Can black draw this?

[Howard E]

[D]8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - -

This is a position from another posting by Walter Eigenmann.
Some engines play a5 and drew the game versus other engines.
Albert Silver, in a smartly played game, showed a win against the reply a5.

But in some of Walter's games Bc8 was played (Zappa2,
for example). Is there a refutation for Bc8, showing a white win?

Maybe Walter or Albert already have this analysis of Bc8 in there
notes. Thanks.

[Howard E]

This one line somehow draws despite, the obvious advantage to white. My guess is that programs do not like Bc8 because of the line given where white takes the h pawn.If anyone spots errors in this line please post.

Does white have an alternative winning plan? If not then I think this would be a good candidate for a "best move" position that is hard for our top programs of today.

Does Zappa2 for example stick with Bc8 or change after a longer think? Any engine play Bc8?


[Event "Computer chess game"]
[Site "HOWARD-PC"]
[Date "2009.12.09"]
[Round "?"]
[White "Howard"]
[Black "Howard"]
[Result "*"]
[BlackElo "2400"]
[Time "15:29:00"]
[WhiteElo "2400"]
[SetUp "1"]
[FEN "8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 4 3"]
[Termination "unterminated"]
[PlyCount "24"]
[WhiteType "human"]
[BlackType "human"]
[Comment "bm Bc8;"]

3. ... Bc8 4. Kd4 Kd6 5. Bxh7 g6 6. h4 Ke6 7. f5+ Kxf5 8. h5 Kg5 9. hxg6
fxg6 10. Bg8 Kf4 11. Bxd5 g5 12. a4 a5 13. Kc5 Ba6 14. b4 axb4 15. Kxb4 *

[playjunior]

This one line somehow draws despite, the obvious advantage to white. My guess is that programs do not like Bc8 because of the line given where white takes the h pawn.If anyone spots errors in this line please post.

Does white have an alternative winning plan? If not then I think this would be a good candidate for a "best move" position that is hard for our top programs of today.

Does Zappa2 for example stick with Bc8 or change after a longer think? Any engine play Bc8?


[Event "Computer chess game"]
[Site "HOWARD-PC"]
[Date "2009.12.09"]
[Round "?"]
[White "Howard"]
[Black "Howard"]
[Result "*"]
[BlackElo "2400"]
[Time "15:29:00"]
[WhiteElo "2400"]
[SetUp "1"]
[FEN "8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 4 3"]
[Termination "unterminated"]
[PlyCount "24"]
[WhiteType "human"]
[BlackType "human"]
[Comment "bm Bc8;"]

3. ... Bc8 4. Kd4 Kd6 5. Bxh7 g6 6. h4 Ke6 7. f5+ Kxf5 8. h5 Kg5 9. hxg6
fxg6 10. Bg8 Kf4 11. Bxd5 g5 12. a4 a5 13. Kc5 Ba6 14. b4 axb4 15. Kxb4 *

Black has two weaknesses, a6 and d5. Usually this is enough for a win, but white's pawn structure in K-side is spoiled so black has some chances.

So a plan would be: fix the weaknesses, and then try to deprive black of counterplay against your k-side. Then, attack weaknesses one after another. Because there are two of them and they are sufficiently far apart, black will be unable to protect BOTH at some moment and will collapse.

So, you can start by playing b4, so that the pawn on a6 is fixed on a white square. Maybe include Kd4 Kd6.
Then maybe push the h2-pawn to h4.
Then put the a-pawn on a5.

And so on


Bxh7 after Kd4 Kd6 is a "move to avoid" IMO.

[Albert Silver]

[D]8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - -

This is a position from another posting by Walter Eigenmann.
Some engines play a5 and drew the game versus other engines.
Albert Silver, in a smartly played game, showed a win against the reply a5.

But in some of Walter's games Bc8 was played (Zappa2,
for example). Is there a refutation for Bc8, showing a white win?

Maybe Walter or Albert already have this analysis of Bc8 in there
notes. Thanks.

