Author: Vincent Diepeveen
Date: 04:24:39 10/30/01
Go up one level in this thread
On October 29, 2001 at 23:27:37, Dann Corbit wrote: >On October 29, 2001 at 22:17:12, K. Burcham wrote: > >> >>in this position, Deep Blue plays 22...g4. >>shredder5 in this position would play 22...Bg6, at depth 17/34 >>what does your program play, with deep search? >> >> >>(1) Kasparov,G - Deep Blue [A07] >>IBM Man vs Machine New York (1), 03.05.1997 >>[John Nunn] >> >> >> [D] 3rr1k1/1p1n1p2/1qp2n1p/p1b1p1pb/4P3/PP2N1PP/1BP2PB1/R1Q1RNK1 b - - 00:00 0 0 44 (0) 1 -0.604 Bc5xe3 Nf1xe3 00:00 0 0 50 (0) 1 -0.076 Re8-e7 00:00 0 0 51 (0) 1 -0.001 Qb6-b5 00:00 0 0 53 (0) 1 0.053 Qb6-a6 00:00 0 0 54 (0) 1 0.095 Nd7-f8 00:00 0 0 71 (0) 1 0.261 Kg8-g7 00:00 0 0 146 (0) 2 -0.355 Kg8-g7 Ne3-f5 Kg7-h7 ++ g8-f8 00:00 0 0 339 (0) 2 -0.348 Kg8-f8 Ne3-f5 ++ a5-a4 00:00 0 0 437 (0) 2 -0.345 a5-a4 Ne3-c4 00:00 0 0 977 (48) 3 0.071 a5-a4 Nf1-d2 Qb6-a6 ++ g8-f8 00:00 0 0 1315 (48) 3 0.096 Kg8-f8 Bb2-c3 Kf8-g7 00:00 0 0 2607 (384) 4 -0.287 Kg8-f8 Bb2-c3 Bc5-d4 Ne3-c4 ++ a5-a4 ++ d7-f8 ++ h5-g6 00:00 0 0 7774 (874) 4 -0.101 Bh5-g6 Ne3-g4 Nf6xg4 h3xg4 00:00 0 0 18465 (3057) 5 -0.100 Bh5-g6 Nf1-d2 Kg8-g7 Nd2-c4 Qb6-c7 ++ g8-f8 00:00 0 0 24519 (3808) 5 -0.006 Kg8-f8 Nf1-d2 Bc5xe3 Re1xe3 Nd7-c5 ++ d7-f8 ++ g8-h7 00:00 0 0 61287 (13085) 6 -0.428 Kg8-f8 Nf1-d2 a5-a4 b3xa4 Bc5-d4 Nd2-c4 Bd4xb2 Nc4xb2 ++ h5-g6 00:01 0 0 88405 (27572) 6 0.016 Bh5-g6 Nf1-d2 Qb6-c7 a3-a4 Kg8-g7 Ne3-f5 Bg6xf5 e4xf5 00:02 0 0 185642 (27572) 7 0.016 Bh5-g6 Nf1-d2 Qb6-c7 a3-a4 Kg8-g7 Ne3-f5 Bg6xf5 e4xf5 00:03 0 0 300465 (30800) 8 -0.209 Bh5-g6 Nf1-d2 Qb6-c7 Ne3-f5 Bg6xf5 e4xf5 Kg8-g 7 Nd2-c4 ++ d7-f8 ++ b6-a7 00:25 0 0 2661523 (950211) 9 0.005 Bh5-g6 Nf1-d2 Kg8-g7 Nd2-c4 Qb6-c7 Bb2-c3 Nd7 -b6 Bc3xa5 Nf6xe4 00:40 0 0 4458027 (1325961) 10 -0.040 Bh5-g6 Nf1-d2 Kg8-g7 Nd2-c4 Qb6-c7 Ne3-f5 Bg6xf5 e4xf5 b7-b5 Nc4-e3 Nd7-b6 01:46 0 0 12454596 (4061616) 11 -0.125 Bh5-g6 Nf1-d2 Kg8-h7 Nd2-c4 Qb6-c7 Ne3-f5 Bg6xf5 e4xf5 e5-e4 Nc4-e3 Nd7-e5 Re1-d1 04:13 0 0 30432823 (7331515) 12 -0.125 Bh5-g6 Nf1-d2 Qb6-c7 Ne3-f5 Kg8-h7 Nd2-c4 Bg6xf5 e4xf5 e5-e4 Kg1-f1 b7-b6 a3-a4 Qc7-b7 11:29 0 0 82720871 (10080646) 13 -0.124 Bh5-g6 Nf1-d2 Qb6-c7 Ne3-f5 Kg8-h7 Nd2-c 4 