advance the a-pawn to a5 , meeting . . . ba5 with b6 and
a new queen or the win of the Black bishop. The White bishop takes
care of both Black pawns.
38.f5 Be l ; 39.Kh5 Bf2 ; 40.g4. Black resigned.
BISHOP + l US. BISHOP + 6

Stonewall formations often lead to endgames with between six and
eight pawns for each side. One side will often have a good bishop, and
the other may have a bad bishop, but unless the weak points in the
enemy structure can be targeted, it is very hard to make progress.
186
BISHOP ENDGAMES
KORCHNOI VS. SHIROV

FIDE Championship Knockout, Groningen, 1996
a b c d e f g h
[2951 White to move
One extra pawn and bishops of opposite color. Is there any reason
to believe that this game shouldn't end in a draw? Black's pawns are all
on light squares. This means that the enemy bishop cannot attack them.
On the other hand, it also leaves the pawns without defense against an
enemy king. The bishop will have to defend the dark squares around
the weak pawns. Still, with the luxury of an extra pawn and no entry
squares for the White king, there is little to worry about. White has a
bad bishop, but the enemy bishop has no meaningful targets. Even if it
gets to h4, the White king can contain it at e2 .
36.Bd8. 36.f5 ! ! would have made things much more difficult.
a b c d e f g h
36 . . . exf5 (36 . . . BgS ; 37.fxe6 fxe6; 38.Bc7 Kf7; 39.a3 ) 37.d5 Kf8
(37 . . . cxdS; 38.c6 and the pawn promotes. ) 38.clxc6 Ke8; 39.Bc7 BgS ;
40.Ke2 Bd8; 4 l .BeS BaS; 42.Kf2 Bd2; 43 .Ke2 Bb4; 44.Bd6 Kd8; 4S.Kf2
h4; 46.Ke2 f6; 47.Kdl aS; 48.Ke2 a4; 49 .Kdl BaS .
187
C ARDOZ A PUBLISHING • ERIC SCHILLER
36 . . . Kg7; 37.Ke2. 37.f5! exf5; 38.d5 K.£8 ; 39.dxc6 Ke8; 40.Bc7 trans
poses to the previous note. 37 . . . Kg6.
a b c d e f g h
[2971 White to move
The advance of the f-pawn is eliminated once and for all by the
Black king. 38.Kd2 Kf5; 39.Bc7.
39.Kc3 loses to the sacrifice 39 . . . Bxf4; 40. exf4 Kxf4. See position
#546.
39 ... Bg7; 40.Kc3 Bf6; 4 l .Kb4? A critical error. White runs out of
patience and tries to invade the queenside. It only takes three moves
for Black to get to the e-pa,vn, however. 4 l .Kd2 Bh4; 42.Ba5 .
4 l . . . Bh4; 42.Ka5 Bf2 . White resigned. 4 2 . . . Bf2. 43.Kxa6 Bxe3;
44.Ba5! 44.Kb7 loses to the simple 44 . . . Bd2 . 44 ... Bxd4; 45.Kb6 Kxf4.
a b c d e f g h
[2981 White to move
The rest of the story is simple. Black will eventually sacrifice the
bishop for the c-pawn, and the central pawn mass will cra,vl up the
board. For example:
46.Kxc6 e3; 4 7 .Kxb5 Ke4; 48.c6 Be5; 49.c7 Bxc7! ; 50.Bxc7 f5; 5 1 .a4
f4; 52.a5 £3 ; 53.a6 f2 ; 54.a7 flQ+.
188
BIS HOP US. KNIGHT
ENDGAMES
Perhaps the best general advice that can be offered when playing
with bishops and knights is to forget about the general wisdom that
bishops are superior. I'm not suggesting merely noting all the excep
tions, but actually abandoning it entirely. Compare the power and the
mobility of each piece in concrete situations . Is it a good bishop or bad
bishop? Is the bishop an effective defender of the pawns? Can the bishop
control the promotions square? For knights, ask whether the knight
can reach pawns which cannot be defended by enemy forces . Check
whether the knight can be dominated by the bishop and be excluded
from the action.
If there are many pawns, the bishops value can be reduced because
there are barriers in the way of some diagonals. If there are very few
pawns, the· bishop is likely to be the better piece. A bishop may not be
able to protect a pawn on the opposite color, but it can take away at
tacking squares from the knight. If the bishop is a dark square bishop
and there is a pawn on a dark square, it can be defended directly. If the
bishop runs on the light squares, indirect defense is possible if the
same pa,vn can only be attacked by a knight which sits on a light square.
Cover the relevant light squares and you protect your pawn.
The knight is better at fighting up close, like a boxer who prefers to
keep attacking with short blows at a short distance. The bishop is a
more strategic fighter who prefers dancing and striking fron1 afar. Evalu
ate each according to the position, taking careful note of the pawn
structure.
189
CARDOZA PUBLIS_HING • ERIC SCHILLER
BISHO P US. KNIGHT + I

When attacking with a single pawn against bishop, it is often better
to have the knight than a bishop of the same color, because a knight
can be used against squares of either color. Of course it is important
that the knight be in a good defensive position. If it is too far away it is
useless.
To win, your main task is to keep the enemy king from getting into
a good defensive position, either at the promotion square or directly in
front of the pawn.
ANALYSIS OF LEKO VS. AKOPIAN

Groningen, 1 996
a b c d e f g h
[299] Black to move
80 . . . Kh4; 8 l .Ne6. Black's king is stranded on the rim, and takes

too much time to get to the queenside. White's plan is very simple. Get
the king to b7, and if Black defends on the gl -a7 diagonal, plant the
knight at b6.
81 . . . Kh5; 82.Ke4 Kg6; 83 .Kd5 Kf6; 84.a6 Ba7. 84 . . . Ke7; 85 .Nd4
threatens Nc6+. 85� . . Kd7; 86.Nb5. The bishop cannot move, and the
king can't get across the c-file. 86 . . . Kc8; 87.Kc6 Kd8; 88.Kb7 etc. 85.Nc7
Ke7; 86.Kc6 Kd8; 87.Kb7 Bgl ; 88.Nd5 Kd7; 89.Nb6+ and there is no
way to stop the pawn.
BISHO P US. KNIGHT + 2

Use one pawn to occupy the enemy king, get the other into a posi
tion where the enemy bishop cannot stop it, and you have a full point.
Easier said than done! Start by getting your king in front of one of your
pawns, as in our example, and you will find the solution is not all that
difficult.
190
BISHO P US. KNIGHT ENDGAM ES
ANALYSIS OF KASPAROV VS. TIMMAN

Linares, 1 993
3
2
a b c d e f g h
[300] Black to nzove
48 . . . Be8. Or 48 . . . Ke4; 49 .Kf6 K£4; 50. g5! Bxf5 5 1 .b5 49.Ne3 Bd7;
50.Nc2 Ke4; 5 l .Na3 Kf3 ; 52.Nc4 Ke4. No better is 52 . . . Bxg4; 53.Ne5+
Ke4; 54.Nxg4 Kd4; 55 .Ne3 ! ; 53 .Nd6+ Ke5; 54.Nf5 Ba4; 55.Ne3 Ke4;
56.Nc4 Kd4; 57.Na3 Kc3 . Time to advance, while the Black king must
catch up.
58.b5 Kb4; 59.b6 Bc6; 60.Kf6 Kc5; 6 l .g5 Kxb6. One pawn down,
another to go. My old buddy, you are moving much too slow.
62.g6 Kc5; 63.Nc2 Be4; 64.g7 Bh7; 65.Ne3 Kd6; 66.Ng4 Kd7; 67.Kf7
Kd6; 68.Nf6 Bc2; 69.Kf8 Bb3 ; 70.Ng4 and White wins.
BISHO P + I US. KNIGHT + 3

Our next example is included to show how difficult some endgame
positions can be. Two stron g Grandmasters had to fi g ure out some
very complicated minor piece vs. pawns positions . This fascinating
endg ame shows both a bishop and a kni ght battling ag ainst mere
pawns-and losing!
19 1
C ARDOZA PUBLIS_HING • ERIC SCHILLER
CONCLUSION OF IBRAGIMOV VS. RUBLEVSKY

3
2
a b c d e f g h
[301] White to move
52.Nxa3 . This brings about the Bishop + 1 vs. Knight + 3 endgame,

White figured that the endgame must be drawn, because Black has a
classic "wrong rook pawn." Usually, if the queening square is the oppo
site color from the bishop, it is impossible to force a win, because the
opposing king will sit on the promotion square, where it usually gets
stalemated. In this game, however, the White king is too far away. 52.e5
Kxh4; 53.f4 Kg4; 54.Ne3+ Kxf4; 55.Nd5+ Kxe5; 56.Nxe7 Ke4; 5 7.Ng8
57.Kb3 a2; 58.Kxa2 h5; 59.Ng8 h4.
3
2
a b c d e f g h
[302] White to move
Black wins! The knight cannot get back to stop the pawn, for ex
ample. 60.Nf6+ Kf3; 6 1 .Nh7 Kf4; 62.Nf6 h3; 63.Nd5+ Kf3 and the h
pawn marches to the promotion square. ) 57 . . . a2; 58.Kb2 Ke3; 59.Nf6
h6; 60.Ng4+ with a draw as both Black pawns fall.
52 . . . Bxa3; 53.Kd3. The king hastens to the kingside. If it can get to
h l , the game should be drawn.
192
BISHO P US. KNIGHT ENDGAMES
53 . . . Kxh4; 54.Ke2 Kh3 ; 55.Kfl .
3
2
a b c d e f g h
[303] Black to move
Almost there, but Black takes away the gl -square.

55 . . . Kh2 ; 56.e5 h5; 57.£4. 57 .e6 h4 58.f4 transposes to the game.
57 . . . h4; 58.£5 Bb2 ! Black threatens to bring the bishop to d4, move the
king to g3, and advance the h-pawn.
59.e6 B£6. 59 . . . Bd4; 60.e7 Kg3 is too slow since White queens first
with 6 l .e8Q and wins. 60.Kf2 h3 ; 6 1 .Kfl Bh4!
3
2
a b c d e f g h
[304] White to move
White resigned, as he must concede control of g2. The pawns can

no longer move without being captured, and if they fall Black returns
to the old plan of . . . Bd4 and . . . Kg3. 6 l . . . Bh4; 62 .Ke2 (62.f6 Bxf6; 63.Kf2
Bh4+; 64.Kfl Kh l ; 65.e7 Bxe7.
193
CARDOZA PUBLIS ft iNG • ERIC SCHILLER
3
2
a b c d e f g h
[305] White to move
66.Kf2 Bc5+; 67.Kfl Kh2; 68.Ke2 Kg2 etc.) 62 . . . Kg2 and the pro
motion of the pawn cannot be delayed for long.
BISHO P + 2 US. KNIGHT + I

Since the knight is not a mobile piece, the most desirable endgames
with a bishop and extra pawn(s) are those where the pawns are far
apart, as in the following example.
CONTINUATION OF SOKOLOV VS. TIMMAN

Bosna '99, Sar�evo, 1 999
3
2
a b c d e f g h
[306] Black to move
4 l ... Nf4; 42 .Be4 Ne2. If the pawn advances, the win is clear. 42 . . . h3
43.gxh3 Nxh3 (43 . . . Ne2 ; 44.a4 Nc3; 45 .a5 Nxe4; 46.Kc6 ! ! )
194
3
2
a b c d e f g h
Black cannot stop the a-pawn!

43.Kc5? 43.a4 is more efficient. The same trick as seen in the pre
vious line is available here. 43 . . . Nc3 ; 44.a5 ! Nxe4; 45 .Kc6! ! 43 . . . Ke6;
44.Bc6.
3
2
a b c d e f g h
[308] Black to move
44 ... Nc3? A critical error. 44 . . . Nf4! would have saved the game.
195
3
2
a b c d e f g h
[309] White to move
One possible continuation is 45 .Kd4 Kd6; 46.Ke3 h3; 4 7.gxh3 Nxh3;

48.Bg2 Ng5; 49.Kd4 Kc7; 50.Kc5 Ne6+; 5 l .Kb5 Nd4+; 52 .Ka6 Ne6;
53.Bd5 Nc5+; 54.Kb5 Nd7; 55.Bc6 Ne5; 56.Bg2 Kb8 draws since the
king gets to a7, after which the knight can just run around. White can
not prevent this. 57.Ka6 (57.Kb6?? Nc4+) 57 . . . Nc4; 58.a4 Nb2; 59.a5
Nc4 and Nxa5.
45 .a4? fails to 45 . . . Nd3+; 46 .Kb6 Nb2; 47.a5 Nc4+; 48.Kb5 Nxa5 !
when 49.Kxa5 h3; 50.g3 h2; 5 l .Bf3 Kf5; 52.Kb4 h l Q! draws, because
after 53.Bxh l Kg4 the last pawn falls.
45.Kd4! The knight is chased back toward White's home rank.
3
2
a b c d e f g h
[310] Black to move
45 . . . Nd l . 45 . . . Na2?? drops the knight to 46.Bd5+. 46.Bf3 ! Nb2.

Again the knight has only one move. 47.Kc3 ! Na4+. No choice here,
either. 48.Kb4.
196
3
2
a b c d e f g h
[311] Black to move
Black resigned. 48 . . . Nb6; 49.Kb5 Nd7.

There are three unpleasant alternatives:
49 . . . Nd5; 50.Bxd5+ Kxd5 5 l .a4.
49 . . . Nc8; 50.Bg4+.
49 . . . Na8; 50.Bxa8.
50.Bg4+. The king and pawn endgame is a simple matter.
BIS HO P + 2 US. KNIG HT + 2

With even pawns, the two big questions are whether a pawn can
promote and whether it is possible to capture a pawn. You must con
cern yourself with the distance between the pawns and their promo
tion squares, and with the health of each pawn. Other considerations
include king centralization, piece placement and presence of rim pawns,
but of course, those are always relevant.
SOKOLOV VS. TIMMAN

Bosna '99, Sar�evo, 1 999
3
2
a b c d e f g h
[312] White to move
197
CARDOZA PUBLIS_HING • ERIC SCHILLER
Both sides have two pawns with no passed pawns. White's bishop is
of the wrong color to promote an h-pawn, if Black lures the pawn to
the h-file by sacrificing a pawn at h3. Nevertheless, White wins, be
cause the b-pawn must fall and Black needs too much time to carry out
the plan.
38.Bd3 . The bishop takes up a position dominating the knight.
38 . . .Kf7. 38 . . . h4; 39.Kd5 Kf7 transposes. 39.Kd5 h4. Necessary, to make
room for the knight at h5, so that it can get back into the game. 40.Kc6
Nh5; 4 1 .Kxb6. We have a complicated Bishop + 2 vs. Knight + 1 endgame
which proves difficult even for these two superstars. See position #306
for the rest.
BISHO P + 2 US. KNIGHT + 3

As the number of pawns increase, we become even more concerned
with passed pawns and the pace of their march up the board. The
bishop is an effective long range defender, so pawns will always require
an escort. A knight must stay close to a pa,vn to assist it either in of
fense or defense.
LEKO VS. AKOPIAN

Groningen, 1996
3
2
a b c d e f g h
[3 13] White to 1nove
White has an extra passed pawn and the knight, though presently
in exile, can get into the game quickly. Black has only one asset, the e
pawn, but it can't get very far. Leko is an excellent technician and brings
home the point cleanly.
69.Nc2 ! e4+; 70.Ke2 Bb6; 7 1 .Ne3+ Ke6; 72.g4!
198
3
2
a b c d e f g h
[314] Black to move
White takes control of f5 and h5, and leaves Black without any pawn
moves. The position is already lost.
72 . . . Bc7. 72 . . . Bc5; 73 .a4 Ke5 is a different plan, going after the h
pawn with the king. But it is a long, slow journey. The winnin g plan
takes a while and involves some precise play.
3
2
a b c d e f g h
[315] White to move
74.a5 Kf4; 75 .Nd5+f Kg3; 76.Nc3 Kxh3; 77.Nxe4. The attack on the
bishop buys precious time. 77 . . . Ba7; 78.Kf3 Bb8 (78 . . . Kh4; 79.Kf4 Bb8+;
80.Kf5 Bc7; 8 l .a6 Bb6; 82 .Nd6. White will play Nc8 and c7. There is
nothing Black can do about the kin gs ide because on . . . h5, White sim
ply pushes the g-pawn until it becomes a queen. ) 79.g5 ! It takes guts to
give up your second-to-last pawn. Good technique is needed to win the
kni ght and pawn vs. bishop endgame. 79 . . . hxg5; 80 .Nxg5+. The poor
position of the Black kin g provides all White needs to win. The method
can be seen in diagram #299.
73.Nf5 Bf4. 73 . . . h5; 74.Ng7+! picks off the h-pawn for free. 74.a4
199
Kd5. The Black king can cope with the a-pawn, but then the kingside
cannot be defended. 75.a5 Kc5; 76.h4! Kb5 .
a b c d e f g h
[316] White to move
With the king so far a\vay, White can afford to sacrifice the knight
and let the connected passed pawns do the rest. 77.Nxh6! Bxh6; 78.g5.
3
2
a b c d e f g h
The bishop cannot cope with the advancing pa,vns.

78 . . . Bf8. 78 . . . Bg7; 79.Ke3 Kxa5 ; 80.Kxe4 Kb5; 8 l .h5 Kc6; 82.h6
B£8; 83.Kf5 Kd7; 84.Kg6 Ba3; 85 .h7 Bb2; 86.Kf7 and the g-pawn de
cides matters .
79.Ke3 Kxa5; 80.Kxe4 Ba3; 8 l .Kf5 Kb5; 82.g6 Bb2 ; 83.h5 Bg7;
84.Kg5 . Black resigned.
200
BISHOP US. K N I G HT EN DGA M ES
IBRAGIMOV VS. RUBLEVSKY

a b c d e f g h
There are two plans. White can eliminate the a-pawn by sacrificing
the knight, as seen in the game, or give up some of the kingside pawns.
In either case a new type of endgame will arise, but the decision must
be made now.
Usually, if the queening square is the opposite color from the bishop,
it is impossible to force a win, because the opposing king will sit on the
promotion square, where it usually gets stalemated. In this game, how
ever, the White king is too far away. 52 .e5 Kxh4; 53.f4 Kg4; 54.Ne3+
Kxf4; 55.Nd5+ Kxe5; 56.Nxe7 Ke4 . The resulting Knight vs. 2 ending
must be handled carefully, and is analyzed in diagram # 1 50 .
Instead, White selected 52.Nxa3. White figured that the endgame
must be dra,vn, because Black has a classic "wrong rook pawn."
52 . . . Bxa3. This brings about the Bishop + 1 vs. 3 endgame, posi
tion #30 1 .
BISHOP + 3 US. K N IG HT + 4
In endgames with 7 or more pawns, passed pawns often need to be
created, and exchanges of pawns will lead to other endings. When the
pawns are split between two flanks, king centralization is important.
The side with the bishop has an advantage as the bishop can attack
enemy pawns and tie do,vn the enemy king to defense. If you have the
knight, placing your pawns on the opposite color of the enemy bishop
will help keep them safe.
201
(72) KASPAROV VS. TIMMAN

Linares, 1993
3
2
a b c d e f g h
37.Kf2 Kf8;38.g3 . White places the pawns on dark squares, where

they will be invulnerable to the Black bishop. 38 . . . Ke7; 39.Ke3 Bc6;
40.h4 Bd7; 4 l .Kf4 Kd6; 42.g4. 42.Nd4! is correct. For some reason
Kasparov didn't get around to playing the move until move 44. The
preliminary measures do not help his game. 42 . . . Kd5; 43 .Nf3 White
wants to encourage Black to play . . . f6, greatly weakening the k.ingside.
42 . . . f6. 42 . . . Kd5; 43.g5 g6 was worth a try.
43.h5. Missing the point again. 43.Nd4 still works. Kasparov ana
lyzed the line thoroughly: 43 . . . g5+; 44.hxg5 fxg5+; 45 .Kxg5 Ke5; 46.Nf5
a5; 47.b4 axb4; 48.axb4 reaching a Bishop + 1 vs. Knight + 2 endgame
examined in diagram #300.
43 ... Ke7; 44.Nd4 K£7; 45.b4.
a b c d e f g h
[320] Black to move
202
BISHOP US. KN I G HT EN DG A M ES
45 ... Ba4. 45 . . . g6 would be a mistake because after 46.hxg6+ Kxg6;

4 7.b5 the king cannot get back to the queenside in time, and the White
g-pawn is safely guarded. 47 . . . Kf7; 48.a4 Ke8 (48 . . . Ke7; 49.Nc6+ Bxc6;
50.bxc6 Kd6; 5 l .Kf5) 49.Nc6! a6; 50.Nb8 axb5; 5 l .Nxd7 bxa4 (5 l . . . Kxd7;
52.axb5) 52.Nxf6+ K£7; 53 .Ne4 a3; 54.Nc3 etc.
46.Nf5 g6. The variation 46 ... a6; 47.Nd6+ Ke6; 48.Ne4 Bb5; 49.a4
Bfl ; 50.Nc3 Kd6; 5 l .b5 axb5; 52.axb5 Kc5; 53.Ne4+ Kxb5; 54.Ng3 was
presented by Kasparov. He understands the position is a 'vin, but you
need to see that the pawn at g7 is lost after Nf5 . 54 . . . Bd3; 55.Nf5 wins
because 55 . . . Bxf5 ; 5 6 . Kxf5 is hopeless . 47.Nd6+ Kg7 ; 48.Nc8 a6;
49.Nd6.
a b c d e f g h
[321] Black to move
Black has to choose a plan here. There is passive defense with . . . Bc6
or . . . Bd7, simplification with . . . gxh5 , and active defense with . . . Kh6.
49 ... Bd l . Kasparov considered this a mistake, and it really barely
qualified as a candidate move. Black cannot afford to clarify the
queenside when the White pawn at g4 is guarded. One alternative is
49 . . . Kh6; 50.Ne4 gxh5 (50 . . . g5+; 51 .Kg3 Kg7; 52.Nd6 Bc6; 53.Nf5+ Kh7;
54.Ne7 Bb5; 55.Nd5 Kg7; 56.Nc7 Be2; 57.a4 f5; 58.gxf5 Kh6; 59 .b5
axb5 ; 60.axb5 Bxb5; 6 l .Nxb5 Kxh5; 62 .Nd6. The f-pawn advances to
f7 and the king takes care of the rest.) 51 .gxh5 Kxh5; 52.Nxf6+ Kg6;
53 .Ke5 Kf7; 54.Nd5 Bb5; 55 .Nc7 Be2; 56.Kd6 gets to the a-pawn.
White has no difficulty dealing with 49 . . . Bd7. 50 .Ne4 f5; 51 .Nc5
Bc8; 52 .h6+ Kxh6; 53.g5+ Kg7; 54.a4 and the b-pawn advances.
49 . . . Bc6 ! ? sets the most problems. 50.Ne4 g5+ (50 . . . f5 is more ac
tive but 5 l .h6+! Kxh6; 52 .gxf5! gxf5; 53.Kxf5 is a win because the White
king and knight combine to with the a-pawn, resulting in a winning
Knight + 2 vs. Bishop endgame. ) 5 l .Kf3 .
203
3
2
a b c d e f g h
[322] Black to move
Black has three defensive tries, but they all come up short.
5 l . . . Kf7; 5 2 . Ke3 Bd7; 53 .Nd6+! Kg7; 54.Kf3 Bc6+; 55.Kg3 Bd7;
56.Ne4 Bb5 (56 . . . £5; 57.gxf5 Bxf5; 58.Nxg5).
5l ... Bd5; 52.a4 Bc6; 53.a5 Bd5; 54.Ke3 Bc6; 55 .Nc5 Bb5; 56.Ne6+
Kh6; 57.Nc7 Bfl ; 58.b5 axb5; 59 .a6 b4; 60.a7 Bg2; 6 l .Kd4! intending
Nd5, cutting off the bishop, so 6 l . . . Bb7; 62.Ne8! b3; 63.Kc3 f5; 64.Nd6
and White wins.
5 l . . . K.h6; 5 2 . Ke3 f5; 5 3 . gxf5 Kxh5; 54.Nd6 K.h6; 55 .Nf7+ K.h5;
56.Ne5 Ba4; 57.f6 etc.
50.Ne8+! K£7; 5 1 .Nxf6! !
3
2
a b c d e f g h
[323] Black to move
The piece sacrifice wins because White will have a protected passed
pawn.
5 1 ... Kxf6; 52.g5+ Kf7; 53.h6. Black resigned. White simply marches
the king to a5 and gets rolling on the queenside . The Black king can
only sit hopelessly and watch the concluding act.
204
BISHOP US. KN I G HT EN DGA M ES
ANALYSIS OF KRAMNIK VS . IVANCHUK

Las Palmas, 1996
3
2
a b c d e f g h
[324] White to move
The h-pawn only has to cross one more light square. It is essential
for White to preserve the knight but how?
38.c4! 3 8.Nfl+ K£2; 39 .Nh2 (39 .Nd2 Ke2 ; 40.Nb l Be4+) 39 . . . Kg3 ;
40.Nfl + Kf2 draws. 38.Nb l loses the knight to 38 . . . Be4+.
38 . . . Bg2. 38 . . . Kxd2; 39.cxd5 cxd5; 40.h6 is out of the question.
39.Nb3. The king is kept out. 39 . . . Bfl. 39 . . . Kd3; 40.Kg7 Kxc4; 4 l .h6
Be4; 42.Nd2+ and the pawn promotes. 40.h6 Bxc4; 4 1 .Nc5. With d3
covered, h7 becomes a threat. 4l . . . Ba2; 42.Kg7 Bb l .
3
2
a b c d e f g h
[325] White to move
The h-pawn is contained, but the queenside pawns fall.

43.Nxa6. 43 .Ne6! ? might have been quicker, intending to plant the
knight at g6 and then promote the pawn. For example 43 . . . Kd2 ; 44.Nf4
Kc l ; 45.Ng6 Bxg6; 46.Kxg6 Kxb2; 47.h7 Kxa3 ; 48.h8Q with an easy
win. 43 . . . Kd4; 44.Nxc7 Kc4. White removes Black's remaining pawn
205
CA RDOZA PU BLISH I NG • ERIC SCH I LLER
and eventually gets the knight to g6. 45 .Ne6 Kb3; 46.Nd4+. 46.Nf8
Kxb2 47.Ng6 'vas more efficient. 46 ... Kxb2; 47.Nxc6 Kxa3; 48.Ne5
Kb4; 49.Ng6.
3
2
a b c d e f g h
[326] Black to move
The barrier is in place and the pawn reaches the goal in a few moves.
49 ... Bxg6; 50.Kxg6. Black resigned.
We turn to an analysis fragment from the same game. The differ
ence here is that the pa,vn is on the g-file.
ANALYSIS OF KRAMNIK VS . IVANCHUK

Las Palmas, 1996
a b c d e f g h
White's knight is dominated, and must inevitably fall to an invad

ing Black king. The question is, can White use the g-pawn to advantage
while the Black king cra,vls to,vard c l ? 44.Kf5 Kd4. White must now
get the g-pawn moving as quickly as possible . There are two ways to go,
advancing the pawn or moving the king.
206
BISHOP US. KNIGHT E N DG A M ES
Getting the king out of the way leads only to a draw. There arc t\VO
logical continuations.
45.Kf4 Kc3 ; 46.Ke3 and now 46 . . . Kb2 ; 47 .Nd3+ Kxa3; 48.g4 Kb3 ;
49.Kd4 Be6; 50.g5 B f5 ; 5 l .Ne5 Kxb4. Then 52.g6 is handled by 5 2 . .. Be6;
53.Nxc6+ Kb5; 54.Na7+ Kb4; 55.g7 a5. White can play 52.Nxc6+ Kb5;
53.Ne7 Be6; 54.g6 where 54 . . . a5; 55 .g7 a4; 56.g8Q Bxg8; 57.Nxg8 a3
dra\vs (57 . . . Kb4; 58.Ne7 a3; 59.Nd5+ Kb3 ; 60.Nc3 �,vins, however), be
cause the knight cannot get back to stop the pawn. The king must
retreat, but then the c-pawn falls . 58 .Kc3 Kxc5.
C lb) 46 . . . Kc2 .
The more complicated line is 45.g4 \Vith the sensible reply 45 . . . Ke3 .
3
2
a b c d e f g h
[328] White to move
46.g5 will allow 46 . . . Kd2 , but White has the strong move 46.Ke5 !
Play might continue 46 ... Bd5? 46 . . . Kd2; 47.Kd4 Be6; 48.Nd3 ! Bxg4;
49.Nd5! wins the c-pawn, \vhich cannot be defended by the bishop.
After 49 . . . Be2 ; 50.Nxc6 Kc2 ; 51 .Ke5! Black is hard-pressed to stop the
invasion. A particularly instructive line is 5 l . . . Bb5; 52.Nd4+ Kb2; 53.Ne6
c6; 54.a4 Bxa4; 55.Nc7 Kc3 ; 56.Nxa6 Bb5; 57.Nb8! Kxb4; 58.Kd6 ! White
will win the c-pawn and Black will have a difficult defense in which the
slightest error can be fatal.
47.a4 Bg8. 47 . . . Kd2; 48.b5! axb5; 49.a5 ! The passed pawn wins .
48.Kf6 Kd2 ; 49.Ke7 Kxc l ; 50.Kd8 Kb2 ; 5 l .Kxc7 Kb3; 52.Kxc6 Kxb4;
53.Kd6 Kxa4; 54.c6!
207
3
2
a b c d e f g h
[329] Black to move
The bishop cannot stop the pawns.
BISHOP + 4 US. K N IG HT + 4
When you have an equal number of pawns and are on the defen
sive, consider jettisoning one of the pawns to reach a drawable pawn
down endgame. Sometimes it is easier to give up a weak pawn than to
expend energy on its defense.
FfACNIK VS. GUREVICH

Jakarta, 1996
a b c d e f g h
[330] Black to move
Black is struggling and has several major problems. The queenside

pawns are vulnerable to an attack by the White knight. The e-pawn
and h-pawns are under immediate attack and both are weak. Gurevich
makes the most of his e-pawn.
40 ... e3! 40 . . . Ke5; 4 l .Nxh7 e3 is too slow: 42.Ng5 e2; 43.Nf3+ Ke4;
208
BISHOP US. KN I G HT EN DGA M ES
44.h5 and the passed pawns advance without hindrance . 4 1 .Kf3 .

4 l .Nxh7? allows 4 l . . . Bxg4 and after 42.Nf6 e 2 ; 43.Kf2 B £3 ; 44.h5 loses
to 44 . . . Ke7; 45 .Ng8+ K£7; 46.Nh6+ Ke6 when the h-pawn falls and the
knight is still in some difficulty.
4 l .Nf3 Kd5 ; 42.Kf4 e2; 43.g5 (43.Ke3 Bxg4; 44.Kxe2 Ke4 wins the
knight. ) 43 . . . Be8; 44.Ke3 Bh5 and the pawn at e2 is taboo because of
Ke4. 4 1 . . . e2 . The pawn must keep on moving. 4 l . . . h6! ? was dismissed
by Ftacnik on the basis of 42 .Nf7+ Kd5 ; 43 .Kxe3 Bxg4; 44.Nxh6 but
there is more to the story. Black still has drawing chances, if he can win
the pawn at a2 . 44 . . . Bdl with two branches:
45.Nf5 is an attempt to keep the bishop from controlling h7, but it
does not work. 45 . . . Bc2; 46.Kf4! Bb l ; 47.h5 (47.Ne3+ Ke6; 48.Nc4 Bxa2;
49.Nxa5 Kf6 will be drawn. The knight cannot leave a5 without aban
doning the b-pawn, and after 50 .Nc6 Bxb3; 5 l .Nxb4 the bishop can be
sacrificed for the remaining pawn.) 47 . . . Ke6; 48.Nd4+ Kd5; 49.Nf3 Bxa2;
50.Nd2! The knight dominates the bishop, and the h-pawn promotes.
50 . . . Ke6; 5 1 .Kg5 Kf7; 52.h6 Kg8; 5 3 . Kg6.
45 .Kd2, followed by 45 . . . Bh5; 46.Nf5, is better. The idea is to use
the king to defend the queenside, get the knight to c4. If Black uses the
king to defend the pawns, then the knight can return to the kingside
and help to advance the h-pawn. 46 . . . Ke5 ; 47.Ne3 Kd4; 48.Nc4 a4 ! ;
49.Ne3 axb3; 50.axb3 Bf7; 5 l .Nc4 i s not likely to get anywhere.
42.Kxe2 Bxg4+.
a b c d e f g h
[331 ] White to move
The game enters another stage (B+3 vs. N+3). 43.Ke3 Bf5. 43 . . . h6 is
another path to the draw. 44.Nf7+ Ke7; 45 .Nxh6 Be6; 46.Kf4 a4; 47.bxa4
Bxa2; 48.Nf5+ Kf6; 49.Nd4 b3; 50.Nf3 b2; 5 l .Nd2 b lQ; 52 .Nxb l Bxb l .
44.Kd4 h5. Black can draw in any case, but advancing the pawn to h5
simplifies matters. The pawn can be guarded by the bishop and if White
209
CA RDOZA PU B LIS_ H I NG • ERIC SCH ILLER
tries to use the combined force of king and knight to get it, the queenside
will be accessible to the Black king.
45.Ne4+ Kc6.
a b c d e f g h
[332] White to nzove
45 . . . Bxe4; 46.Kxe4 Ke6; 47.Kf4 Kf6 is a drawn endgame.

46.Nd2 Bh7; 47.Nc4 Kb5.
a b c d e f g h
[333] White to move
Black has solved all the problen1s, with all pawns securely defended.
After a few repetitions, the players agree to a draw.
48.Nd6+ Kb6; 49.Nc4+ Kb5; 50.Nd6+ Kc6; 5 1 .Nc4. Drawn.
BISHOP + 4 US. KN I GHT + 5

A single static factor may be of great significance when a bishop
faces a knight. In the next game, White correctly understands the need
to switch pawns from dark squares to light squares, but chooses the
wrong pawn to advance.
210
B ISHOP US. KN I G HT E NDG A M ES
SHAKED VS. KOBALIJA

Hoogovens (Group B), Wijk aan Zee, 1998
3
2
a b c d e f g h
{ 3341 White to move
White is in considerable difficulty. The pawn at e3 is under attack,

and the Black king can infiltrate along the diagonal leading to b1 .
37.e4?? 37 .b3 ! was the only way to save the game. White failed to
appreciate that once the pawn reaches a4, the game is effectively over.
37 . . . Kf5 ; 38 .Nh6+ Ke4; 39.Nfi Bxe3; 40.Nxd6+ Kxd4; 4 l .Nc4 is not to
be feared because Black has the undermining tactic 4 l . . . a4! ; 42.Nxe3
axb3 ! ; 43.Nd l Kd3 (43 . . . e5; 44.Kgl e4; 45 .Kf2 Kd3; 46.Ke l e3; 47.Nb2+
Kc2; 48.Na4 b2; 49.Nxb2 Kxb2; 50.Ke2 is a simple win.) 44.Nb2+ Kc3;
45.Na4+ Kb4; 46.Nb2 Kc3 ; 47.Na4+ with a draw.
37 . . . a4; 38.Nf2. 38 .e5 d5; 39.Kg l Be l cleans up on the queenside.
38 . . . Bc3 ! !
3
2
a b c d e f g h
{ 3351 White to move
The bishop cannot be captured because then White cannot stop

the a-pawn from promoting. White resigned. There was no point in
waiting for 39.bxc3 a3 40.Nd3 a2 .
21 1
BISHOP + 5 US. K N IGHT + 4

If you have the knight, you must be careful when going after rim
pawns, because a lot of time will be needed to get the knight to the rim
and back to the center. In this example, White can grab a pawn or two
quickly, but that is a dangerous road.
ANALYSIS OF KRAMNIK VS. IVANCHUK

Las Palmas, 1996
3
2
a b c d e f g h
36 . . . Bg4 requires patience on White's part. 37.Kg5 . (37.Nxa6? Ke6;

38.Nxc7+ Kxe5 ; 39.Kg5 Bdl ; 40.b4 Ke4! gets to the base of the pawn
chain too quickly. 4 1 .c4 Kd4; 42 .c5 Kc3 ; 43.b5 ! cxb5; 44.Nxb5+ Kc4. )
3 7 . . . Be2 ; 38.b4 Kf7; 39.a4 Ke7; 40.a5 K£7.
3
2
a b c d e f g h
[337] White to move
How can White make progress now?