I don't know about a refutation per se, but I played it out against Rybka, same conditions as in the previous game. This time it was characterized by Black getting into zugzwang a couple of times. Note that this sort of thing is a fairly common theme in same-colored bishop endgames. I owe thanks to a master colleague for teaching me to play these a few years back. I am including a few diagrams to help illustrate the problems Black ends up facing.

Essentially, after Bc8 Black will find a bishop that is not only the classic "bad" bishop but it is also reduced to an extremely passive role. Black will be forced to play some king-side pawn moves, opening up further weaknesses, until he just runs out of moves.

[Event "Computer chess game"]
[Site "ALBERT-PC"]
[Date "2009.12.10"]
[Round "?"]
[White "Albert"]
[Black "Rybka 3"]
[Result "*"]
[Time "14:35:37"]
[WhiteElo "2400"]
[TimeControl "900"]
[SetUp "1"]
[FEN "8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 0 1"]
[Termination "unterminated"]
[PlyCount "63"]
[WhiteType "human"]
[BlackType "program"]

1. ... Bc8 2. b4 g6 {(g7g6 a2a4 Ke7d6 Kc3d4 f7f6 b4b5 a6a5 h2h4 Bc8e6 Bd3b1 Be6d7 f4f5 g6xf5 f3f4 Bd7c8 Bb1c2 Bc8e6) -0.72/20 18} 3. a4 Kd6 {(Ke7d6 Kc3d4 f7f6 b4b5 a6a5 h2h4 Bc8e6 Bd3b1 Be6d7 f4f5 g6xf5 f3f4 Bd7c8 Bb1c2 Bc8e6) -0.72/21 6} 4. Kd4 f6 {(f7f6 Bd3f1 h7h6 b4b5 a6a5) -0.88/23 54} 5. h4 Bb7 {(Bc8b7 b4b5 a6xb5 Bd3xb5 Bb7a8 a4a5 Kd6c7 Bb5a4 Ba8b7 Ba4e8) -1.28/18 14} 6. b5 a5 {(a6a5 f4f5 g6xf5 Bd3xf5 h7h6 f3f4 Bb7a8 Bf5g6 Ba8b7 Bg6f7 Bb7a8 Bf7h5 Ba8b7 Bh5g4 Bb7a8) -1.60/20 17} 7. f5 gxf5 {(g6xf5 Bd3xf5 h7h6 f3f4 Bb7a8 Bf5c8 Kd6c7 Bc8a6 Kc7d7 Kd4c5 d5d4 b5b6 Ba8f3 Kc5xd4 Kd7c6 b6b7 Kc6c7 Kd4c5 h6h5 Kc5b5 Bf3c6+ Kb5xa5) -3.20/22 50} 8. Bxf5 h6 {(h7h6 f3f4 Bb7a8 Bf5c8 Kd6c7 Bc8a6 Kc7b6 f4f5 Kb6a7 Ba6c8) -3.20/23 6} 9. f4 Ba8 {(Bb7a8 Bf5c8 Kd6c7 Bc8a6 Kc7d6 f4f5 Kd6d7 Kd4c5 d5d4 b5b6 Ba8e4 Kc5xd4 Be4xf5 Kd4c5 Bf5e4 Kc5b5 f6f5 