Bg6xf5 e4xf5 e5-e4 Kg1-f1 b7-b6 a3-a4 Nd7-f8 46:04 0 0 327781499 (111597415) 14 -0.078 Bh5-g6 Nf1-d2 Qb6-c7 a3-a4 Kg8-h7 Ne3- c4 Nd7-b6 Nc4xa5 Bc5-b4 c2-c3 Bb4xa5 b3-b4 Nb6-d7 b4xa5 Nd7-c5 h3-h4 Diep doesn't consider the bad move g5-g4. >> >>"The computer desperately seeks to disturb White's plan. > >The computer has no concept of desperation. > >>Although this move >>creates further kingisde weaknesses, it enable Black to develop some piece >>activity. This is the critical phase. Everybody who has played a computer knows >>the scenario: you get a strategically winning position, the computer makes some >>desperate tactical lunge, you make a couple of inaccuracies and suddenly the >>machine is all over you". (comment by Frederic Friedel) >No argument there! For those who tactical mistakes that's definitely true. For GMs that's not exactly true though. They got their GM because they can play a perfect game! >ChessMaster 8000 output: > >Time Depth Score Positions Moves >0:00 1/3 -0.04 2059 1...Bd4 2. Nd2 Bg6 >0:00 2/6 0.02 42104 1...Bd4 2. Nd2 Bg6 3. c3 Bc5 4. > Qc2 Nf8 >0:01 3/7 0.16 124547 1...Bd4 2. Nd2 Bg6 3. c3 Bc5 4. > Qc2 Nf8 5. Ndc4 >0:03 3/7 0.09 268740 1...g4 2. h4 Bg6 3. Qd2 Nf8 4. > Qe2 a4 5. Nd2 Bxe3 6. fxe3 axb3 > 7. cxb3 >0:06 3/8 0.10 598533 1...g4 2. Nd2 gxh3 3. Bxh3 Qc7 > 4. c4 Bd4 5. Nf5 Nc5 6. Nxh6+ Kg7 >0:10 3/8 0.02 981004 1...Bg6 2. Nd2 Qc7 3. Bc3 b5 4. > b4 Bd4 5. Qb2 >0:18 4/9 0.15 1798706 1...Bg6 2. Nd2 Qc7 3. Bc3 Bd4 4. > Qb2 Nb6 5. Rad1 Bxc3 6. Qxc3 >1:10 4/10 0.14 6627204 1...Bg6 2. Nd2 Qc7 3. Bc3 Bd4 4. > Qb2 Nb6 5. Rad1 Qe7 6. Nf5 Bxf5 > 7. exf5 Bxc3 8. Qxc3 >4:05 5/11 0.11 24328715 1...Bg6 2. Nd2 Qc7 3. Qd1 Bd4 4. > c3 Ba7 5. b4 Nb6 6. Qe2 Qd6 7. > Nec4 Nxc4 8. Nxc4 >14:02 6/12 0.03 88355386 1...Bg6 2. Nd2 Qc7 3. Bc3 Bd4 4. > Qb2 Nc5 5. Nf5 Bxf5 6. exf5 Qd7 > 7. b4 Bxc3 8. Qxc3 Qxd2 9. Qxc5 >I suspect that more interesting would be the analysis of these: > [D]3rr1k1/1p1n1p2/1qp2n1p/p1b1p2b/4P1p1/PP2N1PP/1BP2PB1/R1Q1RNK1 w - - position after g4? DIEP sees this as advantage for white directly from the start: 00:00 0 0 56 (0) 1 0.622 h3xg4 Bh5xg4 00:00 0 0 75 (0) 1 0.654 Ne3-f5 00:00 0 0 1065 (0) 2 -2.149 Ne3-f5 Bc5xf2 Kg1-h2 Bf2xe1 Qc1xe1 ++ h3-g4 