Code, algorithms, languages, construction...
-
ernest
- Posts: 247
- Joined: Thu Sep 02, 2010 10:33 am
Post
by ernest » Sun Apr 10, 2011 2:29 pm
I left it running a little longer...
Code: Select all
.....
1...a5 2.a4 Kc8 3.Qg4 Kb8 4.Be2 Ka7 5.Rf1 Qg6
=/+ (-0.27 !) Depth: 26/9 00:06:00 284mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Ne7 Qe6 9.Nf5 Kb8 10.Rh4 Qd7 11.c3 Ba3 12.Rh6 Rde8 13.Ra1 Nxe4 14.Qxe4 Qxf5 15.Qxf5 Rxf5 16.Rxa3 Rf2+ 17.Kg3 Rxe2 18.Rxh7 Re3+ 19.Kg4
=/+ (-0.36) Depth: 26/36 00:08:32 401mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Ne7 Qe6 9.Nf5 Kb8 10.Rh4 Qd7 11.c3 Ba3 12.Rh6 Rde8 13.Ra1 Nxe4 14.Qxe4 Qxf5 15.Qxf5 Rxf5 16.Rxa3 Rf2+ 17.Kg3 Rxe2 18.Rxh7 Re3+ 19.Kg4
=/+ (-0.36) Depth: 27/36 00:09:13 434mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Rh4 Ne6 9.Rh6 Nf4 10.Bc4 Qd7 11.c3 Rxf5 12.cxb4 Rxg5 13.bxa5
= (-0.15 --) Depth: 28/24 00:10:09 477mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Rh4 Ne6 9.Rh6 Qd7 10.Bg4 Bc5 11.Qc3 Bb4 12.Qe3
= (0.00) Depth: 28/24 00:14:59 698mN
1...a5 2.Nf5 a4 3.Rf1 axb3 4.axb3 Bb4 5.Qd1 Qe6 6.Be2 Kc8 7.c3 Ba3 8.b4 Kb8 9.Qb1 Nc8 10.Qd3 Ncb6 11.Rb1 c5 12.Rgg1 Rc8 13.Ne7
= (0.00) Depth: 29/24 00:48:22 2281mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.Kh1 Rad8 8.h4 Qe6 9.Bg4 Kb8 10.Nh6 Qg6 11.Bf5 Qh5 12.Bg4 Qe8
= (-0.13) Depth: 29/24 01:01:22 2887mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.h4 Rad8 8.h5 Rh8 9.Qh3 Kb8 10.Kh1 Qe6 11.Bg4 Qg8 12.Be2 Qf8 13.c4 Bb4 14.Qh2 Qe8
= (-0.25) Depth: 30/27 01:35:11 4466mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.h4 Rad8 8.h5 Rh8 9.Bg4 Kb8 10.Be2 Qf8 11.Bf3 Rg8 12.Be2 Qe8 13.Bf3 Nc8 14.Qh4
= (-0.23) Depth: 31/26 01:58:52 5594mN
[/size]
-
ernest
- Posts: 247
- Joined: Thu Sep 02, 2010 10:33 am
Post
by ernest » Sun Apr 10, 2011 2:48 pm
Uly wrote:r2k1r2/pp1n1q1p/1npP4/4p1P1/4P3/1PbB2NP/P1P1Q1RK/3R4 b - -
(Black to move)
Gran2k at 1CPU (so the result is reproducible):
.14/12 0:00 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (287.084) 680
15/12 0:00 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (426.958) 718
16/12 0:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (833.391) 750
17/12 0:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (1.493.995) 758
18/12 0:03 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (2.550.949) 769
19/12 0:04 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (3.886.852) 779
19/11 0:09 -0.11++ 1...a5 2.Nf5 a4 3.Qe3 axb3 4.cxb3 Bb4 5.Rc2 Qe6 6.Be2 Nd5 (7.460.300) 763
19/11 0:12 -0.22++ 1...a5 2.Nf5 a4 3.Qe3 axb3 4.cxb3 Bb4 5.Rc2 Qe6 6.Be2 Nd5 (9.357.355) 761
20/12 0:26 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (20.495.558) 763
Strange, Uly, I can't reproduce your output !
