Foreword by
Karsten Müller
2011
Russell Enterprises, Inc.
Milford, CT USA
Tragicomedy in the Endgame
Instructive Mistakes of the Master
by
Mark Dvoretsky
ISBN: 978-1-936490-04-2
© Copyright 2011
Mark Dvoretsky
All Rights Reserved
No part of this book may be used, reproduced, stored in a retrieval system or transmitted in any manner or form whatsoever or by any means, electronic, electrostatic, magnetic tape, photocopying, recording or otherwise, without the express written permission from the publisher except in the case of brief quotations embodied in critical articles or reviews.
Published by:
Russell Enterprises, Inc.
P.O. Box 3131
Milford, CT 06460 USA
http://www.russell-enterprises.com
info@russell-enterprises.com
Cover design by Janel Lowrance
Printed in the United States of America
Signs & Symbols
Foreword
Introduction
Chapter 1: “Swimming” in Theory
Rook and Pawn (in the Opponent’s Half of the Board) versus Rook
Rook and Pawn (in Its Own Half of the Board) versus Rook
Rook and Rook’s Pawn versus Rook. a- and h-pawns
f- and h-pawns
Lasker’s Idea
An Extra Pawn on the Flank
Rook against Pawns
Rook versus Minor Piece
Rook and Bishop versus Rook
Queen and Pawn versus Queen
Queen versus Rook
Chapter 2: The King in the Endgame
Activity of the King
Selecting the Path for the King
Shouldering
Cutting Off the King
Chapter 3: The Strength of Pawns
The Dangerous Passed Pawn
The Pawn Break
Pawn Races
The Outside Passed Pawns
Separated Passed Pawns
Connected Passed Pawns
Pawn Structure
Chapter 4: Zugzwang
Don’t Fall into Zugzwang!
Reciprocal Zugzwang
Reserve Tempi
Chapter 5: In Search of Salvation
The Fortress
Stalemate
Perpetual Check, Perpetual Pursuit
Chapter 6: Tactics
Mate
Double Attack
Various Tactics
Traps
Chapter 7: Piece Play Maneuvers and Exchanges
Piece Placement
Exchanges
Simplification to a Pawn Endgame
Chapter 8: Technique
Prophylaxis
Winning or Losing a Tempo
Move Order
Carelessness
Chapter 9: Premature End to the Struggle
Unjustified Capitulation
Agreeing to a Draw in a Won Position
Conclusion
Index of Players
! a good move
!! a brilliant or hard-to-find move
? a weak move, a mistake
?? a blunder
!? a move worth considering
?! a doubtful move
the only move
= the position is equal
+= White’s position is somewhat better
+- White has the advantage
+– White has a won position
=+ Black’s position is somewhat better
-+ Black has the advantage
–+ Black has a won position
∞ the position in unclear
=∞ with compensation for the material
# mate
zugzwang
m match
wm world championship match
zt zonal tournament
izt interzonal tournament
ct candidates tournament
cm candidates match
ch championship
ch(1) first league championship
wch world championship
ech European championship
f finals
sf semi-finals
qf quarter-finals
ol Olympiad
tt team competition
jr junior or youth competition
sim simultaneous exhibition
W? A position that may be used as a solving exercise, White to move
B? A position that may be used a solving exercise, Black to move
(D) See the next diagram
Nowadays, many players neglect studying the endgame and focus mainly on opening preparation. I think that this approach is flawed, as time invested in studying the endgame will repay high dividends. The point is that not only will your technique improve – you will have the psychological advantage of looking forward to reaching and enjoying an endgame – but your overall understanding of the game of chess itself will improve, as in the last phase of the game, the real potential of every single piece may be seen much more clearly than in a complicated middlegame position.
So the real question is, in my opinion, how to navigate the expansive ocean of endgames, and how to learn the important positions, methods and principles which can act as lighthouses to guide you. The main approach is certainly to pick up a reference work like Dvoretsky’s excellent Endgame Manual and study it thoroughly. Of course, without having studied the principles of endings, even strong players may find themselves adrift at sea, searching for the correct way to weather the storms. In fact this book is based on the well-known maxim that one can benefit from one’s mistakes. And that also definitely applies to the mistakes made by others, as this books so aptly demonstrates.
Mark Dvoretsky is ideally qualified for this project because of his vast experience in this area and his large collection of excellent examples which have been tested with some of his exceptionally strong students. In this book, he deals mostly with rook endgames as they occur most frequently; many valuable half-points can be saved by being familiar with certain positions and methods in these endings. But an insight into the essence of the errors regarding theoretical aspects of the endgame is of course not all this work has to offer. Dvoretsky also looks at principles such as prophylaxis, whose importance can hardly be overestimated, or the question of the when and how to exchange – especially simplification into a pawn endgame – where many blunders often occur.
To get maximum benefit from this book, I advise you to just look at the diagrams first, contemplating the possible strategy and moves. Only read on when you have reached a conclusion about which move you would play and why. Then not only your understanding of the important final phase of the royal game will improve but so will your overall results as well.
Grandmaster Dr. Karsten Müller
Hamburg
October 2010
Many chessplayers have read Dvoretsky’s Endgame Manual (“DEM”). My new book will help them recall important theoretical positions and technical methods, as well as study several sections on endgame theory in depth (for example, rook and bishop versus rook, or rook endings with f- and h-pawns). It compels one to think about the reasons for endgame errors that everyone – from ordinary amateurs to prominent grandmasters – has made, with a view of avoiding similar mistakes.
For those who are not familiar with DEM, the present volume will serve as an introduction to the fascinating world of endgames and, possibly, prompt them to take up the systematic study of endgame theory. I am firmly convinced – contrary to popular opinion – that such an undertaking is much more effective than endless reworking and memorization of opening material. And, as a result, there will be a significant increase in rating and improved tournament results.
