Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chess Analysis Project (prabable for Dann Corbit)

Author: Pete R.

Date: 15:03:13 11/17/99

Go up one level in this thread


If it's correct that Dr. Hyatt has imported the CAP data into his Crafty, then
it would be desirable to get regular dumps of any games that Crafty loses (or
perhaps even draws depending on the number) so that the weak move can be
identified and the CAP data created or updated in the variation played.  That
way the analysis is focussed on the openings that Crafty actually plays against
good opponents, and those openings are tuned at the point that they need tuning,
i.e. in the area where Crafty deviates from best play.  This way there is a
feedback mechanism from actual under-fire playing conditions, and the book wlll
improve immediately.  Difficult cases where a particular variation is causing
repeated trouble can even be hand-analyzed and/or pounded at extreme time
control.  It would also be fun and motivating for C.A.P. members to see Crafty
do better as a result. :)



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.