Author: Robert Hyatt
Date: 06:58:50 08/21/05
Go up one level in this thread
On August 20, 2005 at 22:13:42, Cesar Contreras wrote: > >>I'm hardly "done" either. Got plenty of ideas for the coming year... I spent >>most of the last year completely re-doing every course I teach as we modernize >>our curriculum. As a result, Crafty suffered mightily last year. This year is >>going to see a lot of work done... > > >How important it's to tune parameters/techniques time to time? > >I think your program it's a reference to many people, so they tune their >parameters and techniques to win your program (and other programs of they >choice). > >Maybe that can be an explanation way some older versions of crafty seems >stronger that recent versions, i mean people fine tune their engine with current >versions of crafty and if you don't tune the parameters frecuently enougth to >figth back, then they have some advantage. > >Congrats the major problem with crafty at the moment is that the eval is fairly fragmented. It is due for a rewrite, which I am going to start on soon. It can certainly be significantly faster, and the "overlap" has got to go. In addition, there are some complex parts of the code that can easily be simplified, such as trying to evaluate distant passed pawns, vs distant pawn majorities, vs other things... It's just time to clean it up, and rewrite parts that are due for serious attention, such as the king safety issues among others...
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.