[fen]8/2n1b1kp/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 w - - 0 33[/fen]
Some engines find the best move almost at once. Some take a very long time. Please post analysis with your engines.
Test Position
Moderators: Harvey Williamson, Steve B, Watchman
Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
-
Carl Bicknell
- Member
- Posts: 252
- Joined: Mon May 04, 2009 10:06 pm
OK, here's what I did:
I made an IDeA project to get to the truth of the position. Using 16 cores I ran it for several hours using 16 instances of Houdini 2.
My claim is that 1 Nf5+! is the best move (1 Nh1 is second best).
The critical line is as follows:
1. Nf5+ gxf5 (1... Kg8 2. Nh6+ Kg7 3. Qg3 Bd6)(1... Kh8 2. Nxe7 Qxe7 3. Bxb4 Qxb4 4. Qf6+ Kg8 5. Qd8+)
2. exf5 Qf7 3.f6+ Bxf6 4. gxf6+ Qxf6 5. Qb7 Nbd5 (5... Qg6+ 6. Ka2 Nbd5 (6... Qd6 7. Bxb4 Qxb4 8. Qxc7+)7. Bc4 Qf7 8. Qc6) 6. Bc4 Qf7 7. Ka1 d3 8. Bxb4 Qf1+ 9. Ka2
Nxb4+ 10. Qxb4
From this, my suspicion is that the engines that do well on this position are probably the ones that highly negatively evaluate black's king position. Of course it's possible that they see right to the end, but my hunch is this is an eval issue.
Since HIARCS does well here maybe you can confirm / deny this...?
Anyway, to answer your question, here are some results from infinite analysis:
Houdini 2 struggles:
= (0.24) Depth: 20/45 00:00:01 23495kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Bf1 Bd6 7.Be1 Ke7 8.Bb5 Ke6 9.Bd2 Ke7 10.Be2 Kf7 11.Bf1 Ke7 12.Be1 Kf7 13.Bg2 Ke6 14.Nh6 Ke7 15.Bf1 Kf8 16.Bd3 Kg7 17.Ng4 Kf7 18.Bb5 Ke6 19.Bd2
= (0.24) Depth: 25/50 00:00:07 202mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Bf1 Bd6 7.Be1 Ke7 8.Bb5 Ke6 9.Bd2 Ke7 10.Be2 Ke6 11.Nf2 Nc5 12.Be1 Kf7 13.Bf3 Ke7 14.Bg4 Ne6 15.Nd3 Nc5 16.Nxb4 Nxe4 17.Nc6+ Kf7 18.Bf3 Nc5 19.b4 Ne6 20.b5
= (0.24) Depth: 26/50 00:00:11 337mN
1.Nh1
= (0.29 ++) Depth: 27/58 00:00:23 752mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Bxc7 11.b4 Kc6 12.Bf1 Bd6 13.Kb3 Bf8 14.Kc4 Kb6 15.Bd2 Bd6 16.Bxf4 exf4 17.Kxd4 Bxb4
+/= (0.39) Depth: 27/63 00:00:39 1312mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Kb6 14.Bd2 Bb8 15.b5 Bd6 16.Bb4 Bxb4 17.Kxb4 Kb7 18.Kc4 Kc7 19.Kb3 Kd7 20.Kc2 Kc7 21.Kd2 Kb6
+/= (0.42) Depth: 28/63 00:01:15 2588mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Bd6 14.b5+ Kb6 15.Bd2 Be7 16.Bb4 Bxb4 17.Kxb4 Kc7 18.Kb3 Kb6 19.Kc4 Kc7 20.Kb3 Kb6
+/= (0.42) Depth: 29/63 00:01:34 3254mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Bd6 14.Bd2 Be7 15.b5+ Kb6 16.Be1 Bd6 17.Bd2 Be7 18.Be1
+/= (0.42) Depth: 30/63 00:02:10 4546mN
1.Nh1 Qf7
+/= (0.36 --) Depth: 31/70 00:07:16 15803mN
1.Nh1 Qc6 2.Nf2 Ne6 3.Nd3 Nc5 4.Qf7+ Kxf7 5.Nxe5+ Ke6 6.Nxc6 Nxe4 7.Be1 Bc5 8.Kc2 Nd5 9.Kd3 Kf5 10.Bf3 Nf4+ 11.Kc4 Ne6 12.Bg2 Nd6+ 13.Kd5 Nf4+ 14.Kxc5 Nxg2 15.Nxd4+ Ke4 16.Bxb4 Nb7+ 17.Kb6 Kxd4 18.Kxb7 Kd5 19.Kb6 Nxh4 20.Be7 Nf3 21.b4 Kc4 22.b3+ Kxb3 23.Kc5 Nd2 24.b5 Ne4+ 25.Kd5