4 l.c4 Bxc4. Forced, or the a-pawn falls. 42.Kxh5 Be2+; 43.Kg5 Bc4;
44.Kf4 Be2; 45.Ke3 Bfl ; 46.Kd4 Ke7.
212
BISHOP US. KN I G HT EN DG A M ES
3
2
a b c d e f g h
[338] White to move
White has a passed pawn and Black's pawns are weak. White must
somehow get the king to £5, with the knight at f4 keeping the enemy
king back.
47.Nd3 ! Ke6. 47 . . . Bxd3; 48.Kxd3 Ke6; 49.Kc4 Kxe5; 50.Kc5 Ke4;
5 l .Kxc6 Kd4; 52.b5! and there will soon be a new White queen.
48.Nf4+.
3
2
a b c d e f g h
[339] Black to move
Black cannot survive with either active or passive defense, but that
is not to say that the point is in hand. White must choose wisely after
48 ...Ke7.
4 9 . Kc5 ? ! does n ' t wo rk . 49 . . . Bb5 ; 5 0 . N h 5 Ke 6 ; 5 l . N g7+ Kf7
(5 l . . . Kxe5; 52 .Ne8 followed by Nxcc7 and Nxb5 with a winning king
and pawn endgame, or the win of both pawns. ) 52.Nf5 Ke6; 53.Nd4+! ?
Kxe5; 54.Nxb5 cxb5 .
213
CARDOZA PU BLISHI NG • ERIC SCHILLER
a b c d e f g h
[3401 White to move
Neither side can win. Black will ignore queenside threats and go
after the b-pawn. 55.Kc6 Kd4; 56.Kxc7 Kc4; 57.Kb6 Kxb4; 58.Kxa6
Kc4; 59.Kb6 b4; 60.a6 b3; 6 l .a7 b2; 62.a8Q b l Q+ saves the game.
Correct is 49.Ke4! and the king gets to f5 . 49 . . . Bc4; 50.Kf5 Bg8;
5 l .Ng6+ Kf7; 52 .e6+ Ke8; 53.Kf6 Bh7; 54.Ne5 Be4; 55.Nd7 Bd3; 56.Nb8
Bb5; 57.e7 places Black in zugzwang. 57 . . . Bc4; 58.Nxc6 Bb5; 59.Nd4
Bd3; 60 .Ne6 and after a preliminary check the pawn will promote.
BISHOP + 5 US. KN IGHT +5

For some reason I seem to get a lot of endgames featuring Bishop
vs. Knight with 5 pawns each, and you'll find three examples here,
against a promising young player, an International Master, and a Grand
master. First, however, let's let the Grandmasters instruct us.
DREYEV VS. VAGANIAN

Budapest, 1996
a b c d e f g h
[341 1 White to move
214
BISHOP US. KN I G HT E NDG A M ES
The king will go to g4 to help the f-pawn move forward without

compromising the pawn structure.
39.Kh3! g5?! A rather desperate plan. Black wants to control the
4th rank with . . . Kg6 and . . . h5 . 39 . . . Ke6; 40. Kg4 Kd5; 4 l .f4 illustrates
the basic plan. How should Black reply? 4 l . . . e4I ? is an interesting try,
but it doesn't hold up . 42.Bxg7 Kc4; 43 .f5 e3 ; 44.Kf3 Kd3 ; 45.Bc3 Kxc3;
46.Kxe3 .
3
2
a b c d e f g h
[342] Black to move
White will win because the knight can't hold back both pawns .
46 . . . Kc4 (46 . . . Nd8; 47.f6 Ne6; 48.£7 Kc4; 49 .c6 Kc5; 50 .c7 wins. ) 47.f6
Nd8; 48.c6I
Unfortunately, the alternative plan 4 1 . . . exf4; 42.Kxf4 g6 meets with
43.Kg5 Kxc5, where 44.Kh6 wins both kingside p awns. 44 . . . Kc4; 45.Bf6.
The best square, effectively dominating the knight. 45 . . . Kb3 ; 46.Kxh7
g5. A finesse to deflect the bishop, so that a 1 is not covered. It does not
save the situation. 47.Bxg5 Kxa3; 48.h4 a5; 49.h5 a4; 50.Kh6 Kb3; 5 l .Bf6
a3 ; 52.Kg7 etc.
40.Kg4 Kg6; 4 l .f4.
3
2
a b c d e f g h
[343} Black to 1nove
21 5
CA RDOZA PU BLISftiNG • ERIC SCH ILLER
It is important to time such moves properly, and this is the moment

for the advance. The threat of f5+ leaves Black with only three choices.
He can capture the f-pawn with either pawn, or toss in . . . h5+. 4 l .h4?!
gxh4 42.Kxh4 h5 .
4 1 . . . h5+. 4 l . . . gxf4; 42 .gxf4 e4; 43 .f5+ K£7; 44.Kf4 etc. is simple
enough but 41 . . . exf4; 42.gxf4 gxf4; 43 .Kxf4 requires analysis.
3
2
a b c d e f g h
[344] Black to move
43 . . Kf7; 44. Kf5 Ke7; 45.a4 ! Otherwise Black might get the a-pawn
.
to a4, which might allow more counterplay. 45 . . . Kf7; 46.Bf6 Nb4. Black
is running out of moves. (46 . . . a5; 47 .h3 Nb4; 48.Bd8 Nc6; 49.Bg5 Nb4;
50.Ke5 Ke8; 5 l .Kd6 and the c-pawn has a green light to proceed.) 47.Bg5
Nc6; 48.h4 a5; 49 .h5 Nd4+; 50.Ke4 Ne6; 5 l . Be3. The bishop has a great
view of both sides of the board and the White king can finish things
off by infiltrating the queenside.
42. Kf3 exf4. 42 . . . g4+ can be played immediately, but it is no better.
43. Ke4 exf4; 44.gxf4 Ne7 (44 . . Kf7; 45.Kd5) 45.Be l ! After Bh4, Black
.
will not be able to keep the White king out.

43 .gxf4 g4+.
3
2
a b c d e f g h
[345] White to move
216
BISHOP US. KN IGHT E NDG A M ES
44.Ke4 h4. For 44 . . . N e7, see the previous note. 45 .£5+ Kg5; 46.£6
Kg6. 46 . . . N d8 runs into a familiar theme: 4 7 .c6 ! Nxc6 48.£7 and it is
mate in 6 .
47.Kf4. Black resigned. Both kingside pawns fall. After 4 7 . . . g3 ! ;
48.hxg3 hxg3; 49 .Kxg3 Kf7; 50.Kf4 Ke6; 51 .Kg5 Kfl; 5 2 .Kf5 Black must
soon move the knight, and then the king '\valks into the queenside.
SCHILLER VS. KAHANE

Internet Chess Club, 1997
3
2
a b c d c f g h
[346] Black to move
Facing one of my students, I obtained this favorable endgame. The

pawn at d5 is very weak, and Black must keep the king at c5 or else the
pawn may fall via a zugzwang. White does not really threaten to cap
ture at f6, because after . . . Nxf6 the h-pawn is weak. This suggests a
move like . . . h6, but Black decides to capture at g5 . Even though it cre
ates a potential passed h-pawn, it presents White with a clear plan of
advancing the f-pawn.
The endgame should by all rights end in a draw. The position is
not simple, however. White's a-pawn is very weak, and although the
knight is all the way over at g8, it can eventually target that '\veakness.
40 ... fxg5.
40 . . . h6? ! is not best because the knight cannot defend the kingside
by itself, and the White king is boxed out. Still, it holds with accurate
play. 41 .gxh6 (4 l .Bf7 Ne7; 42.gxf6 gxf6; 43 .e4 d4! ) 41 . . . Nxh6; 42.e4!
Kd6 ! (42 ... dxe4+?; 43 .Kxe4 Kc4; 44.h5 Kb3 ; 45 .a4 Kb4; 46 .Kd5 Kb3;
47.Ke6 Kb4; 48.Ke7 Nf5+; 49. Kf8 etc.) 43.Kd4 (43 .h5 ! ? Ke5 ! ) 43 ... dxe4;
44.Kxe4 Ke6; 45 .Kd4 Nf5+; 46.Kc5 Nxh4; 47.Bc6 Ke5 ; 48.Kb5 K£4;
49.Kxa5 Nxf3; 50.Bxf3 Kxf3; 51 . Kb4 g5 ; 5 2 . a4 g4; 53.a5 g3 ; 54.a6 g2 ;
55 .a7 gl Q; 56.a8Q+ which can be drawn.
211
CA RDOZA PUBLIS H I NG • ERIC SCH ILLER
40 . . . Kd6 ! is the simplest defense.
3
2
a b c d e f g h
[347] White to move
4 l .Kd4 ( 4 l .e4 Ke5) 4 l . . . Ne7; 42.gxf6 Nf5+! Cleanest, but not nec
essary. (42 . . . gxf6; 43 .e4 h6; 44.exd5 Nxd5; 45 .Bf7 Ne7; 46.Ke4 Kc5 ;
47.f4 Kb5 ; 48.Bb3 Kc5 also draws . ) 43. Kd3 gxf6; 44.h5 Nh4; 45.f4 Nf5
is no worse for Black. The bishop must stay in position to defend the
h5-pawn against . . . Ng3 . Black can just shuttle the king around c5-d6-e6.
4 1 .hxg5 Ne7.
3
2
a b c d e f g h
[348] White to move
42.£4. 42 .e4 is premature. Once the pawn advances , the central

pawn s tructure is resolved and White has too little to work with.
42 . . . dxe4+; 4 3 . Kxe4 (43.fxe4 h6; 44.gxh6 gxh6 should be drawn. )
4 3 . . . Nc8; 44. Ke5 Nd6; 45.g6 hxg6; 46.Bxg6 Nc4+; 47.Ke6 Nxa3; 48.Kf7
a4; 49.Kxg7 Nb5; 50.Bc2 a3; 5 l .Bb3 Kb4; 52 .Ba2 Nc3; 53.f4 Nxa2; 54.f5
Nc3; 55.f6 a2 ; 56.£7 a l Q; 57.f8Q+ is a draw.
42 ... Nf5; 43.Bd7. 43 .Bf7 Nd6; 44.Bg8 h5; 45 . Ke2 h4; 46.Kf3 h3;
47.Be6 h2; 48.Kg2 Nc4; 49.Kxh2 Nxa3 ; 50.f5 Kd6; 5 l .Bf7 Nc2 ; 52.f6
218
BISHOP US. KNIGHT EN DG A M ES
gxf6; 53 .g6 Ke7; 54.Bxd5 K£8; 55.e4 Kg7; 56 .Bf7 a4 57.Kg3 a3; 58.Kf4
Nb4; 59.Kg4 a2 ; 60.Bxa2 Nxa2; 6 l .e5 draws.
43 ... Nd6.
3
2
a b c d e f g h
[349] White to move
Black abandons the kingside and goes after the a-pawn. An ambi
tious strategy, but Black failed to follow through on his next move.
44.Bg4? A time pressure blunder. White now should have to fight
to draw. 44.a4 Nc4 (44 . . . h6; 45 .gxh6 gxh6; 46. Ke2 h5; 47 . Kf3 Nc4;
48.Bb5 is also headed for a split point. ) 45. Be8 Nd6; 46.Bh5 Nc4; 47.Be8
would have drawn.
44. .. Ne4?? Black had to boldly grab the a-pawn, and count on the
advance of his own a-pawn to cope with the f-pawn. 44 . . . Nc41 would
have salvaged the game.
3
2
a b c d e f g h
[350] White to move
White does not want to play 45 .e4 because of 45 . . . dxe4+; 46.Kxe4

Nxa3; 47.Be6 a4; 48.f5 Kd6; 49.Bg8 NbS. 45 .Be6 comes into consider
ation, intending to go after the h-pawn. Black will escape with 45 . . . Nxa3,
even after 46.f5.
219
CA RDOZA PU BLISH I NG • ERIC SCH ILLER
3
2
a b c d e f g h
[351 ] Black to move
46 . . . Kd6! Forces White to put the plan into action right away. 4 7. Bg8
a4; 48.Bxh7 Nc4; 49.f6 gxf6; 50.g6 a3 ; 5 l .g7 a2 ; 52 .g8Q a l Q; 53.Qd8+
Kc6; 54. Qc8+ with a draw.
Let's see what actually happened.
3
2
a b c d e f g h
[352] White to move
45.Bf5 Nd6; 46.Bxh7 Nc4. Black could have had this position with
the h-pawn still on the board, had he played 44 . . . Nc4!
4 7 .f5 Nxa3; 48.f6 gxf6; 49.g6. Black resigns .
220
BISHOP US. KN IGHT E NDG A M ES
SCHILLER VS. LEVIN

Groningen GM Open, Holland, 1997
a b c d e f g h
Using my general guideline that \Vith three open files in the center
a knight can holc;l off a bishop, I engineered this endgame after achiev
ing an inferior position. I instinctively felt it could be held, but had to
play very carefully.
38.£4! 38.Kxe3 Bxc4; 39.bxc4+ Kxc4 wins for Black, of course. 38.g4
g5 fixes the kingside pawns on light squares and this could lead to
serious problems later.
38 . . . e2. Otherwise I can play g4 and Black cannot counter with
. . . g5. 39.Kxe2; Bxc4+; 40.bxc4+ Kxc4.
3
2
a b c d e f g h
[3541 White to move
The king and pawn endgame is drawn.

4 1 .g4! b5 ! Black gets rid of the remaining White queenside pa\vn
and both sides end up with queens.
42 .£5.
22 1
CARDOZA PU B LISJII NG • E RIC SCH I LLER
3
2
a b c d e f g h
[355] Black to move
42 . Kd2?? would lose to the simple 42 . . . bxa4; 43 .Kc2 Kd5 and Black
WillS.
42 . . . bxa4; 43 .f6 a3; 44.f7 a2; 45.f8Q alQ; 46.Qf7+ Kc5; 47.Qc7+.
3
2
a b c d e f g h
[356] Black to move
4 7 . . . Kd5; 48.Qd7+ Kc4; 49.Qc6+. Draw agreed, as perpetual check

is available.
222
BISHOP US. KNIGHT E NDG A M ES
SCHILLER VS. BISGUIER

American Open, Los Angeles, 1995
a b c d e f g h
{3571 White to move
Everyone talks about the advantages of bishop vs . knight in open

positions, but as this endgame shows, a knight can be a perfectly ad
equate defender in many cases. It may be an oversimplification to state
that if there are three or less open files between pawn islands, the
knight can handle the defense with multiple pawns on the board. In
the present position, White must play actively and carefully to hold
such positions. Because of some of the openings I play, I wind up in
these positions fairly frequently, and I don't think I have ever lost one,
even against titled opponents such as the Grandmaster playing Black
here.
29.Kc3. Start by centralizing the king, not so much to use it aggres
sively but rather to support the knight which needs safe squares in the
center.
29 . . . Kd6; 30.Kd4.
a b c d e f g h
{3581 Black to move
223
30 . . . h6. Why . . . h6, instead of the . . . f6 preferred by most computer

programs? The answer is simple, though hard for a computer to appre
ciate. After . . . h6, the kingside pawns can only be effectively attacked by
Nf5, whereas after . . . f6 the h5-square can be used. It is easier for Black
to defend f5 than h5.
3 l .Ne4+ Kc6; 32 .Ng3. Heading for h5 . 32 ... Kb5. Black threatens to
invade the queenside, the only practical plan. White returns the knight
to e4, as kingside adventures would be more dangerous. 33.Ne4. 33.Nh5
g6; 34.Nf6 Kb4; 35.Nd5+ Ka3 ; 36.Nc3 b5 gives Black a strong initiative.
33 . . . Kc6. 33 . . . Kb4 runs into 34.Nd6. 34.Ng3 Back to the old plan!
34 . . . a5; 35.Nh5 g6; 36.Nf4 b5; 37.Nd3 Be6.
a b c d e f g h
[359] White to move
White has nothing to worry about, but the following moves sim
plify matters.
38.b4 axb4; 39.Nxb4+ Kb6; 40.a3. Now the a-pawn is safe on a dark
square, and the White king is in position to infiltrate on the kingside.
Black responds by tying down both kings on the queenside. 40 . . .Ka5;
4 l .Kc3 Ka4; 42.Kb2 Bc4; 43.Nc6 Bd5; 44.Ne5.
3
2
a b c d e f g h
[360] Black to move
224
BISHOP US. KNIGHT E NDG A M ES
Black can achieve nothing, so eliminates the remaining queenside

pawns and settles for the half-point.
44 ... b4; 45.axb4 Kxb4; 46.f3. Draw agreed. Should the position
have been dra,vn earlier? Probably, but Grandmasters tend to torture
their opponents for a while in endgames, and it is easy to go wrong.
For this reason some players try to avoid endgames like these when
facing stronger opposition, but it doesn't make sense. Know how to
draw a drawn game and Kasparov himself can't beat you !
BISHOP + 6 US. KN I GHT + 6

The following game features a pawn structure which can easily arise
in a Classical Caro-Kann, though in this case it comes from a Vienna
Game. The bishop clearly has the advantage over the knight. The knights
is tied down to defense of a pawn which is blockaded by the king, when
the king cannot be displaced by a knight check.
ZEIGLER VS. SCHILLER

USA Today Network, 1990
3
2
a b c d e f g h
[361] Black to move
Black has a good endgame, with bishop vs. knight and a weak pawn
at e5. The first thing to consider is the pawn endgame after exchang
ing at £3 . Once that was rejected, the alternative of driving the king
away from f4 came under examination, and that led to victory.
26 ... g5+! 26 . . . Bx£3 ; 27.gxf3 f6; 28. exf6 Kxf6 is considered in dia
gram #83. 27.hxg6 fxg6. 27 . . . Bxf3 is a cheap trick, foiled by 28 .g7!
28.Ne l g5+.
225
CA RDOZA PU BLISaiNG • ERIC SCH ILLER
a b c d e f g h
This is certainly the easier plan for Black. The e-pawn falls immedi
ately.
29 .Kg4 Kxe5; 30.Kh5 Be4. The classic theme of bishop dominat
ing knight. Black does not need the h-pa,vn.
a b c d e f g h
3 l .Kxh6 Kf4; 32.g3+? 32.Kh5 g4; 33.Kh4 g3 ; 34.Kh3 at least throws

a few obstacles in Black's path by tying down the Black king. The
zugz,vang trick is now the only way to \vin. 34 . . . b6; 35.a3 a6; 36.a4 aS;
37.Kh4 Bf5; 38.Kh5 (38 .Nf3 Bc2; 39 .Nd2 Ke3; 40.Nfl+ Kf2; 4 1 .Nxg3
Bxb3; 42 .Nf5 Kxg2; 43 .Ne3+ Kf3 and again domination is a winning
technique. ) 38 . . . Bg4+; 39.Kh4 Bd l ; 40.Nd3+ Ke4; 4 1 .Nc l Kd4; 42 .Kxg3
Kc3; 43.Kf4 Kd2 ; 44.g4 Kxc l ; 45 .g5 Bxb3; 46.g6 Bxc4; 47.g7 etc.
32 ... Kg4. A little torture. White 'vill soon be in zugz,vang, since
neither the knight nor king dare move from their spots. 33 .a3 a5. White
resigned.
226
BISHOP US. K N I G HT EN DGA M ES
PSAKHIS VS. ALEKSANDROV

Rubinstein Memorial, Polanica Zdroj, 1997
3
2
a b c d e f g h
[3641 White to move
The game was agreed drawn here. Let's see why. 34.gxf5 exf5.
a b c d e f g h
[3651 White to move
35.b5 cxb5. 35 . . . Kc7; 36.b6+ Kb7; 37.Kc3 and the king invades via
e5 and d6. 36.Kc3 Kc7; 37.Kb4 g6; 38.Bxb5.
221
3
2
a b c d e f g h
[366] Black to move
38 ... Bc6; 39.Be2 Kd7; 40.Kc3 Ke6; 4 1 .Kd4 Kf6; 42.Bf3 Ke6.
3
2
a b c d e f g h
[367] White to move
43.Bg2 BaS.
B ISHOP + l US. K N I G HT + l
Our final game presents a contrast to the previous one. Here the
bishop cannot compensate for the bad pawn structure. Is it a "good"
bishop because it has great scope from e6, or a "bad" one because
there are 5 pawns on light squares?
228
BISHOP US. KNIGHT EN DGA M ES
KRAMNIK VS . IVANCHUK
Las Palmas, 1996
3
2
a b c d e f g h
[368] White to move
Black has the bishop, but also suffers from a weak pawn structure.
Far from thinking about a win, Black must be concerned with saving
his skin! The notes to the game are based on those by the winner.
White might be tempted to close the queenside by building a fence
at a4, b3 and c4. After all, the bishop can't move into position to attack
the anchor at b3, and in any case the knight or king can defend from
d2 . Kramnik calculates that the time is right to operate on the kingside,
while the Black king is far away. A road must be cleared to g5.
27 .f4! The purpose of this move is to free up the g3 square so that
the king can infiltrate on the dark squares of the kingside. In addition,
the f3-square can be used by the knight. This forces Black to exchange
pawns. 27 ... exf4. 27 . . . f6; 28.fxe5 fxe5 ; 29.Nf3 wins the pawn.
28.gxf4 Kc8; 29.Kg3!? 2 9 . Ke3 Kd7; 30.Kd4 was another good ex
ploitation of the dark squares.
29 ... Kd7; 30.Kh4 f6. The White king must not be allowed to reach
g5 !
229
CARDOZA PU BLISH I NG • ERIC SCH ILLER
a b c d e f g h
[369] White to move
3 1 .e5! This reopens the road. 3 1 ... fxe5. 3 l . . . Ke7 ! ? only delays the
exchange, since White can pile on the pressure with. 32.Ne4.
3
2
a b c d e f g h
[3 70} Black to move
Then 32 . . . fxe5 (32 . . . f5 ; 33.Nc5 Bc4; 34.Kg5 K£7; 35.g3 a5 ; 36.Nb7

a4; 37.Nc5 Bb5 allows White to demolish the kingside with 38.e6+ Kg7;
39.e7 K£7; 40.e8Q+ Kxe8; 4 l .Kxg6 and the remaining Black pawns fall
quickly. ) 33.fxe5 Bd5; 34.Nc5 ! Bxg2; 35.Kg5 ! The a-pawn is of no con
sequence, for the moment. 35 . . . Bh3; 36.Kxg6 is all Kramnik gives, but
it doesn't seem that simple. We examine it in position #336.
32.fxe5 Bd5; 33.g3.
230
BISHOP US. K N IG HT EN DGA M ES
a b c d e f g h
Black tries a different defense, going after the weak e-pawn.

33 . . . Ke6; 34.Kg5. Black was faced with an important decision here.
34 . . . Kxe5.
3 4 . . Kf7 i s superior, according to Kramnik. It deserves further study:
.
35.c4 Bg2 with a choice of 36.e6, 36.a4 or 36.Nb3. Let's take a look.
36.e6+?! Kxe6; 37.Kxg6 picks off the h-pawn. Black would then have
to race the king to the queenside. 3 7 . . . Ke5 ; 3 8 . Kxh5 Kd4; 39.g4 Kd3.
a b c d e f g h
[372] White to move
White loses the c-pawn, but keeps the enemy king away from his
other possessions. Only for a moment, however. 40.Nb3 Kxc4; 4 l .Nc l
(4l .Na5+ Kb5; 42.Nb7 c5) 4 1 . . . Bf3; 42.Kg5 Kd4; 43 .Kf4 Bd5; 44.g5 B£7;
45 .Kf5 Ke3 and it is White who is in trouble.
36.a4?! The knight needs room to roam, and to earn it, the bishop
at g2 must be displaced. But Black can transpose to the last note with.
36 . . . Ke6! ; 37.Kxg6 Kxe5; 38.Kxh5 Kd4.
36.Nb3 is perhaps the best try. 36 . . . Bfl ; 3 7.c5 Ke6; 38.Kxg6 Kxe5;
23 1
CARDOZA PU BLIS H I N G • ERIC SCH ILLER
39.Kxh5 Be2+; 40.Kg5 (40 .g4? Kf4) 40 . . . Kd5 ; 4 l .Nc l ! Bc4; 42.b3 ! This
picks up a tempo over the immediate b4, which would leave the knight
at c l dominated. 42 . . . Bfl ; 43.b4 Bc4 with an interesting endgame which
is considered in diagram #327.
35.Kxg6 h4.
3
2
a b c d e f g h
[373} White to move
A familiar strategic situation, with Black trying desperately to get

into the queenside. Notice how well the pawns at g3 and c3 coordinate
with the knight at d2 to keep the enemy king out. The knight is only
partially dominate; it can use fl to pivot to e3 or h2. Therefore, there is
no need to fear its loss. The knight also makes a significant contribu
tion by controlling f3 . The h-pawn cannot be saved. Black sacrifices it
to send the White king into exile on the edge of the board, allowing
more room for queenside maneuvers and keeping the White king fur
ther from the vulnerable queenside pawns .
After 36.gxh4 Kf4; 37.h5 Ke3. White won as shown in diagram
#324.
232
ROOK EN DG A M ES
A few words on rook endgames . That's a fine joke! My first serious
study of rook endgames was the classic book by Levenfish and Smyslov,
which remains after almost 30 years an invaluable introduction to the
art of playing with rooks and pawns . That book was just a couple of
hundred pages, and the authors noted that a complete course on rook
endgames would have to be at least three times the size. They underes
timated the task. Over 300 pages are needed for John Nunn to survey
positions with a rook for each side and a single pawn!
For our practical purposes, there are a few basic positions you need
to know. After that, just keep your rook active at all times, even if it
may cost you a pawn to do so. Experience and study will improve your
understanding of rook endings . As these are the most common
endgames, you will often need to handle them properly to collect your
fair share of points.
The most important positions are those with a single pawn. You
must master these positions, because they are very difficult to figure
out at the board. I had an important tournament game against an Ex
pert (Candidate Master) a while ago which should have been a simple
draw. In fact, the only reason I hadn't offered one is that I was in a funk
because I had misevaluated the endgame and thrown away a very prom
ising position . I just mechanically played with the single pawn, waiting
for my opponent to reach the Philidor position (diagram #382). When
we got to the critical point, he failed to apply the well-known drawing
technique and I was re\varded with a full point for my persistence.
ROOK US. I
A rook beats a lone pawn if the king can attack the pawn before it
promotes. There is a way to win even after promotion, in special cases
which you will see below. It is very important to know which positions
can be drawn by this method, because it allows you to sacrifice your
rook for an enemy pawn and still hold the game.
233
CA R DOZA PU B LI�H I NG • ERIC SCH I LLER
YAKOVICH VS. KOSTEN

Hastings Congress, England, 1993
3
2
a b c d e f g h
{3 74] Black to move
The position is a draw because the White king supports the pawn
and the Black king is too far away. Black will have to sacrifice the rook
for the pawn and settle for a draw.
LOPEZ VS. ALVARADO

Peruvian Regional Championship, Lima, 1999
3
2
a b c d e f g h
[375] Black to move
This time the king is too close, even though it is still behind the
pawn and the path to g3 is closed.
65 ... g3; 66.Rg8+ Kh4; 67.Kf4 Kh3; 68.Rxg3+ Black resigned.
ROOK US. 2
A classic game will illustrate the critical race between a rook and
connected passed pawns . If you have the pawn, get it to the 6th rank
234
ROOK EN DG A M ES
and you will usually win, unless the opposing king is able to help with
the defense . Once the pawns are on the 61h, the rest is easy to calculate.
The endgame skill required is the advancement of the pawns.
TARRASCH VS. JANOWSKY

Ostende, 1907
3
2
a b c d e f g h
Black has just played 78 . . . Ra4 and threatens to place the rook on
the b-file and get a queen. White must sacrifice the rook to stay in the
game.
79.Rxb2. 79 .f6 Ra l ; 80.Rxb2 Kxb2; 8 l .g5 Rf1 reaches the same
position as the game. 79 ... Kxb2 .
a b c d e f g h
80.f6! White has a winning position. The rook alone cannot handle
the pawns. Note that the pawn at g4 is taboo, because the f-pawn then
advances and Black can do nothing to stop it. 80.g5 Rg4 ; 8 l .g6 Kc3;
82. Kf3 Rgl ; 83 .Ke4 Kd2; 84.Ke5 Ke3 ; 85.f6 Rxg6; 86.f7 Rg5+; 87.Ke6
235
CA RDOZA PU BLISH I N G • ERIC SCH ILLER
Rg6+; 88.Ke5 Rg5+; 89.Kf6 Rgl draws . 80 ... Ra l . Black should have
played this back at move 78, saving a crucial tempo.
8 1 .g5 Rfl ; 82.Kd4.
3
2
a b c d e f g h
The plan is obvious. White must defend the f-pawn with the king,
freeing the g-pawn from that responsibility. Black's king is too far away
to have any effect.
82 . . . Kb3; 83.Ke5 Kc4; 84.g6! Now that the f-pawn is defended, the
g-pawn can advance. 84 . . . Re l+; 85.Kd6. The king must take shelter in
front of the pawns, since otherwise Black can simply check from a safe
distance.
a b c d e f g h
85 ... Rd l+. 85 . . . Rgl is the preference of computers, but Tarrasch

had it all worked out. 86.g7 Kd4; 87.Kc6 Kc4 (87 . . . Rg6; 88.Kb5 ! ) 88.Kd7!
Kd5; 89 .Ke8 Ke6; 90.£7 Ra l ; 9 l .f8N+! K£6; 92.g8Q would have been a
prettier finish, according to Tarrasch.
86.Ke6 Re i+; 87.Kf7. Black resigned.
236
ROOK EN DGA M ES
ANALYSIS OF PANCHENKO VS. BRASLAVSKY

Yalta, 1996
3
2
a b c d e f g h
Black has problems getting the pawns up the board.

52 ... h4 is met by 53.Kb5. Black can send the pawns forward imme
diately or can bring the king closer as an escort. In either case, White
will win.
53 . . . g5; 54.Kc4 Kf6; 55.Rh8 Kf5; ·s6. Kd3 Kg4; 57.Ke3 Kg3; 58.Ke2
Kg2 ; 59.Rh5 h3; 60.Rxg5+ Kh l ; 6 l .Kf2 h2; (6 l . . . Kh2 ; 62 .Rh5 Kh l ;
63.Rxh3#) 62 . Kg3 ! Otherwise it is stalemate. 62 . . . Kgl ; 63.Kh3+ Kh l ;
64. Ra5 Kgl ; 65 .Ral+ Kf2 ; 66 .Kxh2 . 5 3 . . . Kf6; 54.Rf8+ Kg5 ; 5 5 . Kc4 Kg4;
56.Rf6 g5; 57.Kd3 Kg3; 58.Ke2 h3; 59.Kfl h2; 60 .Rh6 g4; 6 l .Rh8 with
a win.
ROOK + I US. ROO K

As already mentioned, a book on these endings takes up over 300
pages, with double columns! All of the positions have been worked out
by computers, those machines that never get bored or tired. As a prac
tical player, you must learn at least the Lucena and Philidor positions,
but the more you study, the better your results will be. In addition to
the endgames presented here, an instructive example is seen in dia
gram #40 1 .
LUCENA POSITION
The Lucena position is one of the most critical in all of endgame
theory. It isn't by Lucena at all. It was in a book called Il Puttino, by Dr.
Alessandro Salvia, published in 1634. B efore you show off your knowl
edge of this arcane fact, you had better learn the position!
237
a b c d e f g h
[381] White to move
White wins as follows. First, you bring the rook to c4, so that it can
get to d4 and drive away the enemy king. Then you slip the king out,
and use the rook to defend it from enemy checks. This is called "build
ing a bridge."
l .Rc4 Ral . l ... Rd2, to stop the check, White takes the a-file with
2 . Ra2, and the king comes to a7, after which the pawn can advance.
2 .Rd4+ Ke7; 3.Kc7 Rei+; 4.Kb6 Rb l+; 5.Rb6. White wins.
PHILIDOR POSITION
This endgame is seen frequently and is the basis for much of the
theory of rook and pawn endgames. White has cut off the enemy king
on the back rank, but Black can hold with accurate defense.
a b c d e f g h
[382] Black to move
Black must defend the 6th rank, so that the White king cannot ap
proach without the pawn occupying e6, which might otherwise be used
as a haven for the king if Black gives check at fl.
238
ROOK EN DGA M ES
Thus l . . . Ra6 must be played. White will oblige by advancing the c

pawn, but first repositions the rook more effectively. 2 .Rb7 Rc6; 3.Ra7
Rb6; 4.e6 Rbl!
3
2
a b c d e f g h
{383] White to 1nove
Black will be able to check along the first rank, and the White king
has no place to hide.
KARPOV VS. YUSUPOV

Linares 199 1
3
2
a b c d e f g h
[384] Black to move
This is perhaps the third most important position in rook endgames.

With rim pawns, it is not possible to hide the king from checks. The
result is a drawn game. One slip, however, can bring Black's game crash
ing down.
73 ... Ka7. 73 . . . Kc7??; 74.Ra8 ! is the famous trick, since 74 ... Rxh7;
75. Ra7+ picks off the rook. 74.Kf4 Kb7; 75 .Kg4 Rh 1 ; 76.Kg5 Rgl+.
239
CA RDOZA PU BLISH ING • ERIC SCH I LLER
a b c d e f g h
[385] White to nzove
The 'Vhite king cannot find shelter so the game 'vill be dra,vn.
KUBALA VS. GIPSLIS

Decin, 1994
E.xperienced players '\ron't hesitate in playing the correct move here.

The cardinal rule of rook endings 'vith a passed pa,vn is to cut off the
enemy king so that the enemy rook has no support in coping 'vith the
advancing pa,vn, 'vhich can be supported by the king. Use your rook to
deal 'vith the king, and your king to deal 'vith the rook!
67.Re2! White can afford to delay by attacking the rook, since 67.Kg3
Rh6; 68.Re2 Re6; 69 .Rxe6+ K.xe6; 70.Kh4 is a 'vinning king and pav;n
endgame, but 'vhy rely on counting 'vhen you can keep the king out
from the start. 67 . . . Rh7; 68.K.f4 Re7? The Grandmaster probably just
'vanted a quick exit. 68 ... Rh8; 69.g5 R£8+; 70.Kg4 'vould only prolong
the agony .
240
ROOK EN DG A M ES
69.Rxe7 Kxe7; 70.Kg5. The king is in front of the pawn, and the
win is trivial. 70 . . . Kf7; 7 1 .Kh6 Kg8; 72 .Kg6 Kh8; 73.g5 Kg8; 74.Kh6
Kh8; 75 .g6 Kg8; 76.g7. Black resigned.
NICKOLOFF VS. KOVACEVIC

Toronto 1989
3
2
a b c d e f g h
[387] White to move
Remember this position, because it figures in a lot of rook and

pawn endgames ! The required technique is bridge-building. Eventu
ally, Black must take control of the g-file so that the king can climb out
and let the pawn promote.
1 7.Rh6 Rf5; 18.Rg6 Ra5; 1 9.Kf1 Ra 1+; 20.Kf2 Rg1 ; 2 1 .Ra6 Rg2+;
22.Kf3 Rg5; 23.Ra2 Rf5+; 24.Kg3 Rg5+; 25.Kh3 and Black saw no need
to contest the decision further. Draw agreed.
SCHUELLER VS. MODR

Austria, 1997
3
2
a b c d e f g h
[388] White to move
24 1
CA RDOZA PU B LI_S H I NG • ERIC SCH ILLER
White has an excellent king position but a somewhat inactive rook.