How much Hash were you using?
-
Ancalagon
- Posts: 15
- Joined: Sat Sep 11, 2010 1:24 pm
Post
by Ancalagon » Sun Apr 10, 2011 10:09 pm
ernest wrote:I left it running a little longer...
Code: Select all
.....
1...a5 2.a4 Kc8 3.Qg4 Kb8 4.Be2 Ka7 5.Rf1 Qg6
=/+ (-0.27 !) Depth: 26/9 00:06:00 284mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Ne7 Qe6 9.Nf5 Kb8 10.Rh4 Qd7 11.c3 Ba3 12.Rh6 Rde8 13.Ra1 Nxe4 14.Qxe4 Qxf5 15.Qxf5 Rxf5 16.Rxa3 Rf2+ 17.Kg3 Rxe2 18.Rxh7 Re3+ 19.Kg4
=/+ (-0.36) Depth: 26/36 00:08:32 401mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Ne7 Qe6 9.Nf5 Kb8 10.Rh4 Qd7 11.c3 Ba3 12.Rh6 Rde8 13.Ra1 Nxe4 14.Qxe4 Qxf5 15.Qxf5 Rxf5 16.Rxa3 Rf2+ 17.Kg3 Rxe2 18.Rxh7 Re3+ 19.Kg4
=/+ (-0.36) Depth: 27/36 00:09:13 434mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Rh4 Ne6 9.Rh6 Nf4 10.Bc4 Qd7 11.c3 Rxf5 12.cxb4 Rxg5 13.bxa5
= (-0.15 --) Depth: 28/24 00:10:09 477mN
1...a5 2.a4 Kc8 3.Nf5 Bb4 4.Qe3 Kb8 5.Be2 Ka7 6.Rg4 Nc5 7.Rf1 Rad8 8.Rh4 Ne6 9.Rh6 Qd7 10.Bg4 Bc5 11.Qc3 Bb4 12.Qe3
= (0.00) Depth: 28/24 00:14:59 698mN
1...a5 2.Nf5 a4 3.Rf1 axb3 4.axb3 Bb4 5.Qd1 Qe6 6.Be2 Kc8 7.c3 Ba3 8.b4 Kb8 9.Qb1 Nc8 10.Qd3 Ncb6 11.Rb1 c5 12.Rgg1 Rc8 13.Ne7
= (0.00) Depth: 29/24 00:48:22 2281mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.Kh1 Rad8 8.h4 Qe6 9.Bg4 Kb8 10.Nh6 Qg6 11.Bf5 Qh5 12.Bg4 Qe8
= (-0.13) Depth: 29/24 01:01:22 2887mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.h4 Rad8 8.h5 Rh8 9.Qh3 Kb8 10.Kh1 Qe6 11.Bg4 Qg8 12.Be2 Qf8 13.c4 Bb4 14.Qh2 Qe8
= (-0.25) Depth: 30/27 01:35:11 4466mN
1...Bb4 2.Nf5 Kc8 3.Qe3 Kb8 4.Be2 a5 5.c3 Bc5 6.Qg3 Ka7 7.h4 Rad8 8.h5 Rh8 9.Bg4 Kb8 10.Be2 Qf8 11.Bf3 Rg8 12.Be2 Qe8 13.Bf3 Nc8 14.Qh4
= (-0.23) Depth: 31/26 01:58:52 5594mN
[/size]
Not much to do with Stockfish but for comparison I tried the position with Rainbow Serpent. It shows the same behaviour for a5, even though the PV search is different and does take longer. There is a False Fail High for a5 only later than with Stockfish, at 26 plies deep. It tries to resolve 1... a5 for almost 10 minutes but fails to produce a score within the search window or a second Fail High. Later Bb4 does fall within the search window and just overtakes the 0.00 score for 1..Qf3. Maybe the score for a5 is just not stable enough. I doubt any change in Razoring would help, in all cases where this occurs, it is much too general a search instability problem for that.
