In a 1993 game played by Wojtkiewitz vs Khalifman, the following endgame appeared on the board. Black to move.

[FEN "5k2/1p4p1/1p2bp1p/3p4/5PP1/P3P2P/1P1K2B1/8 b - - 0 1"]

Stockfish gives +0.9 for white.

Is the position a win for White or a draw?

  • if sleepy's answer is satisfying, you should consider accepting with the tick so that this question is considered solved.
    – Evargalo
    Commented May 28, 2021 at 13:27

1 Answer 1


I have seen this position in Dvoretsky's endgame manual. According to his analysis, white is winning after Ke7, Kc3 Kd6, a4!, eventually getting the pawns to b5 and f5, king to d4 and breaking through with e4 at the right moment.

Update. Actually, I have checked it, and in the book the position is already after Ke7, Kc3 Kd6. Here white has to play a4 to stop b5. Importantly, earlier b5 does not work for black either, because white has Kc3-b4 winning the b5 pawn.

Update 2. As @Evargalo correctly pointed out, the key to this position is the pawn ending arising after 1. ... Ke7 2. Kc3 Kd6 3. a4 g5 4. Kd4 Bf7 5. Bf3 Be6 6. f5 Bf7 7. b4 Be8 8. b5 Bf7 9. Bd1 Be8 10. Bb3 Bf7 11. e4 bg8 12. Ba2 Bf7 (probably a time-winning pass) 13. Bxd5 Bxd5 14. exd5 Kc7.

[FEN "8/1pk5/1p3p1p/1P1P1Pp1/P2K2P1/7P/8/8 w - - 0 1"]

Here the game continued: 15. Kc3 (aiming at Kb4 and a5 to enable Kc5 and d6) Kd6 16. Kc4 (triangulation). Now:

(1) 16. ... Ke5 17. a5 bxa5 18. Kc5 a4 19. d6 b6+ 20. Kc6 a3 21. d7 a2 22. d8Q a1Q 23. Qd6+ Ke4 24. Kxb6 and apparently white is winning thanks to the b-pawn. White gets his queen to c6, king to c8 and push the pawn. Black resigned after 10 more moves;

(2) 16. ... Kd7 17. Kb4 Kd6 18. a5 Kxd5 19. a6 bxa6 20. bxa6 Kc6 21. Ka4 b5+ 22. Ka5;

(3) 16. ... Kd7 17. Kb4 Kd6 18. a5 bxa5+ 19. Kxa5 Kxd5 20. Kb6 Kc4 21. Kxb7 Kxb5 22. Kc7 and the pawn endgame is winning.

Actually, at first I thought it is an example of impressive calculation from the white player, but then maybe, it all can be found by elimination. You see that getting your pawn to b5 is your best bet to win the bishop endgame, then you have to go to the pawn endgame because otherwise it is a fortress, and then it is a draw unless you go for the queen endgame where the b-pawn is your only chance. So maybe it is an example of how avoid unnecessary calculations.

  • Very interesting. Can you show some variations ? Notably, I cannot see exactly how White breaks through if after e4 (in a position like w: Kd4,Ph3g4f5b5a4,Bf3 or Bb3 / b: Kd6, Bf7,Ph6g7f6d5b7b6) Black keeps waiting with Bf7-g8-f7. Then ed5 leads to a simple fortress and even in the pawn endgame in case of Bxd5 Bxd5 ed5 Kd7 the entry squares e5 and c5 are controled and a5?? ba5 Kc5 a4 even loses. I certainly miss smg obvious, or that Dvoretsky explains.
    – Evargalo
    Commented May 25, 2021 at 7:04
  • To complement, even if White starts with his bishop on b3, enters the pawn endgame by Bd5 Bd5 ed5 Kd7, sacs his extra pawn with d5-d6 and use the free moves of his h-pawn to force entry on d5, yet Black can resist with Ke7-d7-e7. Any g4-g5 is ignored and a4-a5 never seems to work. I am really puzzled about how Dvoretsky expects to win.
    – Evargalo
    Commented May 25, 2021 at 7:49
  • Is ...b5 really a threat though?
    – David
    Commented May 25, 2021 at 8:54
  • @David I suppose the idea is that after black plays b5 and b6, he keeps his king on d6 and there is no zugzwang because white bishop cannot attack both d5 and b5 at the same time. And Kc3 is met with Kc5, so b5 does not fall either. But I agree it is very far from being obvious.
    – sleepy
    Commented May 25, 2021 at 9:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.