Author: Uri Blass
Date: 12:27:28 05/04/02
Go up one level in this thread
On May 04, 2002 at 14:33:37, Roy Eassa wrote: >On May 03, 2002 at 22:48:44, Robert Hyatt wrote: > >>On May 02, 2002 at 13:22:52, Roy Eassa wrote: >> >>> >>>But king safety could help an engine find it SOONER, right? >>> >> >> >>I don't think so, but based only on a quick look. IE Crafty fails high >>to +3 quickly. That is not a positional score but a real win in material... >> > > >There's still a misunderstanding here. I DO know that it's a tactical score, >and that good positional factors will not result in such a high score, but I >thought these factors *could* help steer the engine to find the tactical win >sooner than otherwise. Not so? Here is the output of my engine about it(It is materialistic and has no king safety knowledge). It needs more than 2 minutes on p800 to find Nxh5+. My excuses for the bad result: 1)It does not use today hash tables and I hope to improve the results after teaching it to use hash tables. 2)Better extension rules may also help. You can see based on the output that after Nxh5+ gxh5 there is a big jump from clearly negative to clearly positive score. It means that I probably need big positional scores if I want to find Nxh5+ for positional reasons and I believe that it is dangerous without doing a lot of testing. depth,evaluation in 1/100 pawns,time in seconds,nodes,main line 7 34 3.95 540868 d2b2 c8g4 b1c2 a4a7 d4e5 d6e5 h1g1 7 34 8.23 1176917 d2b2 c8g4 b1c2 a4a7 d4e5 d6e5 h1g1 8 21 20.31 2855514 d2b2 a4a6 h1h2 c8g4 h2g1 g4e2 g3e2 g7g8 8 21 37.25 5281641 d2b2 a4a6 h1h2 c8g4 h2g1 g4e2 g3e2 g7g8 9 18 53.23 7467765 d2b2 a4a6 b1c2 c8g4 e2f4 h7f8 d4c5 b6c5 g3e2 9 19 121.54 17651445 g3h5 9 332 204.46 29933234 g3h5 g6h5 f2f6 e5f6 d4f6 g7g6 e4e5 c5d3 b1d3 c4d3 d2d3 g6h6 f6g5 h7g5 f1f6 h6g7 d3g6 g7h8 g6e8 h8h7 h4g5 a4a2 e8h5 9 332 204.49 29936778 g3h5 g6h5 f2f6 e5f6 d4f6 g7g6 e4e5 c5d3 b1d3 c4d3 d2d3 g6h6 f6g5 h7g5 f1f6 h6g7 d3g6 g7h8 g6e8 h8h7 h4g5 a4a2 e8h5 10 362 770.84 113910163 g3h5 After Nxh5+ gxh5 6 -234 0.73 103739 d2e3 c8g4 h1g1 g7g8 d4c5 d6c5 6 -233 1.28 189790 e2c1 6 -232 1.71 256705 e2c1 e5d4 d2d4 a4a6 d4d1 c8g4 6 -232 2.30 336866 e2c1 e5d4 d2d4 a4a6 d4d1 c8g4 7 -227 3.57 508197 e2c1 c8g4 b1c2 a4a3 d4e5 d6e5 h1g1 7 -226 7.08 1011452 f2f6 7 -179 8.38 1194193 f2f6 h7f6 d2g5 g7h7 f1f6 e5f6 g5h5 h7g7 h5e8 f6d4 e2d4 7 -179 8.65 1235784 f2f6 h7f6 d2g5 g7h7 f1f6 e5f6 g5h5 h7g7 h5e8 f6d4 e2d4 8 155 37.69 5619948 f2f6 h7f6 d2g5 g7h7 f1f6 e5f6 d4f6 e8e7 g5h5 h7g8 f6e7 8 155 38.80 5803772 f2f6 h7f6 d2g5 g7h7 f1f6 e5f6 d4f6 e8e7 g5h5 h7g8 f6e7 9 881 492.62 69933579 f2f6 g7g8 f6g6 g8h8 d4e5 d6e5 g6b6 a4a5 e2g3 c8g4 b6h6 h8g8 g3h5 9 881 494.05 70173328 f2f6 g7g8 f6g6 g8h8 d4e5 d6e5 g6b6 a4a5 e2g3 c8g4 b6h6 h8g8 g3h5 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.