Kb5xa5 f5f4 Ba6f1 Be4b7 Ka5b5 Kd7d6 a4a5 f4f3 a5a6 Bb7d5 a6a7 Bd5b7 Bf1d3 f3f2) -3.24/23 14} 10. Bc8 Kc7 {(Kd6c7 Bc8a6 Kc7d6 f4f5 Kd6c7 Kd4c5 d5d4 b5b6+ Kc7b8 Ba6d3 Ba8f3 Kc5xd4 Bf3d1 Bd3b5 Bd1c2 Bb5d7 Kb8b7 Kd4c5 Bc2d3 Bd7c6+ Kb7b8 Bc6d5 Kb8c8 Bd5e6+ Kc8b8) -3.35/24 21} 11. Ba6 Kd6 {(Kc7d6 f4f5 Kd6c7 Kd4c5 d5d4 b5b6+ Kc7b8 Ba6d3 Ba8f3 Kc5xd4 Bf3d1 Bd3b5 Bd1c2 Bb5d7 Kb8b7 Kd4c5 Bc2d3 Bd7c6+ Kb7c8 Bc6b5 Bd3xf5) -3.44/24 22} 12. f5 Kc7 {(Kd6c7 Kd4c5 d5d4 b5b6+ Kc7b8 Ba6d3 Ba8f3
Kc5xd4 Bf3d1 Bd3b5 Bd1c2 Bb5d7 Kb8b7 Kd4c5 Bc2d3 Bd7c6+ Kb7c8 Bc6b5 Bd3xf5 Bb5a6+) -3.44/22 13} 13. Kc5 d4 {(d5d4 b5b6+ Kc7b8 Ba6d3) -4.24/24 24} 14. b6+ Kb8 {(Kc7b8 Ba6d3) -4.24/25 93} 15. Bd3 Bf3 {(Ba8f3) -4.24/23 0} 16. Kb5 Kb7 {(Kb8b7 Kb5xa5 h6h5 Bd3a6+ Kb7c6 Ba6b5+ Kc6b7 Bb5d3 Bf3d5 Bd3a6+ Kb7c6 Ba6b5+ Kc6b7 Bb5e2 Bd5e4 Be2xh5 d4d3 Bh5d1 d3d2 h4h5 Be4xf5 h5h6 Bf5h7 Bd1f3+ Kb7b8 Ka5b4 Bh7e4 Bf3d1) -5.03/20 12} 17. Kxa5 h5 {(h6h5 Bd3a6+ Kb7c6 Ba6b5+ Kc6b7 Bb5d3 Bf3d5 Bd3a6+ Kb7c6 Ba6b5+ Kc6b7 Bb5e2 Bd5g2
Be2xh5) -5.03/20 42} 18. Ba6+ Kc6 {(Kb7c6 Ba6b5+ Kc6b7 Bb5d3 Bf3d5 Bd3a6+ Kb7c6 Ba6b5+ Kc6b7 Bb5e2 Bd5g2 Be2xh5 d4d3 Bh5d1 Bg2e4 h4h5) -5.04/20 71} 19. Bb5+ Kb7 {(Kc6b7 Bb5d3) -5.12/20 30} 20. Bd3 Bd5 {(Bf3d5 Bd3a6+) -5.12/19 23} 21. Ba6+ Kc6 {(Kb7c6 Ba6b5+) -5.12/18 30} 22. Be2 Kb7 {(Kc6b7 Be2xh5 Bd5e4 Bh5d1) -5.04/14 11} 23. Bxh5 d3 {(d4d3 Bh5d1) -5.12/14 5} 24. Bd1 d2 {(d3d2 h4h5) -5.59/13 45} 25. h5 Be4 {(Bd5e4 h5h6) -6.42/14 12} 26. h6 Bxf5 {(Be4xf5 Ka5b5) -6.42/14 4} 27. Kb5 Bd3+ {(Bf5d3+ Kb5c5) -6.33/12
6} 28. Kc5 Ka6 {(Kb7a6 a4a5) -6.33/11 3} 29. a5 Kb7 {(Ka6b7 Kc5d4) -7.60/13 4} 30. Kd4 Bf5 {(Bd3f5 Kd4c3) -8.30/13 13} 31. Ke3 Ka6 {(Kb7a6 Ke3xd2) -8.59/13 5} 32. Kxd2 Bg6 {(Bf5g6 Bd1e2+) -9.08/13 8} *

After 9...Ba8

[D]b7/8/3k1p1p/pP1p1B2/P2K1P1P/8/8/8 w - - 1 10

Here the zugzwang is easy to see coming. Black cannot play the loose pawns since they will be immediately captured by the bishop.