+/= (0.38) Depth: 31/70 00:12:51 28517mN, tb=23
HIARCS got this really quickly on my first run this morning, but on my second run later in the day it got distracted by 1.Nh1:
New game
8/2n1b1kp/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 w - - 0 1
Analysis by HIARCS 13.2 MP:
1.Nf5+
+/= (0.33) Depth: 15/32 00:00:01 3345kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8
+/= (0.33) Depth: 15/33 00:00:01 4280kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Qe8 6.Qxe8 Nxe8 7.f6
+/= (0.36) Depth: 16/40 00:00:02 8248kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Qe8 6.Qxe8 Nxe8 7.f6 Bc5 8.Bf3 Nd6 9.h5 Kf7 10.Kd3 Nf5
+/= (0.36) Depth: 17/42 00:00:03 13441kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Ncd5 6.Bd3 Qc7+ 7.Qxc7
+/= (0.38) Depth: 18/42 00:00:04 21894kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Ncd5 6.Bd3 Qc7+ 7.Qxc7 Nxc7 8.Be4 Na6 9.Bf3 Bd6 10.h5 Kf7 11.g6+ hxg6 12.fxg6+ Kg7 13.Be2 Nc7 14.Kd3 Nb5
+/= (0.35) Depth: 19/44 00:00:07 35617kN, tb=2
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.f6 Qd5 6.Qg3 Qf5+ 7.Bd3 Bd6
= (0.25) Depth: 20/47 00:00:13 74827kN, tb=19
1.Nh1
= (0.26) Depth: 20/47 00:00:14 80709kN, tb=19
1.Nh1 Qc6 2.Nf2 Ne6 3.Nd3 Nc5 4.Qf7+
+/= (0.36) Depth: 20/47 00:00:17 97836kN, tb=19
1.Nh1 Bd6 2.Nf2 Qe7 3.Ng4 Ne6 4.Nf6 Nf4 5.Bb5 Qe6 6.Ka2 Kh8 7.Qf2 Nh3 8.Qe1 Be7 9.Ne8 Nf4 10.Bxb4 Kg8 11.Bd2
+/= (0.44) Depth: 24/53 00:01:32 554mN, tb=111
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kc2 Kf8 14.Kd1 Ke7 15.Ke2 Kd7 16.Bg8 Nb5
+/= (0.44) Depth: 25/53 00:02:24 890mN, tb=354
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kb1 Bf8 14.Kc2
+/= (0.44) Depth: 26/53 00:03:46 1431mN, tb=1481
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kb1 Bf8 14.Ka2 Bd6
+/= (0.44) Depth: 27/57 00:06:32 2559mN, tb=2456
1.Nf5+
+/= (0.45) Depth: 27/62 00:11:09 4647mN, tb=17605
1.Nf5+
+/= (0.69 ++) Depth: 27/62 00:22:21 9844mN, tb=32037
I made an IDeA project to get to the truth of the position. Using 16 cores I ran it for several hours using 16 instances of Houdini 2.
My claim is that 1 Nf5+! is the best move (1 Nh1 is second best).
The critical line is as follows:
1. Nf5+ gxf5 (1... Kg8 2. Nh6+ Kg7 3. Qg3 Bd6)(1... Kh8 2. Nxe7 Qxe7 3. Bxb4 Qxb4 4. Qf6+ Kg8 5. Qd8+)
2. exf5 Qf7 3.f6+ Bxf6 4. gxf6+ Qxf6 5. Qb7 Nbd5 (5... Qg6+ 6. Ka2 Nbd5 (6... Qd6 7. Bxb4 Qxb4 8. Qxc7+)7. Bc4 Qf7 8. Qc6) 6. Bc4 Qf7 7. Ka1 d3 8. Bxb4 Qf1+ 9. Ka2
Nxb4+ 10. Qxb4
From this, my suspicion is that the engines that do well on this position are probably the ones that highly negatively evaluate black's king position. Of course it's possible that they see right to the end, but my hunch is this is an eval issue.
Since HIARCS does well here maybe you can confirm / deny this...?
