Author: Uri Blass
Date: 06:47:58 08/26/02
Go up one level in this thread
On August 26, 2002 at 09:38:52, Johan Hutting wrote: >On August 26, 2002 at 08:15:36, Uri Blass wrote: > >>>I don't have a random .ctg handy here so I can't test it, but an old, buggy >>>version of celes (rated 1800- or so) once scored 4/20 vs deep fritz in a 5 0 >>>match while both used deepfritz.ctg. >> >>it seems impossible to me to score so well against deep fritz. > >It was quite simple actually, Celes just had a plus when out of book, fritz >decided to draw 8 times while that build of Celes didn't comprehend 3fold >repetition yet. I admit the result is rather excessive, but it does point out >the problem. Would you rather start with -1, -2 or worse out of book or with >-0.5..0.5? Note that good books also give you solid positions with few >positional weaknesses. I understand now. It is not going to be possible in more than 90% of the cases against movei even with 1.e4 a6 because even in this worse position I do not expect repetition. Even if celes does not know about repetition it is going to try to improve the position and not to repeat. 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.