Endgame Explorations 9: Grigoriev

Noam Elkies

Thanks to Mike Glinski for filling in for me during my absence. Following the example of his column on the Platov brothers, I shall devote each of the next few columns to the work of a single composer.

This column's featured composer is Nikolai Dmitriyevich Grigoriev, whose thorough investigations of both the artistic and the practical aspects of pawn endgames form the basis of the modern theory of these endgames. In The Oxford Companion to Chess (by D. Hooper and K. Whyld, Oxford University Press 1984) we find that Grigoriev, born in 1895, was also a very strong tournament player, winning the Moscow Championship four times during the 1920's and coming fifth in the 1920 USSR Championship. But it is in endgame composition and analysis that Grigoriev made his most enduring contributions, and in his specialty, the pure pawn endgame, he had no equal. Two years before Grigoriev's premature death in 1938 from complications of an appendicitis operation, the French magazine La Stratégie organized a tourney for endgame studies with two pawns against one, and Grigoriev ran away with ten of the twelve awards, including this composition which shared first place:

(diagram) [White to play and win]

To support the advance of the e-pawn White must give up the h-pawn and answer ... Kxh4 with Kf4. To ensure that this is possible White engages in a delicate struggle of corresponding squares:

1 Kg3! Ke4 2 Kg2!
A mutual Zugzwang. White to move only draws after 3 Kf1 (3 Kg3 Ke3) 3 ... Kf5! 4 Kf2 Kf4! and again it's mutual Zugzwang, since after 5 e3+ (or K-any) 5 ... Kg4, White no longer has 6 Ke3. Likewise 2 Kh2? fails to 2 ... Kd4! 3 Kg1 Ke5! etc.
2 ... Ke3 3 Kf1 Ke4 4 Ke1!
Again, not 4 Kf2? Kf4 with mutual Zugzwang. The White king heads to the other side of the e-pawn.
4 ... Ke3 5 Kd1 Kf4 6 Kd2 Ke4
Now 6 ... Kg4 7 Ke3 allows White to attain his goal. Nor would 5 ... Ke4 6 Kd2 Kf4 have helped because 7 Kd3 would preserve the king's access to f4.
7 e3 Kf3 8 Kd3 Kg3 9 Ke4 Kg4 10 Ke5 Kxh4
This can no longer be postponed, but White must still play carefully.
11 Kf4 Kh3 12 e4 Kg2 13 e5!
Not 12 Kf3? h4! or 13 Kg5? Kg3! drawing. Now White reaches a theoretically won queen endgame.
13 ... h4 14 e6 h3 15 e7 h2 16 e8/Q h1/Q 17 Qe2+ Kg1
Or 17 ... Kh3 18 Qg4+ Kh2 19 Qg3#.
18 Kg3
And White wins.

Our remaining examples come from 1234 Modern End-Game Studies, compiled in 1938 by M. A. Sutherland and H. M. Lommer and reprinted in 1968 by Dover Publications. We begin with another pawn endgame, this a second prize winner in Shakhmatny Listok, 1929:

(diagram) [White to play and draw]

White is nominally a pawn ahead here, but taking account of the tripled b-pawns effectively puts Black a pawn ahead, and with normal play Black would win by using the king to advance the d-pawn and at the right moment trading that pawn for White's triplets to reach a theoretical win of king and pawn against king. Instead White finds a way of using the tripled pawns to reach an unusual and surprising draw:

1 Kg2 Kc7 2 Kf3 Kd7
Not 2 ... Kb6 3 Ke4 Kxb5 4 Kd5 Kxb4 5 Kxd6 Kxb3 (or 5 ... b5 6 Kc6) 6 Kc5 and the last pawn falls.
3 Kf4!
We will see that the immediate 3 Ke4? loses to 3 ... Ke6, so White triangulates, but not with 3 Ke3? d5! 4 Kd4 Ke6! which wins.
3 ... Ke6
Or 3 ... Ke7 4 Kf5 Kf7 5 b6 and 6 Ke4 draws -- see below.
4 Ke4!
And now 4 ... b6 5 Kd4 d5 6 Ke3! Ke5 7 Kd3 d4 8 Kc4! Ke4 stalemates, as does the "echo" line 4 ... d5+ 5 Kd4 Kd6 (5 ... b6 6 Ke3! transposes) 6 b6 Ke6 7 b5 Kd6 8 b4 Ke6 9 Kc5! Ke5.

(diagram) [White to play and win]

Another two-against-one pawn endgame, one that has since been rediscovered independently several times since its first appearance in 1930. It seems incredible that White must move his king to h8 to win!

1 b3
Not 1 b4? Ka6 tying the White king to the eighth rank, nor 1 Kb8? b4! forcing off all the pawns because 2 c4?? even loses to 2 ... b3!
1 ... Ka5!
Or 1 ... b4 2 c4 Ka6 transposing, but this gives White an extra chance to go wrong. Not, however, 1 ... Ka6? 2 b4! Kb6 3 Kb8 Kc6 4 Ka7 and wins at once.
2 Kb8!
2 b4+? Ka6! again draws, while 2 Ka7(b7)? b4 3.c4 is stalemate!
2 ... b4
Forced, since 2 ... Ka6 3 Kc7 or 2 ... Kb6 3 b4 lose more quickly.
3 c4 Kb6
So Black has the opposition after all, but now the c-pawn is passed instead of backward and demands some of the Black king's attention.
4 Kc8 Kc6 5 Kd8 Kd6 6 Ke8 Ke6 7 Kf8 Kf6 8 Kg8 Kg6 9 Kh8! Kf6!
Not 9 ... Kh6? 10 c5 and promotes, so the White king has escaped the eighth rank, but Black still has a horizontal opposition.
10 Kh7 Kf7 11 Kh6 Kf6 12 Kh5 Kf5 13 Kh4 Kf4 14 Kh3! Kf5
Again Black dare not follow with 14 ... Kf3? 15 c5, so the White king finally gets back to the Queenside:
15 Kg3 Kg5 16 Kf3 Kf5 17 Ke3 Ke5 18 Kd3
And since 18 ... Kd5 is illegal, Black must allow 19 Kd4, 20 c5 and 21 Kc4, winning.

And finally we see Grigoriev venture outside the realm of the pawn endgame in the following study (second prize, Shakhmatny, 1928):

(diagram) [White to play and draw]

We correctly expect a pin-stalemate, but the pin does not come on the first rank!

1 a8/Q!
Not 1 Kxh2 Kf3 and White gets mated.
1 ... Qxa8+ 2 b7 Qa7
White gets his stalemate sooner after 2 ... Qxd8 3 b8/Q+ Qxb8 4 Bg3+ or 2 ... Qb8 3 Bg3+!
3 Bf2! Qb8
Not 3 ... Qa1+? 4 Kxh2 Qb2 5 b8/Q+ Qxb8 6 Bg3+ and wins. Note that without the pawn at h7, Black would prevail here with 4 ... Qh8+ 5 Kg2 Qg8+! and 6 ... Qxd8.
4 Bg3+! Kxg3 5 Nc6 Qxb7
And it's stalemate again, thanks to the new pin on the long diagonal!

Next column, we will turn from the subtleties of Grigoriev to the heroics of A. A. Troitzky.


Noam Elkies is Professor of Mathematics at Harvard University. This article originally appeared in both Chess Horizons and the American Chess Journal website.

Next column: Troitzky.


Posted by permission of the author, 7-13-2004.
This page last modified 7-13-2004 by wdj.