The win involves activating the rook, advancing the pawn, and keeping
the king safe from checks. The last task is the key to getting the full
point.
79.Rh 2. The rook must be active. Keeping Black's king off the sev
enth rank is not important. 79 . . . Ke7; 80.Rf2. The king is cut off on the
f-file. 80 . . . Rg3 ; 8 l .Kh6 ! This frees the g-pawn to move. 8 1 . . . Rh3+;
82.Kg7! The pawn can now advance one square.
82 . . . Ke6; 83.g6 Rg3 . 83 . . . Ra3 ; 84. Kg8 Ra8+; 85 .Rf8 Ra2 ; 86.g7 Rg2;
87.Rf4 Ke5; 88.Rh4 Kf6; 89.Rh6+ Ke5; 90.Kh8 forces Black to give up
the rook. 84.Kh7. White repeats the technique, to get the pawn to g7.
84 . . . Rh3+; 85.Kg8 Rg3; 86.g7.
3
2
a b c d e f g h
[389] Black to move
This is the target position. The pawn has advanced as far as it can,
and the threat is K£8 followed by g8Q. Black must prevent this.
86 . . . Ke7; 87.Re2+. The king must abandon its coverage of £8, and
White will be able to promote .
87 . . . Kf6; 88.Kf8 Black resigned.
242
ROOK EN DG A M ES

3
2
a b c d e f g h
[390] Black to move
White's king is cut off on the rank, but a draw is still available.
1 2 . . . Rg4?! ; 1 3 .Ra8 Rg3; 14.Ra2 Rg2; 1 5 .Ra3+ Kh2 ; 1 6.Kf4 h3; 1 7.Ral
Rgl ; 1 8. Ra2+ Kh l ; 1 9 .Kf3 h2; 20.Rb2 is a draw. Black cannot make
progress.
13.Ra8 Rc2.
3
2
a b c d e f g h
[391 } White to move
Gulko did not find the clearest path to the draw here. As is often
the case, the most efficient move is to bring the king in closer to the
enemy pawn. Instead of his move, bringing the rook to a4, he points
out that he should have played
14.Kg5 since 14 . . . Rg2+; 15.Kh5 leaves Black with no way to ad
vance the pawn.
243
CARDOZA PU B LI_S H I N G • ERIC SCH ILLER
3
2
a b c d e f g h
[392] Black to move
Agreed drawn.
ROO K + I US. ROOK + I

T\vo of the most crucial features of a rook endgame \vith a pawn
for each side is the possibility of sacrificing the rook for a pawn, reach
ing a dra\ving position as shown in the section on Rook vs. 1 , and
sacrificing (losing) a pawn, reaching Rook and 1 vs. Rook. You must
always consider these options while advancing your pawn or attacking
an enemy pawn.
(63F) ANALYSIS OF GULKO VS. YEPISHIN

3
2
a b c d e f g h
[393] White to move
White has to keep the king on the fourth or fifth rank to preserve
the dra\v. First let's see \vhat happens if he doesn't. 16.Kc3? Kg3; 1 7.Rh8
f3; 1 8 .Rg8+ Kf4; 1 9 . Rf8+ Ke3 ; 20.e6 Rei+! This drives the White king
244
ROOK EN DGA M ES
one file further from the e-pawn, and more importantly lets the rook
drop back to defend. 2 l . Kb2 Rc7 ! ; 22.Rf7 ReS; 23 . e7 ReS and Black
can promote the pawn.
1 6.Kc4 is acceptable. 16 . . . Kg3; 1 7.Rh8 £3; 1 8.e6 where 18 . . . f2!; 1 9 .Rf8
Rd6 ( 1 9 . . . Re l ; 20.Kd5 fl Q; 2 l .Rxfl Rxfl ; 22.e7 Rei; 23.Kd6 is also
drawn. ) 20 .e7 Re6; 2 l .Rg8+ Kh2 ; 22.Rh8+ Kgl ; 23.Rg8+ draws because
23 . . . Kfl allows; 24.e8Q Rxe8; 25.Rxe8 Kgl ; 26.Rf8 . On the other hand,
18 . . . Rd6; 1 9 .e7 f2; 20.e8Q (20.Rg8+ Kf4; 2 l .Rf8+ Kg3; 22.e8Q) 20 . . . fl Q+
is actually bad for Black. 2 l .Kc5 ! robs Black of any further checks, un
less the rook is sacrificed. 2 l . . . Rd5+ (2 l . . . Rf6; 22.Qe3+ Q£3; 23.Qg5+
wins the rook, because 23 . . Kf2; 24.Rh2+ Ke 1 ; 25.Qd2+ Kfl ; 26. Qc l +
.
forces mate. ) 22.Kxd5 Qd3+; 23 .Ke6 Qe4+; 24.Kf7 Qf5+; 25.Kg7 and
the king is safe, because any check on the g-file allows . . . Qg6.
16.Kc5 ! is the simplest path to the draw, keeping the king in a
position to guard the e-pawn. 16 . . . Kg3 ; 1 7.Rh8 £3; 1 8.e6 Re i ; 19.Kd6
f2; 20.Rf8 flQ; 2 1 .Rxfl Rxfl ; 22.e7 Re l ; 23.Kd7.
a b c d e f g h
[394] Black to move
Draw.
ROOK + 2 US. ROOK

Except when the pawns include a rim pawn and a pawn two files
away, White will usually win with rook and two pawns against a lone
enemy rook, though there are some exceptions. The first two examples
are devoted to typical drawing motifs . The key exception will be pre
sented first. The third game involves Billy Colias, a talented young player
who died at a young age, displaying excellent technique bringing home
the point. Sometimes, as our last example shows, there are stalemating
ideas if corner squares are involved.
245
CA RDOZA PUB LI$ H I NG • ERIC SCH I LLER
KARPOV VS. YUSUPOV

Linares, 199 1
3
2
a b c d e f g h
[395] White to tnove
Things start off predictably as White prepares to advance the h

pawn.
6 l .Rh8 Rg5; 62.Ke4 Rc5; 63.K.f4 Rc4+; 64.Ke5 Rc5+; 65.Ke6 Rg5;
66.Kf7 Rc5; 67.Rh7.
3
2
a b c d e f g h
[396] Black to move
White places his final hope in a single rook pawn endgame. There
is nothing better, although the endgame is drawn with proper play.
(67 .h6 Rc6; 68.h7 Rc7+; 69.Ke6 Rc6+; 70.Kd5 Rh6; 7l .Kc5 Rh5+; 72.Kc6
Rh6+; 73 .Kc7 Rh l t No progress will be made. Of course Black must
never capture the a-pawn, or Ra8+ will win.)
67 . . . Kxa6; 68.h6 Rc7+. A draw was agreed here, because if the king
comes to the queenside to avoid checks, the pawn cannot advance:
69.Kg6 Rc6+; 70.K.f5 Rc5+; 7 1 .Ke4 Rh5; 72.Rh8 Kb7; 73.h7.
246
ROO K EN DGA M ES
3
2
a b c d e f g h
[397] Black to move
You should make sure you know how to defend correctly here. The
continuation is seen in diagram #384.
The other situation you need to be familiar with is when one side
has a rim pawn and a pawn two files away, that is, either a- and c-pawns
or f- and h-pawns. This position is not always drawn, but there are
situations where no win is possible with correct defense.
It's worth mentioning that this endgame doesn't arise all that of
ten. Both sides often try to avoid it because it can be very difficult to
play. If you are defending the weaker side, you might well head for it if
there is no better drawing chance in sight. Assuming that the stronger
side has h- and f-pawns and is playing White, try to keep your king at g7
and rook at b 1 , using the rook as actively as you can and try to avoid
getting your king stuck on the h-file.
GLIGORIC, VS. SMYSLOV

Moscow, 1947
3
2
a b c d e f g h
[398] White to move
24l
CA RDOZA PU B LISH I NG • ERIC SCH ILLER
This is representative of a good, solid defense. The rook at al is far

from the enemy king so that it can give check safely. The king guards f6
and h6. It is interesting that in his own commentary, Smyslov, who
would become the world' s leading expert on rook endgames, doesn't
even consider an alternative. Ho'v else is White to make progress? Is
there any configuration which will be even more auspicious for the
advance of the pawn?
99.h6+.
3
2
a b c d e f g h
[399] Black to move
As usual, a pawn advance changes the strategy. Now White will not
be able to use the h6-square.
99 . . . Kh7; 1 00.Rd6 Ra2; 1 0 1 .Kg5 Rg2+.
3
2
a b c d e f g h
White's plan is to sacrifice the h-pawn in order to achieve a posi

tion where the f-pawn can advance. The defender must play carefully!
1 02.Kf6 Kxh6; 1 03.Ke7+ Kh7; 104.£5 Re2+; 1 05 .Re6 Ra2. Black
allows the pawn to advance. As the Philidor position taught us, the
pawn on the sixth can often be a liability.
248
ROOK ENDGA M ES
106.£6 Ra8!
3
2
a b c d e f g h
[4011 White to move
This is a move to remember, as careless play is fatal, for example

( 1 06 . . . Ra7+; 1 07.Kf8 Kg6 ( 1 07 . . . Ra8+; 1 08.Re8 Rxe8+; 1 09.Kxe8 Kg6;
1 1 0.£7) 1 08.£7+! Kh7; 1 09.Re 1 Ra8+ ( 1 09 . . . Kg6; 1 1 0.Kg8 ! Rx£7; 1 1 l .Rgl +
Kf6; 1 1 2.Rfl+ wins the rook. ) 1 1 0 . Ke7 Ra7+; 1 1 1 .Kf6 Ra6+; 1 1 2 . Re6
Ra8; 1 1 3.Re8 Ra6+; 1 1 4.Ke5 and Black will soon run out of checks .)
107.Kf7! 1 07.£7 Kg7; 1 08.Re 1 Ra7+; 1 09.Ke6 Rx£7 draws. 107 ... Kh6;
108.Re l Ra7+; 109.Re7 Ra8!
3
2
a b c d e f g h
[4021 White to move
White cannot make progress. The game continued 1 1 0.Rd7 Kh7;

l l l .Rd l Ra7+; 1 1 2 .Ke6 Ra6+; 1 1 3.Rd6 Ra8; 1 14.Rd4 Kg8; 1 15.Rg4+
Kf8 where a draw was agreed.
249
CA RDOZA PU B LIS H I NG • ERIC SCH I LLER
MARCONNET VS. COLIAS

USA, 1989
3
2
a b c d e f g h
[403] Black to move
82 . . . g5 . The presence of the pawn at g5 is crucial to the winning

plan. One square further would be one step too far, and the Black king
would be unable to find shelter from checks. The pawn plays a dual
role as defender of the king, and, at the right moment, a defender for
the rook.
83.Rf8 Kh4; 84.Rh8+ Kg4.
a b c d e f g h
{404] White to move
The king is safe here. Black can of course drive the king back to the
first rank, but there is another squeeze play which is equally effective.
85 .Rf8 Rb l . Although the White king is stalemated, there is no
way to get rid of the rook with a check that does not result in lifting the
stalemate. In this case, 86.Rf4+ Kxf4! lets the White king move to h3,
where it is quickly checkmated.
86.Rh8 Rb2+; 87.Kgl g2; 88.Rg8. 88.Rf8 Kg3 will terminate the
250
ROOK EN DGA M ES
game with . . . Rb 1 , but White can prolong it with 89.Rb8 Rf2 (89 . . . Rxb8??
is stalemate !) 88 . . . Kf3; 89.Kh2. 89 .Rxg5 Rb 1 +; 90 .Kh2 Rh 1 # is a basic
checkmating pattern. 89 . . . Rb4! ; 90.Rf8+. 90.Rxg5 Rh4+; 9 1 .Kgl is a
study-like parallel to the previous note.
90 . .. Rf4!
3
2
a b c d e f g h
[4051 White to move
The threat of . . . Rh4+ forces the exchange.

9 l .Rxf4+ gxf4. 9 l . . . Kxf4; 92.Kxg2 Kg4 takes much longer. 92.Kgl
Ke2. White resigned.
SOMAK VS. MEERA

Goodricke International, Calcutta, 1999
3
2
a b c d e f g h
[4061 White to move
Things appear hopeless for White. T\vo connected passed pawns

are screaming down the flank. Capturing the pawn leads to a lost
endgame as White cannot maintain the opposition. Not willing to con
cede, White reconfigures the position to set up the old stalemate cheapo!
66.Rfl+! Kg4; 67.Kh l .
25 1
CARDOZA PU BLISH I NG • ERIC SCH I LLER
3
2
a b c d e f g h
[407] Black to move
If the Black king winds up at f2 or f3, the position is stalemate.

White can therefore check at either of those squares without fear of
capture .
67 . . . Rh6; 68.Rf7 g5; 69.Rf2 Rh3. 69 . . . Kg2 provides an opportunity
for 70.Rf3+! because 70 Kxf3 is stalemate. 70.Rg2+ Kf4; 7 1 .Rf2+ Ke4.
. . .
If the kings seeks shelter at h5 or h6, White can play the rook to the 8th
rank and threaten Rh8+. If the Black king is at h5 and moves to g4,
then an exchange of rooks also leads to stalemate, 72.Rg2 Rh5; 73.Re2+
Kf4; 74.Rf2+ Kg4; 75.Rg2+ Kh4. Black is tired of the game, and hopes
for a blunder in a drawn king and pawn endgame.
76.Rxh2+ Kg4; 77.Rxh5 Kxh5; 78.Kgl Kh4; 79.Kh2 Kg4; 80.Kg2
Kf4; 8 1 .Kf2 . Drawn.
ROOK + 2 US. ROO K + I

When the opposing pawns are close together, squeezing out a win
is very difficult, often impossible. If there is distance between the pawns,
there is more room to maneuver and the defensive task is more diffi
cult. Connected passed pawns are best of all, of course, because that
often creates winning chances via a rook sacrifice, leading to positions
seen in the Rook vs. 2 section.
252
ROOK EN DGAM ES
OLSEN VS. JAKOBSEN

COPEHAGEN CITY CHAMPIONSHIPS
Denmark, 1999
3
2
a b c d e f g h
[409] Black to move
This is a fairly straightforward \vin, the only thing you need to watch
out for is a potential stalemate.
54 ... Ra2+; 55.Kg l . 55.Kg3 h2; 56.Rh6 Kg7; 57.Rh4 a4 and Black
wins since the rook can't cope with both pawns. 55 ... Ke7; 56.£4 Rc2.
56 . . . a4 is more direct. 57.£5 a4; 58.£6+ Kf7; 59.Ra6 Ra2; 60.Rb6 a3;
6 l .Rb3 Rg2+. 6 l . . . h2+; 62. Kh l Kxf6 falls for the old trick: 63 .Rf3+ Ke5;
64.Re3+ Kd4; 65 .Rd3+ Kc4; 66.Rc3+ Kb4; 67.Rb3+ Ka4; 68.Rb4+ Ka5;
69 .Rb5+ and stalemate is inevitable.
62.Kh l a2; 63.Ra3 Rb2; 64.Ra7+ Kx£6; 65.Ra6+ Ke7. White re
signed.
ROOK + 2 US. ROOK + 2

From this point forward generalizations become difficult. It all
depends on the health of the pawn structure, the activity of the rooks,
and effective piece placement. We'll start with a position that is rela
tively balanced.
253
LOPEZ VS. ALVARADO

3
2
a b c d e f g h
White has a small advantage because of the outside passed pawn

on the b-file. Black has no immediate concerns, as the White king can
not cross the c-file. Black must decide whether it will be wise to ad
vance the e-pawn, bringing it closer to the promotion square but allow
ing the enemy king to come to the aid of the b-pawn.
56.b4 e5+. Although this plan is not without risk, it should draw if
Black follows up correctly. 57.Kd5 Rd7+; 58.Kc5.
3
2
a b c d e f g h
Where should Black move the rook? Passive defense ( c7), active
defense (d l ) or aggressive counterplay (d4)?
58 . . . Rd4? Two alternatives need to be considered.
58 . . . Rd l ? ! ; 59.b5 Rcl +; 60.Kd6 Rd l+; 6 l .Kc6 Rc l +; 62.Kb7 Rc4.
Black prepares to take the g-pawn, after which the rook can be sacri
ficed for the White pawn, preferably on b7 or b8 so that the king is too
254
ROOK ENDG A M ES
far away from the kingside. Unfortunately, this plan is too slo\v. 63.b6
Rxg4; 64.Kc6 Rb4; 65 .b7 Rxb7; 66.Kxb7 e4; 67.Kc6 Kf5; 68 .Kd5 and
White wins.
58 . . . Rc7+! offers the best hope of defense. 59.Kd6 Rc4; 60 .b5 Rxg4;
6 l .Rf8+ Kg6; 62.Kxe5 Rb4; 63.Rb8 g4; 64.b6 g3; 65 .b7 g2; 66.Rg8+ and
both pawns disappear. 59.Kb6 is an option, but careful defense holds:
59 . . . Rc4; 60.b5 Rxg4; 6 1 .Kc5 ( 6 l .Rc8 tries to build a bridge, but the
construction does not meet the deadline. 6 l . . . Rb4; 62.Rc5 e4; 63.Kc6
e3; 64.Rc2 Ke5; 65 .Rg2 Kf4; 66 .b6 g4; 67.b7 Kf3; 68.Rg l e2; 69.Kc7
Kf2 and Black wins. ) 6l . . . Rg2; 62.b6 Rb2; 63.Kc6 Rc2+; 64.Kb7 e4 !
White can't easily advance the pawn. 65. Rf8+ Kg6; 66.Re8 Kf5; 67.Ka6
Ra2+; 68.Kb5 Rb2+; 69 . Kc5 Rc2+ and the king cannot hide.
59.b5 Rxg4; 60.b6 Rgl ; 6 1 .b7.
a b c d e f g h
White has succeeded in marching the b-pawn so quickly that Black

has to give up the rook right away, and worse, the White king is within
striking distance of the e-pawn and is in the square of each pawn.
6 1 ... Rb l ; 62.b8Q Rxb8; 63.Rxb8 g4; 64.Kd5 Kg5; 65.Kxe5 . The
pawns are quickly gobbled up by the White forces . The end is simple
enough to work out, but if you need help, see position #375.
255
CA RDOZA PU BLIS H I NG • ERIC SCH ILLER

a b c d e f g h
Black wins primarily because the White king must race to stop the
c-pawn, and the two remaining pawns fall. It must be played with abso
lute precision because it wins by a single tempo!
8 ... c3; 9.Rf8 Kc5; 10.Ke4 Kc4; l l .Rc8+ Kb3; 12.Rb8+ Ka2; 13 .Rc8.
1 3 . Kd3 c2 ; 1 4 . Rc8 transposes . 13 . . . c2 ; 14.Kd3 Kb l ; 1 5 .Rb8+ Kc l ;
1 6.Rf8. 1 6.Rc8 loses to 1 6 . . . Kd l ; 1 7. Rf8 Rd2+.
1 6 . . . Rd2+; 1 7.Kc3 Kd l ; 18.Rfl+ Ke2 ; 1 9 . Rc l Kxe3 ; 20.Rxc2 .
20.Re l + Re2 ; 2 l .Rc l Kf3 ; 22.Rgl Rg� ; 23.Rc l Kxg3 and Black wins.
20 . . . Rxc2+; 2 l .Kxc2.
3
2
a b c d e f g h
[414] Black to move
Black wins because he is close enough to the g-pawn to grab it and

take up a post at g2 before White can slip his king onto g l .
2 l . . . Kf2; 22.Kd2 Kxg3; 23 .Ke l Kg2.
256
ROOK ENDG A M ES
DE FIRMIAN VS. GALLAGHER

Biel Open, Switzerland, 1993
3
2
a b c d e f g h
[415] Black to move
The obvious question here is whether Black can afford to exchange

rooks . The resulting endgame is simple, but there is a little trick involv
ing the h-pawn that you need to know.
56 . . . Rxb7! ; 57.Rxb7 K.xb7; 58.K.xd5 Kc7; 59.Ke6 Kd8; 60.Kf5.
a b c d e f g h
This is the position you had to see from the first diagram. Black
converts the White g-pawn into an h-pawn, after which the rook pawn
endgame is drawn.
60 . . . h3 ! ; 6 1 .gxh3 Ke7; 62.Kg6 Kf8. Drawn.
ROOK + 3 US. ROO K + 2

Situations with an extra passed pa,vn are fairly common. If each
side has pawns on one flank and a single pawn on the other, the weak
pawn will be an obvious target. Since passivity in a rook endgame can
25l
CA RDOZA PUB LIS H I NG • ERIC SCH I LLER
be fatal, it is sometimes worthwhile to give up the pawn in order to

keep the rook active. You must know how to defend against distant
passed pawns if this strategy is to succeed. Generally, you will keep
your rook behind the passed pawn, as that makes it difficult to pro
mote.
GULKO VS. YEPISHIN

3
2
a b c d e f g h
This is a very practical rook endgame. White faces a difficult de

fense. Black can play . . . Rc l and advance the c-pawn to c2. Then, if the
Black king can come to the queenside, the king can escort the pawn
\Vhile the rook defends the kingside . Life is not so simple, however.
Black's king can't leave the kingside, for the moment. The f-pawn and
h-pawn are weak. Black must consider the possibility that while the c
pawn advances to c2, White may pick off the kingside pawns, perhaps
creating a drawn rook vs. pawns endgame.
In order to save the game, White must play with great precision. I
only have access to the end of this game, with some helpful analysis
from Gulko, so the move numbers start with one.
l .Ke2 ! l .Rc6 lets Black cut off the king with l . . . Rc2 . Then the c
and h-pawns advance, tying down White's pieces. 2.Rc8 c4; 3 . Rc6 h5;
4.Rc8 K£6; 5 .Rf8+ Ke5; 6.Re8+ Kd6; 7. Rf8 Rh2 ! Black gives up a pawn,
and allows his king to be cut off north of the meridian. Yet this is the
only way to win cleanly. (7 . . . Kc5; 8.Rxf5+ Kb4 gives the pawn a royal
escort, but yields a draw to 9 .Rxh5 c3; l O .RhS Rd2; l l .Rc8 c2 ; 1 2 .g4
Kb3 ; 1 3 .g5 Kb2; 14.Kf4 c l Q; 1 5 . Rxc l Kxc l ; 1 6.g6 Rg2 ; 1 7.Kf5 Kd2;
1 8 .e4 Kd3; 1 9.e5 Kd4; 20.Kf6 ! Rf2+; 2 1 .Ke7 Kxe5 ; 22.g7 Rg2; 23.Kf7
with a draw. ) 8.Rxf5 leading to a Rook + 2 vs . Rook + 2 endgame which
cannot be held. See diagram #4 1 3 .
258
ROOK EN DGA M ES
3
2
a b c d e f g h
l . . . h5. l . . . Re4; 2.Rc6 c4; 3 . Kf3 h5; 4.Kf2 leads to a draw. 4 . . . h4;
5.gxh4+ Kxh4; 6.Ke2 Kg4; 7.Rg6+ Kh4; 8.Kd2 Rg4; 9.Rh6+ Kg5 (9 . . . Kg3;
1 0.Rf6 Rg5; l l .Rc6 Rg4; 1 2.Rf6 with a draw by repetition soon to come.)
I O.Re6 and now 1 0 ... Re4?; l l .Rxe4 fxe4 leads to a King +2 vs. King + 1
endgame that cannot be won. See position #27 for further analysis.
2 .Rc6 Rc2+; 3 .Kd3 . The king stays close to the passed c-pawn.
3 ... Rg2; 4.Rxc5 Kg4.
3
2
a b c d e f g h
[419] Black to move
Black must not be tempted by 4 . . . Rxg3, because this gives White an

escape with 5.Ke2 h4; 6.Kf2 Kg4; 7.Rc4+ Kh3; 8.Rf4 Rg5; 9.e4 fxe4;
1 0 . Rxe4.
259
CA RDOZA PUBLISH I NG • ERIC SCH ILLER
3
2
a b c d e f g h
Black cannot make progress in this elementary drawn position.

5.Rc8. White could also move the rook to a5 , but in the end the
same position will be reached.
5 . . . Kxg3 ; 6.Rg8+ Kh3. 6 . . . Kf3; 7.Rh8 Rg5; 8.Kd2 Kg3; 9.Ke2 h4;
l O.Kfl h3; l l . Kgl K£3+; 1 2 .Kh2 Kxe3; 1 3 .Ra8 ( 1 3 .Rxh3+ Kf2 is more
problematic because the rook cannot leave the h-file without allowing
checkmate or an unfavorable exchange of rooks) is one of the h- + f
pawn endgames that can be held.
3
2
a b c d e f g h
[421 ] Black to 1nove
The position is a draw, provided that White continues to play accu

rately. 1 3 . . . Rh5 ; 1 4 . Ra3+ Kf2 ; 1 5 . Ra2+ Kf3 ; 1 6 . Ra3+ Kf4; 1 7. Ra4+
( 1 7.Rxh3?? Rxh3+; 1 8 .Kxh3 Kf3 wins ) 1 7 . . . Kg5 ; 1 8. Ra3 Kg4; 1 9.Rg3+
Kf4; 20.Ra3 and no progress is being made.
7.Rf8 Rf2 ; 8.Rh8 h4.
260
ROOK ENDGA M ES
3
2
a b c d e f g h
White must now play decisively.

9.e4! Rf3+. 9 . . . fxe4+; I O .Kxe4 Kg3; l l .Rg8+ Kh2; 1 2 .Rh8 h3; 1 3.Ke3
Rf7; 1 4. Ke2 Kg2 ; 1 5 . Rg8+ Kh l ; 1 6 . Rh8 h2 reaches an important
endgame, analyzed in Nickoloff-Kovacevic in diagram #387. I O.Kd4
Rf4. 1 0 . . . f4; l l .e5 Re3; 1 2 .Rf8 Kg3 ; 1 3. Rg8+ Kf3; 1 4 . Rh8 forces Black
to choose between a draw by repetition and 1 4 . . . Re l ; 1 5 . Rxh4 Rd l+
leads to a critical position in a Rook + 1 vs. Rook + 1 endgame, ana
lyzed in diagram #393 .
l l .Ke5 Rxe4+; 12.Kxf5 Rc4 reaches diagram #390 .
a b c d e f g h
26 1
ANDERSSON VS. HUBNER

Kloosters, Ter Apel, 1997
3
2
a b c d e f g h
[424] Black to move
White's plan is simple . The king will come to f4 eventually, and the
e-pawn will fall. First, White needs to maneuver the rook to the 5th
rank, so that it can come to f5 , and protect the pawn at f2 .
5 4 . . . Rb4; 55.Re5 Kg6; 56.Rf5 Rb8; 57.Kf4 ReS; 58.Re5 Rxe5. Tan
tamount to resignation, but Black has no chances for survival. 59.Kxe5
Kg5 ; 60.Kxe4 Kxg4; 6 l .Kd3 . Black resigned.

Most of the endgames with three kingside pawns each will be drawn,
unless there are serious structural ""eaknesses. It is a fairly common
situation, since in most openings both sides castle on the kingside and
are reluctant to part with any piece of the pa,vn barrier that protects
the king. In rook endings, as in most others, a particularly favorable
king position gives rise to winning chances.
262
ROOK EN DGA M ES
SERPER VS. SCHWARTZMAN

USA, 1996
3
2
a b c d e f g h
[425] Black to move
Two Grandmasters battle in a rook endgame with no real pawn

weaknesses. Surely the game will end in a draw? Not if there are tacti
cal mines on the battlefield!
I . . . Kf6? l . . . Ke5 ! draws. 2.Re3+ (2.Kg7 f4; 3.Rxf4 Rxf4; 4.gxf4+ Kxf4;
5.Kxg6 Kg4; 6.f4 Kxf4; 7.Kxh5 Kf5 with the famous rook pawn draw. )
2 . . . Kf6! Black doesn't want to get cut off on the queenside. 3 .f4 ( 3 . Ra3
f4! ; 4.Rf3 Kf5 transposes to the 2 . Kg7 line.) 3 Kf7; 4.Kh7 Kf8! and the
. . .
king just commutes between f8 and f7. Note that 4 . . . Kf6? would be a
costly mistake because after 5 . Kg8 we reach the horrible conclusion of
the actual game, picking it up at move 4.
2 .Rf4! K£7 . 2 ... Rxf4 loses by force: 3 .gxf4 K£7; 4.Kh7 Kf6; 5 . Kg8
Ke6; 6.Kg7 Kd5; 7.Kxg6 Ke4; 8.Kxh5 Kxf4; 9 . Kg6 and a new queen is
four steps away. 3.Kh7 Kf6; 4.Kg8!
3
2
a b c d e f g h
[426] Black to move
263
Black is in zugzwang.
4 . . . g5 . 4 . . . Ke5; 5. Rxg4 fxg4; 6 . Kf7 Kf5; 7.Kg7 g5; 8 . Kh6 ! gxh4;
9.gxh4 g3 ; 1 0.fxg3 Kg4; l l .Kg6 Kxg3 ; 1 2. Kxh5 Kf4; 1 3. Kg6 etc.
5.Rxg4 hxg4. 5 ... fxg4; 6.Kh7 Kf5 ; 7.Kh6 gxh4; 8.gxh4 wins as in
the previous note . 6.h5.
3
2
a b c d e f g h
Black can't prevent a ne'v queen, and gamely tries to get to the
kingside.
6 . . . Ke5 ; 7.h6 Ke4; 8.h7 Kf3 ; 9.h8Q Kx£2 ; IO.Qe5. Black resigned,
rather than suffer through 1 0 . . . Kg2 ; l l .K£7 f4; 1 2 .gxf4 gxf4; 1 3 .Qxf4
g3; 1 4. Kg6 Kh3; 1 5 .Kh5 g2; 1 6. Qh4#.
NOGUEIRAS VS. HJARTARSON

Biel Intezonal, Switzerland, 1993
3
2
a b c d e f g h
White must keep rooks on here. He went down in flames after

39.Rf2? Rxf2 ! ; 40.Kxf2 leading to the pa,vn endgame shown in dia
gram #36 .
264
ROOK EN DGA M ES
Here is how he should have played: . 39.Rf5 ! ? Rxb2; 40 .Rc5 ! (40 .g4
b4; 4 l .cxb4 c3; 42.Rc5 c2 gives Black the win. ) 40 . . . Kg6; 4 l . Kfl Rb3;
42.Kf2 with two branches .
4 2 . . . Rxc3; 43.Rxb5 only manages to draw.
a b c d e f g h
43 . . . Rd3; 44.Rc5 c3 (44 . . . Rd4; 45.Ke3 Rg4; 46.Kf3 h5 is a draw be

cause the pawn at h5 is too vulnerable) 45 .g4 Rh3; 46.Ke2 h6; 47.Kdl
Rg3; 48.Kc2 Rxg4; 49.Kxc3 h5 ; 50.Kd3 h4; 5 l .Ke3 h3; 52.Rc2 Kg5;
53.Kf3 draws.
42 . . . Kf6; 43.g4 Ke6 does force White to play with extreme care.
a b c d e f g h
44.g5?! Kd6; 45 .Rf5 Kc6; 46.Rf7 Rxc3 ; 47.Rxh7 Rc2+; 48.Kf3 Rc l

followed by . . . Rgl , with a win, or 49. Kf2 c3; 50 .g6 c2 with the threat of
. . . Rfl + and promotion with check. Safer is 44.Ke2 Kd6; 45 .Rh5 Rxc3
36.Rxb5! with play similar to that of the 42 . . . Rxc3 line .
265

The normal method of winning a 7-pawn endgame is to reduce it
to an endgame with fewer pawns, selecting those with the most prom
ising chances . If the Black pawns are healthy, especially if they are at f7-
g6-h5, the win is remote, if it exists at all. With one weak pawn, how
ever, everything changes .
CONTINUATION OF ANDERSSON VS. HUBNER

a b c d e f g h
[431] White to move
Amazingly, Black is still tied down. The rook must stay on the fourth
rank to protect the pawn, and the king must guard h6.
5 1 .Kg3 Rb4. Black will simply slide the rook back and forth, so
White must take action on the kingside. 52.h4! This forces an exchange
on the h-file which brings about a winning Rook + 3 vs. Rook + 2
endgame. 52 ... gxh4+; 53.Kxh4 Ra4; 54.Kg3 ! See position #424 for the
rest if you can't figure it out yourself.

Doubled pawns are always a liability, but backward pawns can be
much worse. In the example, Black has a real problem at h7, which will
require passive defense if it is to be held. Because White has a doubled
pawn, Black can afford to give it up, if greater mobility will result.
266
ROOK ENDG A M ES
PANCHENKO VS. BRASLAVSKY

Yalta, 1996
8
7
3
2
a b c d e f g h
[432] Black to move
The game continued 32 . . . Rf7. 32 . . . Rf5 is a much more active plan .

33.Rh2 Rxg5 (33 . . . Rf7 returns to the game with a meaningless extra
tempo for White . ) 34.Rxh7 Rg4 ! ; 35 .b3 (35 . Kc3 b5; 36.c5 Rg3+; 37.Kc2
Rg2+; 38.Kb3 Kd8! ) 35 . . . b5; 36.Rc7 bxc4; 37.Rxc6 Rg2+; 38.Kc3 cxb3;
39.Kxb3 Kd7; 40.b5 Rg4 (or 40 . . . g5) .
33.Kb3 Rd7. A poor station for the rook. After all, White i s not
about to exchange. The d-file lacks access to the 5th rank and a check
on d3 can be countered by attacking the rook. All in all a very poor
choice. 33 . . Rf5 is similar to the previous note but 33 . . . Re7 was a lesser
.
evil.
34.Rh2. The rook at d7 is paralyzed. 34 . . . Kf8. The king goes to
protect the weakling, but his majesty is really needed on the queenside.
3
2
a b c d e f g h
267
CA R DOZA PUB LIS H I NG • E RIC SCH ILLER
35.c5 ! Black gets access to d5, but the pawn is defended and there
is a clear path for the White king. 35 ... Kg8; 36.Ka4! Rd5; 37.Rg2 h5! ?
Black takes advantage o f the opportunity to get this pawn to a safe
square. It is now a protected passed pawn. But there is a price: the g
pawn is weak. This seems irrelevant as long as the Black rook is at d5,
but can it stay there? What about the weakling at e7?
38.Ka5 Rd4. Black threatens . . . Rg4. 38 . . . Rd7; 3 9 . Kb6 Kg7; 40.Re2 !
Black cannot make progress. 39.Re2 K£7; 40.b5! cxb5; 4 l .Kxb5. Be
cause Black's king is cut off, the White king and pawns can force the
win of the enemy rook. 4 l . . . Rd5; 42.Rg2 Ke7; 43.Kb6 Kd8; 44.b4 Kc8.
The king finally defends the queenside, but the pawn at g6 is helpless.
a b c d e f g h
[4341 White to move
45.b5. 45. Rf2 ! is the most direct win. Black can't afford to take the
pawn. 45 . . . Rxg5; 46.Rf8+ Kd7; 47.Kxb7 Rg4; 48.c6+ Ke7; 49.c7 Rxb4+;
50.Ka6 Rc4; 5 l .c8Q Rxc8; 52.Rxc8 wins, despite the long journey that
awaits the White king.
3
2
a b c d e f g h
[4351 Black to tnove
268
ROOK ENDG A M ES
You can find the analysis in position #380.

45 ... Rd8. 45 ... Kd8; 46.Ra2 (46.Kxb7 Rxc5 ; 47.b6 Rc4; 48.Rd2+ Ke7;
49.Kb8 Ke6; 50.Kb7 Kf5 ; 5 l .Ra2 ! Kxg5; 5 2 . Ka7 h4; 53.b7 Rb4; 54.b8Q
Rxb8; 55 .Kxb8 h3; 56.Kc7 Kh4; 57.Kd6 Kg3; 58.Ra3+ Kg2; 59.Ra2+
Kgl ; 60.Ra l + Kg2=) 46 . . . Rd7; 47.Ra7 h4; 48.Rxb7 Rxb7+; 49.Kxb7 h3;
50.c6 h2; 5 l .c7+ Ke7; 52.c8Q h l Q+; 53.Qc6 looks very bad for Black.
46.Ra2 Kb8; 4 7 .Rf2 Kc8; 48.Rf6! The weakness at g6 proves fatal.
48 . . . Rg8; 49.Rf7. White wins .

