Author: Severi Salminen
Date: 12:00:27 01/05/01
Go up one level in this thread
>I just reuse the score from the last iteration and center on that with an >aspiration windows of one third of a pawn. I do the +1 -1 parts too. At first I >found them "ugly", but as far as I can see, it is necesary. I also use the score from previous iterations. Only the first 1 ply search uses infinite window. If you would leave those +1 or -1 away you are telling the search fonction that you allready know a move that gives you that score. But this is not true and if the true score indeed is the same as the first try returned we are in trouble. They look ugly but are mathematically correct. Severi
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.