Wrong color Bishop, test position

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
User avatar
michiguel
Posts: 6389
Joined: Thu Mar 09, 2006 7:30 pm
Location: Chicago, Illinois, USA
Contact:

Wrong color Bishop, test position

Post by michiguel » Sun Jan 26, 2014 10:37 pm

[D]8/1Bn5/P2k3p/8/2K2pp1/8/6PP/8 b - - 2 44 bm f3

This was in Crafty Gaviota, CCT16. Gaviota let itself got into trouble for lacking some knowledge that Crafty has, but it was saved by the bell based on another piece of knowledge.

f3 draws
Not so difficult, but it is instructive. Some engines may not see the draw.

Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6

zullil
Posts: 5684
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: Wrong color Bishop, test position

Post by zullil » Sun Jan 26, 2014 11:50 pm

michiguel wrote: Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6
Latest Stockfish (16 threads, 4 GB hash, no tablebases) avoids a6 but doesn't see a win here, at least so far:

Code: Select all

info depth 52 seldepth 73 score cp 18 nodes 80067282816 nps 29358926 time 2727187 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c5 b7d5 c5d7 d5e4 d7c5 e4f3

User avatar
michiguel
Posts: 6389
Joined: Thu Mar 09, 2006 7:30 pm
Location: Chicago, Illinois, USA
Contact:

Re: Wrong color Bishop, test position

Post by michiguel » Mon Jan 27, 2014 12:06 am

zullil wrote:
michiguel wrote: Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6
Latest Stockfish (16 threads, 4 GB hash, no tablebases) avoids a6 but doesn't see a win here, at least so far:

Code: Select all

info depth 52 seldepth 73 score cp 18 nodes 80067282816 nps 29358926 time 2727187 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c5 b7d5 c5d7 d5e4 d7c5 e4f3
Yes, I am not sure it is ultimately a win because the wrong color bishop problem could show up at any time, but at least white should keep the chances alive.

Miguel

zullil
Posts: 5684
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: Wrong color Bishop, test position

Post by zullil » Mon Jan 27, 2014 12:08 am

zullil wrote:
michiguel wrote: Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6
Latest Stockfish (16 threads, 4 GB hash, no tablebases) avoids a6 but doesn't see a win here, at least so far:

Code: Select all

info depth 52 seldepth 73 score cp 18 nodes 80067282816 nps 29358926 time 2727187 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c5 b7d5 c5d7 d5e4 d7c5 e4f3

Code: Select all

info depth 53 seldepth 73 score cp 13 nodes 113334100351 nps 30229972 time 3749064 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c7 b7e4 c7a6 e4h1 a6c5 h1d5 c5d7 c4d4 d7c5 d5c4 d6c6 a5a6 c5a6 c4a6 f4f3 a6c4 c6d6 d4e4 f3f2 e4f5 d6c5 c4f1 c5d6 f5g6 d6e5 g6g5 e5e6 f1c4
Hit 30M nps :)

zullil
Posts: 5684
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: Wrong color Bishop, test position

Post by zullil » Mon Jan 27, 2014 12:27 am

zullil wrote:
zullil wrote:
michiguel wrote: Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6
Latest Stockfish (16 threads, 4 GB hash, no tablebases) avoids a6 but doesn't see a win here, at least so far:

Code: Select all

info depth 52 seldepth 73 score cp 18 nodes 80067282816 nps 29358926 time 2727187 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c5 b7d5 c5d7 d5e4 d7c5 e4f3

Code: Select all

info depth 53 seldepth 73 score cp 13 nodes 113334100351 nps 30229972 time 3749064 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c7 b7e4 c7a6 e4h1 a6c5 h1d5 c5d7 c4d4 d7c5 d5c4 d6c6 a5a6 c5a6 c4a6 f4f3 a6c4 c6d6 d4e4 f3f2 e4f5 d6c5 c4f1 c5d6 f5g6 d6e5 g6g5 e5e6 f1c4
Hit 30M nps :)

Code: Select all

info depth 54 seldepth 73 score cp 13 nodes 127861146394 nps 30412775 time 4204192 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c7 b7e4 c7a6 e4h1 a6c5 h1d5 c5a6 d5g2 a6c5 g2d5

info depth 55 seldepth 73 score cp 13 nodes 152434339903 nps 30435353 time 5008463 multipv 1 pv a8e4 e6c5 e4d5 h6h5 h2h4 g4h3 g2h3 h5h4 d5f3 c5a6 f3b7 a6c7 b7e4 c7a6 e4h1 a6c5 c4b4 c5a6 b4b5 a6c5 b5c4 c5a6 h1e4 a6c5

bob
Posts: 20642
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: Wrong color Bishop, test position

Post by bob » Mon Jan 27, 2014 4:38 am

michiguel wrote:[D]8/1Bn5/P2k3p/8/2K2pp1/8/6PP/8 b - - 2 44 bm f3

This was in Crafty Gaviota, CCT16. Gaviota let itself got into trouble for lacking some knowledge that Crafty has, but it was saved by the bell based on another piece of knowledge.

f3 draws
Not so difficult, but it is instructive. Some engines may not see the draw.

Probably here, a6 should be avoided.