After 11...Kd6 comes zugzwang no.1 with 12.f5

[D]b7/8/B2k1p1p/pP1p4/P2K1P1P/8/8/8 w - - 5 12

The here is another crucial point, though Black is clearly lost already.

[D]8/1k6/1P3p1p/K4P2/P2p3P/3B1b2/8/8 b - - 0 17

Black played 17...h5 which loses the h-pawn obviously, and will provide a passed pawn on the opposite wing, however there aren't any real alternatives. If 17...Bd5 or another waiting move, white just plays 18.Kb5 Bc6+ 19.Kc5 Bxa4 20.Be4+ Kb8, and the white king cleans up all the black pawns. If 17...Bc6 then 18.Ba6+ Kb8 19.Kb4 and again white is free to capture all black's pawns.

I don't claim this is proof, merely that it is one way to win it.

Albert

[Howard E]

Thanks, I'll explore these themes you suggest in greater detail.
Do you have a line where white definitely wins? I realize it is a nice plan and advantage for white but the position has some quirky elements, as you have pointed out. In the meantime I'll attempt to find a winning white line as you suggest(Junior2010 by the way played similarly with the white pawns on a5 and b4 but go no knockout punch).

[Albert Silver]

Thanks, I'll explore these themes you suggest in greater detail.
Do you have a line where white definitely wins? I realize it is a nice plan and advantage for white but the position has some quirky elements, as you have pointed out. In the meantime I'll attempt to find a winning white line as you suggest(Junior2010 by the way played similarly with the white pawns on a5 and b4 but go no knockout punch).

It is hard to say, because some moves require changes. 4...f6 means I cannot enter e5 if the Black king leaves d6. What does this change? Well, my previous plan to strangle black was a5, b5, Bxb5, and a6 but now he has Kc6 after a5, preventing b5 with no white king penetration in e5, so it is much less obvious. Hence playing b5 instead of a5.

The overall theme is unchanged though. Black has three terrible problems that white will seek to press to convert:

1) the pawn endgame is plain dead, which says a lot. The passed a-pawn would decided it easily with no bishops.

2) The black king and bishop are tied down to passive roles, mostly waiting and praying.

3) Black's d5 pawn. Thanks to it, Black quickly finds its bishop reduced to moving back and forth on two squares, which accelerates the zugzwang.

Albert

[Howard E]

Thanks. I appreciate the detail you give from your game analysis. It is also nice to observe that our powerful chess engines sometimes miss these deep kinds of analysis.

[jarkkop]

8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 0 1


1...Bb5 2.Bxb5 axb5 3.Kd4 Kd6 4.f5 h5 5.a3 h4 6.h3 Kc6 7.b3 Kd6 8.a4 Kc6 9.f4 b4 10.f6 g6 11.Ke5 Kc5 12.a5 d4 13.a6 Kb6
= (0.25) Depth: 19/38 00:00:07 10180kN
1...Bb5 2.Bxb5 axb5 3.Kd4 Kd6 4.f5 Kc6 5.a3 Kd6 6.b3 Kc6 7.a4 bxa4 8.bxa4 Kd6 9.f4 Kc6 10.h3 Kb6 11.Kxd5 Ka5 12.Ke5 f6+ 13.Ke4 Kxa4
² (0.33) Depth: 20/38 00:00:13 18895kN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 Bc8 5.a5 Kc6 6.Bb5+ Kb7 7.Kc5 Ka7 8.Kxd5 Bb7+ 9.Bc6 Ba6 10.Kc5 Be2 11.Be4 Ka6 12.Kb4 h5 13.h4 f6 14.f5 Bf1 15.Bd5 Bd3
± (0.71) Depth: 21/56 00:02:53 247mN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 Bc8 5.a5 Kc6 6.Bxh7 Ba6 7.Bg8 f6 8.f5 Be2 9.f4 Bd3 10.Be6 d4 11.h4 Be4 12.h5
² (0.63) Depth: 22/56 00:03:21 286mN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 f5 5.a5 Be6 6.Kb5 Kc7 7.a6 Kb8 8.Kb6 d4 9.Kc5 Ka7 10.Kxd4 g6 11.Ke5 Bd7 12.Bc4 Bc6 13.Be2 Be8 14.Kd4 Bc6 15.Kc5 Be8
² (0.66) Depth: 22/56 00:04:02 343mN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 g6 5.a5 f5 6.a6 Bc6 7.Be2 Bd7 8.Bf1 Bc6 9.Bd3 Kc7
± (0.71) Depth: 23/56 00:05:05 432mN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 g6 5.a5 Be6 6.Kb5 Kc7 7.a6 f5 8.Kc5 Kb8 9.Kb6 d4 10.Kc5 Ka7 11.Kxd4 Bb3 12.Be2 Bg8 13.Kc5
² (0.68) Depth: 24/56 00:07:33 652mN
1...a5 2.b4 axb4+ 3.Kxb4 Kd6 4.a4 g6 5.a5 Be6 6.Kb5 Kc7 7.a6 f5 8.Kc5 Kb8 9.Kb6 d4 10.Be2 Bb3 11.Kc5 d3 12.Bxd3 Ka7 13.Be2 Be6 14.Kd4 Bg8 15.Bf1 Be6 16.h4 Bc8 17.Ke5
± (0.75) Depth: 25/58 00:19:06 1697mN

