Author: James Swafford
Date: 11:31:58 01/05/02
Go up one level in this thread
On January 05, 2002 at 12:42:18, John Coffey wrote: > >>> >>>Severi >> >>You are right. I just think the idea of use history moves as killer moves but >>forget totally about attributes of history moves :) > > >Can someone explain the difference between history and killers? > >john coffey 'Killer moves' are just a list of two or three moves per ply of search that have caused beta cutoffs. When you start your search, you can try the killer moves stored for that ply of search in hopes of getting a quick cutoff and minimizing move generation. The history heuristic, as explained in another post, does not entail storing moves. Each time you get a beta cutoff, update a global array 'history[64][64]' by incrementing it in some fashion. A simple approach is to increment history[source][dest] by one, but you may experiment with other techniques that take depth into account. After you generate moves, you can use the moves' history score as a guide when trying to figure out which one to play first. Typically this only applies to non-captures. -- James
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.