Author: Bas Hamstra
Date: 02:08:19 03/16/04
Go up one level in this thread
On March 15, 2004 at 16:44:59, Uri Blass wrote: >On March 15, 2004 at 16:24:16, Ralf Schäfer wrote: > >>On March 15, 2004 at 14:28:57, Dann Corbit wrote: >> >>>On March 15, 2004 at 14:20:19, Ralf Schäfer wrote: >>> >>>>On March 14, 2004 at 20:28:15, Bas Hamstra wrote: >>>> >>>>>This copy & paste from Shredder 5 is soooo cool. This is a tough position from >>>>>BS2830. I am curious how other engines do here. Tao 5.4 used to solve this in 2 >>>>>minutes but the latest version has a hard time here. >>>>> >>>>>[D] 2r1r3/p3bk1p/1pnqpppB/3n4/3P2Q1/PB3N2/1P3PPP/3RR1K1 w - - >>>>> >>>>> >>>>>Ciao! >>>>> >>>>>Bas. >>>> >>>>Hello Bas, >>>> >>>>I'm still surprised that my own engine Cheetah finds this move in 7 seconds at >>>>ply 7 - maybe 7 is its lucky number ;-) >>>> >>>>Here is the thinking output: >>>> >>>>Cheetah 0.73 WB2 64 MB: >>>> 1 00:00 0,85 Lb3xd5 >>>> 1 00:00 2,15 Lb3c2 >>>> 2 00:01 0,79 Lb3c2 Te8g8 >>>> 3 00:02 0,83 Lb3c2 Te8g8 Te1e2 >>>> 4 00:02 0,38 Lb3c2 Le7f8 Lh6d2 Lf8g7 >>>> 4 00:02 0,47 Lb3xd5 Dd6xd5 Td1d3 Te8g8 >>>> 4 00:02 0,49 Lb3a4 Le7f8 La4xc6 Tc8xc6 >>>> 4 00:03 0,53 Dg4h4 Kf7g8 Lb3c2 f6f5 >>>> 5 00:03 0,61 Dg4h4 Dd6d7 Lb3xd5 Dd7xd5 Dh4g4 >>>> 6 00:04 0,49 Dg4h4 Dd6d7 Td1c1 Kf7g8 Lh6d2 g6g5 >>>> 7 00:05 0,61 Dg4h4 Kf7g8 Td1c1 f6f5 Sf3g5 Le7xg5 Lh6xg5 >>>> 7 00:07 0,62 Te1xe6 Dd6xe6 Sf3g5+ f6xg5 Dg4f3+ De6f5 Lb3xd5+ Kf7f6 Ld5xc6 >>>>Df5xf3 >>>> 8 00:09 0,74 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 Sc6e5 d4xe5 Dd6c6 >>>> 9 00:11 1,15 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 Dd6d8 Dg4e6+ Kf7g7 Lb3xd5 >>>>Sc6a5 >>>>10 00:13 2,01 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 Te1xe7 >>>>10 00:17 2,01 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 >>>>11 00:21 0,90 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 Ld5xc6 Tc8xc6 >>>>11 00:52 0,90 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 >>>>12 00:54 1,16 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 Ld5xc6 Tc8xc6 >>>>12 02:56 1,16 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Lf4xd6 f5xg4 Lb3xd5+ >>>>Kf7f8 >>>>13 03:34 1,59 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >>>>Kf7f6 Lb3xd5 Sc6d8 b2b4 >>>>13 11:50 1,59 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >>>>Kf7f6 >>>>14 14:31 1,80 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >>>>Kf7f6 Lb3xd5 Sc6d8 g2g3 >>>>14 54:43 1,80 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >>>>Kf7f6 >>>>15 01:05:54 1,60 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 h7h5 Dg4xc8 Te8xc8 Lf4xd6 >>>>Le7xd6 Lb3xd5+ Kf7g7 >>> >>>Can you dump your evaluation terms to say why your engine picked the move so >>>quickly? >>> >>>I am guessing king safety? >> >>Hi Dan, >> >>well, I think that its king safety will help to find the move fast, but maybe >>this is not the only reason. After playing the moves from ply 15 Cheetah sees >>the passed pawn on the d-file, at this point the king-safety is turned off >>because there are too little pieces on the board (Cheetah needs 4 pieces or 2 >>pieces and a queen for doing king safety) >> >>Cheetah's king-safety is a lighter version from Rebel's king safety which is >>described on his page, so it is interesting if Rebel can find this move too... >> >>best wishes >>Ralf > >It seems that you have big king safety scores that help you to fins the move >fast by evaluation a position when white is a rook down for 2 pawns as better >for white based on king safety. > >13 11:50 1,59 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >Kf7f6 >14 14:31 1,80 Te1xe6 Dd6xe6 Td1e1 De6d6 Lh6f4 f6f5 Dg4h3 Dd6xf4 Dh3xh7+ >Kf7f6 Lb3xd5 Sc6d8 g2g3 > >Uri Maybe that helps him find it quickly. But eventually this wins material... Bas.
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.