r2k1r2/pp1n1q1p/1npP4/4p1P1/4P3/1PbB2NP/P1P1Q1RK/3R4 b - -
Engine: Rainbow Serpent 2.0 Build 116 (Athlon 2009 MHz, 256 MB)
by Romstad, Costalba, Kiiski
1/01 0:00 +0.36 1...Ke8 (152) 0
2/02 0:00 -0.24 1...Ke8 2.Nf5 (105.385) 192
2/06 0:00 +0.36 1...Qf3 2.Rf1 Qxe2 3.Rxf8+ Nxf8
4.Nxe2 (107.897) 197
3/05 0:00 -0.12 1...Qf3 2.Nf5 Qxe2 3.Bxe2 Nc5 (125.542) 217
3/10 0:01 -0.04 1...Ke8 2.Nf5 Nc5 3.Qf2 Nxd3 4.Rxd3 Bd4
5.Qh4 c5 6.Rf3 (356.569) 335
4/07 0:01 -0.04 1...Ke8 2.Rf1 Qg6 3.Nf5 Nc5 4.a3 Rc8 (431.922) 350
5/10 0:03 -0.04 1...Ke8 2.Rf1 Qg6 3.Nf5 Nc5 4.h4 Qf7
5.Nh6 Qxf1 6.Qh5+ (1.324.217) 430
6/06 0:03 -0.08 1...Ke8 2.Rf1 Qg6 3.Nf5 Rc8 4.h4 (1.705.038) 445
6/05 0:03 +0.07++ 1...Qf3 2.Nf5 Qxe2 3.Bxe2 Nc5 (1.708.842) 444
6/05 0:03 +0.19++ 1...Qf3 2.Nf5 Qxe2 3.Bxe2 Nc5 (1.709.166) 444
6/07 0:04 +0.24 1...Qf3 2.Qxf3 Rxf3 3.Nf5 Nc5 4.Be2 Rf4 (1.796.848) 447
7/10 0:04 +0.24 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Rd3 Bd4 6.Nf5 (1.846.232) 447
8/09 0:04 +0.19 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Nf5 Nc5 (1.975.690) 451
9/16 0:04 +0.19 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Nc5 6.Bf5 Ke8 7.Bxh7 Kd7
8.Bf5+ Ne6 9.g6 (2.067.847) 454
10/12 0:04 +0.12 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Rg8 6.Ne2 Bb4 7.Bf5 (2.126.747) 456
11/13 0:05 +0.08 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bg4 Bb4
5.Nf5 Nc5 6.Re2 Rf8 7.Re3 Kd7 (2.685.579) 470
12/14 0:06 +0.04 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Rg8 6.h4 Ke8 7.Ne2 Bb4 8.Bf5 (3.289.260) 481
13/18 0:07 +0.12++ 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Nc5 6.Bf5 Bb4 7.Bxh7 Kd7 8.g6 Ne6
9.Nf5 Nf4 10.g7 (3.700.890) 488
13/14 0:08 0.00-- 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Bb4 6.Bf5 Rh8 7.Rf2 Rf8 8.Rg2 (4.124.795) 493
13/15 0:08 +0.04 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf7 4.Bh5 Rf8
5.Bg4 Bb4 6.Bf5 Rh8 7.Rf2 Rf8 8.Ne2 Nc8 (4.355.539) 496
14/20 0:09 +0.12++ 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Nh5 Nc8 7.Be6 Rf3 8.Bg4 Re3
9.Ng7 Nxd6 10.Nf5 Nxf5 11.Bxf5 (4.931.635) 501
14/13 0:10 0.00-- 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (5.119.850) 504
14/13 0:10 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (5.211.126) 503
15/13 0:10 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (5.491.129) 506
16/13 0:11 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (6.041.179) 508
17/13 0:13 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (6.861.969) 510
18/13 0:17 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (9.274.878) 516
19/13 0:20 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (10.817.825) 516
20/13 0:24 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (12.659.434) 516
21/13 0:30 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (15.731.854) 516
22/13 0:42 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (21.873.040) 513