I have called those instances in which serious endgame errors occurred “tragicomedies” – sad for the chessplayers themselves, but amusing for spectators. As was noted in DEM, in which a “tragicomedy” section finished many chapters, such episodes will serve as an excellent warning against ignoring elementary theory. Moreover, they become etched into one’s memory, thereby contributing to a better understanding of those endgame concepts.
In every tournament, and in fact in virtually every round, there are more and more new examples of instructive endgame errors. If desired, a multi-volume encyclopedia could be released on this theme. Several authors, for example Leonid Verkhovsky or Adrian Mikhalchishin, would specialize in these endings. I have not tried to familiarize the reader with fresh examples, although the book does contain a number of them – on the contrary, I have gladly made use of fragments which have been previously published. Neither the quantity nor newness of the material to be studied is that important. The level of comprehension that the reader acquires from the lessons in this book is much more critical. Therefore many insufficiently significant examples from my collection were not included in the book.
Let me note other self-imposed restrictions. Inasmuch as the book is basically an introduction (or, on the other hand, a supplement) to DEM, to avoid repetition, I have chosen not to include examples from the Manual. Preference was give to examples from practice by very strong, or at least rather well-known chessplayers. With rare exceptions, fragments from games with rapid time controls or blitz games were not used, as chessplayers are capable of a lot of nonsense when their flag is hanging, and it is senseless to include it in the ideological base. However nowadays it is almost always necessary to play endings in severe time trouble. The principle reason for this is the radical change in the time controls as well as elimination of adjournments. This is further aggravated by FIDE’s pernicious policy of having games played at as quick a pace as possible, which inevitably leads to superficiality and poorly reasoned decisions. This in turn results in the diminution of chess as an art, depriving the games played of both ideological and aesthetic value. But I should also note that under such difficult conditions, a good working knowledge and understanding of endings becomes especially important. In fact, in endgames played in time trouble, it will be possible to regularly outplay less experienced opponents.
The following material should be studied by actively trying to solve, on your own, the problems posed in these positions. Next to many diagrams appearing at a critical move you will find a question mark, indicating that it may be used as an exercise. The majority of exercises are elementary, but some will pose quite a challenge.
Mark Dvoretsky
Moscow
November 2010
An analysis of the possible causes of endgame mistakes must naturally begin with a weak grasp of theory. And this is not only because of what can be seen from a great number of tragicomic episodes from chess praxis. The parsing (for experienced players – repetition) of some of the most important theoretical positions and ideas will be useful in other parts of the book and it will be necessary to refer to them repeatedly.
At times, all the information about endgame theory seems to be stored in innumerable endgame encyclopedias and reference books. No one – including the authors of these kinds of books – can possibly know or remember all of this theory. As far as theory is concerned, I only understand the core knowledge necessary for the skilled, practical chessplayer. My overall objective when writing DEM was to identify and as best as possible offer a precise explanation of this core theory.
In the majority of the different endgames, the amount of theoretical knowledge necessary for the practical player is not great. However, this is not so with rook endings. Here it is necessary to master more information than all of the other parts combined. It is also the most important and difficult section and that is where we will therefore begin.
“The long and short side.” A center pawn divides the board into two unequal parts: one part contains four files, the other three.
If the weaker side’s king does not manage to get in front of the pawn, then it should try to get on the short side, so as to be able to check from the long side with the rook.
The pursuit of the enemy king by a rook is effective when, between the rook and pawn, there are at least three open lines (files, as in this case, or ranks). If there is less distance, the king will approach the rook without straying too far from its pawn, maintaining the possibility of defending it if necessary.
Let’s take a look how these principles are applied (and alas, frequently not applied) in practice.
1. Svidler – Pelletier
Biel 2001
The black king is positioned where it should be, on the short side. However, if it were to be cut off along the g-file (51.Rf7-g8!!), the fact that it is on the short side would not matter: White would advance his king and pawn without any difficulty.
But those kinds of moves are usually are made only by some especially sharp blitz players. In a normal game, the black king inevitably gets to g6 (or g7). That is why it is important to prevent the only effective defensive plan, checking from the long side. With this in mind, the rook must take the a-file (the black rook will be too close to the king and pawn on the b-file).
51.Ra7! Rf2+ 52.Ke6 Kg6 53.Ra8!, etc., would have won. We will become more familiar with this kind of position in the analysis of the next example
Peter Svidler chose 51.Kf6?? (also wrong is 51.e6?? Rf2+ 52.Ke5 R×f7 53.ef Kg7 54.Ke6 Kf8=) 51...Rf2+ 52.Ke7 Ra2!
The black rook has taken up the long side and the position is now drawn.
53.Rf1 Kg6! (not allowing the king to be cut off by 54.Rg1 – now Black is ready to start with the checks from the side) 54.Rd1 Ra7+ 55.Rd7 Ra8 56.Rc7 Kg7 57.e6 Kg6 58.Kd6 Ra6+ 59.Kd7 Ra8 Draw.
White made the same kind of mistake as in the previous ending.
2. Bolland – Euwe
Weston 1924
72.Kd7?? Ra3 (Black is able to save himself inasmuch as his rook is on the long side) 73.Rc2 Ra5 74.e6 Ra7+ 75.Rc7 Ra8! 76.Rc1 Ra7+ 77.Kd8 Kf6 78.e7 Ra8+ Draw.
Seizing the important a-file with the rook would have decided matters.
72.Ra2! Rb3 73.Ra7+ Kg6 (73...Kf8 74.Ra8+ Kg7 75.e6+–) 74.Ke7! (premature is 74.e6? in view of 74...Kf6! 75.Rf7+ Kg6 76.Rf8 Rb6+ 77.Kd7 Rb7+ 78.Kc6 Ra7=, and the black rook occupies the long side.) 74...Rb8 (74...Rh3 75.Ra6+ and 76.e6) 75.e6 Kg7
An important theoretical position has arisen, one which is well worth noting. I reproduce my comments from DEM. The black rook is only two files from the pawn and this situation provides White with real chances for success.
