Author: Uri Blass
Date: 01:04:30 01/03/01
Go up one level in this thread
On January 03, 2001 at 03:35:57, Pete Galati wrote: >On January 02, 2001 at 02:45:27, Ferdinand S. Mosca wrote: > >>On January 02, 2001 at 01:24:15, Steve Timson wrote: >> >>>This position is from a game Chester played on ICC vs. a human. >>> >>>[D] 1r3rk1/pbqnnppp/4p3/2ppP3/p2P4/B1PB1N2/2P2PPP/R2QR1K1 w - - 0 15 >>> >>>Chester's opponent played Bxh7 and went on to win. Chester has a very hard time >>>understanding this position. It doesn't realize Bxh7 is a good move until a few >>>moves down the resulting sequence (Bxh7 Kxh7 Ng5+ Kg6 and it sees the swing >>>after Qg4). How do other programs do here? Do they see Bxh7 in a reasonable >>>time period? >>> >>> - Steve >> >> >> >>Hi Steve, >> >>Yace on P3_450 found Bxh7 in less than 27 seconds. >>This position is a bit hard for programs since it gives up material. >> >>Dinan > >With the Dos Comet on a Celeron700 in Windows, approx 18mb hash, it took 77 >seconds. I have doubts that this is very difficult for a program at all. They >can see the King getting trapped real fast. I just wish I could see the rest of >what Comet is thinking, but that King appears to be in big trouble. > >Pete > >Comet-B.27 Wed Jan 3 02:12:29 2001 > >depth score sec nodes pv > 1. +0.19 0 947 d3b5 > 2. +0.15 0 3466 a3c5 d7c5 d4c5 c7c5 a1a4 > 3. +0.15 0 10709 a3c5 d7c5 d4c5 c7c5 a1a4 > 4. +0.11 0 12140 a3c5 d7c5 d4c5 f8c8 f3g5 > 4. +0.11 0 41242 a3c5 d7c5 d4c5 f8c8 f3g5 > 5. +0.02 0 49891 a3c5 d7c5 d4c5 f8c8 a1a4 c7c5 d1d2 > 5. +0.11 0 80031 d1d2 f8e8 d4c5 e8c8 a3b4 d7c5 b4c5 > 5. +0.11 1 133031 d1d2 f8e8 d4c5 e8c8 a3b4 d7c5 b4c5 > 6. +0.14 2 245899 d1d2 f8e8 a3c5 d7c5 d4c5 c7a5 e1b1 > 6. +0.14 5 766414 d1d2 f8e8 a3c5 d7c5 d4c5 c7a5 e1b1 > 7. +0.13 7 1067950 d1d2 f8c8 e1b1 e7g6 a3c5 d7c5 d3g6 > 7. +0.13 12 1694460 d1d2 f8c8 e1b1 e7g6 a3c5 d7c5 d3g6 > 8. +0.04 24 3132774 d1d2 f8e8 e1b1 b7a8 d2g5 b8b1 a1b1 > 8! +0.26 77 11278271 d3h7 g8h7 f3g5 h7h6 d1g4 e7g6 g5e6 > 8! +0.53 79 11826970 d3h7 g8h7 f3g5 h7h6 d1g4 c5d4 a3e7 > 8. +0.56 83 12775906 d3h7 g8h7 f3g5 h7g6 d1g4 f7f5 a3b4 > 8. +0.56 83 12798067 d3h7 g8h7 f3g5 h7g6 d1g4 f7f5 a3b4 > 9? +0.42 85 13314268 d3h7 g8h7 f3g5 h7g6 d1g4 f7f5 g4f5 > 9? +0.15 113 18995944 d3h7 g8h7 d4c5 > 9. +0.15 114 19120060 d3h7 g8h7 d4c5 > >==> My Move: d3h7 in 1:59 CometB27(64 mbytes hash under chessbase) had a fail low without solving it. It is a bad point to stop it. If you give Comet more time it is going to solve the fail low and see a negative score for Bxh7+ and it is going to change its mind to Qd2 Comet can find Bh7 again at depth 10 and it fail high again and again to get +1.61 score after 4:47 on pIII800. It fails high again at depth 11 and it can see +1.97 score after 6:15 and +1.99 at depth 12 after 9:11. It fails high again at depth 13 with score of +2.12 after 13:38 I agree that the position is not very hard but your result for Comet is too optimistic because you stopped it before it had an opportunity to see the tactical reason. The times are not exactly correct and may be slightly smaller because the p800 did another task(this post) Uri
This page took 0 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.