23/22 1:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rc3 8.Nf5 a5
9.Ng7 a4 10.Ne6+ Kc8 11.Bf5 axb3
12.cxb3 (31.313.523) 507
24/12 1:40 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (50.844.167) 508
25/12 2:34 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (77.621.971) 502
26/12 3:29 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (104.882.549) 500
26/02 5:56 +0.16++ 1...a5 2.Nf5 (175.504.282) 491
26/12 15:47 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (465.416.739) 491
27/12 16:32 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (487.527.003) 491
28/12 19:52 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 a5
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (586.611.491) 491
28/09 140:30 +0.15 1...Bb4 2.Nf5 Nc5 3.Qg4 Ne6 4.Rgg1 Nf4
5.Be2 Kc8 (3.930.399.101) 466
best move: Bc3-b4 time: 142:29.000 min n/s: 465.468 nodes: 3.979.256.126
-
Ancalagon
- Posts: 15
- Joined: Sat Sep 11, 2010 1:24 pm
Post
by Ancalagon » Mon Apr 11, 2011 9:53 pm
This is with a new version. The Fail High is resolved correctly, at a bit higher depth than Uly's Stockfish. If I'm correct the main variation with 1... Bb4 can also be slightly below 0.00, and in that case 1... a5 is preferred. But I did not go to extreme depths here:
r2k1r2/pp1n1q1p/1npP4/4p1P1/4P3/1PbB2NP/P1P1Q1RK/3R4 b - -
Engine: Rainbow Serpent 2.0 Build 118 (Athlon 2009 MHz, 256 MB)
by Tord Romstad, Marco Costalba, Joona Kiiski
1/01 0:00 +0.36 1...Ke8 (152) 0
2/02 0:00 -0.24 1...Ke8 2.Nf5 (105.475) 198
2/04 0:00 -0.16 1...Qf3 2.Nf5 Qxe2 3.Bxe2 (107.951) 191
3/05 0:00 -0.12 1...Qf3 2.Nf5 Qxe2 3.Bxe2 Nc5 (112.610) 194
3/10 0:01 -0.04 1...Ke8 2.Nf5 Nc5 3.Qf2 Ne6 4.Be2 Bd4
5.Qg3 Nf4 6.Rxd4 (321.167) 297
4/10 0:01 -0.03 1...Ke8 2.Nf5 Nc5 3.Qf2 Ne6 4.Be2 Bd4
5.Qg3 Nf4 6.Rxd4 (384.646) 311
4/07 0:01 +0.56 1...Qf4 2.Qh5 Qf3 3.Rgg1 Qxh5 4.Nxh5 Rf2+ (498.626) 328
5/06 0:01 -0.92 1...Qf4 2.Rf1 Qxg5 3.Nf5 Qf6 4.Rg7 (547.478) 333
5/05 0:02 -0.08 1...Ke8 2.Rf1 Qg6 3.Nf5 Bb4 (1.093.508) 374
5/07 0:03 +0.12 1...Qf3 2.Nf5 Qxe2 3.Bxe2 Nc5 4.Bf3 Bd4 (1.127.979) 375
6/07 0:03 +0.40 1...Qf3 2.Rf1 Qxe2 3.Rxf8+ Nxf8
4.Nxe2 Bb4 (1.219.171) 378
7/08 0:03 +0.28 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf4 4.Bg4 a5
5.Ne2 (1.495.864) 389
8/14 0:04 +0.12 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc5
5.Bf5 Bd4 6.Bxh7 Kd7 7.Ne2 Kxd6 8.g6 (1.808.945) 396
9/19 0:04 +0.12 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Nh5 Nc8 7.Be6 Rf8 8.Bxd7 Kxd7
9.Nf6+ Rxf6 10.gxf6 Nxd6 (1.920.347) 400
10/19 0:05 -0.03-- 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Ne2 Nc8 7.g6 hxg6 8.Rxg6 Nxd6
9.Rgxd6 Bxd6 10.Rxd6 Kc7 (2.117.428) 405
10/12 0:06 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Bd4 6.Bxh7 Nxd6 7.c3 (2.701.583) 411
11/13 0:06 +0.08 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.c4 Nc8 7.Be6 Rf4 (2.903.506) 415
