Author: Uri Blass
Date: 14:15:24 06/08/04
Go up one level in this thread
On June 08, 2004 at 14:26:45, Tord Romstad wrote: >On June 08, 2004 at 13:36:27, Fabien Letouzey wrote: > >> >>On June 08, 2004 at 13:25:59, Joachim Rang wrote: >> >>>thanks for the test. I think they are in fact about equal. >> >>>regards Joachim >> >>Don't worry, they won't be equal anymore when Gothmog 1.0 is out :) > >Why do you think so? As I explained to you yesterday, the "improvements" I >have done since Gothmog 0.4.8 do not seem to have been effective. > >In fact, Gothmog seems to have degenerated into a highly complicated random >move generator. No changes I make, even very dramatic ones, have *any* >measurable effect on the playing strength. The only effect of changes to >the search or eval is to give different seeds to the random move generator. >The moves played are different, but neither stronger nor weaker. :-( > >Tord I suspect that it can be significantly improved by only changing weights. There is no reason to believe that the numbers that you have for evaluation and search are the best. 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.