However, the immediate 76.Ra1? (with the threat of 77.Rg1+) does not work: 76...Rb7+ 77.Kd8 Rb8+ 78.Kc7 Rb2 (threatening 79...Kf8 or 79...Kf6) 80.Rf1 Ra2! 81.e7 Ra7+ with a draw, since the rook has managed to check from the long side as needed.
To win, his opponent must be on move. In fact, on 76...Rc8, 77.Ra1 decides matters; on 76...Rb1, the white rook occupies the important a8-square; bad is 76...Kg8 77.Kf6 Rf8+78.Rf7. There remains only 76...Kg6, but that move, as we shall see, worsens Black’s position.
76.Kd6+!
But not 76.Kd7? Kf6 77.e7 Kf7=.
76...Kf6 77.Kd7 Kg7 (77...Rb1 78.e7; 77...Kg6 78.Ra1) 78.Ke7!
White has reached his goal by triangulation.
78...Kg6
On 78...Rb1, 79.Ra8! Rb7+ (79...Rb2 80.Ke8 Rh2 81.Ra7+ Kf6 82.e7 Rh8+ 83.Kd7) 80.Kd6 Rb6+ (80...Kf6 81.Rf8+ Kg7 82.e7) 81.Kd7 Rb7+ 82.Kc6 Re7 83.Kd6 Rb7 84.e7 decides.
79.Ra1! Rb7+ 80.Kd8
Also good is 80.Kd6.
80...Rb8+
If 80...Kf6 instead, the route to victory is instructive: 81.e7! Rb8+ (81... R×e7 82.Rf1+) 82.Kc7 Re8 83.Kd6! Rb8 84.Rf1+ Kg7 85.Kc7 Ra8 86.Ra1!+–.
81.Kc7 Rb2
82.Re1!
The point! With the king on g7, Black would reply 82...Kf8, but now the pawn cannot be stopped.
82...Rc2+ 83.Kd7 Rd2+ 84.Ke8 Ra2 85.e7+–
“Philidor’s Position” – Two Methods of Defense. This discussion is about those situations in which there is a white pawn on the fifth rank.
3. F. Portisch – Biro
Zalakaros 1996
With the white king on e5, a defensive system demonstrated back in the 18th century by the great French chessplayer François-André Philidor may be employed: 52...Rh6!. The rook moves along the sixth rank until White plays f5-f6, after which it drops back to the first rank and starts checking from the rear. The king will not be able to be protected from them on f6.
In practice, the weaker side often does not manage to take the sixth rank in a timely manner. Then another defensive method is necessary: placing the rook behind the passed pawn.
With the king on e6, 52...Rf2! draws easily: 53.Kf6 Kg8! (the king of course goes to the short side) 54.Ra8+ Kh7. The basic idea behind the defense is shown after 55.Ke6 Kg7! – and thanks to the rook’s position on f1, the pawn cannot advance and the moves must be repeated: 56.Ra7+ Kf8 57.Kf6 Kg8! 58.Ra8+ Kh7, etc. If White plays 55.Rf8, preparing 56.Ke7, then the rook goes to the long side, 55...Ra1 with an obvious draw.
52...Re2+?
The Hungarian chessplayer playing Black is not familiar with the established method of defense. However, the position still remains drawn.
53.Kf6
Necessary was 53...Kg8! 54.Ra8+ Kh7 55.Kf7 Rb2!. There are three free files between the b-file, from which the black rook will operate, and the white pawn, and this distance is sufficient. With a center pawn (e.g., if the kings and pawn were moved one file to the left) Black would lose, as we saw in the analysis of the previous example.
53...Re8?
The decisive mistake. Passive defense, with the rook on the eighth rank, is a safe approach against a knight’s pawn, but not against a bishop or center pawn.
54.Kg6 Re1
Or 54...Kg8 55.f6 Rb8 56.Rg7+ Kf8 57.Rh7 Kg8 58.f7+.
55.Ra8+ Black resigned.
4. Burn – Spielmann
San Sebastian 1911
Knowledge of the second method of defense with Philidor’s Position suggests a simple route to a draw: 62...Rc4! 63. K×c6 Kb8.
As we realized in the analysis of the previous example, 62...Rb4+ 63. K×c6 Kb8 also does not lose. The king goes to the short side and the rook occupies either the h- or g-file. And in general, upon almost any other normal move, the position would remain drawn – except for the move played in the game.
62...Re7??
Now White could have won with 63.Rg8+! Kd7 64.Kb7 Re1 65.Rg7+ Kd8 66. K×c6 Rc1.
When the king is on the long side, positioning the rook behind a bishop’s pawn (or knight’s pawn) does not save Black in view of 67.Rg8+ Ke7 68.Rc8! followed by 69.Kb7. Black is unable to prevent the pawn promoting as it is not possible to effectuate flank checks – the short side is too short. The pawn reaches the seventh rank and what is known as “Lucena’s Position” arises, which, with the king on the long side, is usually won without difficulty several ways.
It is important to note that with a center pawn, a draw may be achieved even with the king on the long side. Let’s take a look at the previous position, but with everything moved one file to the right.
Here 67.Rh8+ Kf7 68.Rd8 does not reach its objective in view of 68...Ra1!. Although the rook and pawn are separated by two ranks, the rook’s position on d8 is rather clumsy. After 69.Kc7 (69.Rc8 Rd1! 70.Kc6 Ke7!) 69...Ra7+ 70.Kb6 Ke7 the draw is obvious.
Let’s return to the game.
63. R×c6+??
White allows the enemy king to remain on the queenside and the position once again become drawn.
63...Kb8 64.Rh6 Rb7+! (the analysis of the previous exampled confirmed that 64...Re8?? 65.c6+– is hopeless) 65.Kc6 Rc7+ 66.Kd6 Kb7 67.Rh8 Rc6+ 68.Kd5 Rg6 69.Rh7+ Kc8 Draw.
