Author: Robert Hyatt
Date: 10:27:09 06/02/04
Go up one level in this thread
On June 01, 2004 at 11:13:30, José Carlos wrote: >On June 01, 2004 at 11:04:52, Sune Fischer wrote: >>There's gotta be a better and more interesting way than to just repeat won >>games, if that's all there is to it then I'll happily swallow my duplicate >>losses with pride. >>;) >> >>-S. > > My learning doesn't repeat won games. It's not the way to go in my opinion. >But it fights well against those who repeat games. My learning works smoothly >tunning the book, adding and removing lines here and there. > I give more details in my reply to Uri above. > > José C. Interesting note in that my learning did not try to repeat won games. My original thinking was that on ICC, when Crafty wins, it is very likely caused by a mistake of the opponent, rather than exceptional moves by Crafty. But when Crafty loses, it is momentous since that is not common against humans. I have just modified my learning to be way more aggressive, in that it will now repeat won games within reason of a little bit of randomness. Should work much better in machine vs machine tournaments and matches, for certain. Change was trivial of course.
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.