Author: Pat King
Date: 10:09:42 12/03/02
Go up one level in this thread
On December 03, 2002 at 11:29:56, Robert Hyatt wrote: >On December 02, 2002 at 22:46:50, Bob Durrett wrote: > [snip] >>If no such algorithm could POSSIBLY exist, then why not? > >Within the framework of alpha/beta there is no way to do this. > >Why? > >Because alpha/beta doesn't search the entire tree. It searches the first branch >at the >root, for the score, then it uses that score to prove that the other root moves >are worse, >without proving exactly how much worse they are. > >And therein lies the problem. To get the second best move, you don't just prove >that it >is worse than the best, you have to search it completely to see exactly how much >worse >it is, and that drives up the time exponentially. > Perhaps there's a middle ground here. You can sort the scores of the fail-lows at the root, cost O(B^2) at worst, and then probe the hash table for the corresponding variations. You'd know the PV was worth (say) exactly 314, variation 1 < 271, variation 2 < 141, etc. As a practical matter, how useful are those less-than scores? [snip] Pat
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.