Most of us have encountered endgames with an extra passed pawn
on one side (usually the queenside) and even pawns on the other flank.
Whether or not the position can be won often depends on king posi
tion and the existence of weak pawns in the defender's camp.
ANDERSSON VS. HUBNER

a b c d e f g h
[436] White to move
In this position White has additional advantages . The White king

has plenty of breathing room on the kingside. The rook is in ideal
position. Black's king has only one move . Nevertheless, the win is not
simple. White will advance the pawn to a7 with the rook at a8 .
39.a4 Ra l . The rook always belongs behind this pawn. 40.a5 Ra2;
4l.a6 Kg6; 42.Ra8 Ra l ; 43 .a7.
269
3
2
a b c d e f g h
[437} Black to move
Already we have an important position. Black needs to keep the

rook at a l to contain the White king, which might otherwise find a way
to the queenside to support the pawn. Black's king must move from g6,
because of the threat of Rb8+.
43 . . . Kg7 . 43 . . . Kf7 ? ? loses to the classical trick 44. Rh8 ! Rxa7;
45.Rxh7+. 44.Kg3 Rgl+; 45 .Kh2 Ra l ; 46.Kg2 h6. The king has no
moves, so Black liberates h7 for future use. Then Black plans to simply
shuttle the king between h7 and g7. 47.Kh2 Kh7; 48.Rf8! The only way
to make progress is to jettison the a-pawn in return for the f-pawn.
A ndersson correctly evaluates the subsequent position as a win .
48 . . . Rxa7; 49.Rxf6 Kg7; 50.Re6 Ra4. We now have an endgame with
all pawns on one side of the board. The analysis continues in diagram
#43 1 .
MALADA VS. GARA

First Saturday March International, Budapest, 1999
3
2
a b c d e f g h
[438} White to move
270
ROOK EN DGA M ES
White to move and lose? draw? win? It all depends on your choice.
42.c5? 42.g4! is the obv ious move, and it is best move. White takes
control of f5 and h5 . Black will soon run out of moves . 42 . . . f6; 43 .h4
Kc6; 44. Rh3 and the h-pawn has a clear path. 42 ... Ra4+; 43.Ke5?? Walk
ing into mate is the most embarrassing thing that can happen to a
Master! Black does not fail to seize her opportunity. [43.Ke3 Kc6; 44 .h4
(44.g4? g5 ! ; 45 .fxg5 Rxg4; 44. Kf3 Rd4; 45.g4 Rd2 ; 46.h4 Rd5; 47.Ke4
a5 gives White a few chances, if he can ever figure out how to get his
rook to the h-file without dropping both queenside pawns . Black can
defend and earn a draw. ) 44 . . . e5 ! ; 45.g3 exf4+; 46.gxf4 a5 is equal.]
43 ... Ke7!
3
2
a b c d e f g h
[439] White to move
White resigned, because . . . f6 is mate!
RIBEIRO VS. REIS

Portuguese Masters, Caldas de Felgueira, 1999
a b c d e f g h
[440] Black to move
27 1
This endgame should be drawn, but it illustrates the dangers that

lay hidden in rook endgames. All Black has to do is follow the tradi
tional advice of centralizing the king. However, he stubbornly resists
doing so, and further complicates matters by advancing his d-pawn to a
position \Vhere it cannot be defended properly. The result is a disaster.
36 . . . d2? I ' m almost inclined to award this a double question mark
if only because the move is so pointless. It could be played at any time,
and there certainly is no rush. The entire plan of going after the pawn
at b3 is misguided. If the pawn remained in place at d3, then the b2-
pawn could be vulnerable. 36 . . . Kg7 ! ; 37.Kg2 Kf6; 38.Kf2 ReS ; 39.Ke3
Rc2; 40.Rd l Rxb2; 4 l .Rxd3 Rxh2; 42.Rd5 a6; 43. Rd6+ Ke5; 44.Rxa6
Rb2 should be drawn.
37.Rdl Rd3; 38.Kg2 b4; 39.Kf2 a5; 40.Ke2 Rxb3; 41 .Rxd2 a4; 42.£4?
This is a terrible move. The White king needs to find shelter on the
kingside, and this advance just gives Black more control of the third
rank. 42 .g4 a3; 43.bxa3 bxa3; 44.Rd8+ Kg7; 45.Ra8 is the correct plan,
with two important branches.
45 . . . Rb2+; 46.Ke3 Rxh2; 47.Rxa3 h5 ; 48.Ra7+ Kf6; 49.Rh7.
a b c d e f g h
49 . . . h4 (49 . . . Kg6; 50.Rxh5 Rxh5; 5 l .gxh5+ Kxh5; 52 .Kf4 is a win

ning endgame. ) 50.f4 h3; 5 l .g5+ K£5; 52.Kf3 Rh l ; 53.Rf7+ Kg6; 54.Rf6+
Kh7; 55.Rh6+ Kg7; 56.Kg3 Rgl+; 57.Kxh3 Rhl+; 58.Kg4 Rxh6; 59.gxh6+
Kxh6; 60.Kf5 is also a win.
45 . . . Kf6 ; 46 . Kf2 Rb2+; 4 7 . Kg3 a2 ; 48.f4 provides real winning
chances because if the Black king races to the queenside, White will
eventually sacrifice the rook for the a-pawn and use the kingside ad
vantage to win. For example, 48 . . . Ke6; 49.£5+ Ke5; 50.Ra4 Rd2; (50 . . . Kd5;
5 l .f6! Ke6; 52.Ra6+ K£7; 53 .h4) 5 l .h3 Rd3+; 52.Kh4 Rd2 ; 53 .Kg5 Rf2;
54.h4 Kd5; 55.h5 Kc5; 56.Kh6 Kb5; 57.Ra3 Kb4; 58.Rxa2 Rxa2; 59.Kxh7
272
ROOK ENDGA M ES
Rg2; 60.f6 Rxg4; 6 1 .£7 Rf4; 62 .Kg6 and the pawns cannot be stopped.
42 . . . h5? Black missed 42 . . . a3 ! ; 43.bxa3 bxa3; 44.Rd8+ Kg7; 45.Ra8
Kg6; 46.Ra6+ K£5 , which should draw. 43.Kf2 Kg7. Finally the king
gets moving, but it is far away.
3
2
a b c d e f g h
[442] White to move
44.Kg2! The king needs to get to h3 . 44 ... h4; 45.Kh3 ! hxg3; 46.hxg3
Kf6. It is too late for 46 . . . a3 ; 47.bxa3 bxa3; 48.Kg4 Rb2 ; 49.Rd7+ Kf6 ;
50.Rd6+ Ke7; 5 l .Ra6 a2 ; 5 2 . Kg5 and Black cannot restrain the White
pawns.
47.Kh4 Kf5 ; 48.Re2 Rxb2 . Desperation, but it does not work.
49.Rxb2 b3; 50.g4+. Black resigned. By the time his king reaches the
queenside White will have a new queen.

With five pawns on each side there are many different pawn con
figurations that will direct the flow of the game. A central theme, how
ever, is the activity of the rooks . Before looking at single rooks in ac
tion, we will start with a position where each side has two rooks . Black's
powerful coordination more than compensates for the weak pawns. As
you play through the game, consider the use of each rook separately as
well as part of the team.
In the single-rook endgames, the rook does not have a powerful
partner. The king, however, can often act in support of the rook.
Whether there is one rook or two, the need to control the initiative is
crucial in endgames with so many pawns . It is always useful to control
the pace of the game, but when there are so many pawns, there are
also many targets. Keep on the attack and your opponent cannot main
tain his balance. Even where defense is possible, it is easy to make a
mistake.
273
CA RDOZA PUB LIS H I NG • ERIC SCH ILLER
RUBLEVSKY VS. GELFAND

Rubinstein Memorial, Polanica Zdroj, 1998
3
2
a b c d e f g h
[443] Black to move
Black has an obvious advantage thanks to the doubled rooks and

the weak White pawns at b3 and f3. On the other hand, his second
rank needs protection and the pawn at e6 is isolated and weak. Racing
toward time control, Black maintains the initiative.
37 . . . R8d5. The threat of mate at hS forces White to open up the
position. 38.g4. 38.Rh l is too passive and the single-rook endgame
after Black grabs the pawn cannot be held. 38 . . . Rh5+; 39.Kg4 Rhxh2 ;
40.Rxh2 Rxh2; 4 1 .Kf4 (4 1 .Rc7?? e5 ! and White has to give up the rook
to prevent mate.) 41 . . . h5; 42.g4 e5+; 43 .Ke4 hxg4; 44.fxg4 Rh4; 45.Rf3+
Ke6; 46.Rg3 g5 and White is too tied down to offer serious resistance.
38 . . . Rb2. The ne"\v threat is Rdd2, doubling on the seventh.
a b c d e f g h
[444] White to ntove
39.Re3? 39.Kg3 demands more skill from Black. 39 . . . Rdd2 ; 40.h4

Rg2+ ( 40 . . . e5; 4 1 . Rg 1 is annoyingly solid. ) 4 1 .Kf4 Rbe2 threatens mate
2 74
ROOK EN DG A M ES
with . . . e5 . Gelfand and Huzman offer 42 .g5+ (42.Re3 e5+; 43.Ke4 Rd2
Another mate threat! 44.Ree l Rd4+; 45.Ke3 h5 when; 46 .g5+ Kf5 ;
47.Rh l is checkmated by 47 . . . Rgd2 and Rdd3 . ) 42 . . . Ke7; 43.Re3 Rxe3;
44.Kxe3 Rb2; 45.Ra3 Rh2 winning a pawn.
39 . Rdd2; 40.Rae l . White counts on activity to compensate for
. .
the pawn. The precarious position of the White king insures that White
cannot hold the initiative. 40.Rh l g5 is winning, according to the ana
lysts, but the win requires some technique . 4 l .Kg3 Rg2+; 42 .Kh3 h6 1
White is at a loss for a good move. It is not exactly zugzwang, but very
uncomfortable as White must shuffle the rook along the third rank.
43 .Rc3 e5 .
3
2
a b c d e f g h
[445] White to move
Black's plan is to set up a pin on the third rank and exploit it by

playing . . . e4. Hard to see in this position, but that's what makes the
endgame so difficult! 44.Rc5 Rge2; 45.Rxa5 Re3; 46.Kg3 Rbxb3; 47.Rfl
e4! ; 48.Rf5+ Ke6; 49.Kg2 Rxf3; 50.Rlxf3 Rxf3 ; 5 l .Rxf3 exf3+; 52 .Kxf3
Kd5 and Black wins l
3
2
a b c d e f g h
[446] Black to move
275
The constant pressure reaps its reward. Black picks off the pawn
with check and the e-pawn remains intact.
40 . . . Rxh2+; 41 .Kg3 Rbg2+; 42.Kf4 g5+; 43.Ke4. Suddenly the rooks
look silly, blocked by the useless king. Gelfand insures that the king
will not leave e4. 43 . . . Rd2 !; 44.Rd3. There is no sensible alternative to
allow a pair of rooks to be exchanged.
44 . . . Rxd3; 45.Kxd3 h5; 46.gxh5 Rxh5. This single-rook endgame
is quite trivial. 47.Kc4 Rh4+; 48.Kc5 Rb4; 49.Re3 Kf5; 50.Kd6 Kf4;
5 1 .Rc3 e5; 52.Kd5 e4; 53.fxe4 Rxe4; 54.Rc7 Rb4. White resigned.
KOLEV VS. MARINKOVIC

Vrnjacka Banja, Yugoslavia, 1990
a b c d e f g h
[447] White to move
This pawn structure developed out of a Benko Gambit. Often Black

has the advantage because the passed c-pawn is protected while the
White a-pawn can be picked off by a coordinated action of king and
rook. Black would like to have the king closer to the a-file, but this,
would allow White to circle around and operate against the kingside
pawns . The h-file is open, but if the h-pawns were off the board the
back rank could be used to get the rook to f8, g8 and h8 . White's pawns
on the 4th rank are weak, but defensible by the king. Black's rook must
retreat to keep pace with the a-pawn. If Black had time to get the king
to the queenside , Black would win. As things stand, White will get the
pawn to the 7th. Can it be protected from attack by the Black king? Yes .
But after i t gets to the 7th, what then?
38.a5 Rb8. 38 . . . Rg3 can be played because of the threat of . . . Rg2+.
39.Ra l Rg2+ (39 . . . Rxg4; 40.a6 and the rook cannot get back in time.)
40.Kf3 Rh2; 4 l .a6 Rh8; 42.a7 Ra8; 43.g5 K£8; 44. Ke3 Ke8; 45.Kd3 Kd7;
46. Kc4 Kc7; 47.Kb5 Kb7.
276
ROOK EN DGAM ES
a b c d e f g h
[448] White to move
The king is at b7, the a-pawn falls, but Black, despite the extra
pawn, is lost. 48.e5 ! Rxa7 (48 . . . c4; 49 .Kxc4 Kb6 ; 50.e6 fxe6 ; 5 l .dxe6
Kc7; 52.Kd5 Kb7; 53 .f5! gxf5; 54.g6 f4; 55.g7 f3; 56.Rb l + Kc7; 57.Rb8
£2; 58 .g8Q fl Q; 59.Qd8#) 49.Rxa7+ Kxa7 .
3
2
a b c d e f g h
{449] White to move
At first glance, or during a long calculation, White might reject this

position as bad. There are no passed pawns and no potential passed
pawns . . . yet! 50.e6 ! fxe6 ; 5 l .dxe6 Kb7; 52.f5. White wins.
39.Kd3 Kf8; 40.a6 Ke8; 4 l .a7 Ra8.
277
CARDOZA PU BLIS H I NG • ERIC SCH ILLER
a b c d e f g h
[450] White to move
The rook is trapped in the corner. Black's king can get to b7, but
the position is still lost. This winning method deserves study, because a
variation like this can easily be improperly rejected because of the cal
culation that Black can win the pawn. White wins by setting up a win
ning pawn endgame.
42.Kc4 Kd7; 43.Kb5 Kc7; 44.Ka6 e6; 45.Rb2.
a b c d e f g h
[451] Black to move
Black resigned. The agony can be prolonged, but death is a fore

gone conclusion. For example 45 ... exd5; 46.Rb7+ Kc8; 47 .exd5 £5; 48.g5
c4; 49.Rg7 c3. (49 . . Rxa7+; 50.Kxa7 c3; 5 l .Rh7 c2; 52.Rh l gets back in
.
time.)
50.Rg8+ Kc7; 5 1 .Rxa8 c2; 52 .Rc8+ Kxc8; 53.a8Q+ Kd7; 54.Qc6+
etc.
278
ROOK EN DG A M ES
ZVEREVA VS. RAKITSKAYA

Sankt Petersburg Women's Championship, Russia, 1999
a b c d e f g h
[452] Black to move
Black has to worry about the weakness of the pawns at b4, f6 and
h5 . He decides to eliminate a pair of f-pawns.
4 7 . . . Rf3 . 4 7 . . . Kg7 is the more conservative option, but it too has
some problems. 48.Rb6 Rd3; 49.Rb7! Kg8; 50 .Rc7! Kg7; 5 l .Rc6 e4;
52.Rc4 e3; 53.Re4 (53.Rxb4 e2; 54.Re4 Rd4 ! ! ; 55.Rxd4 e l Q; 56.Rf4
with a fortress draw? No! 56 . . . Qa5; 57.Kxh5 Qxa2 ; 58.b4 Qe2+; 59.Rg4+
Kh7; 60.Kh4 Kh6 ! ) 53 . . . Kh6; 54.Re7 Rd4+; 55.g4 hxg4; 56.hxg4 Rd2;
57.Rxe3 Rxa2; 58.Re7 Rh2+; 59.Kg3 Rb2; 60.Rxf7 Rxb3+; 6 l .Kf4.
3
2
a b c d e f g h
[453] Black to move
White wins the f-pawn and Black's queenside play is too slow:
6l . . . Rb l ; 62 .Rxf6+ Kg7; 63 .Rb6 ! b3; 64. Kg5 Kf7; 65.Rb7+ Ke8; 66. Kf6
Kd8; 67.g5 b2 68 .g6 etc.
48.Rxf6 Kg7; 49.Rb6 Rxf5; 50.Rxb4.
279
CARDOZA PUBLIS H I NG • E RIC SCH I LLER
3
2
a b c d e f g h
[454] Black to move
Black has freed the game, and is just one pawn down, but the con
nected passed pawns are a real problem.
50 . . . Rf2 ; 5 1 .a4 Ra2; 52.Kxh5. White's patience has been rewarded
with a two pawn advantage. This should be a quick win, but it takes
almost thirty more moves. 52 . . . Kf6; 53.Kg4? White missed 53.Rb6+!
Kf5; 54.g4+ Ke4 (54 . . . Kf4; 55.Rf6+ Ke3; 56.Rxf7) 55 .h4 Kd4; 56.Rf6
Rb2 ; 57.Rxf7 Rxb3; 58 .g5 e4; 59 .g6 Rg3; 60.a5 e3; 6 1 .Re7 Rg2; 62.g7
e2; 63.Kh6! White wins.
53 . .. Ke6; 54.Kf3. Why on earth is White retreating. Just push the
h-pawn, since . . . Rh2 is always met by Rh4. 54 ... Kd5; 55.h4 f5; 56.h5
e4+; 57.Ke3 Rg2; 58.Rb5+ Ke6.
3
2
a b c d e f g h
[455] White to move
White's poor play has led to some technical problems .

59.Rb6+ Ke5; 60.Rg6 Rb2. White is going to have to give up one of
the pawns. Which one? The connected passed pawns on the queenside
are nice, but the pawn at g3 is the only barrier to the Black f-pawn.
280
ROOK EN DGA M ES
White makes a good choice.

6 l .Rb6. 6 l .a5 is also strong. 6l . . . Rg2; 62.Rb5+! Kf6; 63.h6. 63 .Kd4!
is not easy to see, but the point is that on 63 . . . Rxg3; 64. Rb6+ Black
must retreat the king, allowing Ke5 , or else 64 . . . Kg5?? which loses pro-
saically to 65 .Rg6+ Kf4; 66.Rxg3 Kxg3; 67.h6 Kf2; 68.h7 e3 ; 69.h8Q e2;
70.Qh4+ Kfl .
63 ... Rxg3+; 64.Kf2 f4; 65.Rh5!
a b c d e f g h
[456] Black to move
Black must retreat the rook to h8 to defend. Both sides have con
nected passed pawns . White is lucky that hers are faster.
65 . . . Rf3+; 66.Kg2 Rg3+; 67.Kh2 Rg8; 68.h7 Rh8; 69.Kg2 Kg6;
70.Rh l .
3
2
a b c d e f g h
{457] Black to move
Black cannot take the pawn.

70 . . . Kg5. 70 . . . Rxh7; 7 l . Rxh7 Kxh7; 72 .a5 and White gets a new
queen. 7 1 .a5. The threat is simply a6-a7-a8 and after . . . Rxa8, then h8
and White has an extra rook. Black can just arrange a few checks .
28 1
7 1 . . . Kg4; 72.Kf2 e3+; 73.Ke2 Kg3; 74.Rgl+ Kh2; 75.Rg7 Kh3; 76.b4.
Black could resign, but perhaps the game was at the end of a sudden
death control. 76 . . . Kh4; 77.b5 Kh5; 78.b6 Kh6; 79.Rgl f3+; 80.Kxf3.
HRVACIC VS. MALADA

First Saturday March International. Budapest, 1 999
3
2
a b c d e f g h
[458] Black to move
At first glance it seems that the position is likely to be drawn. The

White kingside pawns are frozen, and the Black king controls the me
ridian. A closer examination reminds us that White's pawns are healthier
and that the Black rook is frozen. In fact, Black has no kingside moves.
If . . . g6, then hxg6 creates more weaknesses. Advancing the f-pawn would
be a blunder when there is a check at e5. White controls the seventh
rank and his king can transfer to the kingside. The conclusion is that
White has some advantage and can try to play for a win. Unless Black
sees a concrete threat, he should be content with King moves or check
ing with the rook.
34 . . . a4? " If it ain't broke, don't fix it! " That's good general advice
in the endgame. 34 . . . Rc5+; 35.Kd3 Rd5+; 36.Ke4 must be calculated, of
course. The question is, what can White achieve after 36 . . . Rg5 even
with the king centralized.
282
ROOK EN DG A M ES
a b c d e f g h
Notice that Black has not moved any pawns since the starting posi
tion. Should the White king go back to the queenside, the position
would be similar to the start, and nothing has been achieved. Black will
now play . . . Kb6-b5-c5 etc. The Black rook cannot be removed except by
sacrificing the g-pawn, or bringing the king to g3. The former is use
less, but the latter is reasonable.
First, let's consider 37.Rb7+ Kc5 . 38.b4+ (38.Rc7 gives up the initia
tive. Black doesn't even need it, and can simply throw a check at e5 and
retreat to g5. ) 38 . . . axb4.
a b c d e f g h
[4601 White to move
White must select either the pawn or the rook for the recapture.
We can easily dismiss 39.Rxb4 because 39 . . . f5+! ; 40.Kf4 ( 40.gxf5 Rxh5 )
40 . . . fxg4; 4 1 .fxg4 g6; 42.hxg6 Rxg6 is an easy draw.
Therefore we need to consider 39.axb4+, where a logical continua
tion is 39 . . . Kc4; 40.Kf4 c5.
283
CA R DOZA PU BLISH I NG • ERIC SCH ILLER
3
2
a b c d e f g h
4 1 .bxc5 (4 l .b5 Kb4; 42 .b6 Kb5; 43. Rb8 Rd5; 44.Rb7 c4; 45 .Rxg7
Kxb6; 46.Rg6 c3! ; 47.Rxf6+ Kb5 and Black wins ! ) 4 l . . . Kxc5; 42.Rd7
Kc4; 43. Kg3 f5; 44.Kh4! g6; 45.Rg7 fxg4; 46.fxg4 gxh5 ; 4 7 .Rxg5 hxg5+;
48.Kxh5 ! with a winning pawn endgame.
The alternative is 37. Kf4, where the best reply is 37 . . . a4! , with three
replies, none of which achieve anything. 38. b4 Kc4 and . . . Kb3 is a
threat.
38.Rb7+ Ka6 ! ; 3 9 . Rb4 axb3; 40.Rxb3 is a very passive rook posi
tion.
38.bxa4 Kxa4; 39.Rb7+ Kb5 is about even. Black has better king
position, White has the outside passed pawn.
In any case, the game continued 35.b4 c5, reaching the following
diagram.
a b c d e f g h
[462] White to move
Black must have reasoned that he could force White to either cap
ture at c5 or allow a capture at b4 without allowing the enemy king to
284
ROO K EN DGA M ES
capture and pick off the a-pawn. Wrong!

36.Rb7+ Kc6?! This gives the rook access to the a-file. The correct
plan is 36 . . . Ka6 ! ; 37.Rc7 Kb6.
8
7
3
2
a b c d e f g h
White cannot win by exchanging rooks, for example 38.Rxc5 Rxc5+;

39.bxc5+ Kxc5 ; 40 . Kd3 Kd5; 4 l .f4 (But not 4 l .Ke3 Kc4; 42.Ke4 Kb3;
43. Kd3 Kxa3 ; 44. Kc3 Ka2 ; 45 . Kc2 a3 ; 46.Kc l Kb3; 47.Kb l Kc3 and
wins. ) 4 l . . . Kc5; 42 .Kc3 Kd5; 43 .Kd3 with a draw.
3 7.Ra7! cxb4+; 38.Kxb4 Rd5 . 3 8 . . . f5 ; 3 9 . Ra5 ! 39.Rxg7 Rd4+;
40.Ka5! Black resigned. 40 . . . Rd3; 4 l .Rg6! Rxa3; 42 .Rxf6+ Kc5 ; 43.Rxh6
Rx£3; 44.Kxa4 etc.; 44 . . . Rf4+; 45.Ka5 Rxg4 and the game will be drawn.
46. Rh8 Rgl ; 47.Rc8+ Kd5; 48 .Kb6 Ke5; 49.Rc6 Kf5; 50.h6 Rhl and
. . . Kg5.
LOPEZ VS. ALVARADO

3
2
a b c d e f g h
285
We pick up the game at this point to illustrate a simple technique

which may not be familiar to less experienced players. Black, faced
with the loss of the b-pawn, tries to overextend the White king by at
tacking the h-and e-pawn in turn. The king eventually escapes, but Black
nevertheless can pick off a pawn.
34 . . . Rh l ; 35.Kg2 Re i ; 36.e4 Re2+; 37.Kh3 Re3 . Black gets one of
the pawns and maintains the material balance, but White has a major
advantage in his outside passed pawn. Enough to win? Not without
help, but in the endgame one often gets a little help from the oppo
nent . .
38.Rxb6 Rxf3.
3
2
a b c d e f g h
[465] White to move
White has to worry about the weak e-pawn, and that should pro
vide enough counterplay for Black. Nevertheless, the defense requires
precision and as we see, the pressure can often lead to errors.
39.Kg2 Re3; 40.Rb4 Re2+; 4 1 .Kh3. This much is straightforward.
Black must now activate the king. This can be accomplished by moving
pawns, but Black decides not to alter the pawn structure just yet. Each
pawn move represents a serious commitment, and it is understandable
that Black doesn't want to make a serious move right after time control.
41 . . . Kh7. 41 . . . h5 is a more active move, tying down the kingside
and perhaps . . . h4 will be available, since White would have to capture
with the pawn, as the king must guard h2. In fact, the position after
40 . . h5 is rather awkward for White, who may well stand worse, given
.
the useless position of the rook at b4.

42.g4! White eliminates the possibility of . . . h5, for the moment.
42 . . . Kg6; 43.Kg3 Kf6; 44.h4.
286
ROOK EN DG A M ES
3
2
a b c d e f g h
[4661 White to move
When the pawns are advanced and the opponent has room to
maneuver behind the lines, wirining prospects are few and far between.
In fact, Black has the better chances here, if he plays correctly. About
this time players must start thinking of king and pawn endgames where
Black has to run to the b-file to stop the pawn, giving White an en
trance into the kingside, perhaps. Black should play the king to e5 or
the rook to e 1 , so that it will not be attacked by K£3 .
Instead, he radically changes the nature of the endgame by thrust
ing the g-pawn forward.
44 . . . g5?; 45.hxg5+ hxg5; 46.Kf3. Just like that, Black's winning
chances are gone, and now he is fighting for a draw.
46 . . . Rc2; 4 7 .Ke3 Rc3+; 48.Kd4.
3
2
a b c d e f g h
[4671 Black to move
Black plays a bit passively now, but the position is solid enough to
allow for that. I would have kept the rook on the 6th rather than re
treating.
287
48 . . . Rc7; 49.e5+ Ke7; 50.Rb8 f6. In a position where your oppo

nent has more mobility, you should try to exchange pawns and open
up lines for your own pieces. Therefore this is the correct plan.
5 1 .exf6+ Kxf6; 52.Rf8+ Ke7; 53.Rg8 Kf6; 54.Rf8+ Ke7; 55.Rg8 Kf6.
3
2
a b c d e f g h
[4681 White to move
A draw would be a reasonable conclusion here, but White decides

to see what can be made of the b-pawn. For the rest, see diagram #4 1 0.

With six pawns each, there will not be more than two open files,
unless there are doubled pawns. More often, there will be only a single
open file. Control of that file will certainly be important. The key to
understanding the control of a single open file is the evaluation of
endgames after an exchange of rooks. Once that is established, you can
determine who may occupy the d-file safely.
AKHMADEYEV VS. NENASHEV

Novgorod Open, Russia 1999
3
2
a b c d e f g h
[4691 White to move
288
ROOK ENDG A M ES
The position looks stable and although the Black king might, per
haps, get to b3, it isn't easy to see how a great deal of progress can be
made.
Because there is a pawn at a4, the consequences of allowing the
king to win the pawn at b2 are grave. Black will need one move to
capture the b-pawn, one to capture the a-pawn, one to get out of the
way, and three to promote. That is a total of six moves . White's kingside
pawns, even with the Black pawns remove, need at least five moves just
for the pawns, and Black needs three captures to free the h-pawn. One
other plan suggests itself: White can allow Black to take the b-pawn at
b2 if he gets the pawn at b5 in return. Then the c-pawn needs only five
moves, and the king is not in the way.
After working all that out, if the exchange of rooks is unfavorable
for White, other plans must be considered. The king can stay in the
center, but at some point may have to commit to a flank. If so, which
one? Black has two invasion paths: . . . Kd5-c4-b3 and . . . Kf5-f4-g3 . The
White king can only prevent one of them. Of course there is no need to
rush things. White can also try to stay in the center and delay a deci
sion. But given that Black will control the d-file, the king must act now
if it wants to get to the queenside. 34.Rd2 Rd5 will transpose to posi
tions considered below, so we'll just look at the move to d3 .
34.Kd3 Rd5+; 35.Kc2. 35.Ke3 is the alternative, keeping the king
in the center. 35 . . . £5 changes the position. The e-pawn will require care.
Hard to see here, where it is guarded by both king and rook, but those
pieces have other duties . 36.Re l Rd6; 37.Re2 Kd5; 38.Rd2+ Kc5 . There
are two branches to follow:
39.Re2 Kc4; 40.Kf2 Re6 is one way to go. After 4 l .Rd2 Kb3 ! The
rook is forced to stay on the second rank. White can either move the
king or try to open up the kingside. 42.g4 fxg4; (42 . . . hxg3+) 43 .fxg4
Rf6+; 44.Ke2 (44.Ke3 Rfl ; 45. Ke4 Rb l ; 46.Kf5 Rxb2; 47.Rxb2+ Kxb2 ;
48.Kxg5 Kxa3; 49 .Kxh4 Kb2 ; 50 .g5 a3; 5 l .g6 a2 ; 52.g7 a l Q; 53 .g8Q
Qel + followed by . . . Qxc3 is much better for Black.) 44 . . . e4; 45 .Ke3
Rf3+; 46.Kxe4 Rxh3; 47.Kf5.
289
CA RDOZA PU BLIS H I NG • ERIC SCH I LLER
a b c d e f g h
[470] Black to move
What do you play here as Black? White is about to get a passed g

pawn. Black must keep a rook on the g-file to restrain it, taking just
enough time to play . . . Rb l if necessary to win . the b-pawn. 47 . . . Rh l ;
48.Kxg5 h3; 49 . Kh4 Ka2 ; 50.g5 Kb3 ; 5 l .Rf2 Rgl ; 52 .Rh2 Rg2 ; 53.Rxh3
Kxb2 ; 54.Rg3 Rh2+; 55 . Kg4 Kxa3 ; 56.c4+ Kb4; 57.g6 ! Black cannot stop
the pawn with the rook, but must press on with the promotion. 57 . . . a3;
58.g7 a2 ; 59 .g8Q a l Q; 60.Qf8+. The only check. 60 . . . Kxc4; 6 l . Qf7+
Kc5 ; 6 2 . Qc7+ Kd5 ; (62 . . . Kb4; 6 3 . Qd6+) 6 3 . Qd7+ with a draw. But
40 . . . Rd5 ! is the easier drawing method. 4 1 .Ke l Kb3 ; 42. Kf2 and Black
cannot make progress .
The alternative is 39.Rxd6, with 39 . . . Kxd6; 40 . Kd2 Kd5; 4 l .Kd3
Kc5; 42 . Kd2 Kc4.
3
2
a b c d e f g h
[471 ] White to 1nove
White can save the b-pawn, but only temporarily. 43.Kc2 e4; 44.fxe4
fxe4; 45.Kd2 Kb3 ; 46.Kc l e3 etc.
35 ... £5.
290
ROOK EN DGA M ES
3
2
a b c d e f g h
[472] White to move
With the White king temporarily cut off, Black can play more am
bitiously on the kingside.
36.b3. It is almost zugzwang! 36.Kc l Kd6; 37.Kc2 Kc5; 38.Kc l Kc4;
39.Kc2 e4; 40.fxe4 Re5 ; 4 1 .Kd2 Rxe4; 42.Rxe4+ fxe4 .
3
2
a b c d e f g h
[473] Black to move
This position would have to be evaluated back at move 36, with the
surprising pawn sacrifice not escaping attention, perhaps in serious
time pressure ! No wonder endgames are so difficult. 43.Ke3 Kb3 ;
44.Kxe4 Kxb2; 45. Kd4 Kb3 ; 46.Kc5 Kxc3; 47.Kxb5 Kb3; 48.Kc5 Kxa3 ;
49.Kc4 Kb2 Black wins. 36.Re 1 ! would have been best. Remember, there
is no clear way for Black to penetrate and therefore structural changes
should be deferred unless they clearly improve matters . Make Black
show some sort of plan!
3 6 . . . Rd8 . B lack should have give n s erious co nsideration to
36 . . . axb3+; 37.Kxb3 Rdl , where 38.Kb4 Rbl +; 39.Kc5 Rb3; 40.Rc2 Rxa3 ;
4 1 .Kxb5 e4! ; 42 .fxe4 fxe4 is not easy for White . Black's king can act in
29 1
support of the e-pawn, so that even if the rook has to be sacrificed for
the c-pawn, there is enough power to draw. Or even win!
a b c d e f g h
[474] White to move
A plausible line is 43.c4 Rb3+; 44. Kc6 Ke5; 45 .c5 e3; 46.Kd7 Rd3+;
47.Ke7 Rd2; 48.Rc l e2; 49.Re l Rc2; 50.c6 Kd4; 5 1 . Kd7 Kd3; 52 .c7
Kd2 ; 53.Rxe2+ Kxe2; 54.c8Q Rxc8; 55.Kxc8. Black wins.
37.Kb2 Rg8.
a b c d e f g h
[475] White to move
White is still hanging in, but feels under pressure (perhaps time
pressure?) to act.
38.c4 bxc4; 39.bxa4? The fatal mistake. White is trying to be too
clever, hoping to use the king to get to the c-pawn. But both a-pawns
are now weak, and Black has a clear kingside majority. 39.bxc4 Rb8+;
40.Ka2 Rb3 is ugly, because White cannot afford to go into any king
and pawn endgames as the kingside maj ority is too dangerous. Still,
Black has work to do after 4 1 .Rc2 e4; 42.fxe4 fxe4; 43.c5 Rb8; 44.c6
ReS; 45.c7 Kd7.
292
ROO K ENDGA M ES
3
2
a b c d e f g h
[476] White to move
Black is going to capture at c7 next move. White could spend a lot

of time deciding which of Black's pawns should be captured after 46.Rc4
or 46.Rc5, but there is a simpler and more ele g ant solution. 46.Kb2 ! ! A
precious tempo is g ained. 46 . . . Rxc7 (46 . . . e3; 47.Kc l Rxc7; 48.Rxc7+
Kxc7; 49. Kdl Kd6; 50.Ke2 Kd5; 5 l .Kxe3 Ke5; 52.Kf3 draw. ) 47.Rxc7+
Kxc7; 48.Kc3 Kd6; 49.Kd4 e3; 50.Kxe3 Ke5; 5 l . Kf3 is the same.
39 . . . Ra8!; 40.Rd2. 40 .Kc3 Rxa4; 4 l .Ra2 Kd5 ; 42.Rd2+ Kc5; 43 .Ra2
e4; 44.fxe4 fxe4 and White is lost. 40 . . . Rxa4; 4 1 .Rd8 e4. 4 l . . . c3+; 42.Kb3
Rc4 is clever, but White can defend. 43.Rd l ! c2 ; 44.Rc l Kd5; 45 .Rxc2
Rxc2; 46.Kxc2 Kc4. 42.fxe4 fxe4; 43.Kc3 Rxa3+; 44.Kxc4 Ke5.
3
2
a b c d e f g h
[477] White to move
The White king will be cut off.

