Author: Uri Blass
Date: 01:38:54 10/31/00
Go up one level in this thread
On October 31, 2000 at 03:11:54, Jouni Uski wrote: >Yes it's funny, that in first try Gandalf won but in next try it only draws >despite of learning! Is this negative learning? > >Jouni No Fritz played a different move because it could not use the permanent brain. If you repeat the same opening then it is more logical to use 0 seconds for moves that you played in the past than to repeat the same calculations so nothing is wrong in gandalf's learning. This learning may be counter productive only against stupid programs that do not learn because they can be lucky and play different moves because they cannot use the permanent brain when the opponent is using 0 seconds. This learning may be productive against programs like mchess that have learning function but try to correct the mistake later in the game and not in the opening book or against programs that are not deterministic so it is not clear that the 0 seconds is a bad idea. 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.