12/14 0:07 +0.04 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rh8 6.Nh5 Nc8 7.c3 Bxd6 8.Nf6 (3.213.781) 418
13/12 0:08 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (3.630.165) 423
14/12 0:09 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (3.916.197) 426
15/12 0:10 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4
5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (4.341.003) 428
16/18 0:14 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Re3 8.Rf2 Be1
9.Rxe1 Rxe1 10.Rf7 (6.384.017) 437
17/13 0:16 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (7.475.741) 441
18/13 0:19 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (8.436.901) 441
19/13 0:22 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (9.793.407) 442
20/13 0:27 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (12.111.177) 441
21/13 0:36 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (15.905.713) 438
22/13 0:50 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (21.899.312) 437
23/13 1:13 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (32.011.345) 434
24/13 1:42 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8
5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (44.174.815) 432
24/09 3:12 +0.12++ 1...a5 2.a3 Qf3 3.Qxf3 Rxf3 4.Nf5 Nc5
5.Be2 Rf4 (81.709.456) 424
24/21 9:45 +0.02 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Nh5 Bd4 8.Nf6 Nxf6
9.gxf6 Nxd6 10.c3 c5 11.cxd4 Nxf5 (248.351.453) 424
25/15 11:05 +0.05 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Bg4 Bb4 8.Nf5 Rf8 (282.956.259) 424
26/22 13:34 +0.12 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Bg4 Bb4 8.Nf5 Rg8
9.Bh5 Rf8 10.Bg4 Nc5 11.Ng3 b5
12.axb5 (346.350.649) 425
27/19 37:14 0.00 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Bg4 Bb4 8.Nf5 Rg8
9.h4 Rf8 10.Re2 Nc5 (935.901.264) 418
28/18 105:20 +0.04 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Bg4 Bb4 8.Nf5 Rg8
9.h4 Rg6 10.Bh5 (2.716.148.924) 429
29/20 122:52 0.00 1...a5 2.a4 Qf3 3.Qxf3 Rxf3 4.Be2 Rf8
5.Bg4 Nc8 6.Bf5 Rh8 7.Bg4 Bb4 8.Nf5 Rg8
9.h4 Nc5 10.Ng3 Nd7 11.Nf5 (3.187.384.768) 432
best move: a7-a5 time: 127:38.125 min n/s: 433.552 nodes: 3.320.197.337
-
Uly
- Posts: 838
- Joined: Thu Jun 10, 2010 5:33 am
Post
by Uly » Tue Apr 12, 2011 10:28 pm
ernest wrote:Strange, Uly, I can't reproduce your output !
How much Hash were you using?
128MB, from this version:
http://www.open-chess.org/stockfish_pa_gtb_gran2k.7z
Now, the wanted behavior is that no matter how much hash one gives the engine, or how many cores one gives Stockfish, if a5 fails high then Stockfish should stick with the move. The idea is that, if Stockfish resolved the fail low after the fail high until the real score is known, it would stick with a5 until it finds that 1...Bb4 is better (note that it doesn't matter if a5 is not the best move, it only matters that it's better than 1...Qf3, the move Stockfish goes back to after wrongly abandoning a5).
