10th Anniversary Edition

Learn from the Legends

By

Mihail Marin

Quality Chess
www.qualitychess.co.uk
 Contents
Key to symbols used & Bibliography 4
Preface 5

1 Akiba Rubinstein’s Rook Endings 11

2 Alexander Alekhine and the Fourth Phase of the Game 83
3 In the Patriarch’s Footsteps 117
4 Tal’s Super Rooks vs. Two Minor Pieces 153
5 Petrosian’s Exchange Sacrifices 189
6 Bobby Fischer’s Pet Bishop 255
7 Karpov’s Opposite Coloured Bishops Endings 307
8 Viktor, the “Non-Existent” Hero 375
9 Quo Vadis, Magnus? 423

Biographies of Our Heroes 457

Name Index 469
Game Index 473
Alphabetical Game Index 476


Preface
 Preface to the third edition
Every time I publish a new book it feels as if I am blessed with yet another child. How else could
I describe my thoughts when holding in my hands something created by me and which bears in
it so much of my intimate self?
I remember very clearly the moment when I expressed this kind of feeling for the first time. In
1987 I took part in my first ever Zonal tournament. Among the 24 participants of the Warsaw
Zonal, I was the only FIDE Master or, in other words, the outsider of the event. Very few would
have taken my chances seriously and even I was mainly aiming for an honourable result, but after
more than two weeks of battling, I emerged as one of the four happy qualifiers for the Interzonal!
My friends encouraged me to exploit the favourable moment to the maximum by annotating
some of my crucial games (including two wins over Kiril Georgiev, then ranked in the top 20 in
the world) for the Chess Informant, a sort of periodical Chess Bible in those times. This was my
official start as a chess writer and I considered it a success since a few months later I received my
free copy of the Informant.
Keeping my enthusiasm to myself is something I was never very good at, so I immediately told
my girlfriend that I felt as if a baby had been born to me. She looked at me with her beautiful,
deep, but on that occasion condescending eyes and after a short hesitation replied: “Why so,
I thought getting an Informant for free was a good thing!” The whole episode ended with a
disappointed “Never mind!” but I now realize that the joy offered by my new “chess babies” has
not lost even a single bit of freshness over all these years.
The book you are holding in your hands is no exception, but since it is not a new book, but
just a new edition of my first effort for Quality Chess, some explanation is needed. First of all,
knowing that ten years after the first edition there might still be interest in what when I grow old
I may call my principal work, feels the same as when a boy confirms the hopes of his parents by
graduating from the best university around.
Besides, there is plenty of new material in it: the Carlsen chapter, a lot of corrections and
additions, so to a certain extent this is a new book. Now is a good moment to express my
gratitude to my friend Karsten Müller, renowned endgame expert and the perfect host for me
during the live ChessBase Endgame Magic show whenever I visit Hamburg! Karsten provided me
with a long list of possible mistakes and inaccuracies from the previous editions. Reviewing the
respective paragraphs and making the needed corrections linked me emotionally with the past,
creating the perfect illusion of rejuvenating myself by ten years.
But I confess that my best moments were those in which I could refute Karsten’s claims. This
does not have anything to do with my innate competitive nature, but with the satisfaction of
proving that the human mind can still be stronger than computers – if only we believe in it!
Now, please allow me to introduce my youngest baby-book: the third edition of my Legends!

Mihail Marin
Bucharest, May 2015
Alexander Alekhine and the
Fourth Phase of the Game
84 Learn from the Legends

