Author: Uri Blass
Date: 15:22:16 08/25/02
Go up one level in this thread
On August 25, 2002 at 17:55:27, Ulrich Tuerke wrote: >On August 25, 2002 at 15:09:33, Gian-Carlo Pascutto wrote: > >>On August 25, 2002 at 14:45:37, Alexander Kure wrote: >> >>>Crafty and SOS were tested with the Fritz 6 book (general.ctg from young >>>talents). Goliath i can't remember. >> >>Thanks for the correction. I was under the impression general.ctg was >>an autotmatically generated book. >> >>If that's true, then it's likely Crafty and SOS are overrated in SSDF. > >The conclusion to me is, that amateur engines should not be tested by SSDF >because a fair comparison is impossible. > >Either the play with the same book as the commercial engines, then they have an >"unfair" advantage. Or they play with a private book which results in an ELO >loss of about 100 points (if it's true what Alex said, and I don't doubt it). > >An amateur does not have the resources to generate a high class book nor will he >be going to pay a chess expert for this. > >So whatever procedure SSDF will choose, it can't be fair. > >That's a real tragedy. > >Uli I disagree. I do not believe that you lose 100 points from using a private book. GCP test a book that is based only on GM games so we are going to get an upper bound because a private book can be better than it(for example you may use statistics about hundreds of Comet's games to decide about the moves that Comet plays in the opening and you can avoid using the big book when there is no danger of repeating the same loss twice to avoid blunders in the big book). 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.