if Bc8 then

2.Kd4 a5 3.a3 Ke6 4.b4 axb4 5.axb4 f5 6.Kc5 d4 7.b5 Bb7 8.Kxd4 Bxf3 9.Kc5 h6 10.h4 g6 11.Bc4+ Kd7 12.b6 Bb7 13.Bb5+
± (0.87) Depth: 21/50 00:00:50 68857kN
2.Kd4 a5 3.a3 Ke6 4.b4 axb4 5.axb4 Bb7 6.Kc5 h6 7.b5 Kd7 8.f5 Kc8 9.f4 f6 10.b6 Kb8 11.h4 Kc8 12.Bb1 Kb8 13.Ba2 Bc8 14.Bxd5 Bxf5 15.h5
± (0.87) Depth: 22/50 00:01:12 100mN
2.Kd4 a5 3.a3 Ke6 4.b4 axb4 5.axb4 Bb7 6.Kc5 h5 7.Bc2 d4 8.Bb3+ Ke7 9.Kxd4 Bxf3 10.b5 f6 11.Bd5 Bg4 12.b6 Kd7 13.Bb7 h4 14.Kc5 Bf5 15.Bf3 Kc8 16.Bc6 Kb8 17.Bd5
± (0.95) Depth: 23/51 00:02:44 234mN
2.Kd4 a5 3.a3 Ke6 4.Kc5 g6 5.b4 axb4 6.axb4 Kd7 7.f5 Kc7 8.fxg6 hxg6 9.f4 Bg4 10.Kxd5 Kb6 11.h4 f5
± (1.03) Depth: 23/58 00:05:16 457mN
2.Kd4 a5 3.a3 Kd6 4.Bxh7 g6 5.h4 Ke6 6.f5+ Kxf5 7.h5 Be6 8.f4 Kxf4 9.hxg6 fxg6 10.Bxg6 Bg4 11.Be8 Be2 12.Bc6 Bg4 13.Bxd5 Bd7 14.Be4 a4
± (1.03) Depth: 23/58 00:06:38 580mN
2.Kd4 a5 3.a3 Kd6 4.Bxh7 g6 5.h4 Ke7 6.f5 Bxf5 7.h5 Kf8 8.h6 a4 9.Kxd5 Be6+ 10.Ke5 Ke7 11.f4 f6+ 12.Kd4 Kf7 13.Ke4 Bf5+ 14.Kd4 Be6
± (1.10) Depth: 24/60 00:10:42 944mN
2.Kd4 a5 3.a3 Kd6 4.Bxh7 g6 5.h4 Ke6 6.f5+ Kxf5 7.h5 Kg5 8.hxg6 fxg6 9.b4 axb4 10.axb4 Bd7 11.Ke5 d4 12.Kxd4 Kf6 13.Bg8 g5 14.Bd5 g4 15.fxg4 Bxg4 16.b5 Ke7 17.b6 Kd6
± (1.06) Depth: 25/60 00:13:47 1225mN