We are all familiar with the magical attraction exerted by chess. Once a person (be it a child or an
adult, it does not matter) has learned it properly he or she will long to play again and again and
will feel the emotion of each game with intensity as if it was a matter of life and death.
Equally remarkable is the fact that even in the darkest periods of the history of mankind chess
has managed to maintain its status as the royal game. Many of the greatest chess masters have
also appeared to enjoy a protected status even when confronted by the most powerful dictators.
A good example is the rather paradoxical attitude of the Soviet regime towards a certain refugee
named Alexander Alexandrovich Alekhine.
Coming from a rich and influential family, Alekhine had every reason to disagree with the
political changes during what could have been the best years of his youth. The new regime took
from him practically all his goods and wealth. There were even rumours that if had he not been
such a strong chess player he would have been killed. After having failed to embark on the last
ship with emigrants departing from Odessa, Alekhine had to find a way of making a living. His
broad education allowed him to gain very good jobs, which he fulfilled conscientiously.
However, in 1921 the chance to play a tournament in France arose and Alekhine chose liberty;
he never returned to his homeland. The future World Champion was strongly focused on his
chess career and avoided making political declarations, but one relatively innocent statement he
apparently made was enough for the Soviet leaders to declare him an enemy of the state.
We now reach the most interesting part of the whole issue. By normal Soviet standards, Alekhine
should have immediately become a non-person whose name would have disappeared from all the
archives, never to be officially mentioned again. However, the leaders seem to have understood
that Alekhine was a unique figure, not only as a player but also as a commentator. His books on
the New York tournaments of 1924 and 1927 were published in Russian (what is more, with the
author’s approval!) as well as many other games annotated by Alekhine.

The man who made Chigorin’s dreams come true

With the passing years Alekhine became, in the official Soviet view, an essential link in the long
chain of glorious Russian players. He was the man who made Chigorin’s dreams come true and
inspired Botvinnik to take over the relay. The other communist countries naturally shared this
point of view. It is quite significant that, in my youth, the three most outstanding chess books
published in Romanian were the translations from Russian of the best game collections of the
following players: Chigorin, Alekhine and Botvinnik.
My whole childhood and, indirectly, my entire chess career were influenced by a book I found
in my father’s library, 300 partide alese ale lui Alehin (Alekhine’s 300 selected games), published
in 1957.
Alekhine’s moves and comments were those of the hero that every child needs for inspiration.
Even the biographical chapter written by Panov was melodramatic enough (though strongly
politicised, the way I see it now) to match any best-selling novel from the Romanian market.
For years I spent countless hours working with that book, or simply reading it. I allowed myself
to be separated from my hero’s games only for such boring activities as my school duties, eating
or sleeping. This attitude was directly encouraged and inspired by the book itself. Panov writes
 Alexander Alekhine and the Fourth Phase of the Game 85

that Alekhine’s thoughts were permanently focused on chess even when he had to do something
different, and that he always carried with him a pocket chess set.

Game collections of strong players are always instructive but are rarely presented in a systematic
way. However, Alekhine’s games were reserved by fate for a different treatment.
Soon after the champion’s death one of the leading Soviet grandmasters, namely Alexander
Kotov, started structuring the wealth of Alekhine’s games according to tactical and strategic
themes. This was intended to help the reader improve his understanding of chess. The final result
of years of work was the two volume book Shakhmatnoe nasledie A. A. Alekhina (Alekhine’s chess
inheritance).
I became acquainted with Kotov’s book much later than with Alekhine’s 300 selected games.
Although I was familiar with most of the games, I understood that Kotov had done a really
great job. His introductions to the chapters as well as the classification of the material are highly
instructive.
It would have been, however, impossible to give deep comments to more than 360 games, so he
limited himself to quoting Alekhine’s own notes or stating some general considerations.
In this chapter, inspired by one of the sections of Kotov’s work, I shall focus on a very interesting
theme: positions where each side has only queen and one rook left (and pawns, of course). I have
used examples cited by Kotov as well as some of his ideas. However, I have focused more than he
did on the critical moments, dug deeper analytically and arranged the games in what seemed to
me the most logical order.

General considerations
What makes these positions so special?