Philidor’s Position has been reached with the rook on the sixth rank.
We have just seen how the weaker side should avoid Lucena’s Position. But this is not always possible.
5. Peng Zhaoqin– Chiburdanidze
Moscow wch 2001
After 65...fg?? 66.hg Black’s position is hopeless. The king is cut off on the long side (however, with a knight’s pawn, there is no short side). It is not possible to prevent the advance of the pawn by reaching Lucena’s Position.
66...Rf1 67.Rg7+ Kf8 68.g5 Rg1 69.Kh7 Rh1+ 70.Kg6 Ra1 71.Rb7 Rg1 72.Rb8+ Ke7 73.Rg8! (a winning method known by us) 73...Rg2 74.Kh7 Rg1 75.g6 Rh1+ 76.Kg7 Re1 77.Ra8 Rg1 78.Ra2 Black resigned.
There is a simple draw after 65...Rf3! 66.Rg7+ Kf8 (also good is 66...Ke8 67.g5 R×h3+ 68.Kg6 f4, but not 66...Ke6? 67.Kg6!+–) 67.g5 R×h3+ 68.Kg6 Ra3. After 69.Rb7 f4 70.Rb8+ Ke7 71.Kg7 the presence on the board of the f-pawn saves Black. And if 69.Rf7+ Kg8 70.R×f5, then either 70...Ra7, intending to drive the enemy king from g7 with check, or 70...Ra8.
As has already been noted, with a knight’s pawn, passive defense, putting the rook on the back rank, assures an easy draw. There is one subtlety: after 71.Rf6 Rb8 72.Rc6 Ra8 73.Kh6 Rb8 74.g6 Ra8 75.Rc7 Rb8 76.Rg7+, 76...Kh8!= must be played, but not 76...Kf8?? 77.Kh7+–, and White has reached Lucena’s Position.
6. Aronian – Carlsen
Moscow 2006
69...Ra1?!
Black needlessly switches from one defensive posture to another. It was simpler to impede the advance of the pawn by keeping the rook in the rear, 69...Re2! (70.Kd6 Kf7!), and only in reply to 70.Re8, go over to attacking from the side, 70...Ra2.
70.Ke7 Ra5 71.e6 Ra7+ 72.Rd7
An important theoretical position has arisen, in which a draw may be achieved without difficulty, thanks to the fact that the rook is on the long side. Nevertheless, there are still a few fine points that are necessary to know, as very many chessplayers in this situation would still lose. This game is no exception.
72...Ra8
A safer method of defense is to control the eighth rank. For if White manages to play Ke8 and e6-e7 with impunity, his position would become won.
However, the rook may retreat to almost any square, e.g., 72...Ra1. The point is that after 73.Ke8+ Black is saved by the active 73...Kf6! 74.e7 Ke6! and the unfortunate position of the white rook is telling. It is important that on 75.Kf8, there is the check 75...Rf1+!.
From the foregoing it is clear that the only losing rook retreat along the a-file is 72...Ra6??. After 73.Ke8+ Kf6 74.e7 Ke6 75.Kf8! it is not possible to check with the rook. Paradoxically, many chessplayers would put the rook on a6! The most well-known case is the ending in the game Capablanca-Menchik (DEM, 9-11).
73.Rd6!?
There is nothing in 73.Rd8 Ra7+ 74.Kd6 Ra6+ 75.Ke5 Ra5+ 76.Rd5 Ra8.
With the rook on d6, Black must keep his rook on the eighth rank. For if Ke8 is permitted, then the active counterplay to be had with Kf6, is not available – the pawn will go to e7 with check!
Bringing the rook closer to the king with 73...Rb8? is not good in view of 74.Rd8 Rb7+ 75.Kd6 Rb6+ 76.Kd7 Rb7+ 77.Kc6+–. This means that there is only one move which does not make matters worse: 73...Kg6!=.
73...Ra7+?? 74.Ke8 Black resigned.
There could follow 74...Ra8+ 75.Rd8 Ra1 76.e7 Ra7 77.Rd1 or 77.Rc8.
What can be said about what happened? The conclusion seems obvious: the young and extremely talented Norwegian grandmaster was not familiar with basic endgame theory, never having studied this ending
However, the next day after the game I spoke with Carlsen’s trainer, grandmaster Peter Heine Nielsen. He assured me that Magnus had studied endgame books. This means that the suggested explanation is not accurate; it is not a question of a lack of basic knowledge. At the same time, this episode should give us cause to reflect on the discrepancy between theory and practical experience, the necessity to apply acquired knowledge in practice.
In mathematical lessons in school, theoretical rules, formulas and method problem-solving are studied. But teachers, not being limited to teaching just theory, offer students a great number of exercises. Without this, the acquired knowledge would simply become dead weight, and most of it would soon be forgotten.
Let’s assume that someone who has never sat behind the wheel of a car, has learned by heart traffic rules and regulations and knows exactly when to turn the wheel and step on the gas. Would this be enough to allow him on the street in traffic? Of course not. It would first be necessary to train, develop and consolidate one’s driving skills.
And so it is with chess. Knowledge of endgame theory does not guarantee skill in actually playing endings. Michel de Montaigne accurately wrote in his well-known Experiences: Knowledge is a dangerous sword; in a weak hand that does not know how to wield it, it gets in its master’s way and wounds him.
As you read endgame books, you become acquainted with many instructive examples, but you do not know which ones are particularly important and demand a deep understanding of their subtleties. (By the way, one of the main systematic concepts in writing DEM was to precisely flag and allocate the key, obligatory endgame material.) To better understand and remember the studied material, it is important to consider additional examples on a given theme, and to work through suitable exercises.