45.Re8+ Kf4; 46.Kd4 Rd3+; 47.Kc4 Rd2; 48.Rf8+ Ke3. White re
si gned.
293
CA RDOZA PU BLIS H I N G • ERIC SCH I LLER
ROOK + l US. ROOK + 6

Extra pawn or better pawn structure? Evaluating the balance is tricky,
but healthy pawns often have more lasting value . Or, to put it another
way, last longer! The example here involves rather extreme deformi
ties, just to drive home the point.
KRASILNIKOV VS. POPOV

N ovgorod Open, Russia 1999
3
2
a b c d e f g h
[478] White to move
White has an extra pawn, but it is hard to imagine a worse pawn

structure. The evaluation of the position really depends on the pawn
at c6. If that falls, Black is clearly better. Since the destruction of the
pawn is more or less inevitable, White wisely uses the time to improve
the position of the king and start hacking at the pawn at e5.
23.Ke2 Kd8; 24.Ke3 Ke7; 25.£4 exf4+; 26.Rxf4 Rxc6; 27.Kd3.
a b c d e f g h
[479] Black to move
294
ROOK EN DG A M ES
Phase one \vas simple enough. White is effectively down a pawn,

but the superior king position and the more active rook provides enough
compensation. Should White be playing for a win? Hardly. But strange
things can happen . . .
27 . . . Rd6+; 28.Ke3 Rd5; 29.Rb4 Ra5?! Pointless. 30.a4 Ke6; 3 l .Rc4
Re5+; 32.Kd3 Rd5+; 33.Ke3.
3
2
a b c d e f g h
[480] Black to move
33 ... c5. This restrains the mobility of White' s rook by taking con
trol of b4 and d4, and also creates a barrier that allows the king to go to
a5 if necessary. That has considerable ramifications for pawn endgames.
On the other hand, the pawn at c5 is more vulnerable to the White
king. All this is relevant only if the queenside is the arena. We must not
forget that it is on the kingside that White has the pawn minority, and
the h-pawn is especially weak.
34.Kf3.
3
2
a b c d e f g h
[481] Black to move
295
The best plan, in my view, is - to defend the pawn at c5 with . . . Kd6,

play . . f5 to control e4, and not worry about the White king entering
.
the kingside because . . . Rd2 will always be devastating. Instead, Black

reduces maneuvering room on the kingside .
34 . . .h5?!; 35.Ke3 Re5+; 36.Kd3 Rd5+; 37.Ke3. One can assume that
somebody was in time pressure, and wanted to get to move forty and
reach time control. This is an important factor in many endgames.
37 ... a5. A useful tempo is thrown away for no reason. White can take
advantage of this, which has significance in the endgame. 38.Re4+.
White agrees to an exchange of rooks.
38 . . . Re5; 39.£4 Rxe4+; 40.Kxe4. Look at the following diagram and
make up your own mind as to his wisdom.
3
2
a b c d e f g h
The position has evolved into a King + 5 vs. King + 5 and is contin
ued in position #83 .
296
ROOK A N D M I NOR PI EC E
EN DG A M ES
Combinations of rooks, minor pieces and pawns complicate mat
ters, and if there are enough of them on the board th e position is more
of a middlegame than an endgame. Nevertheless, there are a few fun
damental facts that ease the task of playing these difficult positions.
Bishops are the natural complements of rooks . Rooks are restricted
to travel on files, bishop use diagonals of one color exclusively. A bishop
can usually cover many squares that are inaccessible to a rook. Knights
are of less help, except when combat takes place in a confined area.
Knights prefer to hang out with queens, where they contribute squares
to a piece that already has scope over ranks and diagonals.
When a rook battles against a single minor piece, the advantage is
substantial. One pawn may, in favorable circumstances, provide suffi
cient compensation, but usually two are needed to balance the extra
firepower of the rook, especially if the defender has a knight. Whether
defending with bishop or knight, the safety of the pawns must be your
greatest concern.
ROOK US. KN IGHT + 2

Many students figure that t'vo connected passed pawns and knight
should defeat a bare rook. True, the side with the rook has virtually no
winning chances, but defense is possible if the king has a good defen
sive position and the rook is active.
297
DJURHUUS VS. FYLUNGEN

Norwegian Team Championship, Langesund, 1999
a b c d e f g h
[483] Black to move
57 ... Rb4; 58.Nd5 Rb3; 59.Nf4 Ra3; 60.Nh3+ Kf6; 6 1 .Nf2 Ke5;
62.Ne4 Rb3; 63 .Ng5 Ra3; 64.Nh3 Rb3 and the knight circles forever.
The game continued 65.Nf4 Rc3; 66.Ne2 Ra3; 67.Nf4 Rc3; 68.Ng2
Ra3; 69.Nh4 Kf6; 70.Nf5 Ke5; 7 1 .Nh4 before the players, no doubt
dizzy by now, agreed to call it a day.
ROOK + BISHOP + 3 US. ROOK + BISHOP + 4

Bishops of opposite color have less defensive value in endgames
with rooks because the attacker can use the bishop effectively in attack
ing pawns or squares that enable to rook to operate with maximum
power. Activity of the rooks is a crucial factor. Our example shows how
painful a passive rook can be.
KRASENKOV VS. KOSTEN

Asti, 1996
3
2
a b c d e f g h
[484] White to move
298
ROOK A N D M INOR PI ECE EN DGA M ES
White is trying to draw the position where he is a pawn down, with

the pawn at d6 in peril. He can either try to save it with Bd5+ and f4 to
prevent . . . Ke6 or . . . Ke5 or . . . Be5 . That's the passive plan. Alternatively,
White can ignore the e-pawn, which is doubled, after all, and get the
rooks off the board, with a bishops of opposite color endgame. Many
would choose the former plan, not fully trusting the bishop endgame
to be drawn. Krasenkov has no hesitation to walk right into it.
33 .Ra7! 33.Bd5+ Kf6; 34.f4 Bh6 just drops the f-pa,vn. 33 . . . Ke6;
34.Ra8!
3
2
a b c d e f g h
[4851 Black to move
Timing of exchanges often determines the fate of a game. A good

endgame tecJ:lnician does not hesitate when a drawing opportunity
arises, but students often miss the chance because they have insuffi
cient faith in their evaluation of the subsequent endgame .
34 . . . Rxa8; 35.Bxa8 Kxd6; 36.Bb7.
3
2
a b c d e f g h
[4861 Black to move
299
CA RDOZA PU BLI�H I NG • ERIC SCH ILLER
There is nothing to fear in this endgame. After a few more moves it

was agreed drawn.
36 . . . g5; 37.Bc8 g4; 38.Kg2 Kc7.
3
2
a b c d e f g h
[487] Black to move
38 . . . Kc7; 39. Ba6 Be5; 40.Bd3 f4; 4 l .Bf5 g3; 42.fxg3 fxg3; 43 .Bd3
and White simply moves the Bishop back and forth.
ROOK + KN IGHT + 2 US. ROOK + KNIGHT + 2

With even material it is unlikely that slo'v positional plans will work
in endgames with rooks and knights. When a tactical possibility pre
sents itself, evaluate it carefully and then seize your chance!
MALJUSH VS. TIMOSHENKO

Minsk, 1996
3
2
a b c d e f g h
[488] Black to move
There are no passed pawns, and it seems that Black cannot avoid
an exchange of knights which will lead to a dead drawn rook endgame.
300
ROOK A N D M INOR PIECE ENDG A M ES
There is, however, a tactical resource here.

l ...f4! An amazing turn of events . Black cannot allow the pawn to
advance to f3, which means that capturing the knight is out of the
question. Capturing the pawn at f4 doesn't help, because Black wins
the exchange with . . . Nxf4+. Both of the exchange-down endgames are
lost. So White must give up the knight for the pawn at g4 and try to get
rid of the remaining pawn.
2.Nxg4!?
2.gxf4+ Nxf4+; 3.Kf2 Nxe2; 4.Kxe2 Ra2+; 5.Kfl Rxh2 is a simple
win. 2 .Nxd5 f3+; 3.Kf2 fxe2; 4.Kxe2 Ra2+; 5.Ke3 Rxh2 is more compli
cated.
3
2
a b c d e f g h
[489] White to move
This position cannot be defended against best play. Black wants to

play . . . Rg2, and White can either defend that square or maneuver the
knight to support the pawn. 6.Nc3 (6.Nf4 Kf5; 7.Kd3 R£2; 8.Ne2 Rf3+
and 9.Kd4 Ra3; 1 0.Nc3 Rxc3[ leads to a winning pa\vn ending for Black.
9.Kd2 Ke4 is pretty ugly, too.) 6 . . . Rg2; 7.Ne2 holds the pawn, but now
the king must stay close to the knight, and the knight cannot move
without dropping th g-pawn. So White can slowly circle around with
the king. 7 . . . Kf6; 8.Kd3 R£2; 9.Ke3 Rf3+ 1 0.Ke4 ( 1 0 .Kd2 Ke5; l l .Ke l
Ke4; 1 2.Kd2 Rf2; 1 3 .Ke l Kf3 with a winning king and pawn endgame
coming soon, for example 1 4 .Nc3 Rc2; 1 5 .Nd5 Kxg3; 1 6.Ne3 Ra2 ;
1 7.Nfl+ Kf3; 1 8.Nd2+ Rxd2 etc.) 1 0 . . . Kg5[ The White king is cut off
and the knight still can't move. l l .Kd4 K£5; 1 2 .Kc4 Ke4; 1 3.Nc3+ Rxc3+!
14.Kxc3 Kf3; 1 5.Kd3 Kxg3; 1 6.Ke2 Kh2 \Vith a new queen.
2...Kxg4; 3.h3+. 3 .Kf2 Nc3; 4.Re8 Nd l+; 5.Ke l Ne3+; 6.Ke2 (6.Kd2)
6 . . . f3+! 7.Kxe3 (7.Kf2 Rfl+; 8.Kxe3 Re i+ is simple, as is 7.Kd2 f2 ! )
7 . . . Re l + and White can give up. 3.gxf4 Nxf4+ and the rook falls.
3 ... Kf5.
30 1
CARDOZA PU BLISH I NG • ERIC SCHI LLER
3
2
a b c d e f g h
[490] White to move
All White needs to do is buy time to play gxf4 and head for a Rook
and Knight vs. Rook endgame. Unfortunately, there is no time!
4.Rf2. 4.Rb2 Ne3+; 5 .Kf3 Ra3; 6.Rb5+ Nd5+! ; 7.Ke2 Ke4! The threat
is Nc3+, and the pawn at g3 is still under fire.
4.Kf2 Nc3; 5.g4+ (5.Rc2 Ndl +; 6.Ke2 fxg3) 5 . . . Kf6; 6.Re8 Nd5;
7.Rd8 Ra2+; 8.Kf3 Ra3+; 9.Kg2 (9.Ke4 Re3+! ! ; 1 0.Kxd5 Rd3+) 9 . . . Rg3+;
1 0 .Kf2 ( 1 0.Kh2 Ke5; l l .Re8+ Kd4; 1 2 . Rd8 Ke4; 1 3.Re8+ K£3; 1 4.Rd8
Ne3; 1 5.Rd3 Ke2; 1 6.Rd4 Nfl +; 1 7.Kh l f3; 1 8.h4 Rh3+; 1 9 .Kgl f2+;
20.Kg2 Rh2#).
4.Rd2 Ne3+; 5.Kf2 Nfl ! ; 6.g4+ Ke4; 7.Re2+ Ne3 and now it is time
to resign, for example 8.Re l (8.Rb2 Rf1 +; 9.Ke2 f3+; 1 0.Kd2 Nc4+)
8 . . . Ra2+; 9.Kgl f3 etc. If instead 7.Rd8 then 7 . . . Nh2 and when White
runs out of checks Black plays . . . Ra2+ and pushes the f-pawn forward.
4. . Ne3+.
.
3
2
a b c d e f g h
3 02
ROOK A N D M INOR PIECE ENDG A M ES
Black still has \Vork to do. Remember, if the pawn at f4 goes, so

does the \vin.
5.Kh2; S.K£3 is a little better, not that it matters . S . . . Nd l ! ; 6.Rfl
(6.g4+ Ke5 ; 7.Re2+ Ne3 and . . . Rfl+ next. ) 6 . . . Ra3+; 7.Ke2 Nc3+; S.Ke l
Ra l +; 9.Kf2 Ne4+; 10.Kg2 Rxfl ; l l .Kxfl Nxg3+. 5 . . . Nfl+; 6.Kg2 Nxg3.
White could resign with a clear conscience, but perhaps this was a time
scramble. (I don't have the game prior to the ending, so don't kno\v
what move number this is .
7.Rb2 Kg5; 8.Rb8 Ra2+; 9.Kf3 Ra3+; 1 0.Kg2 Ne4; l l .Rb5+ K£6;
12.h4 f3+; 13.Kh2 f2; 14.Rb8 Rh3+! Black finishes the game elegantly.
15.Kg2. 1 5 .Kxh3 fl Q+ 15 . . . Rg3+; 16.Kh2 fiN+!
3
2
a b c d e f g h
[492] White to move
Now White resigned, because being mated by two knights, after

. . . Nf2, is just too much humiliation.
ROOK + B ISHOP + 4 US. ROOK + K N I G HT + 4

When bishops face knights, it usually boils down to a question of
whether the bishop is good or bad. A bad bishop is just as big a prob
lem in endgames \Vith rooks as it is on its own.
303
POLGAR VS. LEKO

Tilburg, 1996
3
2
a b c d e f g h
{493] Black to move
Black is obviously better, with a wonderful strong knight vs. bad

bishop. White's d-pawn is blockaded. Black's rook has an active posi
tion. Still, how is Black going to make progress? The key is the weak
ness of the pawn at g3, which at first glance seems to be more than
adequately defended by the pawn at h2 .
42 . . . h4! This powerful stroke exposes the weakness. White is more
or less forced to capture the pawn. 43.gxh4. 43 .Rg2 hxg3; 44.hxg3
(44.Rxg3 Re2; 45 .h3 Ne3 is very strong for Black.) 44 . . . Ke6; 45.Kd2
Rb3; 46.Kc2 R£3 (46 . . . Ne3+ is too clever by half. 47.Kxb3 Nxg2; 48.Kb4
Ne3; 49.Kxb5 Nf5; 50.Kc6 Nxg3; 5 1 .d5+ K£7; 52.d6 and the pawn is
unstoppable. ) 47.Kd2 Kf5; 48.Ke2 Kg4; 49.Bd6 Ne3; 50.Rgl Nf5 is a
triple fork.
43 ... Rh3; 44.Re2 Rxh4.
3
2
a b c d e f g h
{494] White to move
304
ROOK A N D M INOR PIECE ENDGA M ES
Black now has a passed pawn to play with, and the rook sits beauti
fully at h4, where it pressures enemy pawns at h2 and d4, and can drop
back to h8 if necessary.
45.Kd l g5. This pawn is advanced first, so that Black can use f4 as
a base of operations for the knight, without needing the rook as an
escort. 46.Ke l . Polgar's defense is based on the idea of defending the
h-pa,vn with the king, allowing the rook to be used more actively.
46 ... Kg6; 47.Kfl Kf5; 48.Kgl Re4.
3
2
a b c d e f g h
[495] Black to move
Black offers and exchange of rooks which must be declined, since

the bad bishop would be of no use in the endgame.
49.Rc2. 49.Rxe4 Kxe4 dooms the b-pawn. 49 ... Re3 ; 50.b4. Sooner
or later Black had to play this move, so that the rook can be involved in
the game.
3
2
a b c d e f g h
[496] Black to move
50 ... Ke4; 50 . . . Rb3 is another path, but it is not as promising be

cause White can activate the king. S l.R£2+ Ke4; 52.Re2+ Kd3; 53.Re6
305
CARDOZA PU BLISH ING • ERIC SCH ILLER
Nxb4; 54.Rxf6 Nd5; 55.Rf5 Rb l+; 56.Kf2 Rb2+ (56 . . . Rfl+; 57.Kxfl Ne3+;
5 8 . Kf2 Nxf5 ; 5 9 . d5 should end in a draw. ) 5 7 . Kg3 Nc3 ! ; 5 8 . Kg4
(58.Rxg5?? Ne4+; 59.Kf4 Nxg5; 60.Kxg5 Rxh2) 58 . . . Ne4. 5 1 .Ra2 g4;
52.Rb2 Nf4. 52 . . . f5 could be played instead.
53.Rb l Re2. Black's pieces converge on the enemy king. White's
bishop remains a mere spectator. 54.Rfl Nd5 .
3
2
a b c d e f g h
[497] White to move
55.Rf2. White doesn't want to exchange rooks, but there is no choice.

55.Kh l f5; 56.Kgl Ne3; 57.Rcl f4; 58 .Bd6 £3; 59.Bg3 Rg2+; 60.Kh l
Rxg3 ! ; 6 l .hxg3 f2 is a cute win. 62 .d5 Kxd5; 63.Rc5+ Ke4; 64.Rc l fl Q+
etc. 55 ... Rxf2. 55 . . . Re l+ also wins elegantly, for example 56.Kg2 Ne3+;
57.Kg3 Rgl+; 58.Kh4 g3! ; 59.Ra2 (59.hxg3 Rh l+) 59 . . . gxh2 ! ; 60.Rxh2
Nf5+.
56.Kxf2 f5; 57.Ke2 f4; 58.Kf2. The queenside pawns cannot be
defended, and White cannot generate any kingside counterplay.
58 ... Kd3; 59.Bd6 Kxd4; 60.Ke2 Nc3+; 6 1 .Kfl Ke3. White resigned.
ROOK + 2 US. KNIG HT + 2

When there are two pawns each, the advantage of the rook de
pends on the defensive capability of the knight. If the defender can use
the knight in both attack and defense, then there are drawing chances.
306
ROOK A N D M INOR PIECE E N DG A M ES
(32) PSAKHIS VS. SMYSLOV

Rostov on Don, 1993
3
2
a b c d e f g h
[499] White to move
White's king is infiltrating and the rook is active, but Black has a
solid formation. How can White get to the g-pa,vn? If the king gets to
f7, it still needs the help of the rook to get the pawn, and that means
leaving the f-pawn unguarded. The h-pawn may be of some concern,
too, since if attacked, the rook will need to sit passively on the h-file.
60.Rb4. Necessary, if the king is to advance. The f-pawn requires
the attention of at least one of White's pieces. 60 ... Kh7; 6 1 .Ke6.
3
2
a b c d e f g h
[500] White to move
The White king approaches, but the rook must remain on the fourth
rank to protect the pawn.
6 1 ... Kg7; 62.Ra4. 62 .Rb7+ Kh6 forces one of White's pieces to re
treat to defend f4. 62 ... Ng3.
307
CA RDOZA PU B LISH ING • ERIC SCH ILLER
3
2
a b c d e f g h
[500] White to move
T his is the ideal defensive formation. The knight will just shuttle
back and forth.
63.Ra8 Nh5; 64.Ke5. 64 .Ra7+ Kh6; 65 .Ra4 Kg7. 64...Ng3. Agreed
drawn.
ROOK + 2 US. KN IGHT + 5

The next example shows the power of a rook holding off a pile of
enemy pawns. T he key is the ability of the rook to move great distances
quickly. T he slow knight can't keep up.
DJURHUUS VS. FYLLINGEN

Norwegian Team Championship, Langesund, 1999
3
2
a b c d e f g h
White has a pair of connected passed pawns among the three pawns
that provide more than enough compensation for the exchange. Black
has to give up the b-pawn for the d-pawn, as passive defense is out of
the question.
308
ROOK A N D M INOR PI ECE E N DGAM ES
5 l ... Rd8. 5 l ... Rc6; 52.Kg3 Rd6 is not as effective. 53.f4+ Kg6 ;
54.Ne7+ K£6; 55.Nf5 holds all the pawns, and 55 ... Rc6 is countered by
56.Kf3 ! (56.d5 Rc3+; 57.Kh4 Rxb3; 58.d6 Rd3 is too slow. White has no
effective continuation and the d-pawn will fall. If 59.Kh5 Rh3+; 60.Nh4
Ke6; 6 l.g5 Kxd6; 62.g6 then 62 ... Ke6; 63.Kg4 Ra3 wins the a-pawn.)
56... Rc3+; 5 7.Ke4 Rxb3; 58.g5+! K£7; 59.Ke5 with two lines worth ex
amining:
59 ... Ra3; 60.Nd6+ Kg6; 6 l.Nc4 Rxa4; 62.Nxb6 Ra2 ; 63.d5 a4; 64.d6
Re2+! ; 65.Kd5 R.d 2+; 66.Kc6 a3; 67.Nd5 (forced) 67 ... a2; 68.d7 a l Q;
69.d8Q Qa4+.
3
2
a b c d e f g h
[502] White to move
70.Kb7 ( 70.Kc5 Rc2+; 7 l.Kd6 Rc6+; 72.Ke5 Qa l+; 73 .Ke4 Rc4+;

74.Kf3 Qd l +; 75.Kg3 Rc3+ and Black wins.; 70.Kb6 Rb2+; 7 1.Kc5 Rb5+;
72.Kd6 Rxd5+! ; 73.Kxd5 Qd l + etc.) 70... Qb5+; 7 1.Qb6+! Qxb6+;
72.Nxb6 and White will reach a Rook vs. Knight endgame which should
result in a draw. 59 ...b5 ! ?; 60.Nd6+ Kg6; 6 l.axb5 a4; 62.f5+ Kxg5 ; 63.f6
Kg6; 64.f7 Kg7; 65.Ke6 a3; 66.Nf5+ Kf8; 67.Nh4! Kg7 (67... Rg3; 68.b6);
68.Nf5+ with a draw.
52.Nxb6 Rxd4; 53.Kg3 Rd3; 54.Nc4 Rxb3 ; 55.Nxa5 Rb4; 56.Nc6
Rxa4 57.Ne7.
309
3
2
a b c d e f g h
[503] Black to move
Even with connected passed pawns White cannot make progress.

See the continuation in diagram #483 .
ROOK + 3 US. KN IG HT + 3
With three pawns each, and the defending pawns split, the poor
knight is no match for a rook.
ARNASON VS. PETURSSON

Reykjavik, 1980
3
2
a b c d e f g h
[504] White to move
Black is up the exchange, and the only compensation for White is

the active position of the king and the possibility of winning the weak
h-pawn. Black will eventually win the pawn at h3, and even if the pawn
at h4 goes, there will still be a passed pawn on the kingside. The supe
rior mobility of a rook compared with a knight allows Black to come
and go on the queenside to deal with any White threats.
310
ROOK A N D M INOR PI ECE EN DGAM ES
46.Nf3. 46.Nxc6 Rb3+; 47.Kc4 Rxh3; 48. a5 Rh l ; 49.a6 Ra l ; 50.a7

h3 and Black gets a new queen. 46 . . . Ra2; 47.Nxh4 Kf7 .
a b c d e f g h
[5051 White to move
48.Nf3 Ke6; 49.Kc4 Rxa4+.
3
2
a b c d e f g h
[5061 White to move
All Black has to do is avoid the exchange of all remaining pawns,

and that is not hard to achieve.
50.Kc5 Ra3! The rook will defend c6 from behind. 5 1 .Ne5. 5 l .Ng5+
K£5; 52 .h4 Rc3+; 53.Kd6 g6 and the collection of White's pawns is just
a matter of time. 5 l . . . Rc3+.
311
CA RDOZA PU BLIS H I N G • ERIC SCH ILLER
3
2
a b c d e f g h
[507]White to move
52.Nc4? This self-pin puts up no resistance at all. 52 ... g5! Black will
have time to grab the h-pawn, and can defend against the d-pawn from
h8 if needed. 53.Kb4 Rxh3; 54.Ne5 Kd5. White could have given up
here, but perhaps time pressure was a factor. 55.Ng6 Kxd4; 56.Nf8
Rh6; 57.Nd7 g4; 58.Nc5 g3; 59.Nb3+ Ke3. White resigned.
ROOK + 3 US. B ISHOP + 4

One of the greatest advantages a rook has over a minor piece is the
ability to contain an enemy king. If it can be restricted to the back rank,
the defender will need to worry about the king combining forces with
the rook to set up a back rank or rim mate.
YAKOVICH VS. KOSTEN

Hastings Congress, England, 1993
3
2
a b c d e f g h
[508] White to move
White has a passed pawn for the exchange, but the king is trapped
312
ROOK A N D M INOR PIECE E N DG A M ES
on the first rank and the f-pawn is \veak. Rather than send his king to
the kingside and risk losing the c-pawn, White keeps control of c2 and
prevents Black from getting behind the pawn.
38.c5 ! Ra5. 38 . . . Rxf2; 39.c6 and Black cannot get the rook back to
defend against the pawn. 39.c6 Rc5; 40.Bd7.
3
2
a b c d e f g h
[509] Black to move
Black has contained the pawn. To try to win he needs to mobilize

both the king and the rook. If the king plays a purely defensive role,
then there is no way to make progress. If the king attacks, however, the
rook will have to remain on the c-file to guard against the advance of
the pawn to the promotion square.
40 ... Kg7; 4 l .Ke2 Kf6; 42 .h4 Ke5 ; 43.Kf3 . White defends the
kingside with the king. This secures the draw. 43 ... Rc3+; 44.Kg4 Kf6;
45.Kf4 Rc4+; 46.Ke3 Ke5; 47.Kf3 f5. Black tries to involve the pawn,
which at least takes away g4 from the White king. 48.Be8 Rc3+; 49.Kg2
Ke4; 50.Bd7 Rc2. 50 ... Rc5; 5 l .Be8 is no better.
3
2
a b c d e f g h
[510] White to move
313
Black has made a lot of progress, but White's defense is solid.

5 l .Bc8!? A little joke. 5 1 . . . Kd3. 5 l ... Rxc6?? wouldn't be so funny for
Black: 52.Bb7 Kd5; 53.Kf3 Kd6; 54.Bxc6 Kxc6; 55.Kf4 and White will
win! 52.Bd7 Ke2 ; 53.h5 !
3
2
a b c d e f g h
[511] Black to move
White has a new drawing plan. The pawn at c6 can be jettisoned if

the kingside pawns are taken care of. 53 ... Ke l ; 54.hxg6 hxg6; 55.Kf3!
3
2
a b c d e f g h
[512] Black to move
The presence of the Black king at e 1 allows White to activate the

king, even though the f-pawn is lost.
55 ... Rxf2+; 56.Ke3 Rc2. Forced, to stop the pawn. 57.Kf4. Headed
to g5. 57 ... Kf2; 58.Be8! Rc4+; 59.Ke5 Kxg3; 60.Bxg6 f4; 6 l .Be4.
314
ROOK A N D M INOR PI ECE ENDG A M ES
3
2
a b c d e f g h
[513] Black to move
Here the game was agreed drawn, because whenever Black advances
the f-pawn, White sacs the bishop and gets to a theoretically drawn
rook vs. pawn endgame. 6 1 ... f3; 62.Bxf3 Kxf3; 63.Kd6. Black will have
to give up the rook to stop the pawn.
ROOK + 5 US. KNIGHT + 6

Our example shows how even a seemingly strong and well-defended
knight cannot easily keep an enemy rook at bay.
KOSTIC VS. GLODEANU

MI Festival, Bucharest, 1999
3
2
a b c d e f g h
[514] Black to move
The endgame does not look too bad for White. After all, the knight
is secure at d4 and there is no way that Black can get the rook to c 1 and
invade the back rank. You might think that Black would have a hard
time finding a winning plan. In fact, the idea is simple, just aim for our
315
CARDOZA PU BLIS H I NG • ERIC SCH ILLER
old friend the zugzwang. First Black will centralize the king.
2S ... g6; 29.g4. White discourages .. f5, which would gain control of
.
important squares at e4 and g4. 29 ... Kg7; 30.Ke2 Kf6; 3 l .h4 h5; 32.g5+
Ke5; 33.Kd3 ReS!; 34.Kd2.
a b c d e f g h
White is forced to defend c l.

34 ... Kd5; 35.b3 b5; 36.Kd l ReS; 37.Nc2. An attempt to complicate
things by threatening Nb4+. 37 ...Ke4. 37... Re4; 38.Nb4+ when 38 ...Rxb4;
39.axb4 Ke4; 40.Ke2 Kf5; 4 l.f3 is a simple win. 3S.Ke2 RdS; 39.f3+.
Not 39.Nd4? Rxd4; 40.exd4 Kxd4; 4 l.Kd2 Ke4. The h-pawn falls, and
the king and pawn endgame is won by a tempo. 42.Ke2 (42.Kc3 Kf4;
43.Kb4 Kg4; 44.Ka5 Kxh4; 45.Kxa6 Kxg5; 46.a4 bxa4; 4 7.bxa4 h4; 48.a5
h3; 49.Kb6 h2; 50.a6 h l Q; 5 l.a7 Qa8) 42 ... Kf4; 43.a4 bxa4; 44.bxa4
Kg4; 45.Ke3 Kxh4; 46.Kf4 Kh3; 47.Kf3 h4 etc.
39 ... Ke5; 40.Nb4 Kd6; 4 l .Nd3. White does not dare to capture the
a-pawn because of ...Ra8. 4l . . . a5; 42.Kd2 a4!
3
2
a b c d e f g h
[514] White to move
316
ROOK A N D M I NOR PI ECE EN DGA M ES
This establishes a weakness at a3 that will not go away. In any kind

of rook endgame it is useful to pin down the enemy pawn on the sec
ond or third rank, so that if it is later captured there is less distance to
the promotion square.
43.b4 Kd5; 44.Nc5 Kc4+; 45.Kc2 ReS! Black wants to encourage
the e-pawn to step forward so that Black can gain access to d4. White
does not oblige. 46.Ne4 Re6.
3
2
a b c d e f g h
[515] White to move
If the White king moves, then Black will sacrifice the exchange and
get the king to b3 or d3 with a \Vinning king and pawn endgame.
47.Nd2+ Kd5; 48.Nfl. If 48.e4+, then Kd4 and the infiltration comes
quickly and brutally. After 48.Ne4! ? Black has to decide whether to sac
the exchange.
Black can, and should, hang on to the rook with 48 . . . Rc6+; 49.Nc3+
Kc4; 50.Ne4 Rc7; 5 l .Nc3 ReS; 52.Ne4 Rc6; 53.Nc3 f6. Now 54.f4 f5
grabs control of the e4-square. 55.Nd l Re6; 56.Nb2+ (56.Kd2 Kb3)
56 ... Kd5; 57.Kd3 pushes the enemy king back the c-file belongs to the
rook. 57 . . . Rc6; 58.Nd l Rc l . 59.Kd2 loses to 59 . . . Ral so White could
try 59.Nc3+.
317
a b c d e f g h
Black must now avoid temptation: 59 . . . Rxc3+?; 60.Kxc3 Ke4; 6 l .Kd2

Kf3; 62 . Kd3 Kg4; 63.e4 fxe4+; 64.Kxe4 Kxh4; 65. Kf3 ! Kh3; 66.f5 and
White wins ! Instead, 59 . . . Kc6! is the simple win, where the rook goes
after the rim pawns.)
The immediate sacrifice 48 . . . Rxe4; 49.fxe4+ Kxe4; 50.Kd2 Kf3;
5 l .Kd3 Kg4 is a draw.
a b c d e f g h
52.Ke4! makes Black work for it. 52 . . . Kxh4; 53.Kf4 Kh3; 54.Kf3 h4;
55 .e4 Kh2 ; 56.Kf2 Kh l (56 . . . h3; 57.e5 Kh l ; 58.Kfl h2??; 59.Kf2 and
Black loses because the f-pawn is forced to move. ) 57.Kfl Kh2; 58.Kf2
Kh3; 59.Kf3 etc.
48 ... Rc6+; 49.Kb2 Rc4. In order to protect the h-pawn, one of the
central pawns must advance.
318
ROOK A N D M INOR PI ECE EN DGAM ES
3
2
a b c d e f g h
[518] White to move
50.£4. 50.e4+ Kd4; 5 l .Nd2 Rc3; 52.f4 Kd3 and Black wins. 50 ... Rc7;
5 1 .Nd2 Re7; 52.e4+. White has no choice. But now the sacrifice works.
52 ... Rxe4! ; 53.Nxe4 Kxe4; 54.Kc3.
3
2
a b c d e f g h
[519] Black to move
54 . . .Ke3 ! 54 . . . Kxf4 will also win, but it is trickier. 55.Kd4 Kg4; 56.Kc5
Kxh4; 57.Kxb5 Kxg5; 58.Kxa4 h4; 59.b5 h3; 60.b6 h2 ; 6 l .b7 h l Q;
62 .b8Q Qe4+; 63.Ka5 Qf5+ and the queens must come off. 64.Ka4
(64.Ka6 Qf6+; 64. Qb5 Qxb5+; 65 .Kxb5 f5; 66.a4 f4; 67.Kc4 £3; 68.Kd3
Kf4; 69 .a5 Kg3 ; 70.a6 f2 and Black queens with check. ) 64 . . . Qf4+.
55.Kc2 Kxf4; 56.Kd3 Kg4. White resigned. It only takes 6 moves
for Black to make a queen, but White needs 8 moves.
ROOK U S . TW O M I NOR PI ECES

Two minor pieces are an advantage against a rook, even with a
pawn less, because the piece can work independently to target weak
nesses. Only one piece is needed for defense, the other is free to roam.
319
CARDOZA PUB LIS H I N G • ERIC SCH I LLER
That assumes that the pieces are not under pressure and cannot be
chased from critical squares. Do not underestimate the power of the
rook!
PAVASOVIC VS. ROMANISHIN
Bled Open, Slovakia, 1 999
a b c d e f g h
[522] White to move
30.e5 ! ? The best try in a bad position. 30 ... c5! 30 . . . fxe5; 3 l .Nf3
Re8; 32.Bc2+ Kh6 (32 . . . e4?; 33.Ng5+ Kg6; 34.Nxe4) 33.Be4 lets White
set up a stubborn defense.
3 l .Bd5. 3 l .exf6 gxf6; 32 .Bd5 Ra6; 33 .Nb3 Rxa2; 34.Nxc5 Rd2;
35.Be4+ Kh6; 36.h4 b4 gives Black some winning chances.
3
2
a b c d e f g h
[521] White to move
One idea is to exchange to rook for a minor piece and the two
kingside pawns, and play with three pawns against a piece. Black can
try to take advantage of the uncoordinated White forces . 37.Kgl Rd4;
38.Bf5 Rd5; 39.Nb3 c5; 40.Kf2 Re5; 4 l .Bd3 and it is hard to see any
way to make progress. There are two alternatives at the previous move:
320
ROOK A N D M INOR PI ECE ENDG A M ES
40 . . . c4 is met by 4 l .Be6 with a draw after the only playable move:

4 l . . . Rh5 ! ! ; 42.gxh5 cxb3; 43.Bxb3 Kxh5. White has the wrong color
bishop.
40 . . . Rxf5+ is a mistake which loses despite the connected passed
pa,vns. 4 1 .gxf5 c4; 42 .Nd2 c3; 43.Nb3.
3
2
a b c d e f g h
[522] Black to move
A knight in this position prevents the pawns from advancing.

43 . . . Kh5; 44. Kg3 c2; 45.Nc l Kh6; 46.Kg4 Kg7; 47.Kf4 Kh6; 48.Ke4 Kh5;
49.Kd5 Kxh4; 50.Ke6 Kg5; 5 l .Nb3 and Black winds up in zugzwang.
3 l . . . Rd8; 32.Be4+ Kg8; 33.Nf3 fxe5; 34.Nxe5.
a b c d e f g h
[523] Black to move
34 ... c4. 34 . . . Re8; 35 .Bd5+. The availability of a bishop check once

again escapes the x-ray attack. 35 . . . Kh8; 36.Nf7+ Kh7; 37.Ng5+ Kh6;
38.Nf7+ is a study-like draw. 35.Nf3 b4; 36.Kg2 b3; 37 .axb3 cxb3; 38.Kf2
b2; 39.Ke2.
32 1
CA RDOZA PUBLISH ING • ERIC SCH ILLER
3
2
a b c d e f g h
[524] Black to move
The rest is simple, just eliminate the bishop and promote the pawn.
39 ... Re8; 40.Nd2 Rxe4!
RODRIGUES VS. GALEGO

Portuguese Masters, Caldas de Felgueira, 1999
3
2
a b c d e f g h
[525] White to move
Just how hard is it to draw with Rook and 1 vs . 2 Knights + 2? That

depends on where the pawns are. At the edge of the board the knights
lack room to maneuver, and the rook can create stubborn defense. In
this example, Black suffers from the need to defend both pawns. The
game is presented without much commentary. Feel free to try to come
up with some effective plan for Black!
45.Ra7+. The weakness at g7 forces the Black king to seek refuge
on the rim.
45 ... Kf8; 46.Ra8+ Kf7; 47.Ra7+ Kg8; 48.Ra8+ Kh7; 49.Rb8.
322
ROOK A N D M INOR PIECE ENDGA M ES
3
2
a b c d e f g h
[526] Black to move
The Black king is paralyzed, and the knight at g6 cannot move until
the h-pawn advances. 49... h3; 50.Kg3.
a b c d e f g h
[527] Black to move
Black is totally paralyzed, and must accept an exchange of pawns.

50 ... Ne6; 5 1 .Kxh3 Nxg5+; 52.Kg4. All White has to do is sacrifice
the rook for the pawn and the game is a technical draw. Black takes
every measure to prevent that. 52 ... Nf7; 5 3.Kf5 ( 53.Rb 7 Nge5+.)
53 ...Ne7+; 54.Ke6 Nc6; 55.Rb l. (55.Rb7?? Nfd8+.) 55 . ..Kg6; 56.Kd5
(56.Rgl+ Ng5+; 5 7.Kd5 Ne7+; 58.Ke5 Kh5.) 56 ... Ne7+; 57.Ke6 Nf5.
323
3
2
a b c d e f g h
[5281 White to move
58.Rb6. 58.Rgl+ Ng5+; 59.Ke5 Nh4; 60.Kf4 Nh£3; 6 l.Rg3 Kh5;

62.Rxf3? Nxf3 ; 63.Kxf3 Kh4 is a winning pawn endgame for Black.
58 ... N5h6; 59.Rb l Ng4; 60.Rb8 Ng5+; 6 l .Ke7 Nf6; 62.Rb6 Nge4;
63.Rb7. Threatening K£8. 63 . . . Nc5; 64.Rb l Nfe4; 65.Rb8. Intending
Rg8, now that Black has no checks. 65 ... Kh7; 66.Rb l . Now Rgl will do
the trick. 66 ... Ng3; 67.Kd6 Nce4+; 68.Ke5 g5.
3
2
a b c d e f g h
Black has finally succeeded in advancing the pawn, but this doesn't
help, as a draw is easily forced.
69.Rb6 Nc5; 70.Kf6. 70.Rb4 would have been more efficient, but
both players were probably exhausted and may have been in the sud
den-death time control. Black could only prevent Rg4 by sticking a knight
at e4, but 70 ...Nce4; 7 l.Rxe4 Nxe4; 72.Kxe4 is a drawn endgame.
70 ...Nce4+; 7 1 .Ke5 Kg7; 72.Rc6 Nd2; 73.Rc3 Nge4; 74.Rc6. 74.Kf5
was a quicker path to the peace agreement. 74 ... g4; 75.Kf4?! Better is
75.Kf5! followed by Rg6. 75 ... g3; 76.Rc l Kg6; 77.Rgl. Finally! 77 ... Kf6;
78.Rxg3 Nxg3; 79.Kxg3. Drawn.
324
ROOK AND MINOR PIECE ENDGAMES
KHALIFMAN VS. IVANCHUK

8
7
3
2
a b c d e f g h
[530] White to move
The advanced e- and h-pawns offer White every prospect of v ic

tory, but how do they advance?
58.Rxc7+! The sacrifice of the rook leaves Black with a useless
bishop. 58...Bxc7; 59.e6+. Black resigned. The Black bishop cannot
assist in the defense, and the king is no match for the pawns.
325
QUEEN ENDGAMES
Queen endgames are essentially long-winded tactical operations.
Other than presenting the trivial advice that you should keep your king
safe and your queen active, there isn't much that can be said. There are
a few critical positions involving queen and pawn on the 7th rank that
you should learn, otherwise, just play carefully!
Queen endings have been studied for a long time. Grandmaster
Yuri Averbakh has devoted much of his career to studying them. With
the aid of powerful computers, John Nunn and others have deepened
our theoretical knowledge of these endgames, but as a practical matter
they remain the most diff icult.
I would advise against spending too much energy trying to prevent
a lone pawn from getting to the 7th. Unless there is a perpetual harass
ment of the enemy king, it isn't realistic. Remember, each pawn ad
vance resets the 50-move rule clock. Concern yourself mostly with the
critical positions involving a queen against a pawn or a queen and
pawn against pawn. These are the positions you need to know at the
chessboard.
QUEEN US. I
A pawn can defend against a queen only if its promotion can be
assured or, importantly, if it is a bishop pawn. Even when the support
ing king is far away, the queen can usually force the enemy king to step
in front of the pawn, and then the attacking king can advance. After
the defender moves, checks are given until the king is once again block
ing the pawn, and then the king edges closer.
326
QUEEN E N DG A M ES
KAVCIC VS. IVANOVIC

Slovakian junior Championship, 1993
3
2
a b c d e f g h
Black wins by allowing White to queen the pawn!