Anyway, to answer your question, here are some results from infinite analysis:
Houdini 2 struggles:
= (0.24) Depth: 20/45 00:00:01 23495kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Bf1 Bd6 7.Be1 Ke7 8.Bb5 Ke6 9.Bd2 Ke7 10.Be2 Kf7 11.Bf1 Ke7 12.Be1 Kf7 13.Bg2 Ke6 14.Nh6 Ke7 15.Bf1 Kf8 16.Bd3 Kg7 17.Ng4 Kf7 18.Bb5 Ke6 19.Bd2
= (0.24) Depth: 25/50 00:00:07 202mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Bf1 Bd6 7.Be1 Ke7 8.Bb5 Ke6 9.Bd2 Ke7 10.Be2 Ke6 11.Nf2 Nc5 12.Be1 Kf7 13.Bf3 Ke7 14.Bg4 Ne6 15.Nd3 Nc5 16.Nxb4 Nxe4 17.Nc6+ Kf7 18.Bf3 Nc5 19.b4 Ne6 20.b5
= (0.24) Depth: 26/50 00:00:11 337mN
1.Nh1
= (0.29 ++) Depth: 27/58 00:00:23 752mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Bxc7 11.b4 Kc6 12.Bf1 Bd6 13.Kb3 Bf8 14.Kc4 Kb6 15.Bd2 Bd6 16.Bxf4 exf4 17.Kxd4 Bxb4
+/= (0.39) Depth: 27/63 00:00:39 1312mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Kb6 14.Bd2 Bb8 15.b5 Bd6 16.Bb4 Bxb4 17.Kxb4 Kb7 18.Kc4 Kc7 19.Kb3 Kd7 20.Kc2 Kc7 21.Kd2 Kb6
+/= (0.42) Depth: 28/63 00:01:15 2588mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Bd6 14.b5+ Kb6 15.Bd2 Be7 16.Bb4 Bxb4 17.Kxb4 Kc7 18.Kb3 Kb6 19.Kc4 Kc7 20.Kb3 Kb6
+/= (0.42) Depth: 29/63 00:01:34 3254mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Bd6 4.Nd3 Nd7 5.Nxb4 Nc5 6.Bd3 Ke6 7.Kc2 Kd7 8.Be1 N5e6 9.Na6 Nf4 10.Nxc7 Nxd3 11.Kxd3 Bxc7 12.Kc4 Kc6 13.b4 Bd6 14.Bd2 Be7 15.b5+ Kb6 16.Be1 Bd6 17.Bd2 Be7 18.Be1
+/= (0.42) Depth: 30/63 00:02:10 4546mN
1.Nh1 Qf7
+/= (0.36 --) Depth: 31/70 00:07:16 15803mN
1.Nh1 Qc6 2.Nf2 Ne6 3.Nd3 Nc5 4.Qf7+ Kxf7 5.Nxe5+ Ke6 6.Nxc6 Nxe4 7.Be1 Bc5 8.Kc2 Nd5 9.Kd3 Kf5 10.Bf3 Nf4+ 11.Kc4 Ne6 12.Bg2 Nd6+ 13.Kd5 Nf4+ 14.Kxc5 Nxg2 15.Nxd4+ Ke4 16.Bxb4 Nb7+ 17.Kb6 Kxd4 18.Kxb7 Kd5 19.Kb6 Nxh4 20.Be7 Nf3 21.b4 Kc4 22.b3+ Kxb3 23.Kc5 Nd2 24.b5 Ne4+ 25.Kd5
+/= (0.38) Depth: 31/70 00:12:51 28517mN, tb=23
HIARCS got this really quickly on my first run this morning, but on my second run later in the day it got distracted by 1.Nh1:
New game
8/2n1b1kp/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 w - - 0 1
Analysis by HIARCS 13.2 MP:
1.Nf5+
+/= (0.33) Depth: 15/32 00:00:01 3345kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8
+/= (0.33) Depth: 15/33 00:00:01 4280kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Qe8 6.Qxe8 Nxe8 7.f6
+/= (0.36) Depth: 16/40 00:00:02 8248kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Qe8 6.Qxe8 Nxe8 7.f6 Bc5 8.Bf3 Nd6 9.h5 Kf7 10.Kd3 Nf5
+/= (0.36) Depth: 17/42 00:00:03 13441kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Ncd5 6.Bd3 Qc7+ 7.Qxc7
+/= (0.38) Depth: 18/42 00:00:04 21894kN