As I noted in the introduction to DEM, the connection between theory and specific practical endings is not that far from being direct and obvious. At times it is not simple to make out, in complicated positions, familiar theoretical patterns and to understand just which of the ideas that have been studied may be necessary to use. In other cases, a position, on the contrary, may be very similar to the theoretical. It is important to notice the slight differences and establish how they will affect the course and outcome of the struggle.
In connection with my previous remarks, I recall an episode at one of the sessions at my school for gifted chessplayers held in Moscow in October 2001. Participating were not only young candidate masters, but also masters and even grandmasters. I was discussing the theory of rook-and-pawn-versus-rook endgames. The students accurately took down everything. They were later presented with the following position (DEM, 9-22).
Were the black rook on e1, a draw would be achieved without any particular problem (“the second method of defense in Philidor’s Position”). But the rook is positioned less favorably, and on move, White will win, although, contrary to expectations, not without some difficulty. He has to precisely consider concrete variations, supported by, of course, the evaluation of core theoretical positions.
7. Rohde – Cramling, D.
Innsbruck wch jr 1977
I divided the students into two-player teams. Teams competed among themselves, playing out this ending. On each team, one member played White, another Black, so the overall chances in these matches were objectively equal. Teams were given time (30 or 45 minutes) to prepare – analysis of a position, reference to notes and moving pieces on the board (but naturally without the help of a computer). Then the matches began.
Under these conditions, it would seem that the task for the chessplayers was not too difficult. Alas, not one of the games finished properly; in each, one of the players (in most cases both) made serious errors, and, as rule, right from the outset. I repeated the training exercise, using another rook endgame from Hector-Krasenkov (DEM, 9-14), with the same sad result.
Three months later a similar experiment was repeated at a training session with the strongest French club at that time, NAO. I reviewed with the club members some of the basic theory of rook endings and then proposed that they play each position against me in a simultaneous exhibition (although without preparing analysis). The results turned out identically: not one of the players was up to the task. And among them, two of France’s leading grandmasters, Etienne Bacrot and Laurent Fressinet...
I do not relate this story casually. It evidently illustrates that the real level of endgame skill, even for very strong chessplayers, is far from desirable. And it is no wonder – they in fact have never systematically studied endgame theory and have had no training in the calculation of specific variations supported by theoretical knowledge.
I do not doubt that we would see the same result if the subject of our study were a serious middlegame problem. Is it really necessary to spend almost an infinite amount of time and effort on openings, as the overwhelming majority of chessplayers do, if a large part of the development of their other chess skills suffers?!
I shall now bring to your attention two “tragicomedies in several acts” in which both players commit errors.
8. Salwe – Burn
St. Petersburg 1909
This position, of course, is absolutely drawn. However, one of the leading chessplayers of the past, Amos Burn, demonstrates his unfamiliarity with basic endgame theory and gradually slips into a lost position.
75.Rf7+ Ke8?!
The first mis-step. More to the point is 75...Kc8! – the king should try for the short side.
76.Rh7 Rd3
A good move, but, in fact, simpler was 76...Ra3 followed by 77...Ra6, bringing about Philidor’s basic defensive method.
77.Ke6 Re3+
It is still not too late to bring the king to the short side with 77...Kd8!?.
78.Kd6 Ra3 79.Rh8+ Kf7 80.Rc8
80...Ra6+?
The losing move. Flank checks from the short side are hopeless. In these types of situations, the position may only be saved by putting the rook behind the pawns: 80...Rd3 81.d5 Rd1
82.Kc6 Ke7! or 82.Rd8 Ra1!. As has already been noted, with a center pawn, this method assures an easy draw even if the king is on the long side.
81.Rc6?
The drawback of this move is that it allows Black to get back to the correct plan: 81...Ra1 82.d5 Rd1!=.
81.Kd7! would have won (though not without some difficulty): 81...Ra5 (nothing changes after 81...Ra7+ 82.Rc7) 82.Rc5! Ra7+ 83.Rc7 Ra5 84.Kd6+ Kf6.
85.d5
Weaker is the recommendation of A. Iglitsky made many years ago, 85.Rc8 Ra6+ 86.Kd7 Ra7+ 87.Kc6?.
The annotator suggested 87...Ra6+? 88.Kb5 Rd6 89.Kc5 Ke7 90.Re8+, not having noticed the subtle refutation 87...Ra4!!=.
85...Ra6+ 86.Rc6 (useless is 86.Kc5 Ra5+ 87.Kc6 Ra6+ 88.Kb5 Ra1)
In reply to 86...Ra1, it makes sense to play 87.Rb6!, so that the king can make it to the c-file more easily. For example, 87...Ra7 (87...Kf7 88.Kc7) 88.Rb1 Ra6+ 89.Kc7 Ra7+ (89...Ke5 90.d6) 90.Rb7 Ra8 91.Kd7 and 92.d6+–.
On 86...Ra7, premature is 87.Rc1 Ra6+ 88.Kc7 Ra7+, and 89.Kb6? lets the win slip in view of 89...Rd7! 90.Kc6 Ke7. It is possible to play either 87.Rb6 or choose 87.Rc5!. Black is in zugzwang; he must worsen the position of either the king or rook: 87...Ra1 (87...Kf5 88.Rc1 Ra6+ 89.Kc7 Ke5 90.d6+–) 88.Kd7 Ra7+ 89.Rc7 Ra8 90.d6+–.
81...Ra8?
Not being familiar with the correct plan of defense (Ra1-d1) allowed his opponent, by playing 82.Kd7!+–, to get back to the lines that were examined. Alas, Georg Salwe did not take advantage of this possibility and squandered his chances to win – Black finally was able to set up Philidor’s Position.
82.Rc7+? Ke8= 83.d5 Kd8 84.Rd7+ Ke8 85.Rc7 (85.Rh7 Ra6+) 85...Kd8 86.Rc6 Ra7 87.Rb6 Rd7+ 88.Kc5 Rc7+ 89.Kd4 Ra7 90.Rh6 Rg7 91.Ra6 Rb7 92.Ra1 Kd7 93.Ke5 Rb6 94.Ra7+ Kd8 95.Kd4 Rh6 96.Kc5 Rg6 97.Rb7 Rh6 98.Rb8+ Kd7 99.Rg8 Draw.