The first impulse is to classify them as endgames, because of the reduced number of pieces
left on the board. On the other hand elements of the middlegame are also present: a queen in
collaboration with a rook can be a devastating force, especially when attacking the enemy king.
This obvious difficulty of strictly classifying these positions caused Romanovsky to baptize
them the fourth stage of the game, along with the opening, the middlegame and the endgame.
However, this definition only offers a limited solution from the abstract point of view of chess
terminology. For practical reasons, it is more important to understand the way the dual character
of the position is reflected in its evaluation and the choice of plans.
The most difficult decision to be made regards the kings’ positions. For the eventuality that
queens may be exchanged, the kings should be ready for rapid centralization. At the same time,
kings are best advised not to abandon their residences while middlegame influences remain. This
obvious contradiction makes long term planning quite difficult for both sides.
For instance, if one of the players attempts a mating attack, he should constantly take into
account that the initially exposed enemy king could instantly become a strong centralized piece
in case of simplifications. In the game fragment presented below I benefited from this interesting
psychological situation by turning a basically losing position into a win.
86 Learn from the Legends

Valentin Stoica – Mihail Marin with 34.axb5† axb5 (Although strategically

undesirable, 34...¢xb5 would be relatively
Eforie 1988
safer. The queenside weaknesses would be
 compensated by the vulnerability of the
     e4-pawn.) 35.£d5† when after 35...¢b6 he
can create decisive threats with 36.¦a1 ¦e8.
     
      
        
        
      
   
      
       
     
Right after the opening my experienced 
opponent (future friend and trainer) carried 37.b4! The black king’s shelter is destroyed
out a nice tactical operation, winning a pawn and the attack is unstoppable. For instance
and forcing my king to travel from g8 to c6. The 37...cxb4 38.£d4† £c5 39.¦a6† winning the
material advantage and attacking chances are queen. Instead of 37.b4, the natural 37.£d2
sufficient reasons to consider White’s position would be less clear because of 37...¢b7! and
objectively winning. However, proving this in White has no clear continuation of the attack.
practice is not easy since there is no immediate This variation was not easy to foresee from
technical solution, while attacking the king the beginning, although after 33.a4 intuition
with a reduced number of forces requires could have led White, step by step, on the
imagination and precise calculation. right path.
Judging from the course of the game, it seems
that my opponent completely underestimated A second best continuation was 33.h4!?
my potential resources for counterplay. They securing the first rank (an important detail,
consist of the queenside majority supported by as we shall see) and preparing a dangerous
the king and would become quite significant candidate for promotion in the endgame. Black
in the case of simplifications. would not be able to organize any counterplay,
since his pieces would be practically tied up to
33.£c8† the defence of the king.
Since the d-file does not offer White a direct If 33...¢b7, with the aim of placing the king
path to the enemy king, he should have opened on a safer square, White could play 34.£d5†
new lines on the queenside. £xd5 35.exd5 with a clear extra tempo
compared to the game, which would have
Therefore, the strongest move was 33.a4!. made his winning chances significantly higher.
Facing the immediate threat of an attack along
the a-file (or, in case of ...bxa4, the b-file) Black 33...¢b6 34.£d8†
has to regroup with 33...£c7. However, this 34.a4 would still have been unpleasant for
move allows White to centralise his queen Black.
 Alexander Alekhine and the Fourth Phase of the Game 87

 37...£a1† 38.¢f2 £f6†.

     Instead, White could have continued
     36.£d3 since 36...£xe4 is not a threat
anymore because of 37.£d6† ¢b7 38.£xc5
    with a clear advantage.
   