[D]B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 3 43 am a6
I went back and looked at the log. I suspect it is drawn BEFORE a6 in fact. Crafty kept bouncing back and forth between a6 and Bd5 but the scores were almost identical. Currently I don't classify the wrong bishop + RP as a 0.0 type draw, but divide the score by 4. I think I am going to revisit this a bit and try to classify endgames a bit more accurately. Call this one a certain draw, where some (bishops of opposite color) are "drawish". And there are some in between those extremes. Needs a little tuning.

User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 7:33 pm
Location: Antalya, Turkey
Contact:

Re: Wrong color Bishop, test position

Post by beachknight » Mon Jan 27, 2014 6:30 pm

Quick analysis of the position with Crafty 23.01 CB:

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.01 3cpu:

43.a6 Nc7 44.Bb7 f3 45.gxf3 Nxa6 46.fxg4 Nc5 47.Bd5 Nd7 48.Kd4 Nf6 49.Bf3 Nxg4 50.Bxg4 h5 51.Bxh5 Ke6 52.h4 Kf6
+/- (1.42) Depth: 20/36 00:00:15 67062kN
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.01:

43.a6 Nc7 44.Bb7 f3 45.gxf3 Nxa6 46.fxg4 Nc7 47.Kd4 Ne6+ 48.Ke3 Ke5 49.Bf3 Nd4 50.Be4 Ne6 51.h4 Nd4 52.Bd3 Ne6 53.Bc2 Nd4 54.Be4
+/- (1.30) Depth: 22/38 00:00:45 204MN
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.01:

43.a6 Nc7 44.Bb7 f3 45.gxf3 Nxa6 46.fxg4 Nc5 47.Ba8 Ne6 48.Kd3 Ke5 49.Ke3 Nd4 50.Be4 Nb5 51.h4 Nc3 52.Bd3 Nd1+ 53.Kf3 Nc3 54.Bg6 Nb5 55.Be4
+/- (1.30) Depth: 24/42 00:02:43 738MN
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.01:

43.a6 Nc7 44.Bb7 f3 45.gxf3 Nxa6 46.Bxa6 gxf3 47.Kd4 f2 48.Ke3 Ke5 49.Kxf2 Kf4 50.Be2 Kf5 51.Kg3 Kg6 52.Kg4 h5+ 53.Kh4 Kh6 54.Bxh5 Kg7 55.Be8 Kf6
+/- (1.38) Depth: 25/45 00:06:12 1652MN
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.01:

43.a6 Nc7 44.Bb7 f3 45.gxf3 Nxa6 46.Bxa6 gxf3 47.Kd4 f2 48.Ke3 Ke5 49.Kxf2 Kf4 50.Be2 Kf5 51.Kg3 Kg6 52.Kg4 h5+ 53.Kh4 Kh6 54.Bxh5 Kg7 55.Kg5 Kh7 56.h4
+/- (1.40) Depth: 26/46 00:11:37 3030MN
(Chess Engines Testing, 27.01.2014)

Both Crafty 23.05 CB versions (with 3cpu) crash. :(
hi, merhaba, hallo HT

User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 7:33 pm
Location: Antalya, Turkey
Contact:

Re: Wrong color Bishop, test position

Post by beachknight » Mon Jan 27, 2014 6:41 pm

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.05 x64 1cpu:

43.Bb7 Ng5 44.Kd4 Ne6+ 45.Kc3 Nc5 46.a6 f3 47.gxf3 Nxa6 48.fxg4 Nc5 49.Bg2 Ke5 50.h3 Ne6 51.Ba8 Nf4 52.h4 Ng6 53.g5
+/- (1.12) Depth: 20/37 00:03:06 332MN, tb=131760
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.05 x64:

43.a6 f3 44.gxf3 Nc7 45.Bb7 Nxa6 46.fxg4 Nc5 47.Bg2 Ne6 48.Bd5 Nf4 49.Bb7 Ke5 50.Bc8 Ke4 51.h4 Ke5 52.Bb7 Ng6 53.g5
+/- (1.18) Depth: 21/36 00:06:43 874MN, tb=270571
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.05 x64:

43.Bb7 Ng5 44.Kd4 Ne6+ 45.Kc3 Nc5 46.a6 Kc7 47.Kd4 Kb6 48.Ke5 f3 49.gxf3 gxf3 50.Bxf3 Nxa6 51.h4 Nc5 52.Kf6 Nd7+ 53.Ke6 Nc5+ 54.Kf5 Kc7
+/- (1.20) Depth: 21/39 00:10:07 1366MN, tb=419263
(Chess Engines Testing, 27.01.2014)

21:
B7/8/3kn2p/P7/2K2pp1/8/6PP/8 w - - 0 1

Analysis by Crafty 23.05 x64:

43.Bb7 Ng5 44.Bc8 f3 45.gxf3 gxf3 46.Kd3 Kc7 47.Bg4 f2 48.Ke2 Kc6 49.Kxf2 Kb5 50.Kg3 Kxa5 51.Kh4 Ne4 52.Bf5 Nd6 53.Be6 Ne4 54.Bf5
+/- (1.21) Depth: 22/41 00:14:12 1936MN, tb=538787
(Chess Engines Testing, 27.01.2014)

Crafty 23.05 x64 2cpu crashes. :(
hi, merhaba, hallo HT

Post Reply