Author: Russell Reagan
Date: 20:59:16 06/02/02
Go up one level in this thread
On June 02, 2002 at 13:36:51, Gian-Carlo Pascutto wrote: >You don't need to store every position to 'solve' chess. If just win/loss/draw >and the move to reach it is enough, a small and finite amount of RAM will do. I think it was you who pointed out that RLE compression would be good for using win/loss/draw tablebases. If you could somehow get a position to map to indexes into a tablebase file, and get the position's results to cluster together well (IE many consecutive wins, then many consecutive losses, then many consecutive draws), you could turn potentially the entire 32-man tablebase data into a handful of bytes. However, I think this is another case of an idea that sounds great, but is probably not practical to implement. For example, if you could create such a function to group results of positions that well, you would have a very fast, almost perfect evaluation function. Russell
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.