66 ...Ke6! ; 67.h8Q. Underpromotion fails to save the day. 67.h8N
Ke7; 68.Nf7 Qxfi+; 69.Kh8 Qf5 and mate follows. 67 ...Qf7#.
RIEDNER VS. LOEBLER

Graz, Austria, 1997
3
2
a b c d e f g h
[532] White to move
Again, Black is allowed to promote his pawn, but mate follows.

76.Ke3 f1Q; 77.Qd2#.
QU EEN US. 2
Two connected passed pawns can defeat a minor piece, often even
a rook, but against a queen they are not as effective.
327
ANALYSIS OF HECTOR VS. SPEELMAN

Roskilde Open, Denmark, 1998
3
2
a b c d e f g h
[533] Black to move
Blackjust brings the king over, since White cannot escort the pawns
safely to the seventh rank.
52 . . . Kc5; 53.Kh5 Kd5; 54.g6 Ke5; 55.g7 Qh7; 56.Kg4 Qf5+; 57.Kh4
Qg6 is a simple winning plan.
QU EEN us. 1 rmore1

Most queen and pawn endgames are quite difficult. The subtleties
of the ending have been studied for centuries, but new discoveries are
constantly being found. The one position you must remember is the
case of queen against bishop pawn. This can be drawn, if the stronger
side's king is not too close. Even a king just a few squares away cannot
achieve victory.
QUEEN AND BISHOP PAWN DRAW
3
2
a b c d e f g h
328
QU EEN EN DGA M ES
Black can easily bring the queen close to the enemy pieces .
1 ... Qg1+; 2.Kh8 Qh2+; 3.Kg7 Qg3+; 4.Kh8 Qh4+; 5.Kg7 Qg5+;
6.Kh8 Qh6+; 7.Kg8 Qg6+; 8.Kh8!
3
2
a b c d e f g h
[535] Black to move
If Black captures the pawn it is stalemate, but otherwise no progress

can be made. The position is a draw.
QUEEN + I US. QU EEN

Winning with an extra pawn in an endgame with one queen each is
very difficult. Usually the pawn can work its way up the board, helped
by the queen which also shelters a king. The tricky part comes when
the pawn reaches the seventh rank, and must be promoted without
allowing a perpetual check.
(13 1) VAN DER WIEL VS. FEDOROWICZ

Graz, 198 1
a b c d e f g h
This endgame is enormously complex. It has been subject to scru

tiny by composers, theoreticians and computers. The machines have
329
CARDOZA PUB LISH I NG • ERIC SCH I LLER
corrected many analyses that have been passed down through genera
tions. Since our task is practical, not theoretical, we will look for the
desirable target positions for the defender. Aim for them and you should
survive. If you have the pleasure of playing with the pawn, try to keep
your opponent from reaching the ideal position.
This game was really a collaborative effort at the 1 98 1 World Youth
Team Championship where I was captain of the American side. John
Fedorowicz was our first board gladiator, having to face the likes of
Garry Kasparov, and had to suffer prolonged torture in the present
game, which was an important one against a strong Dutch team. It was
played back in the days of adjournments, and had three overnight analy
sis sessions to assist the players, who nevertheless, partly out of fatigue,
sometimes strayed from the paths. So the game is mostly a song with
out words. We'll just be paying attention to a few critical positions.
66.Qd5 Qc3+; 67.Kd7 Qg7+; 68.Kc6 Qc3+; 69.Kb5 Qb2+; 70.Kc5
Qa3+; 7 l .Kd4 Qa7+; 72.Kd3 Qa3+; 73.Ke2 Qe7+; 74.Kd l Qa3 ; 75.Qc4
Qd6+; 76.Kc2 Qg6+; 77.Qd3 Qc6+; 78.Kdl Qa8.
a b c d e f g h
[536] White to move
We were prepared for this eventuality. White is going to be able to

advance the pawn. In fact, we knew that the pawn would sooner or
later march up the board. It was a slow trip, taking several days!
79.a3 Kg2; 80.Kc2 Kf2 ; 8 l .Qd4+ Kf3; 82.a4 Qe4+; 83.Kc3 Qc6+;
84.Qc4 Qf6+; 85.Kb3 Qb6+; 86.Qb5 Qe3+; 87.Kc4 Qe4+; 88.Kc5 Qe5+;
89.Kc6 Qe6+; 90.Kc7 Qe7+; 9 l .Qd7 Qe5+; 92.Kb7 Qb2+; 93.Qb5 Qg7+;
94.Ka6 Qg6+; 95.Ka7 Qgl+; 96.Qb6 Qal ; 97.a5 Ke2; 98.Qe6+ Kfl ;
99.Qf5+ Ke l ; IOO.Kb6 Qb2+; IOI.Kc6 Qc3+; 1 02.Kd6 Qb4+; 103.Qc5
Qb8+; 1 04.Ke7 Qb7+; 1 05.Kf6 Qf3+; 1 06.Qf5 Qa8; 107.Qe6+ Kfl ;
1 08.a6 Kgl ; 1 09.Qb6+ Kh2; IIO.Qc7+ Kgl; l l l .Qc5+ Kh2; 1 1 2.a7.
3 30
QUEEN EN DG A M ES
3
2
a b c d e f g h
[537] Black to move
Well, the pawn is at a7. The defender remains in place. The queen
will have to chase the enemy king around for a while. Remember, the
50-move rule time bomb is ticking! From here on, just concentrate on
the position of the Black king. Don't let it stray too far from the corner.
1 1 2 . . . Qh8+; 1 13.Kg5 Qg7+; 1 14.Kf5 Qh7+; 1 15 .Kf4 Qf7+; 1 1 6.Ke3
Qb3+; 1 1 7 .Ke4 Qb7+; 1 18.Kd3 Qa6+; 1 19.Ke3 Qe6+; 1 20.Kd4 Qg4+;
1 2 1 .Kc3 Qf3+; 1 22.Kd2 Qg2+; 123.Ke 1 Qe4+; 124.Kd 1 Qf3+; 1 25 .Kc 1
Qh 1+; 126.Kh2 Qb7+; 127.Ka3 Qf3+; 128.Kb4 Qe4+; 1 29.Kb5 Qe2+;
130.Kc6 Qe8+; 1 3 1 .Kb6 Qe6+; 132.Kh7 Qe4+; 133.Kc7 Qh7+; 134.Kc6
Qe4+; 135.Qd5 Qa4+; 136.Kb7 Qb4+; 137.Kc7 Qe7+; 138.Qd7 Qc5+;
1 39.Kb8 Qf8+; 140.Qc8 Qb4+; 141 .Ka8.
a b c d e f g h
[538] Black to move
Careful now, after 1 4 l . . . Qd5+; 1 42.Qb7 Qe8+?? White wins with

143.Qb8+! 141 ... Kg 1 ! ; 142.Qc 1+ K£2 .
33 1
CA RDOZA PUB LISH I N G • ERIC SCH I LLER
3
2
a b c d e f g h
[539] White to move
Johnny van der Wiel has a good sense of humor, so he brought the
4-day marathon to an end with 143 .Qb2+ Qxb2. Stalemate.
QUEEN + 3 US. 2
Even when the king is in the safest position, for example fianchettoed
on the kings ide while controlling the h l -aS diagonal, a win is not al
ways possible. However, such a formation offers the best practical
chances. In our example, the weakness of the defending pawns lets
White apply the straightforward winning technique. This is a pattern
you should remember, as it can often be used to win a queen endgame
with an outside passed pawn.
BABULA VS. MALLAHI

Elista Olympiad, 1 998
3
2
a b c d e f g h
[540] Black to move
Obviously Black is not going to exchange queens to reach a hope

less pawn endgame.
332
QU EEN EN DGA M ES
35 ... Qa5 ; 3 6 . Qe4 Kg7 ; 3 7 . Qc6. The queen belongs on the

queenside, stationed somewhere along the a8-h l diagonal to prevent
Black from giving check. 37 ... Qb4; 38.Qb5. An important consider
ation in such positions is that the check on the long diagonal does not
lead any,vhere for Black unless the kingside pawn barrier has been
disrupted. In this case, the king will be safe at h2. White will eventually
have to retake the long diagonal but this can be done by advancing the
pawn to a6 and placing the queen at b7.
38 ... Qe4+; 39.Kh2 Qf3; 40.Qg5+! With a series of checks White
centralizes the queen and gets to d4. 40 ... Kf8; 4 l .Qd8+ Kg7; 42.Qd4+.
3
2
a b c d e f g h
Now the f-pawn and a-pawn are defended. The rest is simple.
42 ... Kg8; 43.a5 Qa8; 44.Qb6 h4. Desperation, but it doesn't pro
long the game for long. 45.gxh4 Qd5; 46.Qb8+ Kh7; 47.a6 Qa2; 48.Qb6
Qc4; 49.Qf6 Qc7+; 50.Kg2 . No more checks . 50 ... Qd7; 5 1 .Qf3 Kh8;
52.Qe4 Qa7; 53.Qb7. The pawn is going to be escorted to the promo
tion square. Black can arrange a couple of spite checks, but that's about
all.
53 ... Qd4; 54.Qb8+ Kh7; 55.a7 Qe4+; 56.Kh2 Qxh4+; 57.Kg l . Black
resigned.
QUEEN + 3 US. QU EEN + 3

The more pawns on the board, the harder it is to keep track of all
the possible pawn endgames. Nevertheless, unless you are sure you can
hold the pawn endgame, never part with your queen if you are defend
ing. The queen is your last, best drawing hope. Exchange it only when
the final result is clear, and at least a draw. In our example, White
makes a terrible miscalculation, and it is unlikely that time pressure
was involved.
333
CA RDOZA PU BLISHING • ERIC SCH I LLER
(75) BOENSCH VS. JUSSUPOV

Altensteig, 1993
3
2
a b c d e f g h
[542] White to move
47.Qg2?? 47.Kgl Qxb4; 48. Qfl+ was the lesser evil. Black will try
to safeguard the king, play . . . Qd4+ and advance the pawn. 48...Ke5;
49.Qe2+ Kd6; 50.Qd3+ Kc7 is a good start. 47 ... Qxg2+; 48.Kxg2.
Now it is just a matter of calculation. Hard to blame White, really.
It takes a pretty deep search to work it all out. 28-ply. Yet it is not a
matter of calculation, if you know the pattern that starts at move 58.
48 ... Ke5; 49.h5 Kd4; 50.Kf3 Kc4; 5 1 .Ke4 b5; 52.Kf5 Kxb4; 53.Kg6
Kc5; 54.Kxg7 b4; 55.Kxh6 b3; 56.Kg7 b2; 57.h6 b l Q; 58.h7.
3
2
a b c d e f g h
[543] Black to move
The rest should be familiar from the Hector-Speelman conclusion

above in position #533.
58 ... Qb7+; 59.Kg8 Qc8+. Or 59 . . . Qd5+; 60.Kg7 Qg5+; 6 l .Kf7 Qh6;
62.Kg8 Qg6+; 63.Kh8 Qfi; 64.g4 Qf8#. 60.Kg7 Qd7+; 6 1 .Kg8 Qe8+;
62.Kg7 Qe7+; 63.Kg8 Qg5+; 64.Kh8 Qg6. Black resigned.
3 34
QUEEN EN DGA M ES
QUEEN + 4 US. QUEEN + 3

Rarely is an author presented with an ideal instructive position, but
here is a gem. This is a perfect example of the winning plan used to
promote an outside passed pawn.
BABULA VS. MALLAHI

Elista Olympiad, 1998
3
2
a b c d e f g h
[544] Black to move
White's position is perfect. The king is safe at g2, and can hide
from checks on the diagonal by moving to h2. The pawns at f2 and a4
are defended by the queen. The threat is simply aS, a6, Qb7 and the
pawn marches on.
32 ... g5 ! Black tries the only reasonable plan, attempting to open
up the kingside so that possibilities for perpetual check can arise.
33 .Qe3 ! A beautiful move. The threat at g5 prevents Black from grab
bing the a-pawn. 33.hxg5 may not eliminate the possibility of a win, but
it is reckless, with Black gaining counterplay after 33 . . . Qe4+ 34.Kh2 h4.
33 ... gxh4. 33 ... Qxa4; 34.Qxg5+ Kf8; 35.Qxh5 is a simple win be
cause the h-pawn cannot be stopped.
34.Qg5+ Kf8; 35.Qxh4.
3 35
CA RDOZA PUBLISHING • ERIC SCH ILLER
3
2
a b c d e f g h
[543] Black to move
White has made considerable progress. It is Black's king, not White's,

h
t at is vulnerable to checks. Black's kingside pawns are now split and
weak. The next phase of the plan is to advance the a-pawn. For the
continuation, see diagram #540.
336
MISCELLANEOUS
ENDGAMES
In this section I present a selection of endgames which don't fall
into the basic categories. There are a variety of themes illustrated, but
the most important consideration is the decision to exchange pieces or
sacrifice pawns. These are the practical problems that will arise at the
chessboard time and time again.
The first game shows the advantages of superior piece placement,
better bishop, and active king. In the analysis snippet from Korchnoi
Shirov, we see a piece fighting against pawns. Ostermeyer vs . Bischoff
starts with a tactic but then moves to a subtle strategic phase. In Toth
vs. Muhari we switch to an endgame which, by itself, is not difficult,
but would be easy to overlook earlier in analysis.
Sahetchian vs. Gurieli is a strange position at the start, but as with
so many complex endgames it resolves into an important endgame with
fewer pieces. Rublevsky exploits greater space and a more active king
against Spraggett. The middle of the board is the center of attention in
Sokolov-Maliscauskas. A hasty capture brings ruin in Goormachtigh
vs. Van Beek.
To end our voyage, I've selected the conclusion of a game between
two of the very best players: Garry Kasparov and Vladimir Kramnik.
Each side is faced with critical choices as exchanges and alternative
captures are offered.
YEPISHIN VS. LUGOVOI

St. Peterburg, Russia, 1 996
There is a lot going on in this position, so it takes a little work to
figure out the correct plan. Obviously Black is tied down, but how are
the White pieces to break through?
337
CA RDOZA PUBLISH ING • ERIC SCH ILLER
a b c d e f g h
[5441 White to move
Two squares deserve special attention. If White can maneuver the

knight to e6, Black is busted. If the bishop gets to e6, then Nb4 will
lead to the win of the d-pawn. A White pawn at f5 would prevent Black
from obtaining any counterplay by . . . f5 . Put these together, and the
solution comes easily.
39.£4! If Black captures, the knight goes to f4 and then e6. Black
can't play . . . f5, because then the g-pawn hangs. 39...Ng6. 39 ... gxf4;
40.Nxf4 Bc7; 4 l .Ne6 Bg3; 42 .Kxb6 and the a-pawn marches. On the
other hand, 39 . . . Bc7; 40.f5 and Black is paralyzed. 40 . . . Kf8; 4 l .Be6 Ke8;
42 .Nb4 etc.
40.£5 Nf4. Black is rightfully desperate. 40 . . . Ne7; 4 l .Be6+ Ke8;
42 .Nb4 etc. 4 l.Nxf4 gxf4.
3
2
a b c d e f g h
[5451 White to move
This bishops of opposite color endgame is won for White. It isn't

just the d-pawn, which can be picked off immediately, but the useless
ness of the Black bishop, which has no way to get into the game be-
338
M ISCELLAN EOUS ENDG A M ES
cause it must stand guard over the pawn at b6. The win isn't difficult,
as you can see by examining the continuation of the game of diagram
#286.
ANALYSIS OF KORCHNOI VS. SHIROV

FIDE Knockout, Groningen, 1996
3
2
a b c d e f g h
[546] White to move
This is a branch of the analysis of the Korchnoi-Shirov game dis

cussed in position #295 . White has an extra bishop, but all of Black's
pawns are on light squares, so there are no targets. Meanwhile, the e
pawn will be advanced, eventually supported by the f-pawn. Black can
restrain the bishop by planting it at f6, but eventually it will become
restless and vacate the square. Note that Black's potential passed pawn
on the queenside keeps the White king from roaming too far.
Before moving on to the main line, let's consider the more active
defense, 4 l .Kb4 e3; 42.Ka5 Ke4; 43.Kb6 tying the Black king to d5. But
the White king cannot help defend, and can do little but watch as Black's
pawns roll forward. 43 . . . Kd5; 44.Bg5 e2; 45.Bh4.
a b c d e f g h
[547] Black to move
3 39
CA RDOZA PUBLISH I N G • ERIC SCH I LLER
45 . . . a5! ; 46.Be l (46.Kxa5 Kxd4; 47.Bf2+ Kc4; 48.Kb6 Kd5; 49.Ka5

f5 ; 50.a4 bxa4; 5 l .Kxa4 f4; 52.Kb4 Ke4; 53.Kc3 Kf3; 54.Be l Ke3 ! ;
55.Bd2+ Kf2 and wins.)
4 1.Bf6! is the correct move. 4 1 ...e3; 42.Kd3 Kf3; 43 .Bh4 f5; 44.Be l
e2; 45.a3 a5!
3
2
a b c d e f g h
White must reject 46.Bxa5 Kf2; 47.Bb4 f4; 48.Ba5 f3; 49.Bb4 e l Q;
50.Bxe l + Kxe l ; 5 l .Ke3 f2 etc. That leaves 46.h4. Now 46 . . . b4? looks
very strong, and you need to look twenty moves ahead to see the prob
lem! 47.axb4 axb4; 48 .Kc2 Kg2; 49.Bxb4 (49.Kd3 Kfl; 50.Bxb4 e l Q;
5 l .Bxe l Kxe l transposes.) 49 . . . Kfl; 50.Kd3 e l Q; 5 1 .Bxe l Kxe l ; 52.Ke3
Kfl ; 53.Kf4 Kg2; 54.Ke5! (54.Kg5? Kg3; 55.Kxh5 f4; 56.Kg5 f3; 57.h5
f2; 58.h6 fl Q; 59.h7 Qf8) 54 . . . Kg3; 55.Kxe6 f4; 56.d5 f3.
3
2
a b c d e f g h
[549] White to move
57 .d6! (57.dxc6 f2 ; 58.c7 fl Q; 59 .c8Q Qh3+) 57 . . . f2; 58.d7 fl Q;

59 .d8Q Qc4+! ; 60 .Kd7 Qd5+; 6 l .Kc8 Qxd8+; 62.Kxd8 Kxh4; 63.Kd7
Kg5; 64.Kxc6 h4; 65.Kd6 h3; 66.c6 h2; 67.c7 h l Q; 68.c8Q with a draw!
340
M ISCELLA N EOUS ENDG A M ES
3
2
a b c d e f g h
[550] Black to move
46 ... a4! ; 47.Kd2 f4! (47...b4 is not as good as it looks. 48.axb4 a3;
49.Kc2 Kg2; 50.b5! Kfl; 5 1 .Bb4 a2; 52.Kb2 cxb5; 53 .c6 f4; 54.c7 £3;
55 .c8Q f2; 56.Qxe6 e l Q; 57.Bxe l fxe l Q; 58.Qxe l+ Kxe l ; 59.d5 and
White wins! ) 48.Kd3 Kg2! Beautiful! Black gives up the prized e-pawn.
49.Kxe2 f3+; 50.Ke3 b4! Finally!
a b c d e f g h
[551] White to move
5 l .Kd2 bxa3; 52.Kc2 f2; 53.Bxf2 Kxf2; 54.Kb l Ke3 with an easy
win for Black. You can see how it is necessary to see very far ahead in
the endgame!
34 1
CA RDOZA PU BLIS H I N G • ERIC SCH I LLER
OSTERMEYER VS. BISCHOFF

Bundesliga, Germany, 1984
3
2
a b c d e f g h
[552] White to move
White has an extra knight and pawn, but Black's pawn is only two
squares away from promotion. If Black manages to eliminate the two
White pawns, the game will likely end in a draw. White finds the cor
rect solution.
5 l .Rb l ! ! Rxc4. 5 l . . . R�b l ; 52 .Nc3+ Kf4; 53 .Nxb l is not an option.
52.Nb4. The knight shields the remaining pawn. 52 ... Rc5; 53.Kfl ! The
king is the appropriate blockader of the g-pawn.
53 ... Ke3; 54.Kg2 Ra5; 55.Ra l Kd2. Black is no longer concerned
about the g-pawn, but wants to eliminate the a-pawn at all costs.
3
2
a b c d e f g h
[553] White to move
White must take stern measures to prevent this.

56.Ra3 ! Kc l ; 57.Nd3+! Kd2. Now the Black king is held at arm's
length and White can afford to capture the g-pawn. The rest is easy.
58.Kxg3 Ke3 ; 59.Nc5+ Ke2. 59 . . . Kd4; 60.Nb3+.
342
M ISCELLA N EOUS EN DGA M ES
60.Nb3 Ra7; 6 1 .Nd4+ Kd2; 62.a5.
3
2
a b c d e f g h
[554] Black to move
The king is permanently cut off and White can bring the king to
the b-file and drive out the defending rook.
TOTH VS. MUHARI

Paks (Under 1 2), 1994
A queen almost always wins against two minor pieces, even if they
are well coordinated. The reason is that the knight cannot simulta
neously defend the bishop and be defended by the bishop. So it is just
a matter of applying pressure . The only thing to avoid is a situation
where the weaker king is in the corner, with bishop and knight teaming
up on the diagonal to keep the king out. You'll see that in the notes
below.
3
2
a b c d e f g h
[555] Black to move
The game would end instantly if Black promotes to a queen since

Qd3 is checkmate. Black can prolong the game by selecting a knight.
343
CA RDOZA PUBLIS H I N G • ERIC SCHillER
65 . .. e 1N!; 66.Qc3+ Kd l ; 67.Qd4+ Bd2; 68.Kb3.
3
2
a b c d e f g h
[556] Black to move
The first pressure point. Black cannot move the knight without
losing it, and the bishop is pinned. For example, 68 . . . Nc2; 69.Qg4+
Kc l ; 70. Qc4! If the king moves to c l , White delivers mate at a l . That
leaves only one move.
68 . . . Ke2 ; 69.Qe4+ Be3 . 69 . . . Kd l ; 70. Qg4+ Kc l ; 7 l . Qe2 Ba5;
72.Qb2+ Kd l ; 73.Qal +. 70.Kc3 . 70.Kc4 is not terrible, but it allows a
small chance to reach the one drawing position, which I will illustrate
here by making indifferent moves for White. 70 . . . Kd2; 7 l .Kd5 Nc2 ;
72.Qg2+ Kc l; 73.Qfl + Kb2; 74.Qb5+ Kal; 75 .Ke4 Bd4; 76.Qa4+ Kb l;
77.Qb3+ Bb2; 78.Kd3 Nd4 the knight guards b2 and attacks g6. The
White king can never approach, and the game is drawn. This forma
tion is unlikely to appear, and can almost always be avoided, but it is a
good source of swindle points if it can be reached.
70 ... Nf3; 7 1 .Qg4 K£2; 72.Qe4. 72 .Kd3?? Ne5+ and Black will win,
if he knows the Bishop vs. Knight endgame. 72 ...Bd4+; 73.Kd3 Be5;
74.Qe2+ Kg3; 75 .Ke4.
3
2
a b c d e f g h
[557] Black to move
344
M ISCELLA N EOUS EN DG A M ES
White wins one of the pieces. Black resigned a few moves later.
SAHETCHIAN VS. GURIELI

Cappelle Ia Grande, France, 1995
3
2
a b c d e f g h
[558] White to move
Clearly White has no real winning chances . With only a single pawn
remaining, and no possibility of bringing the king to attack the g-pawn,
all that can be hoped for is a draw. Black has a rook and three pawns
for the bishops, and two pawns are both connected and passed. White
must activate the king.
52.Kf2 ReS; 53.Bb2. The bishop heads to a3, where it can restrain
the pawns more easily. 53 ... Re7. 53 . . . d5; 54.Ba3 c4; 55 .Bc5 brings the
bishop to a good defensive square at d4, blocking the pawns. 54.Kf3
d5; 55.Ba3 Rc7; 56.Be6 d4; 57.Bc4.
3
2
a b c d e f g h
[559] Black to move
White blockades at c3, and the bishop lacks mobility at a. The main
asset is the active king, which heads straight for d5 .
345
CARDOZA PUBLIS H I N G • ERIC SCH ILLER
57 ...Kg6; 58.Ke4 h5. White's last pawn is eliminated and Black has
three passed pawns as a result. 59.gxh5+ Kxh5; 60.Kd5 g4; 6 l.Bxc5.
a b c d e f g h
[560] Black to move
Black has a difficult choice here. Either the g-pawn advances or the
rook must be sacrificed.
6 1 ...Rxc5+. 6l ... g3; 62 .Bfl is correct, defending g2 . It is still a close
race, but White survives. 62 .Bxd4? Rd7+! ; 63.Ke4 Rxd4+; 64.Kxd4 g2
is a trap White must avoid. 62.Bd3 does not work as in the game, be
cause 62 . . . Rxc5+; 63.Kxc5 g2 gets a new queen. ) 62 . . . Rxc5+; 63.Kxc5
d3; 64.Kd4 d2; 65 .Be2+ Kh4; 66.Ke3 g2; 67.Kf2 Kh3; 68.Kgl escapes
with a draw.
62.Kxc5 g3 ; 63.Bd5.
a b c d e f g h
[561] Black to move
This was not possible in the 6 l ... g3 line. For the remainder of the
game, see position # 1 95.
346
M ISCELLA N EOUS EN DG A M ES
RUBLEVSKY VS. SPRAGGETT

3
2
a b c d e f g h
[562] White to move
This position sure looks like a draw. After all, the material is even
and there are few pawns left on the board. White's advantage is deci
sive, however. The key asset is the weakness of the pawn at b6 com
pared with the relative invulnerability of the pawn at c4. How can the
twice-defended b-pawn be weak? Just watch!
55.Ne4! The knight attacks the bishop, but also eyes c5. 55 ... Be3.
55 . . . Bc5 leaves the knight with no retreat after the White king ap
proaches with 56.Ke6. After 56 . . . Ke8; 57.Nd6+! Black must give up the
bishop for the knight, as otherwise his own knight falls. 57 . . . Bxd6;
58.Bxd6 g6; 59.Bc7 and the threat of Kd6-c6, with the resulting win of
the b-pawn, can only be countered by 59 . . . Nc5+ but then the White
king infiltrates the kingside. 60.Kf6 Ne4+; 6 1 .Kxg6 Nd2; 62.Kh7 Nxc4;
63.g6 and the knight cannot get back to stop the pawn.
56.Ke6.
3
2
a b c d e f g h
[563] Black to move
347
Black has a problem with the knight. The best he can do is ex
change it. 56 . . . Nc5+; 57.Nxc5 bxc5. 57 . . . Bxc5; 58.Bd6+ leads to an eas
ily winning king and pawn endgame. 58.g6.
3
2
a b c d e f g h
[564] Black to move
Continued in diagram #248 .
SOKOLOV VS. MALISAUSKAS

Biel Interzonal, 1993
3
2
a b c d e f g h
[565] White to move
White faces a strategic choice here . Black is prepared to advance

the pawn to e5. White can prevent this by placing his own pawn there,
but is that the right plan? There are no immediate tactical consequences
here . The evaluation of the endgame pawn structure is critical.
29.e5! White insures that Black will have a weak pawn at e6 which
cannot be defended by a bishop. If it is defended by the knight later in
the game, White can use the bishop to remove the knight. Therefore
the king will need to remain in the area of the weak pawn. At the same
348
M ISCELLA N EOUS ENDG A M ES
time, White can work on the kingside, attempting to weaken the light
squares at g6 and h 7, \vhich are also difficult to defend.
29 . . . Bb4. It doesn't really matter where the bishop goes. 30.g3. White
prepares a kingside advance. 30 . . . Ke7; 3 1 .f4.
3
2
a b c d e f g h
[566] Black to move
3 1 . . . h5? This seriously weakens the pawn at g6, and was surely un
necessary.
3 1 . . . Ba3 ! ; 32.Kg2 Bb4; 33.Kf3 Kf7 is more solid, for example 34.g4
with two significant branches .
First we must consider 34 . . . Bd2; 35.Ke4 Nd5; 36 .£5! gxf5+; 37.gxf5
Nf4; 38.Bc4 ! Ke7; 39.Bxa7 exf5+; 40 .Kxf5 Ng6 (40 . . . Nxh3; 4 l .e6! Bb4;
42 .Be3! and the knight is dominated. 42 . . . h6; 43 .Bxh6 Ng l ; 44.Bg5+
Ke8; 45.Bb5+ Kf8; 46.Kg6 ! ! Be7; 47.Bh6+ Kg8; 48.Be8 Bf8; 49 .Bf7+
Kh8 with the elegant finish 50.e7! Bxe7; 5 l .Bg7#).
The other branch is 34 . . . h5; 35 .gxh5 gxh5; 36.£5 ! exf5; 37.Kf4 Bc3 ! ?;
38.Bxc3 (38.Bxb6 axb6; 39.Kxf5 h4! ; 40.Bc4+ Ke7; 4 1 .e6 Bd4; 42.Kg4
Bf2 is a dra\v. ) 38 . . . Nd5+; 39.Kxf5 Nxc3; 40.Bc4+ (40.Ba6 h4 ! ) 40 . . . Ke7;
4 1 .Kg5 a5; 42.Kxh5 a4; 43 . Kg6 a3 ; 44.h4 a2; 45.Bxa2 Nxa2; 46.h5 and
there is a miracle draw in 46 . . . Nc l ; 47 .h6 Nd3 ; 48.h7 Nxe5+; 49 .Kg7
Nf7.
32.Kf2 Kf7; 33.Kf3 Ba3 ; 34.Bd3 Bb4.
349
CA RDOZA PUB LISH I NG • ERIC SCH I LLER
3
2
a b c d e f g h
[567] White to move
It is White's turn to misplay the position.

35.Bc2?! 35.g4! was the winning move.
3
2
a b c d e f g h
[568] Black to move
35 . . .hxg4+ (35 . . . Bd2; 36.gxh5 gxh5; 37.Ke4 Bb4; 38.£5 exf5+; 39.Kxf5
is bad for Black.) 36.hxg4 Ba3 (36 . . . Nc8; 37.Bc4 Ke7; 38.Ba6 Nb6; 39.Bd3
K.£7; 40.Bf2 Nc8; 41 .Bc4 Ke7; 42.Bh4+ Kd7; 43.Bb5+ Kc7; 44.Bd3 Ne7;
45 .Bxe7 Bxe7; 46.Bxg6 and the kingside pawns prevail. ) Analysis by
Ftacnik. 37.Bf2 Bb4; 38.Bc2 Ba3; 39.Bb3 Ke7 avoids the advance of the
f-pawn but the king can be chased. (39 . . . Bb2; 40.f5 Bxe5; 4 1 .fxe6+ Ke7;
42.Bc5+ Bd6; 43.Bxd6+ Kxd6; 44.Kf4 Nc8; 45.Kg5 Ne7; 46.Kf6 a6; 4 7.g5
a5; 48.Kf7 and the knight must move, after which White grabs the g
pawn.) 40 .Bh4+ Kd7 when the g-pawn is doomed. 41.Bc2 a5; 42 .Bxg6
a4; 43.Bc2 Kc6; 44.Bd8 Bf8; 45 .Bxb6 Kxb6; 46.Bxa4 wins.
35 ... Nc8! The knight will be needed at e7, to protect the weakling
at g6. 36.Bb3 a5; 37.g4 hxg4+; 38.hxg4 Ne7!
350
M I SCELLA N EOUS E N DG A M ES
a b c d e f g h
[569] White to move
White's task is much more difficult now that the knight defends
the kingside.
39.Ke4 Bd2; 40.Bh2 Bh4; 4 1 .Bc l Be l ; 42.Kd4 B£2+; 43.Kd3 Be l .
This may b e safest but i t i s not best. 43 . . . Nc6 ! i s a better try, activating
the knight with pressure at e5, and, importantly, access to d4, which
covers f5 and e6. Consider the following variations: .
The obvious move is 44. Ke4, where 44 . . . Ne7 (44 . . . Ke7; 45. Bd2)
45 .Bd2 Nc6; 46.f5 looks strong, but it isn't so simple. 46 ... gxf5+; 47.gxf5
Nd4!; 48.fxe6+; (48.Bxe6+ Nxe6; 49 .Bxa5 Nc5+; 50.Kd5 Nd3; 5 l .Bc3
Nxe5 ; 52.Bxe5 Bh4 leads to a draw. ) 48 . . . Ke8 ! (48 . . . Nxe6; 49 .Kf5 )
49 .Ba4+ Ke7; 50.Bxa5 Nxe6 and Black holds .
White might try 44 .Bd2 ! ? instead, for example 44 . . . Bb6; 45.Ke4
Bc7; 46.f5 gxf5+; 47.gxf5 Nxe5; 48.fxe6+ K£6; 49.Bf4 Bd6; 50.Kd5 Bb8;
5 1 .Bg3 Bc7; 52.Bh4+ Kg7; 53.e7 wins.
44.Bc4 Bb4; 45 .Kc2! The White king will circle round on the
queenside. 45 ... a4. This prevents White's plan, but creates new options
as the a-pawn becomes weaker. 46.Kd3 ! a3; 47.Ke4 Nc6. 4 7 . . . Bc5; 48.Bd2
Nc6 (48 . . . Bb6; 49.Bb4 and a3 beckons.) 49.f5. 48.f5 gxf5+; 49.gxf5 Nd8;
50.fxe6+! Kg6. 50 . . . Kf8 is also countered by 5 1 . Kd5 . 5 1 .Kd5 ! Be7;
52.Bd2 Kg7. 52 . . . Kf5; 53.Bd3+ Kg4; 54.Bb l intending . . . Ba5-c7-d6.
53.Ba5 Kf8; 54.Bc7 Ke8; 55.Ba2.
351
CARDOZA PUB LISH IN G • ERIC SCH I LLER
3
2
a b c d e f g h
[570] Black to move
_ With the a-pawn blockaded, Black has to worry about zugzwangs.