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.Kc2 Ncd5 6.Bd3 Qc7+ 7.Qxc7 Nxc7 8.Be4 Na6 9.Bf3 Bd6 10.h5 Kf7 11.g6+ hxg6 12.fxg6+ Kg7 13.Be2 Nc7 14.Kd3 Nb5
+/= (0.35) Depth: 19/44 00:00:07 35617kN, tb=2
1.Nf5+ gxf5 2.exf5 Qf7 3.Qe4 Bf8 4.Qxe5+ Kg8 5.f6 Qd5 6.Qg3 Qf5+ 7.Bd3 Bd6
= (0.25) Depth: 20/47 00:00:13 74827kN, tb=19
1.Nh1
= (0.26) Depth: 20/47 00:00:14 80709kN, tb=19
1.Nh1 Qc6 2.Nf2 Ne6 3.Nd3 Nc5 4.Qf7+
+/= (0.36) Depth: 20/47 00:00:17 97836kN, tb=19
1.Nh1 Bd6 2.Nf2 Qe7 3.Ng4 Ne6 4.Nf6 Nf4 5.Bb5 Qe6 6.Ka2 Kh8 7.Qf2 Nh3 8.Qe1 Be7 9.Ne8 Nf4 10.Bxb4 Kg8 11.Bd2
+/= (0.44) Depth: 24/53 00:01:32 554mN, tb=111
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kc2 Kf8 14.Kd1 Ke7 15.Ke2 Kd7 16.Bg8 Nb5
+/= (0.44) Depth: 25/53 00:02:24 890mN, tb=354
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kb1 Bf8 14.Kc2
+/= (0.44) Depth: 26/53 00:03:46 1431mN, tb=1481
1.Nh1 Qf7 2.Nf2 Qxf3 3.Bxf3 Na6 4.Be2 Nb8 5.Bb5 Nc8 6.Nd3 Na7 7.Bc4 Nbc6 8.Bd5 Bd6 9.Kc1 Be7 10.Kc2 Bd6 11.Kd1 Be7 12.Kc1 Bd6 13.Kb1 Bf8 14.Ka2 Bd6
+/= (0.44) Depth: 27/57 00:06:32 2559mN, tb=2456
1.Nf5+
+/= (0.45) Depth: 27/62 00:11:09 4647mN, tb=17605
1.Nf5+
+/= (0.69 ++) Depth: 27/62 00:22:21 9844mN, tb=32037
-
Carl Bicknell
- Member
- Posts: 252
- Joined: Mon May 04, 2009 10:06 pm
Critter 1.6a is pretty good
Analysis by Critter 1.6a 64-bit:
= (0.18) Depth: 18/36 00:00:01 12837kN
1.Nh1 Bd6 2.Nf2 Qe7 3.Bc4 Ne6 4.Ng4 Qf7 5.Qf6+ Qxf6 6.Nxf6 Kf7 7.Kc2 Nxc4 8.bxc4 h5 9.Nd5 Nc5 10.Bxb4 Nxe4 11.Bxd6 Nxd6 12.Kd3 Nf5 13.b4 Nxh4 14.b5 Ke6
= (0.23) Depth: 19/39 00:00:02 19222kN
1.Nh1 Bd6 2.Nf2 Qe7 3.Bc4 Ne6 4.Ng4 Qf7 5.Qf6+ Qxf6 6.Nxf6 Kf7 7.Kc2 Nxc4 8.bxc4 Be7 9.Ng4 Bd6 10.Nf2 Nf4 11.b3 Ke6 12.Nd3 Ng2 13.c5
+/= (0.31 ++) Depth: 20/47 00:00:03 38943kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Be1 Bd6 7.Bf1 Ke7 8.Nf2 Nc5 9.Bd2 Kf7 10.Bg2 Nb7 11.Nd3
+/= (0.32) Depth: 20/47 00:00:04 53616kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kg7 7.Be1 Bd6 8.Ng4 Kf7 9.Bf1 Ke7 10.Nf2 Nc5 11.Bb5 Kf7 12.Bc6 Ne6 13.Nd3 Nf4
= (0.27) Depth: 21/47 00:00:05 60856kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Kg7 8.Ng4 Bd6 9.Nf2 Nc5 10.Bc6 Ne6 11.Nd3 Nf4 12.Nxb4 Nc8
= (0.29) Depth: 22/47 00:00:06 76985kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Kg7 8.Ng4 Bd6 9.Nf2 Nc5 10.Bc6 Ne6 11.Nd3 Nf4 12.Nxb4 Ng2 13.Nd3 Nxh4
= (0.27) Depth: 23/50 00:00:09 111mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Bd6 8.Ng4 Kf7 9.Nf2 Nc5 10.Bf1 Ke6 11.Be2 Kf7 12.Bf1
+/= (0.36 ++) Depth: 24/55 00:00:19 262mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Ke7 14.Ke4 Kf8 15.Be6 Ke7 16.Bg8 Kf8 17.Be6