    36.exd5
Finding himself in time pressure, my
    opponent must have been seduced by the
   idea that the immediate 36...¢c7, planning
to block the pawn from d7, is not possible
    because of the fork on d6. As the further
 course of the game will prove, this detail is not
sufficient for a win.
34...¢b7!?
Vali’s intention might have been to repeat 
moves once in order to approach the time
control, but he only helped me improve my
    
king’s position. The threat ...£xe4 became real    
now.    
35.£d5†?   
This move allows the biggest part of the     
advantage to slip away.
Failing to find an attacking plan, Vali    
decides to give the game a technical character.   
He might have hoped that the concession of
exchanging queens would be compensated by
   
the creation of a strong passed pawn, supported 
by the rook. This overlooks that the previously 36...¦e2
exposed black king will easily block it. This is During the game I did not hesitate for a
an important detail because the black rook will moment to activate my rook.
be allowed to start dangerous counterplay.
However, the more solid 36...¦d7, intending
35.¦e1, defending the pawn, should have been to block the pawn with the king on d6, would
preferred. In this case, Black would have faced have been entirely playable.
a difficult defence but no immediate danger of A possible continuation is 37.h4 ¢c7
losing. 38.d6†!?. Allowing the king to reach the
d6-square would mean giving up any winning
35...£xd5 attempts. In fact, I would already prefer Black’s
I had no interest in continuing to play with position.
the queens on the board. After 35...¢b6, the 38...¦xd6
queen incursion 36.£a8 would lead nowhere This capture is also forced. Otherwise the
because of 36...¦e8 when the queen has black rook will stay passive for the rest of the
to return to d5, since 37.¦d8 even loses to game.
88 Learn from the Legends

39.¦xd6 ¢xd6 40.¢f2 to win space with 42.a4 ¢e5 43.a5 ¢d5 44.h4
White’s kingside majority is an important ¢e5 45.¢e3 ¢d5 and the threat of playing
positional factor, but it is compensated by ...c4 followed by ...¢xc4 winning, forces White
the black king’s superior activity. to repeat the position with 46.¢d3.
40...¢e5 41.¢e3 ¢f5
Threatening to attack the pawns with ...¢g4.

42.¢f3     
  +  
        
       
   
       
        
    t
      
     
 37.d6 ¢c8 38.¦d5 ¦c2
42...a5 While talking to my opponent fifteen years
After general pawn exchanges the most later, I considered that there was a good chance
probable outcome would be a position with his wounds had healed so I finally opened the
white king and a-pawn versus black king. subject of this dramatic game.
Since the black a-pawn is likely to be the last Among other interesting things, Vali told
one to fall, it is useful to advance it as far as me that he considers 38...¦xa2 to be sufficient
possible in order to allow the black king to for a draw as well, suggesting the following
reach the b8-square in time. variation:
43.g4† 39.¦xc5† ¢d7 40.¦c7† ¢xd6 41.¦xg7 ¦a3
The careless 43.a4?? even loses to 43...c4 42.h4 ¦xb3 43.h5 ¦e3!
44.bxc4 bxa4. The most efficient way to fight against
43...¢e5 44.¢e3 a4 45.bxa4 bxa4 46.a3 c4 White’s advancing passed pawn.
47.h5 ¢f6 48.¢f4 c3 49.g5† ¢e6 50.¢e3 44.h6 ¦e1†!
¢f5 An important intermediate check.
It is easy to establish that the black king will 45.¢h2 ¦e6 46.h7 ¦h6† 47.¢g3 ¢e5 48.¢g4
reach the b8-square in time. ¢f6
Black is just in time to eliminate the main
White can try a different move order, starting source of danger.
with 37.¢f2, when for a while play continues
in similar fashion: During the game I did not want to give up
37...¢c7 38.d6† ¦xd6 39.¦xd6 ¢xd6 40.¢e3 my c5-pawn so easily, as if anticipating the
¢e5 41.¢d3 ¢d5 further course of the game. Vali even asked me
Compared with the previous variation, the whether I was playing for a win at this point of
kingside has not yet been weakened. This allows the game, but such a thought was beyond my
the king to stay closer to the other wing in order ambition.
 Alexander Alekhine and the Fourth Phase of the Game 89

 44...c3 45.h7 c2 46.h8=£ c1=£†

    And here is where the unfavourable position
     of the white king comes into account. If
Black had promoted without check, his king
    would have been quickly mated.
   47.¢h2

     
    
        
     
         
     
39.d7†     
With this nervous move exposing the passed   
pawn even more, White makes things easier     
for his opponent.