55 ... Nb7; 56.Bd6!
3
2
a b c d e f g h
56 ... Na5. 56 . . . Nxd6; 57.exd6 is similar to the game. 57.Kc5 Nb7+;

58.Kc6 Nxd6. 58 . . . Na5+; 59.Kb6 Bd8+; 60 .Bc7 Bxc7+; 6 l .Kxc7 and the
knight is trapped. 59.exd6 Bh4; 60.Kc7 Bg3; 6 l.Bc4. Black resigned.
The position is hopeless, as even the desperate 6 l . . . a2 would lose to
62.e7 ! ; 6 l .Bc4 Bh4; 62 .d7+.
3 52
M ISCELLA N EOUS EN DGA M ES
GOORMACHTIGH VS. VAN BEEK

Belgian Teams, Antwerp, 1999
a b c d e f g h
[572] White to move
The pawn at h6 is dangling and White cannot resist the temptation

to take it right away. After all, the bishop is attacked, why not eat on the
run? That's what White does with 26.Bxh6?
26.Ke3 is the right move. With the outside passed pawn on the a
file, Black can't afford to trade pieces . The bishop can have lunch on
the next turn. 26 . . . Ra4 (26 . .. Re4+; 27.Rxe4 fxe4; 28.Kxe4 is hopeless
for Black. ) 27.Bxh6 (27.Rxc6?? Re4+ wins the bishop. ) 27 . . . Bxh6+;
28.Rxh6 f4+; 29.Kd3 Bxf3; 30.gxf3 Rxa2; 3 l .Rxc6 Rxf2; 32.Ke4 Rxh2;
33.Kxf4 RhS is a draw with reasonable defense.
26 ... Re4+; 27 .Rxe4 fxe4; 28.Bxg7 ex£3+; 29.gxf3 Kxg7.
a b c d e f g h
[573] White to move
How things have changed! White has three pawns for the position
but they are all cripples . White can only stop the c-pawn temporarily,
watching hopelessly as the pawns drop off.
353
30.a4 Kf6; 3 l .a5 Ke5; 32.Ke3 Kd6; 33.£4 Bg6; 34.h4 c5; 35 .Kd2
Kc6; 36.Kc3 Kb5; 37.£5 Bxf5; 38.a6 Kxa6; 39.Kc4 Kb6; 40.h5 Kc6;
4 l .h6 Kd6; 42.£4 Kc6. A final zugzwang terminates the game.
KASPAROV VS. KRAMNIK

Intel PCA Tournament, New York, 1994
a b c d e f g h
[574] White to move
White is forced to enter the endgame. White's h-pawn may be a

factor in the long term but Black's queenside pawns are imposing, too .
This type o f pawn structure i s typical o f the Semi-Slav opening. White
is down the exchange, but has plenty of compensation.
23.Qxc6+ Rxc6; 24.Rd l . There was another plan to be considered.
24.Bf4 ReS; 25.Nd6+ Rxd6; 26.Bxd6 Kb7; 27.Re l Kc6; 28.Bf4 Nf6; 29.h4
Nd5; 30.Bcl with two branches:
30 . . . c3 ; 3 1 .bxc3 Nxc3 ; 32 .h5 Nxa2; 33.Bg5 c4 was analyzed by
Kasparov:
a b c d e f g h
[575] White to move
3 54
M ISCELLA N EOUS ENDGA M ES
34.Rdl b3; 35 .Rd8 b2; 36.Rxe8 b l Q+; 37.Kh2 Q£5; 38.Rd8 Qxf2+;
39.Kh3 Qfl+; 40.Kg4 Qe2+; (40 . . . f5+; 4l .Kh4 Qhl#) 4l .Kf4 Qf2+; 42.Ke5
Qb2+; 43 .Kf5 Qf2+; 44.Bf4 Qc5+; 45.Kf6. 34.h6 leads to 34 . . . c3; 35 .h7
b3 (35 . . . c2 ; 36. Bh6 Nc3 ; 37. Bf8 Nd l ; 38.h8Q c l Q; 3 9 . Qf6+ Kb7;
40.Qd6! ) ; 36.Bf6 Kd7; 37.h8Q Rxh8; 38.Bxh8 Ke8; 39.Kg2 c2; 40.Bb2
a5; 4l .Ba3 and Black cannot advance the pawns.
The other plan is 30 . . . Kd7; 3 1 .h5 Rxe7; 32.Rdl Kc6; 33.h6 f5; 34.Kg2
c3; 35.bxc3 Nxc3; 36.Rh l Rh7; 37.Rh"5 Nxa2; 38.Be3 c4; 39.Rxf5 c3;
40.g4 b3; 41 .g5 b2; 42 .Rc5+ Kd6; 43 .Rb5. Black can win a piece with
43 . . . c2; 44.Rxb2 c l Q; 45 .Bxcl Nxc l but the kingside pawns win. 46.Rc2
Nb3; 47.Rc3 and the rook lift is the key to victory. 47 . . . Nc5; 48.g6!
Rxh6; 49 .g7 Rg6+; 50.Rg3 Rxg7; 5 1 .Rxg7 a5 ; 52.Ra7 a4; 53.f4 with a
simple win.
Black will need both king and knight to advance the a-pawn, and
meanwhile the White king helps the f-pawn up the board. Then, when
the king and knight are far away from g8, the rook is sacrificed for the
pawn and a new queen arrives shortly thereafter.
24 ... Re8; 25 .Nd6+! Rxd6; 26.Rxd6 f6; 27.Bxf6.
a b c d e f g h
[576] Black to move
After a forced sequence, Black is faced with a decision. Should the

bishop be captured right away? Obviously it is important to get rid of
it, but in an endgame with pawns every tempo counts, and Black takes
the opportunity to improve his king position.
27 ... Kc7!; 28.Re6 Nxf6; 29.Rxf6 Rxe7; 30.Kfl . With the king still
at c8, White would play Rc6+ and either win the c-pawn or obtain a
winning pawn endgame. Since the White king can contain the Black
pawns, Black can't afford to do battle against three connected passed
pawns.
355
30 ... Re4; 3 1 .Rf4. Another critical exchange opportunity. Black has

no choice, but fortunately the resulting endgame is drawn. 3 1 .Rf3 Rd4;
32.Ke2 a5; 33.h4 a4; 34.h5 a3; 35.bxa3 c3; 36.axb4 cxb4; 37.h6 c2 looks
bad for White, faced with Black queening first and then checkmating.
There is a miracle draw, however! 38.Rf7+! Kb6; 39.Rf6+ Kb5. If the
king retreats, more checks come. The point is that Black can never
afford to interpose a rook because an exchange of rooks with check
leads to a queen endgame that is drawn. 40.Rf5+ Kb6 with a draw. or
40...Kc4; 4l.Rf8 Rd5; 42.Rc8+ Rc5; 43.Rxc5+ Kxc5 ; 44.h7 c l Q; 45 .h8Q,
which is also headed for a draw.
3 l ... Rxf4; 32.gxf4 Kd6; 33.Ke2 a5; 34.a4 c3; 35.bxc3 b3.
3
2
a b c d e f g h
[578] White to move
The player realized that the king and pawn endgame is drawn, so
the game concluded here. For the benefit of those who do not know
the defense, here it is: 36.Kd3 36.c4?? loses to 36...b236 ... c4+! ; 37.Kd2.
Obviously the king cannot capture. Now the b-pawn is safe. 37 ... Ke6;
38.h4 Kf5; 39.h5. Neither side can advance the pawns.
39 ... Kf6; 40.f3 Kf5; 4 l.Kc l etc. With this we conclude our journey
through the collection of competition-level endgames. The next sec
tion presents elementary endgames for the benefit of those who are
not yet familiar with the necessary winning and drawing techniques.
Most readers can skip to the quiz positions and workouts. I hope you
have enjoyed these examples of endgame play and that you have learned
enough to really raise your level of skill. May it bring you success at the
board!
3 56
ELEMENTARY ENDGAMES
This section is presented as a reference. The winning techniques
are shown without irrelevant branches. All you need to do is remem
ber the basic strategy.
WIN N I NG A N D DRAWI N G : TH E RULES

The most elementary aspects of the endgame involve the rules of
chess . There are three possible outcomes to a completed game. The
game can be decisive with a win for one player, or conclude in a draw.
• a game is won when:
• a player checkmates the enemy king.
• a player voluntarily resigns .
• a player exceeds the time control or violates a rule and is for
feited by the arbiter.
A game is drawn when.
• a player is in stalemate.
• both players agree to the draw.
• the identical position is about to appear or has appeared on the
chessboard three times.
• the last fifty consecutive moves have been made by each player
without the movement of any pawn and without the capture of
any p1ece.
• a checkmate cannot be constructed by any possible series of legal
moves, even with the most unskilled play.
Not all of these situations are relevant to endgame study. Players
can resign or agree to a draw in all sorts of positions, but this is never
obligatory. We must assume that in theory, the players will obey the
rules and make the required number of moves in the allotted time.
Our only concerns are:
3 57
CARDOZA PUBLISH I NG • ERIC SCH I LLER
• checkmate.
• stalemate .
• 50-move rule.
• minimum mating material.
The 50-move rule can be relevant in the amount of efficiency re
quired to gain victory in a particular position. As a practical matter, it
is most important in endgames such as Bishop + Knight vs. bare King,
since a f�w errors can make it impossible for the side with the advan
tage to win.
The requirement that minimum mating material be on the board
avoids the tedium of defending, for example, a bare king against king
and knight. Move the pieces around all you like, you can't set up check
mate.
When you have a pawn on the board, you always have mating mate
rial! That's because the pawn can promote to a queen, rook, bishop or
knight. So you just need to know the minimum checkmating force if
you are attacking a bare king (a situation where the player has a king
and no other pieces or pawns). The king and any of the following can
effect a checkmate:
• Queen.
• Rook.
• 2 Bishops.
• Bishop and Knight.
That's all. A king and two knights cannot force checkmate. When
you are defending a bad endgame, you should try to reduce the enemy
forces so that the minimum mating material requirement cannot be
met.
In the next five sections you will see how to force checkmate with
various combinations of pieces. Most readers can skip these sections,
though the Bishop + Knight vs. bare King endgame is a bit tricky and
hard to figure out at the board, so you might want to refresh your
understanding of that one.
CHECKMATE WITH QUEEN AND ROOK US. KING

This position illustrates the simple checkmating idea of driving the
enemy king to the edge of the board. With each move, White takes
away an additional rank without allowing the king to escape.
3 58
ELE M ENTA RY E N DG A M ES
SCHOLTES VS. OPPONENT

German Under- 1 1 Championship, Torgelow, 1994
3
2
a b c d e f g h
[579] Black to move
6 1 ... Kd3; 62.Qe7 Kd4; 63 .Rd8+ Kc4; 64.Qc7+ Kb5; 65.Rb8+ Ka6;
66.Qb6#.
CH ECKMATE WITH QU EEN US. KING

Checkmating the king with a lone queen is really quite simple. You'll
need to use both pieces . The main idea is to limit the moves of the
enemy king until it is forced to the edge of the board, and then bring
the king into position to support a checkmate.
OPPONENT VS. CRISPIN

Cardoza US Open, Kona, 1998
3
2
a b c d e f g h
[580] Black to move
74 ... Qf5; 75.Kh4 Qg6; 76.Kh3 Kf2 ; 77.Kh4 Kf3; 78.Kh3 Qg3#.
3 59
C H ECKMATE WITH ROOK US. KING

With just a rook, the stronger side needs to make good use of the
king in order to achieve checkmate. When the kings are on the same
rank or file, with one square separating them, then the enemy king
can be forced to retreat one rank (or file). The job of the attacking king
is to force this situation as frequently as possible. The weak king will
try to attack the rook, but the rook just flies to the other side of the
board.
OPPONENT VS. DIEHLE

German Under- 1 3 Championship, Osnabrock, 1994
3
2
a b c d e f g h
[581] Black to move
63 ... Rf8+; 64.Ke3 Kg2; 65.Ke4 Kg3; 66.Ke5 Kg4; 67.Ke6 Kg5 ;
6S.Ke7 Rfl ; 69.Ke6 Rf2; 70.Ke5 Re2+; 7 l .Kd4 Kf5; 72.Kd3 ReS; 73.Kd4
Re7; 74.Kd5 Rd7+; 75.Kc6 Rd l ; 76.Kc5 Ke4; 77.Kc4 Re i+; 7S.Kb3
ReS; 79.Kb4 Kd5; SO.Kb5 RbS+; S l .Ka6 Kc5; S2.Ka7 Rb l ; S3.KaS Kc6;
S4.Ka7 Rb2; S5.KaS Kc7; S6.Ka7 Ra2#.
CHECKM ATE WITH 2 BISHOPS

The technique for checkmating with two bishops is simple and not
much effort is needed. Just drive the enemy king to the edge of the
board, and then use a pattern like this.
3 60
ELE M ENTARY ENDG A M ES
OPPONENT VS. MILANOLLO

Feldbach, Germany, 1997
3
2
a b c d e f g h
[582] Black to move
72 ... Bd3; 73.Kc l Be2; 74.Kb l Bd2 White resigned.
3
2
a b c d e f g h
[583] Black to move
74 . . . Bd2; 75.Kal Bdl (75 . . . Bd3?? is stalemate.) 76.Kb l Bc2+; 77.Kal

Bc3#.
CH ECKMATE WITH BISHOP + KNIGHT

Knowing how to mate with bishop against knight is very important.
It is often possible to bail out of a bad game by sacrificing to reach such
an endgame hoping that a few errors will delay checkmate until after
the 50-move rule can be applied. This example features two of the best
players in the world doing battle without benefit of sight of the board.
To deliver checkmate it will be necessary to force the king into a corner
of the board \Vhich is of the same color as the bishop. That means a8 or
hl in the present example.
36 1
CARDOZA PUB LISHING • ERIC SCH I LLER
LJUBOJEVIC VS. POLGAR

Melody Amber Blindfold, Monte Carlo, 1994
3
2
a b c d e f g h
[584] Black to move
The first goal is the usual one of chasing the king onto the rim.
There is no need to concern yourself with the particular destination on
the rim. It is often necessary to drive the king from the wrong corner
(here a l or h8) to the mating square, and this can be accomplished
without too much effort if you know the pattern.
84 ...Kf6. White must retreat toward the a-file or 8th rank. Remem
ber that in pawnless endgames the rims have equal status and a retreat
is any move in the direction of the rim, as opposed to the center. 85.Kc5.
Of the available options ( c5, c7, d7), this square keeps the king furthest
from the rim. When defending, that's what you want to do.
85 ... Ke5; 86.Kc4 Bd5+; 87.Kd3 Nf4+; 88.Ke3 Be4.
3
2
a b c d e f g h
[585] White to move
This configurations forces White to retreat to the second rank.

89.Kd2 Kd4.
362
ELEMENTARY EN DGA M ES
3
2
a b c d e f g h
[586] White to move
This formation drives the king onto the rim, and phase one is com
plete.
90.Kc l Kc3 .
3
2
a b c d e f g h
[587] White to move
When the kings stand in opposition, the bishop is used to remove

one of White's options, leav ing only a single legal move.
9 l .Kd l . The difficulty is that Black cannot maintain control the
diagonal escape square. Here, e2 is under the watchful eye of the knight,
but the bishop cannot maneuver to control f2 , which is a dark square.
Therefore the knight would have to control f2 while the bishop watches
e2. Note that the Black king cannot allow the enemy to get too far
away. 9 l ... Bc2+. 9 l . . . Kd3; 92 .Kc l does not make progress.
92.Ke l Kd3 ; 93.Kf2 Ke4; 94.Kg3 Bd l .
363
3
2
a b c d e f g h
[588] White to move
This is one form of "the cage." The White king is contained to the
.
area e l -e4-h4-h l , and many of those squares can be covered by Black's
forces.
95.Kf2 Nd3+; 96.Kg3 Ke3; 97.Kh4 Kf4! The king cannot get past
h5, and is now driven toward the corner. 98.Kh3 Ne l ; 99.Kh4. 99.Kh2
is mated as follows: 99 . . . Kg4; I OO.Kgl Kg3 ; I O I .Kfl Nd3; 1 02.Kgl Be2;
1 03.Kh l N£2+; 1 04.Kgl Nh3+; 1 05.Kh l Bf3#. 99 ... Ng2+; 1 00.Kh3 Kf3;
1 0 l .Kh2.
3
2
a b c d e f g h
[589] Black to move
Black has a pleasant choice of forced mates. The key moves are
Nf4, Be2, and K£2 . It doesn't matter which is played first.
1 0 1 . . . Kf2; 1 02.Kh3 Be2; 103 .Kh2 Bg4. Taking away h3 . 1 04.Kh l .
364
ELE M ENTARY EN DGAM ES
3
2
a b c d e f g h
[5901 Black to move
If Black chooses correctly, it is mate in 3 ! 1 04 ... Ne3; 105.Kh2 Nfl+;

1 06.Kh l Bf3#.
3
2
a b c d e f g h
[5911 Checkmate
This is the position you are aiming for. It is an easy pattern to re
member, a stack of knight, king and bishop.
ROOK US. PAW N

There is quite a bit of theory on rook vs. pawn, but almost all of it
can be figured out easily at the board. Positions where promotion is
allowed are the ones you need to know.
365
CARDOZA PUB LISHI N G • ERIC SCH ILLER
LAMORELLE VS. BERNARD

French Teams, 1996
a b c d e f g h
[592] White to move
To avoid instant mate (Ra8) White must promote his pawn to a

knight.
59.h8N+ Kf6. White resigns. Any knight move allows capture, and
the only King move allows Ra8 mate.
The next example is a little trickier.
a b c d e f g h
[593] White to move
366
ELEM ENTARY ENDG A M ES
STA LEM ATE

Stalemate is the situation where it is a player's turn to move but
there are no legal moves, and there is no check. There have been many
examples throughout the book, but here is a surprising one that a World
Champion fell into.
KARPOV VS. POLGAR

Game-30 Match, Budapest, 1998
3
2
a b c d e f g h
[594] White to move
Karpov played 75.Qf3?? and the game ended in stalemate!

Now that you have finished the refresher course in basic endgames,
you are ready to tackle the endgames presented in the first part of the
book.
36l
QUIZZES
KAMSKY VS. SHIROV
3
2
a b c d e f g h
[595] White to move
Evaluate the position after 63.Kg3 Kxc5 .
NOGUEIRAS VS. HJARTARSON
3
2
a b c d e f g h
[596] White to move
3 68
QU IZZES
White can offer an exchange of rooks at f2. Is that a good idea?

Should Black accept?
PSAKHIS VS. SMYSLOV
3
2
a b c d e f g h
[597] Black to move.
What is the most efficient way to secure a draw?
TARRASCH VS. JANOWSKY
3
2
a b c d e f g h
[598] White to move
Black has just played 78 Ra4 and threatens to place the rook on
. . .
the b-file and get a queen. What should White do, and what is the
proper result of the game?
369
BOENSCH VS. JUSSUPOV

Altensteig, Germany, 1993
3
2
a b c d e f g h
[5991 White to move
Should White exchange queens with 47.Qg2?
COMAS FABREGO VS. BABURIN

Capablanca Memorial, Havana, 1 999
3
2
a b c d e f g h
[6001 White to move
Is it safe for White to capture the h-pawn? What do you think of the
position after 48.Rxh3 Rxh3; 49.Kxh3?
3 70
QUIZZES

Danish Championship, Aarhus, 1999
3
2
a b c d e f g h
[601] White to move
Does White win by capturing at h6?
WONGWICHIT VS. MCKAY

Australian Junior Championship, 1 999
3
2
a b c d e f g h
[602] White to move
Does White win by pushing the e-pawn? Evaluate 48.e5 dxe5+;

49.Ke4. See (55) for the solution.
37 1
CA RDOZA PUB LISH I NG • ERIC SCH I LLER
KARPOV VS. J. POLGAR
3
2
a b c d e f g h
[603] White to move
What move do you not play here? See diagram #594 for the answer.
GOORMACHTIGH VS. VAN BEEK
3
2
a b c d e f g h
[604] White to move
White decided to capture the h-pawn. Was that a good idea?
3 72
QU IZZES
DE WAAL VS. SADKOWSKI
a b c d e f g h
[6051 Black to move
Exchange rooks? Evaluate 38 ... Rf6; 39.Rxf6+ Kxf6; 40.Kf4.
LOPEZ VS. ALVARADO
3
2
a b c d e f g h
[6061 Black to move
Where should Black move the rook? Passive defense (c7), active
defense (d l ) or aggressive counterplay (d4)?
373
HRVACIC VS. MALADA
3
2
a b c d e f g h
[607] White to move
Which way do you recapture?
HRVACIC VS. MALADA
a b c d e f g h
[608] White to move
Evaluate the endgame after 37.Rxc5 Rxc5 38.bxc5 Kxc5.
3 'l4
QUIZZES
AKHMADEYEV VS. NENASHEV
3
2
a b c d e f g h
[609] White to move
This is another position where correct evaluation of the pawn

endgame leads to a comfortable, peaceful life . What do you play for
White?
LETAY VS. SCHUBERT
3
2
a b c d e f g h
White has just captured at c5 . Black can recapture \Vith either the
pawn or the rook, or can play a finesse, bringing the king to e7. One
line is 48 . . . bxc5; 49.Ke2 Ke7; 50.Ra6 leaves Black with only one active
plan. 50 . . . Rb7 can mix things up but White simply plays 5 l .Kd3 f5;
52.Kc3. 48 . . . Rxc5 ; 49.Rd7+ Ke6; 50.Rxa7 Rxc4; 5 l . Rb7 Rc6. This is
fairly predictable. Black has few winning prospects, unless White coop
erates. What do you think of the other options?
3 15
WORKOUTS
To improve your endgame understanding, try to solve each of the

following endgames as fully as possible . These workouts will test you
ability to evaluate positions which can arise from middlegame exchanges
or even opening variations. The three examples are representative of
practical problems facing a player from the start of the game, or even
earlier, if opening preparation is taken into account.
The first workout shows World Champions from different eras strug
gling with similar positions . You are presented with a game from the
1 999 FIDE championship which bears an eerie resemblance to a classic
game from the 1 927 contest between Alekhine and Capablanca. In the
second example, you must carefully evaluate an endgame which results
from exchanges. In the third case, you must evaluate an endgame that
arises from a relatively forced opening variation and determine whether
or not it can be won.
1 . A Classic Rook and Pawn Endgame

When one side has a healthy passed pawn and the remaining pawns
are arranged symmetrically on the opposite flank, achieving a draw
requires careful play and a very active rook. Play through the following
endgame and look for possible improvements for the defender.
The game ""�11 be repeated later with some helpful hints, but before
moving on to that, consider the methods of play demonstrated in two
similar positions.
376
WORKOUTS
AKOPIAN VS. GEORGIEV

FIDE World Championship, Las Vegas 1 999
3
2
a b c d e f g h
[611] Black to move
32 ... Ra2; 33.Kf3 Kg7; 34.Ke3 Ral ; 35.Ra6 Ra2; 36.Ra4 Ra l ; 37.Ra6
Ra2; 38.a4 Ra3+; 39.Kd4 f6; 40.Ra7+ Kh6; 4 l.a5 g5; 42.Kc5 gxh4;
43.gxh4 Ra4; 44.a6 Ra2; 45.Ra8 Kg6; 46.Kb6 Rb2+; 47.Ka7 Rxf2;
48.Rb8 Rf4; 49.Rb5 Rxh4; 50.Kb6 Re4; 5 l .a7 ReS; 52.Ra5 h4; 53 .a8Q
Rxa8; 54.Rxa8 Kg5; 55.Kc5 h3; 56.Rh8 Kg4; 57.Kd4 Kg3; 58.Ke3 Kg2;
59.Ke2 h2; 60.Rg8+ Kh3; 6 l .Kf2 h lN+; 62.Kf3 Kh2; 63.Rg2+ Kh3 ;
64.Rg6 Kh2; 65.Rxf6 Kgl ; 66.Rg6+. Black resigned.
Can you tell where Black went wrong? Is the endgame actually hope
less? Try to pinpoint the errors, if any. Then look at the classic encoun
ter from the final and decisive game of the 1 927 World Champion
ship. Here you will find some important defensive ideas, but also some
instructive errors .
ALEKHINE VS. CAPABLANCA

World Championship Match Buenos Aires (34), 1 92 7
3
2
a b c d e f g h
[612] White to move
3 'l 'l
CARDOZA PU BLISH I N G • ERIC SCH ILLER
5 1 .a5 Ra6. 5 1 . . . Rf5; 52.Ra4 and the Black rook must scurry back to
prevent the pawn from promoting. 52 . . . Rd5; 53.a6 Rd8; 54.a7 Ra8;
55 .Kf3 Kf6; 56.Ke4 Ke6; 57.Kd4 Kd6; 58.Ra6+ Kc7; 59.Kc5 Kb7; 60.Kb5
and Black has must at some point capture on a7, but the result is a loss
by a tempo. 60 . . . Rxa7; 6 1 .Rxa7+ Kxa7; 62.Kc6 when the White king
gets to the kingside. 52.Rd5 Rf6; 53.Rd4 Ra6; 54.Ra4 Kf6; 55.Kf3 .
a b c d e f g h
[613} Black to move
The king marches to the center, and then to the queenside. 55 ... Ke5.
Black's king rushes to the queenside, hoping to stand in the way of the
pawn. A king is a better blockader than a rookf 56.Ke3 h5; 57.Kd3
Kd5; 58.Kc3 Kc5; 59.Ra2 Kb5; 60.Kb3. White repeated the position to
gain time on the clock and reach time control.
60 ... Kc5; 6 l .Kc3 Kb5; 62.Kd4.
3
2
a b c d e f g h
[614J Black to 1nove
62 ... Rd6+. 62 . . . Kb4 loses to 63 .Ra l and the rook stays just out of
reach. 63.Ke5. White's king heads to the kingside, to attack the pawns .
63 ... Re6+; 64.Kf4 Ka6; 65.Kg5 Re5+; 66.Kh6 Rf5.
378
WORKOUTS
a b c d e f g h
[615} White to move
67.£4. 67.Kg7 was more efficient. Alekhine gives 67 . . . Rf3; 68.Kg8

Rf6; 69.Kf8! Rf3; 70.Kg7 Rf5; 7 l .f4. 67 ... Rc5; 68.Ra3 Rc7; 69.Kg7 Rd7.
The rook tries to defend the pawn at f7. Not so simple!
3
2
a b c d e f g h
[61 6} White to move
70.£5?! 70.Kf6 was stronger. 70 . .. Rc7; 7 1 .f5 gxf5; 72 .Kxf5 ReS+;

73 .Kf6 Rc7; 74.Rf3 Kxa5; 75.Rf5+ would have won cleanly, according
to Alekhine.
70 ... gxf5; 7 1 .Kh6 f4; 72.gxf4 Rd5; 73.Kg7 Rf5; 74.Ra4 Kb5; 75.Re4.
3 79
CARDOZA PUB LISHING • ERIC SCH ILLER
a b c d e f g h
This is the key. White abandons the passed pawn but since the
Black king is so far away, the Black rook is no match for the White
forces. 75 ... Ka6; 76.Kh6 Rxa5. Or 76 . . . Kb7; 77.Re5 Rxf4; 78 .Kg5 Rfl ;
79.Kxh5 f5; 80.Kg5 f4; 8 1 .Rf5 f3; 82.Kg4 and White gets the f-pawn.
77.Re5 Ral ; 78.Kxh5.
3
2
a b c d e f g h
78 ... Rgl ; 79.Rg5 Rh l ; 80.Rf5 Kb6; S l .Rx£7 Kc6; 82.Re7. Black

resigned. White won the game, and the World Championship title!
This type of endgame has been discussed by theoreticians, too.

Consider the drawing method shown by the great Reuben Fine in his
Basic Chess Endings.
380
WORKOUTS
a b c d e f g h
[619} White to move
l .Kf3 h5 2.h4. 2.Ke4 Rxf2; 3.Rc8 Ra2; 4.Rc6 (4.Ra8? Rxh2; 5.a7
Ra2 and Black will win.) 4 . . . Rxh2; 5.Kd5 Ra2; 6.Kc5 g5; 7.Kb6 h4; 8.gxh4
gxh4; 9.a7 h3; 1 0. Kb7 h2; l l .Rcl Rb2+; 1 2 . Ka6 Ra2+; 1 3.Kb7 is even.
2 ... Kf6; 3.Ke3. 3.Ke4 Rxf2; 4.Rc8 Re2+! ; 5.Kd5?? Ra2; 6.Rc6+ Kf5;
7.Kc5 Kg4; 8.Kb6 Kxg3 when Black has all the winning chances.
3 ... Kf5.
3
2
a b c d e f g h
[620} White to move
4.£3 Ra3+; 5 .Kd4 Rxf3; 6.Rf8 Ra3 ; 7.Rxf7+ Kg4; 8.Rf6 Kxg3 ;
9.Rxg6+ Kxh4; I O.Kc5 Kh3; l l .Kb6 h4; 1 2.a7 Kh2 ; 13.Kh7 Rxa7+;
14.Kxa7 h3; 15.Kb6 Kh l .
38 1
CARDOZA PU B LIS H I N G • ERIC SCH I LLER
3
2
a b c d e f g h
[ 621} White to move
1 6.Kc5 h2 forces the rook from the g-file, or else stalemate results.
You should now have enough background to re-examine the

Akopian-Georgiev game, which is presented below with some light notes
to guide your analysis.
AKOPIAN VS. GEORGIEV

FIDE World Championship, Las Vegas 1 999
3
2
a b c d e f g h
[622} Black to move
32 . .. Ra2; 33.Kf3 Kg7; 34.Ke3 Ra l . This is too slow. Black should

have activated the king. 34 . . Kf6; 35.Ra7 Ke5 .
.
3 82
WORKOUTS
a b c d e f g h
[623] White to move
There are three branches to consider:

36.£3 Rg2; 37.Rxfi Rxg3; 38.a4 Rgl ; 39.a5 Ral ; 40.Ra7 Ra3+; 4 l .Ke2
Ra4! is no problem for Black.
36.a4 Ra3+; 37.Ke2 f6; 38.a5 Kf5; 39.a6 Kg4; 40 .Ra8 f5; 4 l .a7 Ra2+;
42 .Kfl (42 . Kd3 Kf3) 42 . . . Kf3 (42 . . . Kh3 ; 43.Rg8 Rxa7; 44.Rxg6 f4;
45 .gxf4 Kxh4; 46.f5 R£7; 47.f6 Kh3; 48.Ke2 h4; 49 .f4 Kh2; 50.Kf3 h3;
5 l .Rh6 Rf8; 52.f5 Rf7; 53.Kg4 Kg2; 54.Rxh3 Rxf6; 55.Ra3 and White
wins. ) 43 .Re8 Ra l+; 44.Re l Rxa7; 45.Re3+ Kg4; 46.Kg2 f4 (46 . . . Ra2;
47.Re6) 47.Re4 Kf5 (47 . . . Rf7; 48.Re5! ) 48.Rxf4+ Ke5 and White, though
better, has few realistic winning chances .
36.Ra5+ is a good practical try. Black can try 36 . . . Kd6; 3 7 .a4 Kc6,
but then 38.Ra7 f5 (38 . . . Kb6; 39.Rxf7 Rxa4; 40.Rf4 and the White
king will infiltrate the queenside. ) 39.Ra6+ Kd5; 40.Rxg6 Rxa4; 4 1 . Rg5
Ke5; 42 .Rxh5 Ra3+; 43 .Ke2 Ke4; 44. Rg5 and White wins by promoting
the h-pawn. A better defense is 36 . . . Ke6; 3 7 .a4 f6, which is not as bad
for Black as the position reached in the game at move 39.
35.Ra6.
3
2
a b c d e f g h
[624] Black to move
383
35 ... Ra2; 36.Ra4 Ral ; 37 .Ra6 Ra2; 38.a4 Ra3+; 39.Kd4 f6; 40.Ra7+
Kh6.
a b c d e f g h
[625} White to move
After time control is reached, Black is left with a very difficult situ
ation, because the White king can quickly approach and defend the a
pawn. 4 1 .a5 g5; 42.Kc5 gxh4; 43.gxh4 Ra4; 44.a6 Ra2 ; 45.Ra8 Kg6;
46.Kb6 Rb2+; 47.Ka7 Rx£2.
3
2
a b _c d e f g h
[626] White to move
48.Rb8. The construction of the bridge begins. 48 ... Rf4; 49.Rb5

Rxh4; 50.Kb6 Re4; 5 l .a7 ReS; 52 .Ra5 h4; 53.a8Q Rxa8; 54.Rxa8.
384
WORKOUTS
3
2
a b c d e f g h
54 ... Kg5; 55.Kc5 h3; 56.Rh8 Kg4; 57 .Kd4 Kg3; 58.Ke3 Kg2; 59.Ke2
h2; 60.Rg8+ Kh3; 6 1 .Kf2 h lN+; 62.Kf3 Kh2 ; 63.Rg2+ Kh3; 64.Rg6
Kh2; 65.Rxf6.
3
2
a b c d e f g h
[628] Black to move
This is the familiar winning position in rook vs . knight. 65 ... Kgl ;

66.Rg6+. Black resigned.
385
2. Evaluation of the Endgame

This game saw two of the Cardoza Whiz kids battling for the title
of United States Cadet Champion. The young players both showed ex
cellent judgement, in a very complex position.
WANG VS. MONT-REYNAUD

US Cadet, Nashville, 1 999
a b c d e f g h
[629] Black to 'move
The position is not so simple. Black must avoid attacks and possible
perpetual checks against the exposed king. 25 ... Rc2. 25 . . . exd6; 26.Qxh6
Rc2; 27.Qg5+ draws.
26.dxc7 Rxd2; 27.Rfdl Rxd l ; 28.Rxdl .
3
2
a b c d e f g h
Black has two extra pawns in the endgame, but his pieces are tied
do\vn. In fact, Black is fighting for a draw!
28 ... Bg4; 29.Rd4 Bc8.
3 86
WORKOUTS
a b c d e f g h
{63 1J White to move
30.Rb4. 30.Rd8 ! ? e6 (30 . . . f6; 3 l .Bf3! K£7; 32.Bh5+ Kg7; 33.Be8!

Be6; 34.c8Q Bxc8; 35.Rxc8 may win for White, unless Black can get rid
of the remaining pawns.) 3 l .Bf3 a6; 32.Bd l ! b5; 33.Bf3 Kg7; 34.Bc6
e5 !
Black needs the move to make room for the bishop to temporarily
escape, so that it can return to c8 to capture the promoted c-pawn.
White will have an extra bishop, just as in the game. 35.Be8 Be6; 36.c8Q
Bxc8; 37.Rxc8 Kf6; 38.Bd7 Rxc8; 39.Bxc8 aS; 40.Bd7 b4 and Black
survives because White cannot hang on to the pawns. 4 1 . Kg2 Kg5;
42.Kf3 f5 ; 43 .Bc6 e4+; 44. Ke3 Kg4; 45 .Kf2 f4; 46.gxf4 Kxf4; 47.Be8
e3+; 48 .Ke2 Ke4 should draw.
30 ... b6. White has provoked this weakness in order to get the bishop
onto the key a4-e8 diagonal. 3 l .Rd4 f5; 32.Rd8 Kf7; 33.Bc6! a6; 34.a4
h5; 35.Kg2 Kg7; 36.Kf3 .
3
2
a b c d e f g h
{632] Black to move
36 ... f4! Black heads for a drawn rook and bishop vs. rook endgames,
\Vith a pair of irrelevant stray Black pa,vns. 37 .gxf4 h4; 38.Kg2.
3 87
3
2
a b c d e f g h
[633] Black to move
38 ... b5 ! The remaining White pawns are eliminated. 39.axb5 axb5;

40.Bxb5 Bb7+; 4 l .Kh2 Bc8; 42.Be8 Be6; 43.c8Q Bxc8; 44.Rxc8 Rxf4.
3
2
a b c d e f g h
[634] White to move
45.Bh5 Kh6; 46.Bdl Rd4 and the game was agreed drawn.
3. The Endgame and the Opening

When preparing an opening repertoire, it is often necessary to care
fully evaluate the winning or drawing chances in endgames that can
arise by force . A. classic example is the Tarrasch Defense to the Queen's
.
Gambit. Your task is to evaluate the following, topical position. By ex

amining the sample continuations, you can come to a conclusion on
the playability of the line. If you believe that Black holds, then there is
nothing to fear from White. On the other hand, if Black's position is
too weak, you wouldn't want to play it.
3 88
WORKOUTS
TARRASCH DEFENSE - CLASSICAL VARIATION

l.d4 d5; 2.c4 e6; 3.Nf3 c5; 4.cxd5 exd5; 5.Nc3 Nf6; 6.g3 Be7; 7.Bg2;
0-0; 8.0-0 Be6; 9.dxc5 Bxc5; I O.Bg5 Nc6; l l .Bxf6 Qxf6; 1 2.Nxd5
Qxb2; 1 3.Nc7 Rad8; 14.Qc l Qxc l ; 15.Raxc l Bb6.
15 . . . Be7 is a significant alternative, but we will concentrate on the
more active continuation.
16.Nxe6 fxe6.
a b c d e f g h
[635] White to move
17 .Rc4. White has explored alternatives here.