+/= (0.34) Depth: 24/55 00:00:19 266mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Kd7 14.Bg8 Ke8 15.b4 Kf8 16.Ba2 Ke8 17.Be6 Kd8 18.Bd5
+/= (0.33) Depth: 25/55 00:00:25 360mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Kd7 14.Bg8 Ke8 15.b4 Kf8 16.Ba2 Ke8 17.Bd5 Kf8 18.Be4 Ke7
+/= (0.33) Depth: 26/55 00:00:42 606mN
1.Nf5+ gxf5 2.exf5 Qf7 3.f6+ Bxf6 4.gxf6+ Qxf6 5.Qb7 Qg6+ 6.Ka1 Qd6 7.Bxb4 Qxb4 8.Qxc7+ Kf6 9.Qd8+ Ke6 10.Bg4+ Kf7 11.Bh5+ Ke6 12.Qg8+ Kd7 13.Qe8+ Kc7 14.Qxe5+ Qd6 15.Qe4 Nd7 16.Qc2+ Kb6 17.Qxh7
+/= (0.41) Depth: 26/57 00:01:50 1726mN
1.Nf5+ gxf5 2.exf5 Qf7 3.f6+ Bxf6 4.gxf6+ Qxf6 5.Qb7 Qg6+ 6.Ka1 Qd6 7.Bxb4 Qxb4 8.Qxc7+ Kf6 9.Qd8+ Ke6 10.Qg8+ Kd7 11.Bg4+ Kc7 12.Qf7+ Kb8 13.Qe8+ Ka7 14.Qxe5 d3 15.Bf3 Ka6 16.Qc3 Qa5+ 17.Qxa5+ Kxa5 18.Kb1 Nd7 19.Be4 d2 20.Kc2 Nf6 21.b4+ Kxb4 22.Bd3 d1R 23.Kxd1 Kb3 24.Kc1 h6
+/= (0.50 ++) Depth: 27/70 00:03:01 2904mN
Analysis by Critter 1.6a 64-bit:
= (0.18) Depth: 18/36 00:00:01 12837kN
1.Nh1 Bd6 2.Nf2 Qe7 3.Bc4 Ne6 4.Ng4 Qf7 5.Qf6+ Qxf6 6.Nxf6 Kf7 7.Kc2 Nxc4 8.bxc4 h5 9.Nd5 Nc5 10.Bxb4 Nxe4 11.Bxd6 Nxd6 12.Kd3 Nf5 13.b4 Nxh4 14.b5 Ke6
= (0.23) Depth: 19/39 00:00:02 19222kN
1.Nh1 Bd6 2.Nf2 Qe7 3.Bc4 Ne6 4.Ng4 Qf7 5.Qf6+ Qxf6 6.Nxf6 Kf7 7.Kc2 Nxc4 8.bxc4 Be7 9.Ng4 Bd6 10.Nf2 Nf4 11.b3 Ke6 12.Nd3 Ng2 13.c5
+/= (0.31 ++) Depth: 20/47 00:00:03 38943kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Be1 Bd6 7.Bf1 Ke7 8.Nf2 Nc5 9.Bd2 Kf7 10.Bg2 Nb7 11.Nd3
+/= (0.32) Depth: 20/47 00:00:04 53616kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kg7 7.Be1 Bd6 8.Ng4 Kf7 9.Bf1 Ke7 10.Nf2 Nc5 11.Bb5 Kf7 12.Bc6 Ne6 13.Nd3 Nf4
= (0.27) Depth: 21/47 00:00:05 60856kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Kg7 8.Ng4 Bd6 9.Nf2 Nc5 10.Bc6 Ne6 11.Nd3 Nf4 12.Nxb4 Nc8
= (0.29) Depth: 22/47 00:00:06 76985kN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Kg7 8.Ng4 Bd6 9.Nf2 Nc5 10.Bc6 Ne6 11.Nd3 Nf4 12.Nxb4 Ng2 13.Nd3 Nxh4
= (0.27) Depth: 23/50 00:00:09 111mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Kc2 Nc5 5.Ng4 Ncd7 6.Nh6+ Kf8 7.Bb5 Bd6 8.Ng4 Kf7 9.Nf2 Nc5 10.Bf1 Ke6 11.Be2 Kf7 12.Bf1
+/= (0.36 ++) Depth: 24/55 00:00:19 262mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Ke7 14.Ke4 Kf8 15.Be6 Ke7 16.Bg8 Kf8 17.Be6
+/= (0.34) Depth: 24/55 00:00:19 266mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Kd7 14.Bg8 Ke8 15.b4 Kf8 16.Ba2 Ke8 17.Be6 Kd8 18.Bd5
+/= (0.33) Depth: 25/55 00:00:25 360mN
1.Nh1 Qf7 2.Qxf7+ Kxf7 3.Nf2 Ne6 4.Nd3 Nd7 5.Kc2 Nec5 6.Nxb4 Nxe4 7.Nc6 Nxd2 8.Kxd2 Bd6 9.Bc4+ Ke8 10.b4 Nb8 11.Nxb8 Bxb8 12.Kd3 Bc7 13.b5 Kd7 14.Bg8 Ke8 15.b4 Kf8 16.Ba2 Ke8 17.Bd5 Kf8 18.Be4 Ke7