After a break of more than 10 moves, we are
Black would have needed to show greater
back in the world of major pieces endgames.
accuracy after 39.¦g5. The drawing path
And again Black seems to be in trouble
would have been quite narrow as we can see
because of his exposed king.
from the following variations.
However, he has the accurate 47...£f4†
39...¢d7 40.¦xg7† ¢xd6 41.h4
48.¦g3 ¦a4! defending the queen and the
Capturing the a6-pawn with 41.¦g6† ¢d5
a-pawn at the same time while also creating the
42.¦xa6 would lose too much time, allowing
threat of simplification with ...£h4†. Neither
Black strong counterplay with 42...c4 when
player can do anything to win.
the only one risking a loss would be White.
41...¦xa2
An interesting attempt to improve over this
The pawn race that will start now is quite
last variation is the immediate 43.h6!? without
exciting, but its outcome will be peaceful.
capturing on c4.
Optically, it might look as if White’s
Black is best advised not to take on b3,
connected pawns will be more dangerous.
because from b1 his newly born queen would
However, the centralized black king has a
have no possibility of checking the white king
significant influence on the evaluation of
after ¢h2, thus leaving her own king exposed
the position. For instance, it prevents the
to the attack. White’s main idea is that after
white rook from holding the c-pawn from
43...c3 44.h7 c2 45.h8=£ c1=£† 46.¢h2
behind. This variation is a clear illustration
£f4† 47.¦g3 his b3-pawn considerably
of how the exchange of queens can reverse
restricts Black’s pieces.
the situation of the kings.
However, the black king now has a reliable
42.h5 c4 43.bxc4 bxc4 44.h6
shelter on the queenside and after the calm
White could stop the enemy pawn with
47...¢c7 his chances would not be worse at all.
44.¦g3 but after 44...¦a5 he will not be
able to defend his own runner, for instance
39...¢d8
45.¦h3 c3 with a draw.
90 Learn from the Legends

 45.¦b2
     There is nothing else left for White.
    45...¦c3
By defending the b3-pawn, Black threatens
    to capture the d7-pawn with his king, when
   the win should not present any difficulties.
46.¦d2 g6! 47.h3 g5
     Now, dominated on both wings, White is in
    zugzwang.
  41...c4
     
     
40.a4??
A horrible mistake just before the time    
control. Seven moves ago this would have been
the winning move, but now it leads to disaster.
   
After exchanging queens, Vali might have     
understood in a flash that a4 instead would   
have been an efficient method of increasing his
advantage. This thought might have followed    
him subconsciously and, under time pressure,   
he could have advanced the pawn as a reflex
caused by disappointment.
   

After a normal continuation, such as 40.h4, 42.h4
the game would have ended in a draw, for 42.bxc4 loses to 42...b3 when the weakness
instance 40...b4 41.¢h2 ¦xa2 42.¦xc5 ¦a3! of White’s back rank will allow the pawn to
43.¦c6 ¢xd7 with complete equality. promote without needing help from the black
king.
40...b4!
Suddenly the threat ...c4 is very strong. 42...cxb3 43.¢h2 b2
Because of the unfavourable position of his Threatening 44...¦c1.
king, White is in big trouble.
44.¦b1 ¢xd7 45.¢g3 ¢e6
41.¦d1?! 0–1
More stubborn would have been: Although its first step was made to the left,
41.¢f1 the king is aiming for b3 in order to support
Play might have continued: the pawn’s promotion. (The path is longer only
41...c4 42.¢e1 cxb3 43.¢d1 ¦c7 44.¦d2 a5! from an optical point of view. Chess geometry
Threatening to transpose to a winning pawn differs from what we learn in mathematics.)
ending with ...¦xd7. Therefore, White resigned.
The immediate 44...¦xd7 would be
premature because of 45.a5!.