Author: Uri Blass
Date: 21:10:12 07/27/02
Go up one level in this thread
On July 27, 2002 at 20:37:48, Vincent Diepeveen wrote: >On July 27, 2002 at 06:33:24, Uri Blass wrote: > >that's a lot of nodes to get 9 ply Uri, >you're not using nullmove? I use null move with R=3 but also extensions for checks that may be more than one ply in part of the cases and the qsearch of this version is buggy(in small part of the cases it return beta instead of returning the right value but that bug did not prevent it to beat gnuchess in a match that peter berger did). I think that the result of the bug in most cases is that the program needs more nodes to get the same depth but usually it does not cause stupid mistakes(otherwise the program had no chance against other programs in chess). I posted later results of a later version when I hope that it has less bugs but still seems to have some bugs because I saw a wrong pv(it seems that there is a mistake in the pv because the score(mate in 2) was correct but the second move in the pv was not correct. If the bug is a bug only in getting the pv in the qsearch then it is not a big problem but I suspect that there are other problems based on watching games and evaluation and I will have to find out. 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.