Author: Robert Hyatt
Date: 20:08:43 02/19/02
Go up one level in this thread
On February 19, 2002 at 15:05:15, Gian-Carlo Pascutto wrote: >On February 19, 2002 at 13:59:52, Robert Hyatt wrote: > >>Alpha/beta works well in _many_ games. But Shogi uses different >>rules. > >I know. But I didn't see anything that would cause a fundamental >problem, i.e. why the a/b approach wouldn't be very effective. I didn't word that very well. I didn't mean to imply that alpha/beta would not be so great for Shogi. I meant to say that alpha/beta would work just fine for Shogi, but the _rest_ of the stuff would not. IE search extensions, evaluation, etc would all need lots of work to fit Shogi... just like it all needs work to fit loser's chess where the game is different. > >>IE alpha/beta works well for loser's chess too, but the >>evaluation has to be far different... > >Yes and no. Proofnumber search works a lot better for finding >winning combinations (very important), but allows no positional play, >so a combination of a/b and proofnumbersearch is needed. > >A program using the combination (or something which pretty much >boils down to it) will spank a classic a/b searcher anyday. > >-- >GCP
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.