So, it seems that at the beginning of the 20th century, chessplayers were less professional than now and did not know theory well. One would think that by the end of the century things would have changed. But look at the following ending.
9. Zhu Chen – Taimanov
Roquebrune 1998
85...Rc1+??
Amazing. The experienced grandmaster Mark Taimanov didn’t know to maintain the rook’s position! There was an elementary draw with 85...Ke7 or 85...Rd3.
86.Kd6 Kc8 87.Rh8+ Kb7 88.Kd7 Rc7+ 89.Ke6 Rg7
90.Re8??
After 90.d6! the pawn cannot be stopped.
90...Rg5?
Only 90...Rg6+ 91.Kd7 Rh6! would have saved Black. The rook would then be positioned on the long side, the maximum distance from the pawn. It was further necessary to seize the h-file, something which both players overlooked.
91.d6 Rg6+
92.Kd7?
92.Ke7! won easily (the threat of 93.d7 does not give Black time for 92...Rh6) 92...Rg7+ (92...Kc6 93.Rc8+ Kd5 94.d7 Rg7+ 95.Ke8 Ke6 does not work in view of 96.Rc6+) 93.Kf6, and not 93...Rh7 94.Re7+, while on 93...Rd7 there follows 94.Ke6.
92...Rg1?
Necessary was 92...Rh6!=.
93.Re7?
Almost any other move reached the goal. 93.Kd8 or 93.Ke7 would escort the pawn to the seventh rank, while 93.Rh8 would have taken control of the long side. Now Black could have drawn by playing 93...Rh1!.
93...Rg8?
The last mistake.
94.Rh7
A position of mutual zugzwang has arisen, one which we previously analyzed in the game Bolland-Euwe. It is Black’s move and he has to allow the worsening of the position of one of his pieces. I remind you that on 94...Kb6 decisive is 95.Rh1 Rg7+ 96.Ke6 Rg6+ 97.Ke7 Rg7+ 98.Kf6 Rg2, and now not 99.Rc1? Rh2!=, but 99.Rd1!+– (this move would not be possible if the king were on b7).
94...Rf8 95.Ke7 Rf1 96.d7
The other alternative is 96.Ke8+ Kc6 97.d7 Re1+ 98.Re7 Rh1 99.Re6+. 96...Re1+ 97.Kd8
97...Kc6
On 97...Rg1 the classic “bridge building” winning method arises – protection on the fourth rank from checks on the file. For example, 98.Rh4 Rb1 99.Re4 Rh1 100.Rb4+ Ka7 101.Kc7 Rc1+ 102.Kd6 Rd1+ 103.Kc6 Rc1+ 104.Kd5 Rd1+ 105.Rd4+–.
However the easiest is probably 98.Rh8 Re1 99.Re8 Rh1 100.Ke7.
98.Kc8 Black resigned.
In these kinds of positions, there is a new way for the weaker side to proceed: frontal attack. There is at least three ranks between a pawn that has not yet crossed the mid-line of the board and the rook on the first rank. As we know, this is enough to successfully harass the enemy king.
10. Filippov – Novikov
Koszalin 1999
The draw is had by 58.Rf1+ Kg4 (58...Ke6 59.Re1) 59.Re1! (59.Rg1+? Kf3–+ is a mistake) 59...Kf4 60.Rf1+ Kg3 61.Re1 Rd5 62.Kc4!.
58.Rh1??
A terrible move, allowing Black to advance the pawn, 58...e4!, winning. The white king is on the long side and (as with flank attacks) with the frontal attack, the rook is too near the pawn.
58...Ke6?? 59.Re1 Rd8 (D)
60.Kc2??
A typical blunder which occurs in many games! White wants to play 61.Rd1, but he is not allowed to do make two moves in a row, and after his opponent’s natural reply, the exchange of rooks is no longer possible.
In similar situations the king should be on the third or fourth rank so as to prevent the enemy rook from defending the pawn from d5. 60.Kc4 led to an easy draw. Also possible was 60.Rh1 e4, and now either 61.Kc2 followed by 62.Rd1, or 61.Rh5 (cutting the king off from the pawn horizontally is one of the useful defensive devices).
60...Kf5 61.Rf1+ Kg4 62.Re1 Kf4 63.Rf1+ Kg3
The king gives itself the maximum amount of activity.
64.Re1 Rd5
Here is where the drawback of White’s 60th move makes itself felt – the rook cannot be attacked: 65.Kc3 Kf2 66.Kc4 Rd4+ or 66.Re4 Kf3–+.
65.Rg1+ Kf3 66.Rg5 Kf4 67.Rh5 Kg4 68.Rh8 e4 69.Kc3 e3 White resigned.
Let’s return to the position in the last diagram and assume that the king is cut off from the pawn not by one, but by two files.
Then Black wins: 60.Kb4 Kf5 61.Rf1+ Kg4 62.Re1 Kf4 63.Rf1+ Kg3 64.Re1 Re8! 65.Kc3 Kf2 66.Re4 Kf3 67.Rh4 (67.Re1 e4–+) 67...Rd8! (of course, not 67...e4? 68.Kd2) 68.Rh3+ Kg4 69.Re3 Kf4 and 70...e4–+.
11. Fischer – Sherwin
Portoroz izt 1958
The question is which of two useful moves (78...Ke6 or 78...Ra8) should be played first?
The answer is simple: 78...Ra8! of course. Black is preparing a frontal attack and the exchange of rooks on f8. On 79.g5, simplest of all is 79...Rf8, although also not bad is 79...Ra4. And on 79.Kh4, there follows 79...Rh8+ 80.Kg5 Rg8+ or 80.Kg3 Rf8!?.
