Author: Torstein Hall
Date: 07:54:33 05/29/03
Go up one level in this thread
On May 29, 2003 at 10:51:24, Jorge Pichard wrote: >On May 29, 2003 at 10:17:30, Tony Hedlund wrote: > >>On May 29, 2003 at 05:06:49, Jorge Pichard wrote: >> >>>I am starting to believe that a lot of the wins of Shredder are do to the good >>>learning and also the extremely optimized Opening Book of Shredder. The only >>>fair way to test the strength of any engine would be if the learning files of >>>the programs being tested are set to zero and to reverse the Opening books after >>>20 games are played between engines. For instance after this match reach 20 >>>games the Shredder7.ctg should be applied to Fritz 8, and Fritz8.ctg should be >>>assigned to Shredder 7.04. >>> >>>Pichard. >> >>we are running Shredder 7.04 under the classical GUI as an UCI engine, using the >>shredder7.bkt book. >> >>Tony > > >In that case after the match reach 20 games it is only fair to disable the books >of both programs for the remaining of the match. Then we can compare if the >opening books is a factor and NOT the evaluation of the engines. This kind of >test will provide us with a lot of valuable data that we can use for comparison >such as: we can compare the margin of games that one program lose from playing >without a Killer book like shredder7.bkt or shrdder7.ctg and at the same time, >how effective the learning feature really is if the learning files of both >programs are set to zero or are disabled for the second half of the match. > >Pichard Openingbook and position learing is part of the program. Not that I think the pos. learning counts for much though. Torstein
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.