[Albert Silver]

8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 0 1

if Bc8 then

2.Kd4 a5 3.a3 Kd6 4.Bxh7 g6 5.h4 Ke6 6.f5+ Kxf5 7.h5 Kg5 8.hxg6 fxg6 9.b4 axb4 10.axb4 Bd7 11.Ke5 d4 12.Kxd4 Kf6 13.Bg8 g5 14.Bd5 g4 15.fxg4 Bxg4 16.b5 Ke7 17.b6 Kd6
± (1.06) Depth: 25/60 00:13:47 1225mN

Yes, this is a pretty clueless engine line. Which one BTW?

[jarkkop]

[D]8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - -

This is a position from another posting by Walter Eigenmann.
Some engines play a5 and drew the game versus other engines.
Albert Silver, in a smartly played game, showed a win against the reply a5.

But in some of Walter's games Bc8 was played (Zappa2,
for example). Is there a refutation for Bc8, showing a white win?

Maybe Walter or Albert already have this analysis of Bc8 in there
notes. Thanks.

I don't know about a refutation per se, but I played it out against Rybka, same conditions as in the previous game. This time it was characterized by Black getting into zugzwang a couple of times. Note that this sort of thing is a fairly common theme in same-colored bishop endgames. I owe thanks to a master colleague for teaching me to play these a few years back. I am including a few diagrams to help illustrate the problems Black ends up facing.

Essentially, after Bc8 Black will find a bishop that is not only the classic "bad" bishop but it is also reduced to an extremely passive role. Black will be forced to play some king-side pawn moves, opening up further weaknesses, until he just runs out of moves.

[Event "Computer chess game"]
[Site "ALBERT-PC"]
[Date "2009.12.10"]
[Round "?"]
[White "Albert"]
[Black "Rybka 3"]
[Result "*"]
[Time "14:35:37"]
[WhiteElo "2400"]
[TimeControl "900"]
[SetUp "1"]
[FEN "8/3bkppp/p7/3p4/5P2/2KB1P2/PP5P/8 b - - 0 1"]
[Termination "unterminated"]
[PlyCount "63"]
[WhiteType "human"]
[BlackType "program"]

1. ... Bc8 2. b4 g6 {(g7g6 a2a4 Ke7d6 Kc3d4 f7f6 b4b5 a6a5 h2h4 Bc8e6 Bd3b1 Be6d7 f4f5 g6xf5 f3f4 Bd7c8 Bb1c2 Bc8e6) -0.72/20 18} 3. a4 Kd6 {(Ke7d6 Kc3d4 f7f6 b4b5 a6a5 h2h4 Bc8e6 Bd3b1 Be6d7 f4f5 g6xf5 f3f4 Bd7c8 Bb1c2 Bc8e6) -0.72/21 6} 4. Kd4 f6 {(f7f6 Bd3f1 h7h6 b4b5 a6a5) -0.88/23 54} 5. h4 Bb7 {(Bc8b7 b4b5 a6xb5 Bd3xb5 Bb7a8 a4a5 Kd6c7 Bb5a4 Ba8b7 Ba4e8) -1.28/18 14} 6. b5

Albert

More resistance with 6. .. axb5

6...axb5 7.Bxb5 h6 8.a5 Ba8 9.a6 g5 10.hxg5 hxg5 11.fxg5 fxg5 12.Ba4 Kc7 13.Bb3 Kd6 14.Ba2 Bc6 15.a7 Ba8 16.Bb1 Kc7 17.Bf5 Kd6 18.Bc8 Kc6 19.Be6 Kd6 20.Bf7 Bb7 21.Bg8 Bc6 22.Bh7
± (0.99) Depth: 25/53 00:05:29 518mN