+/= (0.33) Depth: 26/55 00:00:42 606mN
1.Nf5+ gxf5 2.exf5 Qf7 3.f6+ Bxf6 4.gxf6+ Qxf6 5.Qb7 Qg6+ 6.Ka1 Qd6 7.Bxb4 Qxb4 8.Qxc7+ Kf6 9.Qd8+ Ke6 10.Bg4+ Kf7 11.Bh5+ Ke6 12.Qg8+ Kd7 13.Qe8+ Kc7 14.Qxe5+ Qd6 15.Qe4 Nd7 16.Qc2+ Kb6 17.Qxh7
+/= (0.41) Depth: 26/57 00:01:50 1726mN
1.Nf5+ gxf5 2.exf5 Qf7 3.f6+ Bxf6 4.gxf6+ Qxf6 5.Qb7 Qg6+ 6.Ka1 Qd6 7.Bxb4 Qxb4 8.Qxc7+ Kf6 9.Qd8+ Ke6 10.Qg8+ Kd7 11.Bg4+ Kc7 12.Qf7+ Kb8 13.Qe8+ Ka7 14.Qxe5 d3 15.Bf3 Ka6 16.Qc3 Qa5+ 17.Qxa5+ Kxa5 18.Kb1 Nd7 19.Be4 d2 20.Kc2 Nf6 21.b4+ Kxb4 22.Bd3 d1R 23.Kxd1 Kb3 24.Kc1 h6
+/= (0.50 ++) Depth: 27/70 00:03:01 2904mN
-
Carl Bicknell
- Member
- Posts: 252
- Joined: Mon May 04, 2009 10:06 pm
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
Good question. Perhaps you should start a new 'blended' test suiteCarl Bicknell wrote:All these engines fail to find 1.Nf5 in 5 minutes:
Rybka 4.1
Rybka 3
Junior 13.3
Deep Fritz 13
Komodo 4
Vitruvius 1.12B
Shredder 12
So in conclusion I don't know whether to add this position to my tactical test suite or my positional one. What do you think??
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
In the original position black had just played Kg7?
[fen]7k/2n1b2p/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 b - - 0 32[/fen]
Rybka for eg. could walk straight into this.
New game
7k/2n1b2p/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 b - - 0 1
Analysis by Deep Rybka 4.1 SSE42 x64:
32...Kg8 33.h5 Qc6 34.hxg6 hxg6
= (-0.19 ++) Depth: 7 00:00:00 8kN
32...Kg8 33.h5 Qd7 34.hxg6 hxg6 35.Bf1 Ne6 36.Bh3 Qd6
= (-0.14) Depth: 7 00:00:00 11kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Ng4 Ne6 36.Qf2 Qf7 37.Qxf7+ Kxf7 38.Nh6+ Ke7 39.Kc2
= (0.04) Depth: 8 00:00:00 46kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Ng4 Ne6 36.Qf2 Qf7 37.Qxf7+ Kxf7 38.Nh6+ Ke7 39.Kc2
= (0.04) Depth: 9 00:00:00 51kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Nd3 Nd7 36.h5 Ne6 37.hxg6 hxg6 38.Qh3 Nf4 39.Qh6+ Kf7
= (0.15) Depth: 10 00:00:00 65kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Nd3 Nd7 36.h5 Ne6 37.hxg6 hxg6 38.Qh3 Ndf8 39.Qh6+ Kg8 40.Nf2 Qb7 41.Qh3 Qf7
= (0.13) Depth: 11 00:00:01 144kN
32...Bd6 33.h5 Kg7 34.hxg6 hxg6 35.Kc2 Be7 36.Qh1 Qc6+ 37.Kb1 Kg8 38.Qh6
= (0.03 ++) Depth: 11 00:00:01 190kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qe6 37.Nf6 Kg7 38.Ne8+ Kg8 39.Nf6+ Kg7 40.Ne8+ Kg8 41.Nf6+ Kg7 42.Ne8+ Kg8 43.Nf6+ Kg7 44.Ne8+ Kg8 45.Nf6+ Kg7 46.Ne8+ Kg8 47.Nf6+ Kg7
= (0.00) Depth: 11 00:00:01 216kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Kg7 39.h5 gxh5