17 .Bh3 launches a direct assault on the e-pawn. 1 7 . . . Rf6 is one plan.
1 8.Ng5 Nd4; 1 9 .Kg2 h6; 20 .Ne4 Rf7; 2 1 .Rfd l Rfd7; 22 .Nc5 Bxc5 ;
23.Rxc5 Kf7; 24.Bg4 Kf6; 25.Rb 1 b6; 26.Rc4 Nf5; 27.e4 Rd4; 28.Rxd4
Rxd4; 29.Re 1 Ne7; 30.e5+ Kfi; 3 1 .Bh5+ g6 gave Black a fine game in
Palatnik-Pinal, Havana 1 985. 1 7 . . . Rfe8 is also a reasonable defense.
1 8.Rfd l ( 1 8.Ng5 Nd4; 1 9.Bg4 Rd5; 20.h4 h6; 2 1 .Ne4 h5; 22.Bf3 Nxf3+;
23.exf3 Rd3; 24.Kg2 e5 ; 25.Rc2 Rf8; 26.Ng5 Rfd8; 27.Re 1 Rd2; 28.Re2
Rxc2; 29.Rxc2 g6 worked out well for Black in Kharitonov-Kouatly,
Royan 1 988.) 1 8 . . . Kfi; 1 9.e3 h6 20.Bf1 Rxd1 ; 2 l .Rxd l Rd8; 22.Rxd8
Bxd8; 23 .Nd4 Nxd4; 24.exd4 Bf6; 25.Bg2 b5; 26 .Bc6 b4; 27.d5 was
drawn in Vilela-Pfleger, Havana 1 982.
1 7 .e3 controls d4, but gives Black time to set up a solid defense.
1 7 . . . h6; 1 8.Rc4 Rd6; 1 9 .Re4 ( 1 9.h4 Rf5; 20.Rb l Kf7; 2 1 .a4 Ke7; 22.Re4
ReS; 23.Bh3 g5; 24.hxg5 hxg5; 25.Bg4 Kf6; 26.Kg2 Ne5; 27.Rhl Nxg4;
28.Rxg4 put too much pressure on the g-pawn in Horvath-Brestian,
Germany 1 992 .) 1 9 . . . Bc5; 20.h4 b6; 2 l .Rb l a5; 22.Rb2 Rd l+; 23.Kh2
was played in Chesarsky-Damjanovic, Tel Aviv 1 99 1 . Black could equal
ize with 23 . . . Nb4 ! ; 24.Rxe6 Ra l but instead Black chose to defend the
e-pawn and fell into a passive position which was quickly lost.
Finally, back at the start of the 20th century, 1 7.Rfd 1 Rxdl +; 1 8.Rxd l
3 89
CA RDOZA PUB LISHI N G • ERIC SCH I LLER
Rd8; 1 9 .Rxd8+ Nxd8 was agreed drawn in Rubinstein-Schlechter,

Prague 1 908.
a b c d e f g h
[636] Black to move
17 . . . h6.
1 7 . . . Nd4; 1 8.Nxd4 Rxd4; 1 9.Rxd4 Bxd4; 20.Bxb7 Rb8; 2 1 .Ba6 is a
difficult endgame for Black, Alburt-Meduna, Decin 1 977.
1 7 ... Rd6; 18.h4 h6; 1 9.Re4 ( 1 9.Rfc1 e5; 20.Re4 Re8; 2 1 .e3 Kf7 ; 22.h5!
Bd8; 23.Bf1 Bf6. Black is playing passively and gets punished. 24 .Bb5
a6; 25.Bxc6 Rxc6; 26.Rxc6 bxc6; 2 7 . Ra4 Ra8; 28.Nd2 Ke6; 29.Nc4 Ra7 ;
30.Ra5 . Black resigned in Kochiyev-Kiik, Espoo 1 990. Can you find an
improvement for Black?) 1 9 . . . Nd4!? ( 1 9 . . . Rfd8; 20.e3 Kf7; 2 1 .Bh3 R8d7;
22.Kg2 Bd8; 23.Rc l Bf6; 24.a4 was a bit better for White in Dementiev
Kuindzhi, Lithuania 1 977.) 20.Nxd4 Bxd4; 2 1 .e3 Bc5; should be fine
for Black. 1 7 . . . Rd5 ; 1 8.Rfc 1 Rfd8; 1 9 .Bh3 K£7; 20.e3 Ke7; 2 1 .Kfl h6;
22.Ke2 g5; 23.R1c2 h5 ; 24.Bg2 Ba5; 25 .g4 hxg4; 26.Rxg4 Kf6; 27.h4
gxh4; 28.Rxh4 e5 gave Black a solid game in Pohl-Behrhorst, Germany
1 984.
18.Re4.
a b c d e f g h
3 90
WORKOUTS
18 ... Rf6. 1 8 . . . Rfe8; 19 .e3 Rd3 ( 1 9 . . . Bc5; 20.Rc 1 b6; 2 1 . Bf1 Rc7;
22.Bb5 Na5; 23.Ne5 Rd2; 24.Ra4 Rd5; 25.Ng6 Rc7; 26.h4 Bd6; 27.Rxc7
Rd 1 +; 28.Kg2 Bxc7; 29.Be2 Rd2; 30.Bh5 would have been fine after
30 . . . Bd6, but Black went horribly wrong and lost after 30 . . . Kf7; 3 1 .Rg4
K£6; 32.Nf8 g5; 33.hxg5+ hxg5 ; 34.Nh7+ Kg7; 35.Nxg5 Kh6; 36.Ne4
Rd5 ; 37 .Bf7 Be5; 38.Rg6+ and Black resigned in Quinteros-Hase from
the 1 980 Argentinian Championship. ) 20.Bh3 e5; 2 1 .Rbl Kf7; 22.Bf1
Rd6; 23.h4 g6; 24.h5 gxh5 ; 25.Be2 Kf6; 26.Kg2 Rg8; 27.Rh4 Kc7;
28.Rxh5 left Black's pawns too weak in Magai-Hassan, Moscow 1 994.
19.h4. 1 9 .e3 Kf7 ( 1 9 . . . Bc5; 2 0 .Rb 1 b6; 2 l . Kfl a5; 22.Rb2 Nb4;
23.Nd4 e5; 24.Nb3 Rd l +; 25 .Ke2 Rxf2+; 26.Kxf2 Nd3+; 27.Ke2 Nxb2;
28.Nxc5 bxc5; 29.Rxe5 Re i +; 30.Kxe l Nd3+; 3 1 .Kd2 Nxe5 was eventu
ally drawn in Nogueiras-Vera, Havana 1 983. ) 20.Rb l Rd5; 2 l .Kfl Ke7;
22.Ke2 Kd6; 23.Rg4 R£7; 24.Rb2 Ne5 ; 25 .Nxe5 Rxe5; 26.Rd2+ Kc7;
27.Rg6 Ref5; 28.f4 R5f6; 29.Rxf6 gxf6; 30.Bh3 Re7; 3 1 .Kf3 Bc5; 32 .e4
e5 was about even in Vukic-Muller, Varna 1 975.
19 ... Rd5.
a b c d e f g h
[638} White to move
20.Bh3 . 20.Rc l Rb5; 2 l .Bh3 K£7; 22.Rdl Ke7; 23.Rd2 Rb4; 24.Rxb4
Nxb4; 25.Kg2 Rf8; 26.Rb2 Nc6; 27.e3 Rd8 and White was unable to
make progress in Maximenko-Vonthron, Odessa 1 990.
20 ... Kf7; 2 1 .Kg2 Ke7; 22.Rc l Kd6; 23.e3 Ne5 ; 24.Nxe5 Rxe5 ;
25.Rxe5 Kxe5.
39 1
CA RDOZA PUB LIS H I N G • ERIC SCH I LLER
a b c d e f g h
[639} White to move
Given the bishops of opposite color, this endgame from Ribli-Espig,

Budapest Tungsram 1 975 should be drawable. Whether you choose to
adopt the variation as Black will depend crucially on your level of com
fort in the endgames presented above.
Some Last Thoughts

You have completed three exercises in endgame workouts. You can
continue your education by examining endgames in other games and
other openings, seeking as much information as you can before mak
ing an evaluation. This activity should be done at least once a week, so
that your endgame skills will improve quickly.
I hope you have enjoyed this excursion into the territory of the
endgame, and that you have acquired at least some of the knowledge
you will need to successfully navigate your endgame challenges in tour
nament play.
392
N EVV - C H E SSC ITY MAGAZI N E
Free Online Chess Magazine
CHESS CITY is a sprawling metropolis of chess information, a magazine
with the latest news and analysis, to gossip, trivia, and fun features. Travel
around the world to visit the most fascinating chess competitions, preview
books long they hit the shelves, and read columns on openings, middlegames,
endings, tactics, strategies, mates , and more .
Extensive excerpts from our books are available online. Visit often, be
cause we'll be adding more features for your pleasure r Improve your chess
knowledge with our articles and features on the opening, middlegame,
endgame, s trategy, tactics checkmates and more ! Whether you' re a begin
ner or master, you'll be able to improve your results with our tips.
Chess is a serious game, but i t is also a lot of fun . Chess City Magazine
p resents trivia, photos, anecdotes, chess art, strange games, trivia and even
a bit of gossip for your amusement and pleasure r Drop by often when you
need a break!
Chess City is a hit - one million times over! Chess players around the
world are tun ing in to this great chess community. We've had more than one
million hits in the first six months and are getting more popular daily!
Go to www .chesscity.com for details
393
CARDOZA PUBLISHING CHESS BOOKS
- OP E N I NGS -
WI N N I N G C H ESS O P E N I N G S by Bill Robertie - S hows conce pts and best open ing
moves of more than 25 essential openi ngs from Black's and White's p e rspectives:
King's Gam bit, Center Game, S cotc h G am e , G i ucco Piano, Vien n a Game, B i shop's
Ope n i n g , Ruy Lopez, Fre nch , Caro- Ka n n , Sici l ian , Alekh i n e , P i rc , Modern , Queen's
Gam bit , N i mzo- l ndian, Que en's I ndian , Dutch , King's I ndian, Benon i , E nglish , B i rd's ,
Reti's, and King's I ndian Attack. Exam p l es from 25 g randmaste rs and champions in
c l u d i ng Fischer and Kas pa rov. 1 44 pages , $9.95
WO R L D C H A M P I O N OPENINGS by Eric Schiller - This serious reference work cov

e rs the essential opening theory and m oves of eve ry maj o r chess opening and va ria
tion as played by all the world champions. Reading as m uch l i ke an encyclopedia of
the. m ust-know open i ngs c rucial to every chess playe r's knowledge as a powerfu l tool
s howi ng the i n s ights , co ncepts and secrets as used by the g reatest players of all time,
World Champion Openings (WCO) cove rs a n astou nding 1 00 c rucial openi ngs i n fu l l
co nceptual detail (with 1 00 actual g a m e s f ro m the c h a m p i o n s themselves) ! A must
ha ve book for serious chess players. 384 pages, $ 1 8 . 95
STA N DA R D C H ESS O P E N I N G S by Eric Schiller - The new defin itive standard on

ope ning ch ess play i n the 20th ce ntu ry, this com p re h e ns ive guide cove rs eve ry i m por
tant chess opening and va riation ever played and cu rre ntly in vog u e . I n al l , m o re than
3 , 000 ope n i ng strategies a re p resented ! D iffe ri ng f rom p revious open i n g books which
rely almost exclusive ly on b a re n otat io n , SCO featu res s u bstantial d iscuss ion and
analysis on each o p e n i ng so that you learn a nd u nd e rstand the concepts behi nd them.
I ncl udes more than 250 com pl etely an notated games ( i ncluding a game rep rese nta
tive of each major openi ng) and m o re than 1 , 000 dia g rams! For modern playe rs at any
l eve l , this i s the sta nda rd refe rence book necessary fo r competitive pl ay. A must have
for serious chess players!!! 768 pages, $24 . 95
U N ORTHODOX C H ESS O P E N I N G S by Eric Schiller - The exciting gu ide to a l l the

maj o r u no rthodox ope n i n gs used by chess p l ayers , contains m o re t han 1 , 500 we i rd ,
contenti ous, controve rsial , u nconventi ona l , arrogant, a n d outright strange open ing strat
egies. From the i r tricky tactical su rprises to their b izarre names, these openi ngs fly i n
t h e face of t rad ition. Yo u ' l l m eet such openi ngs a s t h e Orangutang, Raptor Variation,
Halloween Gambit, Double D uck, Frankenstein-Dracula Variation , and even the D ru n ken
King! These openi ng s a re a sexy and exotic way to spice up a game and a g reat
weapon to spring on u ns uspecting and often u n p repa red opponents . More than 750
diagrams s how essential pos itions. 528 pages , $24.95
G A M B IT O P E N I N G R E P E RTO I R E FOR WHITE by Eric Schiller - Chessplayers who

e njoy attacki ng from the ve ry fi rst move are rewa rd ed h e re with a powe rfu l re pertoire
of b ri l l i ant gamb its . Starting off with 1 . e4 or 1 .d4 and th en using such s harp weapons
such as the Goring G amb it (Accepted and Decli ned) , Halasz Gam bit, Alap i n G am b it ,
U lysses G a m b i t , Short Attack and many mo re , t o put g reat pressu re o n opponents ,
Schi l l er p resents a complete attacking repertoire to use agai nst the most popular de
fenses , i ncluding the Sicil ian , F rench , Scandi navia n , Caro-Ka n n , P i rc , Alekh i n e , and
oth er Open Game positions . 1 92 pages , $ 1 4 . 9 5 .
394
CA RDOZA PU BLIS H I N G CH ESS BOOKS
GAM B IT O P E N I N G R E P E RTOI R E FOR BLAC K by Eric Schiller - Fo r playe rs that

l i ke exciting no-holds-barred chess, this ve rsati le gam b it repertoire shows B lack how
to take charg e with agg ressive attacking defe nses agai nst any o rthodox fi rst Wh ite
ope n i n g move ; 1 . e4, 1 . d4 and 1 . c4 . Learn the Scandi navian G am b it agai nst 1 . e4 , the
Schara G am bit and Q u e en's G am b it Dec l i ned va ri ations agai nst 1 . d4, and some flank
and unort hodox gambits also . Black learns the secrets of seizing the i n itiative from
Wh ite's hands, usually by i nvesting a pawn or two, to b eg i n powerful attacks that can
send White to early d efeat. 1 76 pages , $ 1 4 . 95 .
COM P LETE D E FE N S E T O Q U E E N PAWN O P E N I N G S by Eric Schiller - T h i s ag

g ress ive co u nterattacking rep e rtoi re cove rs Black opening systems agai nst vi rtually
eve ry ch ess ope n i n g except fo r 1 . e4 (i ncl u d i n g most flan k gam es) , based on the excit
ing and powe rfu l Tarrasch Defense , an open i ng that h e l ped b ring Championsh ip titles
to Kasparov and Spassky. B l ack l ea rns to effective ly use the Cl assical Ta rrasch , Sym
metrical Tarrasch , Asymmetrical Tarrasch , Marshall and Ta rrasch Gambits , and Ta rrasch
without Nc3, to ac h i eve an early equal ity or eve n an outright advantage in the fi rst few
moves . 288 pages, $ 1 6 . 9 5 .
COMP LETE D E F E N S E T O KI N G PAWN O P E N I N G S by Eric Schiller - Learn a co m

pl ete d efe nsive syste m agai nst 1 . e4. This powerful re pertoi re not o n ly l i mits White's
abil ity to obtai n any significant opening advantage but a l l ows Black to adopt the fl ex
ible Caro-Kan n fo rmatio n , the favo rite weapon of many of the g reatest chess playe rs .
All White's options are explai n ed i n deta i l , and a p l a n is g i v e n fo r Black to com bat t h e m
a l l . Analys i s is up-to-date and backed b y exam ples d rawn from g a m e s o f t o p stars.
Detailed i ndex lets you fol low the opening from the point of a specific playe r, o r through
its h isto ry. 240 pages, $ 1 6.95.
SECRETS OF THE SICI LIAN DRAGON by GM Eduard Gufeld and Eric Schiller - The
m ig hty Dragon Va riation of the Sicilian Defense is one of the most exciting openi ngs in
chess. Eve rything from open i ng piece formation to the endgam e , i ncluding clear ex
planations of all the key strategic and tactical ideas , is cove red i n full conceptual detai l .
I nstead of mem orizing a j u n gle o f variations, you learn t h e really i m p o rtant ideas be
h i n d the op e n i n g , and how to adapt them at the chessboard . Special secti ons on the
h e roes of the D ragon s how how the g reatest playe rs handle the ope n i ng . T h e most
i nstructive book on the D ragon writte n ! 208 pages, $ 1 4.95.
H Y P E R M O D E R N OPE N I N G R E P E RTO I R E FOR WH ITE by Eric Schiller - I nstead of

placing pawns in the center of the board as t raditional openings advise, this co mplete
ope n i n g re pertoire fo r W h ite shows you how to st un opponents by "al low ing" B l ack to
occu py the cente r with its pawn s , wh i l e b u i l d i n g a c ru s h i n g phalanx from the flanks,
ready to s mash the cente r apart with Black's s l i g htest m istake . White's approach is
si mple to lea rn-because Wh ite almost always deve l ops pi eces in the same manner
but can be used agai nst all d efenses no matte r what B lack pl ays ! Plentifu l diagrams
and exp lanations i l l ust rate eve ry concept, with games f ro m the g reatest playe rs show
i ng the principles in act i o n .The Reti and English openings, which form the basis of the
Hyp e rmodern, l ead to l ively games with b ri l l iant sacrifices and s u btle maneuveri n g .
304 pages, $ 1 6 . 95 .
SECRETS O F TH E KI N G 'S I N D I AN by Eduard Gufeld and Eric Schiller - T h e Ki ng's

I ndian is the s i ng l e m ost popular opening and offe rs g reat opport u n ities fo r spectac u
lar attacks and clever d efenses . Readers l ea rn the fundamental concepts , critical ideas,
and h idden resou rces along with the open ing traps and typ ical tactical and strategic
m istakes. All major variations are covered , i ncluding the Classical , Petros ian, Saemisc h ,
395
CARDOZA PUB LIS H I NG CH ESS BOOKS
Ave rbakh , Fo u r Pawns , Fianchetto and u nconventional l i nes . P l aye rs learn how the
strate gies and tactics were applied in the b ri l l iant games of the m ost famous playe rs ,
how they can apply them t o th e i r own gam e . 240 pages, $ 1 4 . 95 .
- M I DDLEGAMEITACTICS/WI N N I N G CONC EPTS -
W I N N I N G C H ESS TACTI CS by Bill Robertie - 1 4 chapte rs of w i n n i n g ideas show the

complete th i n ki n g b e h i n d eve ry tactical concept: p i ns , single and double fo rks , double
attacks , s kewers , d iscove red and double c h ecks , m u ltiple th reats - and oth e r crus h i n g
tactics t o gain an i m mediate e d g e ove r o ppon ents . Learn the power tools o f tactical
p l ay to beco m e a bette r player. I nc l u des g u i d e to chess notation. 1 28 pages , $ 9 . 9 5 .
3 0 3 TRIC KY CH ESS TACTICS Fred Wilson a n d Bruce A /berston - Both a fasci nati ng
challenge and g reat t raining too l , this is a fun and e nte rta i n i n g co l l ection of two and
th ree move tactical s u rprises for the advanced beginner, intermediate , and expert playe r.
Tactics a re arranged by d ifficu lty so that a pl aye r may m eas u re p rog ress as he ad
vances from simple to th e complex positi ons . The examples, d rawn from actual games ,
i l l u strate a wide range of chess tactics from old classics right up to the 1 990's . 1 92
pages, $ 1 2 . 95.
1 0 MOST COMMON C H ESS M I STA KES and How to Fix Them by Larry Evans
This fasci nati ng col l ection of 2 1 8 e rro rs , ove rsights , and outright b l u nd e rs , not only
shows th e price g reat playe rs pay for violating basic principles , b ut how to avo id these
mistakes i n you r own game. Yo u ' l l be challenged to choose between two moves , the
right one , or the one actually p l ayed. F rom negl ecting d evelopment, king safety, mis
j u d g i ng th reats , and p rematu re attac ks , to impulsiveness , snatc h i n g pawns, and basic
i n attentio n , you receive a complete cou rse i n where you can go w rong and how to fix
it. 256 pages , $ 1 4 . 95 .
E N CYC LO P EDIA O F C H ESS W I S D O M , T h e Essentia l Concepts a nd Strategies

of Smart Chess P lay by Eric Schiller - The most i m portant concepts , strategies, tac
tics , wisdo m , and th i n king that eve ry chessplaye r m ust know, plus the gold n u gg ets of
knowledge b e h i n d eve ry attack and defe nse is col lected tog eth e r in one vol u m e . Step
by-step , fro m ope n i n g , m i d d l e and en dgame strategy, to psychological warfare and
tournament tactics , Sch i l l e r s hows th e t h i n king behind each essential concept, and
th rough examples , diagrams , and d iscussions, shows its i m pact on the game . 432
pages , $ 1 9 . 95 .
W O R L D C H A M P I O N C O M B I N ATIONS by Raymond Keene and Eric Schiller - Learn

th e i nsig hts, concepts , and m oves of the g reatest co mbinations eve r by the best play
e rs of all ti m e . From Mo rphy to A l e kh i n e , to Fisch e r to Kasparov, the i n cred ible combi
nations and b rill iant sacrifices of th e 1 3 World Champions a re col lected h e re i n the
most i ns i g htf u l co m binations book yet. Packed with fasci n ating stratagems, 5 0 anno
tated games , and g reat p ractical advice for yo u r own games, this is a g reat compan ion
g u ide to World Champion Openings a nd the oth e r titles i n the World Champion series.
264 pages , $ 1 6 . 9 5 .
WO RLD C H A M P I O N TACTI CS by Leonid Shamkovich and Eric Schiller - T h e authors

show how the g reatest playe rs who eve r l ived used the i r enti re arse nal of tactical
weapons to b ring opponents to the i r knees . Packed with fasci n ating st ratag ems , 50
fu l ly an notated games , and mo re than 200 d i ag rams, players l earn not o n ly the think
i n g and game plan behind the moves of the champions , b ut the i nsights that will allow
them to use these b ri l l iancies in thei r own games . Each tactical concept is f u l ly ex
plai ned with examples and game situations from the champions themselves . 304 pages,
$ 1 8 .95.
396
CA RDOZA PUB LISH ING CH ESS BOOKS
WO RLD CHAMPION C H ESS M ATC H ES by Eduard Gufeld and Efim Lazere v - The
exciti ng h i g h l ig hts and g reat games of eve ry Wo rld Champions h i p chess m atch are
reco u nted in G ufeld's l ively style, with an ecdotes, l ittl e-known sto ries and h u n d reds of
key positio n diagrams. G ufeld does much m o re than p resent the i n-d epth analysis and
th i n king b e h i nd the m oves of the g reatest of the g reats: G u ru s , parap hsycologists ,
dirty tricks, psychological wa rfare and all the side shows that have b rought worldwide
media attention to each contest are b rought to l ife. G ufeld s h o u l d know these secrets ;
he has played with the ve ry best fo r decades , a nd has to his c redit the scal ps of th ree
fo rm er world champions and e i g ht wo rld championship can didates. 304 pages, $ 1 8 . 95
AW ESOM E C H ESS MOVES By Eric Schiller - This collection of b ri l l iant ideas from
real tou rname nts a re not just re g u l ar com b i nati ons or tacti cal swindl es , but m oves of
stu n n i n g o ri g i nal ity. Sch i l l e r has selected 1 00 awesome m oves, a nd t h ro ug h game
pos itions, examples, and clearly explai ned concepts , s hows players how to i m p rove
the i r g rasp of deep positio nal u nderstand ings and swas h b uckl ing tactics. Readers
l earn how to re i nfo rce their g ut i nsti ncts to not just reach for the best move , but the
·
inspired move . 224 pages , $ 1 8 . 9 5 .

- B EG I N N I N G A N D G E N ERAL C H ESS BOO KS -
TH E BASICS OF WI N N I NG C H ESS by Jacob Cantrell - A g reat fi rst book of chess , in

one easy read i n g , beg i n ner's l ea rn the m oves , pieces, basic ru les and principles of
play, standard ope n i n gs , and both Alge b raic and E n g l i s h chess notation . The basic
i deas of the wi n n i ng concepts and strategies of m iddle and end game play a re also
shown . I ncl udes example games of champi ons . 64 pages, $4 .95.
B E G I N N I N G CH ESS P LAY by Bill Robertie - Step-by-step app roach uses 1 1 3 dia

g rams to teac h the basics of chess: ope n ing , middle and endgame strategies, p ri n
ciples o f development , pawn st ruct u re , checkmates , openings and defenses , how to
write and re ad c hess notatio n , j o i n a c h ess c l u b , p lay in tou rnaments , use a c hess
clock, and get rated. Two annotated games i l l l u strate strategic th i n king fo r easy learn
ing. 1 44 pages , $9 . 95
WHIZ KIDS TEAC H C H ESS Eric Schiller & the Whiz Kids - Today's g reatest you n g
stars , s o m e pe rhaps to be futu re world champions, p resent a fasci nating look a t t h e
world o f chess. Each tells of the i r s uccesses , fai l u re s , world trave ls, a n d love o f chess ,
show off their best m oves , and ad mit to thei r m ost e m barrass ing b l u nders . This is
m o re than j ust a fasci nating l ook at prodigies l i ke Vi nay Bhat and I ri na Krush , it's also
a p ri m e r featu ri ng d i agrams, explanations, and w i n n i n g ideas for yo u n g playe rs . 1 44
oversized pages , $ 1 4 . 9 5 .
K E E N E O N C H ESS by Raymond Keene - Com pl ete step-by-step co u rse s hows how

to p lay and deepen one's understa nding of c hess w h i l e keeping the game fun and
exciting. Fasc i nating c hapte rs on chess h e roes and l essons one can lea rn from these
g reats , basic c hess ope n i n g s , strategy, tactics , the best games of chess ever p laye d ,
and the histo ry o f c h ess rou n d o ut a player's education. R eaders a l s o learn how t o use
ch ess notation and all the bas i c conce pts of game play - cast l i n g , pawn promoti o n ,
putting an oppon ent i nto check, t h e five ways of d rawi n g o r stalemati ng games , en
passant, actual checkmate, and much more . 320 pages , $ 1 8 . 95 .
- MATES & E N D G A M ES -
303 TR I C KY C H EC KMATES by Fred Wilson and Bruce Alberston - Both a fasci nating
challenge and g reat tra i n i n g too l , this co l l ection of two , three and bonus fou r move
checkmates is g reat fo r advanced begi n n i n g , i nte rm ediate and exp e rt p l ayers . Mates
397
CARDOZA PU B L-I SH ING CH ESS BOOKS
a re i n order of difficulty, fro m the s i m pl e to very co mplex positions . Learn the standard
patte rns and st ratagems fo r cornering the king: corridor and s u pp o rt mates , attraction
and def l ection sacrifices, pins and a n n i h i l at i o n , the q u i et m ove , and the d readed
zugzwang. Examples, d rawn from actual games, i l l u strate a wide range of chess tac
tics from classics rig ht up to the 1 990's . 1 92 pages , $ 1 2 . 95 .
MASTER C H E C KMATE STRATEGY by Bill Robertie - Learn t h e basic com b i nations,

plus advance d , s u rp rising and u nconventional mates , the most effective pieces needed
to w i n , and h ow to mate oppon ents with j ust a pawn advantage . also, how to work two
rooks i nto an u nstoppable attack; how to wield a q u een adva ntage with deadly i ntent;
how to coordinate pieces of d iffe ri n g strengths i nto i nd efensible positions of the i r op
ponents ; when it's best to have a knight, and when a bishop to wi n . 1 44 pages , $ 9 . 95
BASIC EN DGAM E STRATEGY: K i n gs, Pawns a n d M i nor Pieces by Bill Robertie

Learn th e mati n g principles and combi nations needed to finish off opponents . Fro m
the fo u r b a s i c checkmates using the Ki ng w i t h t h e queen , roo k, two bis hops, a n d
bis hop/knight com bi nations, t o the King/pawn , King/Kn ight and Ki ng/Bishop endgames ,
you ' l l learn the essentials of translati ng small edges i nto decis ive checkmates . Learn
the 50-move ru l e , and the com b i nations of pi eces that can't force a mate against a
lone Ki n g . 1 44 pages , $ 1 2 . 9 5 .
B A S I C E N DGAME STRATEGY: Rooks and Queens by Bill Robertie - The co mpan

ion g uide to Basic Endgame : Kings, Pawns and Minor Pieces, you ' l l l ea rn the basic
mati ng p ri nciples and com b i nations of the Queen and Rook with King, how to turn
m i dd l egame advantages i nto victo ries, by creating passed pawn s , using the King as a
weapon, cleari ng the way for rook m ates , and othe r endgame com b in ations . 1 44 pages ,
$ 1 2 . 95 .
639 ESSENTI AL E N D G A M E POSITI O N S by Eric Schiller - From basic checkmates

to sophisticated double-rook endgames, every im portant endgame concept is explained .
To pics i nc l u d e every key combi nation of king and pawn endgames , bishops, knights,
roo ks , and queens, plus tricky endgames with no pawns. The th i nking behind eve ry
position is explained in words ( u n l i ke diagram-only books) so that p l aye rs lea rn which
positions a re w i n n i n g , wh ich a re d rawn , and which can not be save d . F requent dia
g rams show sta rting and targ et pos itions, so readers can visual ize end goals and
ste e r the m iddlegame to a s uccessful concl usion. 400 pages , $ 1 8 . 95 .
398
CARDOZA PUBLISHING ONLINE
For the latest in Chess Software and Books
by the World's Best Chess Writers
www. cardozapub.com
To find out about our l atest chess and backgammon publications , to order
books and software from third partie s , or simply to keep aw are of our latest publica
tions and activi ties in the ches s worl d:
1 . Go online: www.cardozapub.com
2. Use E-Mai l : c ardozapub @ ao l .com
3 . Call toll free : 800-5 77 -WINS (800-5 77-9467)
4. Write : C ardoza Publishing, 1 32 Hastings Street, Brooklyn, NY 1 1 235
Use order form below to order Cardoza Publishing products .
r - - - - - - - - - - - - - - - - - - - - - - - - �
USE TH I S ORDER FORM TO GET YOU R CH ESS BOOKS !

YES! I want to be a bette r chess player! R u sh me these books : (Write in choices below) :
Qu antity You r Book Order P rice
MAKE CH ECKS TQ: Su btotal

Cardoza Pub l ishing
1 32 Hasti ngs Street
Postage/H a n d l i n g : Fi rst Item $5 00
Brooklyn , NY 1 1 235 Additional Postage
CHARGE BY PHQN E :
Total A m o u n t D u e
Toll- Free: 1 -800-577-W I N S
Local Pho n e : 7 1 8-743-5229
Fax O rders : 7 1 8-743-8284
E-Mail Orders: Cardoza P ub @ aol . com
S H I PP I NG CHARGES: For U S o rders , i nclude $5.00 postage/handling 1 st boo k o rdered; for each
additional book, add $ 1 . 00 . For Canada/M exico , double above amou nts , q uadruple (4X) fo r all other
cou ntries . Orders outside U . S . , money o rder payable i n U.S. dolla rs o n U . S . bank only.
NAM E -------
ADDRESS --
---
C ITY ------- STATE _______ ZIP _____
639 Essent1a1
30 day money back g u a rantee!
� - - - - - - - - - - - - - - - - - - - - - - - - �
399
CARDOZA PUBLI S H I NG ONLI N E
For the Latest in Chess Software and Books
by the World 's Best Chess Writers
www.cardozapub.com
To find out about our latest chess and backgammon publications, to

order books and software from third parties, or simply to keep aware
of our latest publications and activities in the chess world:
1 . Go online: www.cardozapub. com

2. Use E-Mail: cardozapub@aol.com
3 . Call toll fre e: 800-5 77-WINS (800-5 77-9467)
4. Write: Cardoza Publishing, 132 Hastings S treet, Brooklyn, NY 1 1 235
We welcome your suggestions and comments, and are glad to have you
as a customer.
Our philosophy is to bring you the best quality chess books from the
top authors and authorities in the chess world, featuring words (as
opposed to hieroglyphics), clear explanations (as opposed to gibberish),
quality presentations (as opposed to books simply slapped together), and
authoritative information. And all this at reasonable prices .
We hope you like the results.
N E\N - CH ESSCITY MAGAZINE

Free Online Chess Magazine
Cnf:ss CITY is a sprawling metropolis of chess information, a magazine with
the latest news and analysis, to gossip, trivia, and fun features . Travel around
the world to visit the most fa scinating chess compe titions, preview books long
they hit the shelves, and read columns on openings, middlegames, endings,
tactics, strategies, mates, and more .
Extensive excerp ts from our books are available online. Visit often, because
\Ve'll be adding more features for your pleasure ! Improve your chess knowl
edge with our articles and features on the opening, middlegame, endgame,
strategy, tactics checkmates and more ! Whether you 're a beginner or master,
you'll be able to improve your results with our tips.
Chess is a serious game , but it. is also a lot of fun. Chess City Magazine
presents trivia, photos, anecdotes, chess art, strange games, trivia and even a
bit of gossip for your amusement and pleasure! Drop by often when you need
a break!
Go to www .c hesscity.com for details
400

Eric Schiller - 639 Essential Endgame Positions

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eric Schiller - 639 Essential Endgame Positions

Uploaded by

Copyright:

Available Formats

CARDOZA PUBLISHING CHESS TITLES

ABOUT THE AUTHOR

OTHER CARDOZA PUBLISHING BOOKS BY ERIC SCHILLER

UNORTHODOX CHESS OPENINGS- The comprehensive guide to every unorthodox opening

COMPLETE DEFENSE TO QUEEN PAWN OPENINGS- T his aggressive counterattacking rep­

Get online now to learn more about upcoming titles! www.cardozapub.com

Copyright © 2000 by Eric Schiller

Library of Congress Catalogue Card No: 98-74281

Web Site - www.cardozapub.com

Write for your free catalogue of gaming and chess books,

equipment, software and computer games.

4. KNIGHT ENDGAMES 104

6. BISHOP US. KNIGHT ENDGAMES 189

l. ROOK ENDGAMES 233

8. ROOK AND MINOR PIECE ENDGAMES 29'l

9. QUEEN ENDGAMES 326

I 0. MISCELLANEOUS ENDGAMES 337

I I. ELEMENTARY ENDGAMES 357

Checkmate with Bishop + Knight 361

12. QUIZZES 368

O nly a small percentage of chess endgames have been definitively

few hundred for the tournament competitor.

Someone whose tactical skills aren't quite up to the task of calculat­

HOW TO USE TH IS BOOK

a famous endgame technician . He used many of the endgames of such

KING + I US. KING

The square of the king determines whether a pawn can be pre­

There is no Black king in this diagram because the point is to mea­

Imagine a line dra\vn from c3 to c7 to g7 to g3 . That \vould form a

PINTER VS. GEORGIEV

This is a classic winning position. The White king will move in

KING + I US. KING + I

COMAS FABREGO VS. BABURIN

ANALYSIS OF GULKO VS. YEPISHIN

This is a win for White regardless of who is on the move. 1 4.Kd5

STUDY BY RICHARD RETI

2 . . . h3; 3.Ke 7 h2 (3 . . . Kb6 ; 4 . Kd7 h2; 5.c7 h l Q; 6 . c8Q is the same as

ADAMSON: "THE JAILOR"

A deceptively simple position . In fact, it is a very complex position .

. 4 . . . Kd5 (4 . . . Kd7 ; 5 . Kg5 Ke6; 6 . Kxh5 Kf6; 7 . g4 transposes . ) 5 . Kg5

This is a very important move . Black's king is not allowed to slip on

Black is now on the horns of a dilemma. Retreating runs into a

Things seem to be under control , but Black has a surprising re­

( 1 3 .h6 Kg8 ; 1 4 .h7+ Kh 8 ; 1 5 . Kh6 with another s talemate . ) 1 3 . . . Kf7;

KING + 2 US. KING + I

KURAJI CA VS. CIFUENTES PARADA

73 . . . Ka8; 74.b5 Kb8; 75.a3! It is important for White to insure that

CONTINUATION OF TOPALOV VS. SHORT

White has doubled pawns , which usually increase in value as other

Black resigned. 69 . . . h3; 70.f7 h2 ; 7 l .f8Q wins because of 7 l . . . h l Q;

ANALYSIS OF GULKO VS. YEPISHIN

White wins the pawn. 1 6 ... Kf6; 1 7.Kxe4 Ke6 .

5 . Kd7 transposes . 5.Kd6 Kc8; 6.Ke7 Kb8; 7.Kd7 Ka8.

9 ... c5; IO.b7+ Ka7; l l .b8Q+ Ka6; 12.Qb6#.

KING + 2 US. KING + 2

KURAJICA VS. CIFUENTES PARADA

ANALYSIS OF SHIROV VS. ALMASI

59.Kh5 ! As usual, accuracy is essential. 5 9 . Kf5 b4; 60 .g6 b3; 6 l .g7

Kb2 ; 66.Qb7+ Ka l ; 6 7 . Qa7+ Qa2 ; 6 8 . Qxc5 is difficult, even \Vith the

The three pawns cannot provide enough counterplay to earn more

KING + 3 US. KING + 2

AVRUKH VS. GUREVICH

52 ... h4. There is no hope in 52 . . . hxg4 ; 5 3 . hxg4 Kf6; 54.Ke4 Ke6 ;

Black resigned. Control of the opposition gives White an easy win.

COMPLETE DEFENSE TO QUEEN PAWN OPENINGS- T his aggressive counterattacking rep

Someone whose tactical skills aren't quite up to the task of calculat

The square of the king determines whether a pawn can be pre

There is no Black king in this diagram because the point is to mea

Things seem to be under control , but Black has a surprising re

The famous pawn-winning tactic is based on control of the opposi

2 8 . Kd4 Kd6 ; 2 9 . a4 ! White preserves a tempo move (b3 ) , which is al