78...Ke6?? 79.Kh4
The pawn will now inevitably cross the mid-line of the board and the position will be reduced to a won Lucena’s Position. Rushing with the immediate 79.g5? is incorrect, as the king remains cut off from the pawn along the rank.
79...Ra8 80.g5 Rh8+ 81.Kg4 Ke7 82.g6 Rf8 83.Rf5! Rh8
Or 83... R×f5 84. K×f5 Ke8 85.Ke6 Kf8 86.Kf6 Kg8 87.g7+–.
84.Kg5 Rh1 85.Rf2 Rh3 86.g7 Rg3+ 87.Kh6 Rh3+ 88.Kg6 Rg3+ 89.Kh7 Rh3+ 90.Kg8 Black resigned.
12. Farago – Horvath
Hungary ch, Budapest 1995
54.Rh2?
It appears that the Hungarian grandmaster does not know the frontal attack method, and consequently does not bring his rook to the first rank.
54...Kc5 55.Rb2?! (55.Rh1!=) 55...Rc1
On 55...b4, White saves himself with the move 56.Rc2!.
A frontal attack is no longer possible, but now White could have arrived at Philidor’s Position with 56.Kd2! Rh1 (56...Rc4 57.Kd3!) 57.Rb3! Kc4 58.Rg3 b4 59.Kc2=.
56.Rb3?? b4–+ 57.Kd2 Rc4 58.Rb1 Kb5 59.Ra1 (59.Rc1 b3! 60. R×c4 K×c4 61.Kd1 Kd3–+) 59...b3 60.Ra8 Kb4 White resigned.
13. Laveryd – Andersson
Katrineholm 1999
The draw is only to be had by 67.Ke4! b5 68.Rb1 Kb6 69.Ke3!=. Although the white king is cut off from the pawn by two files, with a knight’s pawn the position is nevertheless drawn. The black king does not have enough maneuvering room – it is restricted by the edge of the board. For example, 69...Rd5 70.Ke4 Rd8 71.Ke3 Ka5 73.Ra1+ Kb4 74.Rb1+ Ka4 75.Ra1+ Kb3 76.Rb1+, etc.
And why are other continuations not good? The point is that Black threatens to use the effective winning of cutting the king off along the rank. Cutting the king off horizontally is frequently much more dangerous than vertically!
For example, the move 67.Ke5? is refuted by 67...Rc6! 68.Rb1 Rc4! 69.Kd5 Ra4 followed by Ka6 and b6-b5.
In the game, there was 67.Rb1? Rd4!–+ 68.Ke5 Ra4 69.Kd5 Ka6 70.Kc6 Rc4+ 71.Kd5 b5 72.Ra1+ Kb6 73.Ra2 Rh4 74.Ra8 b4 75.Rb8+ Ka5 76.Kc5 Ka4 77.Rg8 Ka3 78.Ra8+ Kb2 79.Ra4 Kc3 White resigned.
14. Taimanov – Larsen
Palma de Mallorca izt 1970
As has just been noted, this kind of position is drawn even with the pawn on g4. So how should Black go about it? The simplest path to a draw is 58...Rg8+! 59.Kf7 (59.Kf6 Kd4) 59...Rg4=.
58...Kd4?! 59.Ra3 Ke4!
A mistake is 59...Rg8+? 60.Kf7, and 60...Rg4 leads to the loss of the rook.
60.g4 Rg8+ 61.Kh5 Rh8+ 62.Kg5 Rg8+ 63.Kh4
This is an alarming situation for Black, since both the advance of the pawn and the cutting off of the king horizontally is threatened. Nevertheless a draw is attainable two different ways.
The simplest decision was the active 63...Kf4! 64.Ra4+ Kf3=, and 65.g5?? leads to mate.
Also possible is 63...Rh8+! 64.Kg3 Ke5 65.Ra6 (65.Rf3 Rg8=) 65...Rh1=. With the king on g3 and the rook on the h-file, the king cannot make forward progress and the rook cannot take up a typical post on h6, while 66.g5 Kf5 leads to the loss of a pawn.
63...Ke5?? 64.Ra6!
In this case cutting the king off horizontally decides the struggle, inasmuch as on 64...Rh8+ there follows not 65.Kg3?, but 65.Kg5 Rg8+ 66.Kh5 Rh8+ 67.Rh6.
64...Kf4 65.Rf6+ Ke5 66.g5 Black resigned.
15. Gligoric – Fischer
Yugoslavia ct 1959
50...Kd7?!
Now since Black must lose a pawn, he will have to resort to a frontal attack. As the white pawn is not far advanced, it is easier to defend. 50...b4! 51.Rd1+ Ke6 52.Rd3 Ra8 is what should have been played, with an obvious draw.
51.b4 Rh3
The rook’s place is on the back rank, but the straightforward 51...Ra8? 52.Rd1+ loses: 52...Ke6 53.Rd4! Rb8+ (53...Ke5 54.Rd7 Ke6 55.Rb7) 54.Kc6 Ke7 55.Rd5 Ke6 56. R×b5 (A frontal attack has not worked out and the matter will resolve itself in a won Lucena’s Position: 52...Kc8 53.Rc1+ Kd7 54.Rc7+ Kd8 (54...Kd6 55.Rb7, inevitably reaching Lucena’s Position) 55.Rh7+–.
52.Rc1
Before taking the pawn, the enemy king has to be cut off.
52...Rh8??
The decisive error. Only 52...Rh5!, saved Black, and if 53.Rc5, then 53...R×c5 54.K×c5 Kc7 55.K×b5 Kb7=.
53.K×b5??