= (0.04) Depth: 12 00:00:01 247kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Qc6 39.Qd1 Be7 40.Qc2
= (0.00) Depth: 13 00:00:01 383kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Qc6 39.Qd1 Be7 40.Qc2
= (0.00) Depth: 14 00:00:02 562kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Be7 38.h5 Qc5 39.Bxf4 exf4 40.Qxf4 Bd6 41.Qf1 gxh5 42.Nxh5 Qxg5 43.Qf6+ Qxf6 44.Nxf6 Be5 45.Nd5
= (0.16) Depth: 15 00:00:06 1152kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Kg7 35.Ng4 Ne6 36.Nf6 Nf4 37.Bb5 Qe6 38.Bxf4 exf4 39.e5 Bxe5 40.Qb7+ Kf8 41.Qxb6 Qxb6 42.Nd7+ Ke7 43.Nxb6 f3 44.Kc2 Bg3 45.h5 gxh5 46.Kd3 h4 47.Nd5+ Ke6
+/= (0.35) Depth: 16 00:00:15 2597kN
32...Kg7 33.Nh1 Qc6 34.Nf2 Ne6 35.Ng4 Qd6 36.Bf1 Bd8 37.Bd3 Be7 38.Qf1 Nd8 39.Nh6 Qc6 40.Qf3 Nd7 41.Ng4 Qc7 42.Bc4 Nb6
+/= (0.34) Depth: 16 00:00:19 3229kN
32...Kg7 33.Nh1 Qf7 34.Nf2 Qxf3 35.Bxf3 Bd6 36.Nd3 Na6 37.Bg4 Kf7 38.Kc2 Ke7 39.Be2 Ke6 40.Bg4+ Ke7 41.Be2 Ke6 42.Bg4+ Ke7 43.Be2 Ke6 44.Bg4+ Ke7 45.Be2 Ke6 46.Bg4+ Ke7 47.Be2 Ke6
+/= (0.31) Depth: 17 00:00:27 4418kN
32...Kg7 33.Nh1 Qf7 34.Nf2 Qxf3 35.Bxf3 Bd6 36.Nd3 Na6 37.Bg4 Kf7 38.Kc2 Ke7 39.Be2 Ke6 40.Bg4+ Ke7 41.Be2 Ke6 42.Bg4+ Ke7 43.Be2 Ke6 44.Bg4+ Ke7 45.Be2 Ke6 46.Bg4+ Ke7 47.Be2 Ke6
+/= (0.31) Depth: 18 00:00:34 5442kN
(Williamson, 11.07.2012)
[fen]7k/2n1b2p/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 b - - 0 32[/fen]
Rybka for eg. could walk straight into this.
New game
7k/2n1b2p/1n2q1p1/4p1P1/1p1pP2P/1P3QN1/1P1BB3/1K6 b - - 0 1
Analysis by Deep Rybka 4.1 SSE42 x64:
32...Kg8 33.h5 Qc6 34.hxg6 hxg6
= (-0.19 ++) Depth: 7 00:00:00 8kN
32...Kg8 33.h5 Qd7 34.hxg6 hxg6 35.Bf1 Ne6 36.Bh3 Qd6
= (-0.14) Depth: 7 00:00:00 11kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Ng4 Ne6 36.Qf2 Qf7 37.Qxf7+ Kxf7 38.Nh6+ Ke7 39.Kc2
= (0.04) Depth: 8 00:00:00 46kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Ng4 Ne6 36.Qf2 Qf7 37.Qxf7+ Kxf7 38.Nh6+ Ke7 39.Kc2
= (0.04) Depth: 9 00:00:00 51kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Nd3 Nd7 36.h5 Ne6 37.hxg6 hxg6 38.Qh3 Nf4 39.Qh6+ Kf7
= (0.15) Depth: 10 00:00:00 65kN
32...Kg7 33.Nh1 Bd6 34.Nf2 Qe7 35.Nd3 Nd7 36.h5 Ne6 37.hxg6 hxg6 38.Qh3 Ndf8 39.Qh6+ Kg8 40.Nf2 Qb7 41.Qh3 Qf7
= (0.13) Depth: 11 00:00:01 144kN
32...Bd6 33.h5 Kg7 34.hxg6 hxg6 35.Kc2 Be7 36.Qh1 Qc6+ 37.Kb1 Kg8 38.Qh6
= (0.03 ++) Depth: 11 00:00:01 190kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qe6 37.Nf6 Kg7 38.Ne8+ Kg8 39.Nf6+ Kg7 40.Ne8+ Kg8 41.Nf6+ Kg7 42.Ne8+ Kg8 43.Nf6+ Kg7 44.Ne8+ Kg8 45.Nf6+ Kg7 46.Ne8+ Kg8 47.Nf6+ Kg7