00:00 0 0 1133 (0) 2 0.622 h3xg4 Bh5xg4 00:00 0 0 2459 (173) 3 1.325 h3xg4 Bc5xe3 Nf1xe3 Bh5xg4 00:00 0 0 6277 (363) 4 0.506 h3xg4 Bh5-g6 b3-b4 a5xb4 a3xb4 Bc5xb4 00:00 0 0 25704 (5654) 5 0.966 h3xg4 Bh5-g6 g4-g5 h6xg5 Ne3-c4 00:00 0 0 56414 (11138) 6 0.462 h3xg4 Nf6xg4 Ne3xg4 Bh5xg4 Nf1-e3 Nd7-f6 00:02 0 0 204394 (48885) 7 0.460 h3xg4 Nf6xg4 Ne3xg4 Bh5xg4 Nf1-e3 h6-h5 Ne3xg4 h5xg4 ++ f1-d2 00:03 0 0 343021 (86294) 7 0.696 Nf1-d2 Qb6-a7 b3-b4 Bc5-f8 Ne3-f5 a5-a4 Nd2-c4 00:05 0 0 512604 (116944) 8 0.400 Nf1-d2 Qb6-a7 Nd2-c4 g4xh3 Bg2xh3 Kg8-f8 Bb2-c 3 b7-b6 ++ h3-g4 00:07 0 0 730577 (186296) 8 0.548 h3xg4 Bh5-g6 b3-b4 a5xb4 a3xb4 Bc5xb4 Ne3-c4 Q b6-c5 Qc1xh6 Qc5xc4 00:25 0 0 2731856 (750383) 9 0.294 h3xg4 Nf6xg4 Bg2-h3 Ng4xe3 Nf1xe3 Nd7-f6 Kg1- g2 Bc5xe3 Qc1xe3 Qb6xe3 Re1xe3 ++ f1-d2 00:33 0 0 3643383 (1015766) 9 0.506 Nf1-d2 Qb6-a7 Bb2-c3 a5-a4 b3xa4 Kg8-f8 h3-h 4 Bc5-d4 Qc1-b2 Qa7xa4 Bc3xd4 e5xd4 Qb2-b4 Qa4xb4 a3xb4 01:07 0 0 7402451 (2116658) 10 0.291 Nf1-d2 Qb6-a7 b3-b4 Bc5xe3 Re1xe3 g4xh3 Bg2 xh3 Kg8-f8 b4xa5 Nf6-g4 ++ b2-c3 02:16 0 0 15253229 (3679456) 10 0.341 Bb2-c3 g4xh3 Bg2xh3 Bh5-f3 Nf1-d2 Nf6xe4 B c3xa5 Qb6xa5 Nd2xf3 Kg8-f8 a3-a4 ++ f1-h2 02:40 0 0 17870834 (4220380) 10 0.354 Nf1-h2 g4xh3 Bg2xh3 Bh5-g6 Bh3xd7 Nf6xd7 N h2-g4 h6-h5 Ng4-h6 Kg8-f8 Nh6-f5 Nd7-f6 ++ a3-a4 03:50 0 0 25753917 (6499422) 11 0.354 Nf1-h2 g4xh3 Bg2xh3 Bh5-g6 Bh3xd7 Nf6xd7 N h2-g4 h6-h5 Ng4-h6 Kg8-f8 Nh6-f5 Nd7-f6 ++ b2-c3 ++ f1-d2 06:18 0 0 42810178 (7893075) 11 0.437 Nf1-d2 Qb6-a7 b3-b4 Bc5xe3 Re1xe3 g4xh3 Bg 2xh3 Nf6-g4 Bh3xg4 Bh5xg4 Nd2-c4 a5-a4 11:29 0 0 76993736 (11384648) 12 0.437 Nf1-d2 Qb6-a7 b3-b4 Bc5xe3 Re1xe3 g4xh3 B g2xh3 a5-a4 Nd2-c4 Kg8-f8 Bh3xd7 Nf6-d5 [D]3rr1k1/1p1n1p2/1qp2nbp/p1b1p1p1/4P3/PP2N1PP/1BP2PB1/R1Q1RNK1 w - - position after Bg6 Obviously the scores here are the same like the first output produced where diep planned to play Bg6 >Which are the aftermaths of the two choices. Often, when a computer will pound >away for days on some position looking for an appropriate bm, it won't find it. >But if you instead have it analyze the best choice, it will find it in 1/100th >of the time. Not sure why it works that way.
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.