Author: Alvaro Jose Povoa Cardoso
Date: 04:32:32 06/26/02
Hi, everyone. I'm trying to construct checkers EGTBs with distance to win/loss information. I use a full byte for each EGTB entry with the 6 rightmost bits for the distance info. I created the 4 piece tables and sadly I detected a few positions witch have slightly larger distance values. For example: a few positions in the 4 piece have wins/losses in 25 moves but it should be in 23 moves instead, so they are 2 moves above the correct values. I've been stuck with this for 3 weeks now and I'm getting a little frustrated. I studyed Jonathan Schaeffer's paper on Retrograde Analysis, but the algorithm in this paper only uses 2 bits for each entry with no distance to win/loss info. Could someone please help me on this problem? Could someone please give some urls or papers on EGTB generation with distance to win/loss? Best regards, Alvaro Cardoso
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.