Author: jonathan Baxter
Date: 15:19:59 10/02/98
Go up one level in this thread
On October 02, 1998 at 04:05:12, Ernst A. Heinz wrote: >On October 01, 1998 at 20:28:55, jonathan Baxter wrote: >> >>Currently I am working on an improved version of KnightCap, with better >>search and a much *simpler* evaluation function. Interesting preliminary >>results >>show that with parameter learning and *just* piece/square tables we can get >>a rating of around 2500 on ICC. > >If I do not grossly misinterpret your PST values in connection with your >above statement, then the final result seems to imply that in your games >on ICC *material* is simply the dominating factor. Material is definitely dominating. But for ordering materially equal positions, the learnt piece/square tables do a pretty good job. > >Have you tried which ratings "KnightCap" achieves with good hand-tuned PSTs? > No, but they'd probably be similar. we have a very good understanding of the "right" values for PSTs, so they are easy for us to tune. Cheers, Jonathan Baxter >=Ernst=
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.