Returning the favor – Svetozar Gligoric allows his opponent a frontal attack which forces the draw. White could have won with either 53.Rc5, or 53.Rc7+! Kd6 (53...Kd8 54.Rc5 Kd7 55.R×b5 or 55.Kb7 followed by 56.R×b5) 54.Rc6+ Kd7 (54...Kd5 55.K×b5 Rb8+ 56.Rb6) 55.K×b5 Rb8+ (55...Rh4 56.Rc1 Rh8 57.Ka6) 56.Rb6 Rh8 57.Rb7+ (57.Ka6) 57...Kc8 58.Ka6 Rh6+ 59.Ka7.
53...Rb8+ 54.Ka4 Ra8+ 55.Kb3 Rc8 56.R×c8 K×c8 57.Kc4 Kb8! Draw.
16. Timoshchenko – Kharitonov
Frunze 1988
The difference between this position and the previous ones is that here the black king is on the short side. Therefore when it is cut off by one file, the need for a frontal attack disappears – Black will be able to rely on flank checks. Therefore necessary is 54.Rg6! Ra8 55.e4.
When a pawn (other than a rook or knight’s pawn) is on the fourth rank and the king is cut off from it by two files, a frontal attack will not work. Let’s look at the position that arises after 55...Kh5 56.Rg1 Re8 57.Kf4 Rf8+ 58.Ke3 Re8.
The winning method that was set out in the final analysis of the Filippov-Novikov ending will not work in the given situation. There the king was on the long side, but here, it is on the short side. Therefore useless is 59.Kd4 Rd8+ 60.Kc5 Re8 61.Kd5 Rd8+ 62.Kc6 Re8 63.Re1? Kg6 64.Kd7 Ra8 (or 64...Re5 65.Kd6 Ra5), and the rook is ready to start checking from the long side.
The correct plan is to make use of mating threats against the king, which is trapped on the edge of the board. But first the king must be made to leave the h5-square where it is ideally placed, while still maintaining the rook on g1. In other words, a position of mutual zugzwang has arisen and Black must be on move. This problem is solved by 59.Rg2 (if White had placed the rook on g2 on the 56th move, and not on g1, 59.Rg1! could have now been played, saving two tempi) 59...Kh4 60.Rg7 (60.e5? does not work: 60...R×e5+ 61.Kf4 Kh3!) 60...Kh5 61.Rg1! (zugzwang) 61...Kh6 (61...Kh4 62.e5! R×e5+ 63.Kf4+–) 62.Kd4 Rd8+ 63.Kc5 Re8 64.Kd5 Rd8+ 65.Ke6 Re8+ 66.Kf6!+–.
Gennady Timoshchenko did not play accurately and let the win slip.
54.Kf4? Kh5 55.e4 Ra1!
Cutting the king off two files from the pawn is no longer possible. The black rook now checks from the rear, to help the king approach the pawn, and then returns to the long side.
56.e5 Rf1+ 57.Ke4 Kg5 58.Kd5
On 58.Ra6!? the simplest reply is 58...Rb1 (58...Re1+ also does not lose: 59.Kd5 Rd1+ 60.Ke6 Rb1!=) 59.Kd5 Kf5 60.Rf6+ Kg5= followed by 61...Ra1.
58...Ra1 59.Ke6 Ra7 60.Kd5 Ra5+ 61.Ke6 Ra7 62.Rc6 Kg6 63.Rd6 Kg5 64.Rd5 Kg6 65.Rd7 Ra6+ 66.Rd6 Ra7 Draw.
17. Ricardi – Kotronias
Erevan ol 1996
We will see the most important ideas in these kinds of positions in the analysis of the following endgame.
It is useful to note right off that the black king is not able to approach the pawn: 75...Ke7?? 76.a7 Kd7(f7) 77.Rh8 winning the rook. This means that it must stay around g7. But what should be done to counter the transfer of the white king to the queenside, where, on a7, it will be sheltered from vertical checks?
The only possible defensive plan is the so-called “Vancura’s Position,” i.e., the transfer of the rook to the sixth rank with the king on g7. Attacking the pawn from the side, Black prevents the enemy rook from leaving the a-file. And if the white king defends the pawn, horizontal checks follow, from which there is no shelter on the queenside.
As will quickly become clear, to realize this plan, it is best to start with a zwischenschach, driving the enemy king back to a less active position.
75...Ra4+! 76.Kf3 Kg7 77.Ke3 Rc4!. On 78.Kd3 there follows 78...Rc6!, and Vancura’s Position has been established.
In the game there followed 75...Kg7?! 76.Kf5.
Grandmaster Vasilias Kotronias was evidently not familiar with the correct method of defense and lost quickly:
76...Kf7? 77.Ke5+– Kg7 (77...Ke7 78.a7! and 79.Rh8) 78.Kd5 Rf1 79.Rb8 Ra1 80.Rb7+ Kf6 81.a7 Black resigned.
Also hopeless is 76...Rc1? 77.Ra7+!, and on any retreat of the king, decisive is 78.Rb7 and 79.a7.
The precise path to saving the game was shown by Peter Romanovsky back in 1950.
76...Ra5+!
On 77.Ke4 there follows 77...Rc5! 78.Ra7+ Kg6! (but not 78...Kf6? in view of 79.Kd4 Rc6 80.Rh7!) 79.Rb7 Ra5 80.Rb6+ (80.a7 Kf6=) 80...Kf7 81.Rh6!? Kg7! (81...Ke7?? 82.a7 Kd7 83.Rh8!+–) 82.Rc6 Kf7 83.Kd4 Ke7, and the black king gets to the queenside just in time.
And if 77.Ke6, then neither 77...Rb5? 78.Ra7+ Kg6 79.Rb7+–, nor 77...Rg5? 78.Ra7+ Kg8 79.Kf6! Ra5 80.Ra8+ Kh7 81.Ke7+– works. Progress is only made by 79...Rh5!! 80.Kd7 (80.Ra7+ Kg8 81.Rf7 Ra5 82.Ra7 Rh5!) 80...Rh6! 81.Kc7 Rf6!, and Vancura’s Position has been set up.
in these situations the f6-square is the best square for the black rook.