-
ernest
- Posts: 247
- Joined: Thu Sep 02, 2010 10:33 am
Post
by ernest » Tue Apr 12, 2011 11:56 pm
Using 128MB hash/1core, and precisely that version
Stockfish_201_PA_GTB_Gran2k_Win32.exe (04/03/2011 10:10)
my analysis differs from yours:
Code: Select all
r2k1r2/pp1n1q1p/1npP4/4p1P1/4P3/1PbB2NP/P1P1Q1RK/3R4 b - - 0 1
Analysis by Stockfish 2.0.1 PA GTB Gran2k:
1...Qf3 2.Qxf3 Rxf3 3.Nf5 Rf4 4.Kg3 Ke8
= (-0.20) Depth: 6/7 00:00:00 4kN
1...Qf3 2.Qxf3 Rxf3 3.Nf5 Rf4 4.Kg3 Ke8 5.Rf2 Rxf2 6.Kxf2
= (-0.23) Depth: 7/10 00:00:00 5kN
1...Qf3 2.Qxf3 Rxf3 3.Nf5 Nc5 4.Be2 Rf4
=/+ (-0.34 !) Depth: 8/7 00:00:00 14kN
1...Qf3 2.Qxf3 Rxf3 3.Nf5 Nc5 4.Be2 Rf4
=/+ (-0.45 !) Depth: 8/7 00:00:00 15kN
1...Qf3 2.Qxf3 Rxf3 3.Rgg1 Bd4 4.Rgf1 Rxf1 5.Rxf1 Be3 6.Rf7 Bxg5 7.Rxh7
=/+ (-0.46) Depth: 8/12 00:00:00 19kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf4 4.Bg4 Bd4 5.Bf5
= (-0.18 --) Depth: 9/8 00:00:00 28kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf4 4.Bg4 Bd4 5.Bf5 Nc8 6.Nh5
= (-0.11) Depth: 9/10 00:00:00 33kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf4 4.Bg4 Bb4 5.c3 Rxg4 6.hxg4 Bxc3 7.Rc2 Bd4 8.Nf5
= (0.01) Depth: 10/14 00:00:00 58kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Nh5 a5 6.Bf5 a4 7.Ng7 Nc5 8.Bxh7 axb3 9.cxb3
= (-0.05) Depth: 11/16 00:00:00 100kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Rd3 Bb4 6.Nf5 Nc5 7.Re3 Nd7 8.Rd3
= (0.00) Depth: 12/14 00:00:00 189kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 13/12 00:00:00 401kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 14/12 00:00:01 803kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 15/12 00:00:01 885kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf3 7.Re2 Bb4 8.Nf5 Nc5 9.Kg2 Rf4
= (-0.11 !) Depth: 16/17 00:00:01 1185kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 16/17 00:00:01 1224kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf3 7.Re2 Bb4 8.Nf5 Nc5 9.Kg2 Rf4
= (-0.11 !) Depth: 17/17 00:00:02 1719kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (-0.22 !) Depth: 17/17 00:00:02 2015kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 17/17 00:00:02 2195kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 18/12 00:00:04 3987kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5
= (0.00) Depth: 19/12 00:00:07 6040kN
1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8
= (0.00) Depth: 20/13 00:00:11 9513kN
1...a5 2.a3 a4 3.Rf1 Qg6 4.Rxf8+ Nxf8 5.Qf2 Nfd7
= (-0.11 !) Depth: 20/13 00:00:16 13578kN
1...a5 2.a3 a4 3.Rf1 Qg6 4.Rxf8+ Nxf8 5.Qf2 Nfd7
= (-0.22 !) Depth: 20/13 00:00:19 15531kN
[/size]
GTB or no GTB, same analysis...
Could it be that the GUI matters?: I use Fritz11 GUI, you use Arena
Or did you change something from the default parameters?
...or maybe you did not clear the hash?...