= (0.00) Depth: 11 00:00:01 216kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Kg7 39.h5 gxh5
= (0.04) Depth: 12 00:00:01 247kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Qc6 39.Qd1 Be7 40.Qc2
= (0.00) Depth: 13 00:00:01 383kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Qc5 38.Bf1 Qc6 39.Qd1 Be7 40.Qc2
= (0.00) Depth: 14 00:00:02 562kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Ne6 35.Ng4 Nf4 36.Bb5 Qc7 37.Nf6 Be7 38.h5 Qc5 39.Bxf4 exf4 40.Qxf4 Bd6 41.Qf1 gxh5 42.Nxh5 Qxg5 43.Qf6+ Qxf6 44.Nxf6 Be5 45.Nd5
= (0.16) Depth: 15 00:00:06 1152kN
32...Bd6 33.Nh1 Qe7 34.Nf2 Kg7 35.Ng4 Ne6 36.Nf6 Nf4 37.Bb5 Qe6 38.Bxf4 exf4 39.e5 Bxe5 40.Qb7+ Kf8 41.Qxb6 Qxb6 42.Nd7+ Ke7 43.Nxb6 f3 44.Kc2 Bg3 45.h5 gxh5 46.Kd3 h4 47.Nd5+ Ke6
+/= (0.35) Depth: 16 00:00:15 2597kN
32...Kg7 33.Nh1 Qc6 34.Nf2 Ne6 35.Ng4 Qd6 36.Bf1 Bd8 37.Bd3 Be7 38.Qf1 Nd8 39.Nh6 Qc6 40.Qf3 Nd7 41.Ng4 Qc7 42.Bc4 Nb6
+/= (0.34) Depth: 16 00:00:19 3229kN
32...Kg7 33.Nh1 Qf7 34.Nf2 Qxf3 35.Bxf3 Bd6 36.Nd3 Na6 37.Bg4 Kf7 38.Kc2 Ke7 39.Be2 Ke6 40.Bg4+ Ke7 41.Be2 Ke6 42.Bg4+ Ke7 43.Be2 Ke6 44.Bg4+ Ke7 45.Be2 Ke6 46.Bg4+ Ke7 47.Be2 Ke6
+/= (0.31) Depth: 17 00:00:27 4418kN
32...Kg7 33.Nh1 Qf7 34.Nf2 Qxf3 35.Bxf3 Bd6 36.Nd3 Na6 37.Bg4 Kf7 38.Kc2 Ke7 39.Be2 Ke6 40.Bg4+ Ke7 41.Be2 Ke6 42.Bg4+ Ke7 43.Be2 Ke6 44.Bg4+ Ke7 45.Be2 Ke6 46.Bg4+ Ke7 47.Be2 Ke6
+/= (0.31) Depth: 18 00:00:34 5442kN
(Williamson, 11.07.2012)
-
Peter Grayson
- Member
- Posts: 208
- Joined: Sat Aug 04, 2007 3:23 pm
- Location: South Wales, UK
As with many MP engines, what you get with H2.0c Pro depends on the threads and hash configuration. In XP Pro 64 , Deep Fritz 11 GUI, I ran this twice in infinite analysis mode with cleared hash on my Q9550 O/C'd to 3.53GHz quad with 2Gb large pages hash and split depth =14. It solved in less than 1 minute on successive infinite analysis tests. So I set up a multi .epd i.e. same position x10. On 10 runs, it solved 6 times between 3.72s/20 ply and 185.30s/27 ply, failing on 4 occasions with 240s allowed solving time.Carl Bicknell wrote:
Houdini 2 struggles
So Houdini 2.0c can solve it exceptionally quickly, even on my relatively slow quad but it can just as easily fail to solve in an allocated time frame. That is the bane of current multithread implementation ... there is little consistency so the only definitive statement can be made using a single thread but even then, hash and GUI can influence outcome.
If you only run it once you can get "brilliantly fast" or "dismally failed".
Emphasised by HIARCS 13.2 MP 2Gb hash solving 3 times from 3.50s/16 ply to 14.34s/19 ply and failing 7 times.
PeterG