-
Uly
- Posts: 838
- Joined: Thu Jun 10, 2010 5:33 am
Post
by Uly » Wed Apr 13, 2011 12:25 am
I'm using the 64bit version, maybe that's it?
-
Uly
- Posts: 838
- Joined: Thu Jun 10, 2010 5:33 am
Post
by Uly » Wed Apr 13, 2011 12:29 am
Well, here's this, if I put Stockfish to analyze the position in infinite analysis, I get what I posted, but if instead, I tell Stockfish to play a move at fixed depth 20, I get this:
13/12 0:00 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (401.432) 694
14/12 0:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (802.910) 704
15/12 0:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (884.772) 690
16/17 0:01 -0.11++ 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf3 7.Re2 Bb4 8.Nf5 Nc5 9.Kg2 Rf4 (1.185.225) 702
16/12 0:01 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (1.224.462) 699
17/17 0:02 -0.11++ 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf3 7.Re2 Bb4 8.Nf5 Nc5 9.Kg2 Rf4 (1.719.026) 714
17/12 0:02 -0.22++ 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (2.014.893) 728
17/12 0:02 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (2.195.492) 739
18/12 0:05 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (3.987.433) 768
19/12 0:07 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Nc8 5.Bf5 Rf7 6.Be6 Rf8 7.Bf5 (6.039.882) 763
20/13 0:12 0.00 1...Qf3 2.Qxf3 Rxf3 3.Be2 Rf8 4.Bg4 Bb4 5.Bf5 Rf7 6.Be6 Rf3 7.Bg4 Rf8 (9.513.215) 781
20/09 0:17 -0.11++ 1...a5 2.a3 a4 3.Rf1 Qg6 4.Rxf8+ Nxf8 5.Qf2 Nfd7 (13.577.799) 769
20/09 0:20 -0.22++ 1...a5 2.a3 a4 3.Rf1 Qg6 4.Rxf8+ Nxf8 5.Qf2 Nfd7 (15.531.302) 768
20/20 0:39 -0.20 1...a5 2.h4 a4 3.Nf5 axb3 4.axb3 Kc8 5.h5 Rh8 6.Qf3 Kb8 7.Be2 Bb2 8.Rg3 Ra2 9.Kg2 Rg8 10.Qf2 Ba3 11.Ne7 (30.371.874) 766
best move: a7-a5 time: 0:39.891 min n/s: 765.474 nodes: 30.511.052
So the behavior of the engine also depends on whether is analysis mode or play mode...
-
ernest
- Posts: 247
- Joined: Thu Sep 02, 2010 10:33 am
Post
by ernest » Wed Apr 13, 2011 1:23 am
Uly wrote:Well, here's this, if I put Stockfish to analyze the position in infinite analysis, I get what I posted, but if instead, I tell Stockfish to play a move at fixed depth 20, I get this:
.................
So the behavior of the engine also depends on whether is analysis mode or play mode...
Good find, Uly!
Now the plot really thickens!...
"play a move at fixed depth 20"... what is this? only possible with Arena?
In Fritz GUI, infinite analysis is... infinite analysis!
So
your "play mode" is
my "analysis mode"???
I must say I am a bit lost here...
-
Uly
- Posts: 838
- Joined: Thu Jun 10, 2010 5:33 am
Post
by Uly » Wed Apr 13, 2011 1:45 am
I'm using Shredder Classic, here, when I put the engine to analyze a position, it goes at it endlessly, but I also can put it in "Human against the Machine mode", there, I can choose time control, I put it at 20ply (so, I make a move, and Stockfish reaches 20 plies and makes a move, then it's my turn again) and force a move (so it's as if I moved) I get different behavior.
If your Fritz GUI gives you that behavior in Infinite analysis, then it doesn't give you the option of analysis mode (which shouldn't matter because the engine's behavior should be the same!) Fritz GUI would use "Play mode" for everything.
So, just GUI quirks, it's still a mystery why Stockfish behaves differently whether is has to "analyze" or "play a move"in Shredder Classic.
