Author: Odd Gunnar Malin
Date: 17:15:28 01/22/03
Go up one level in this thread
On January 22, 2003 at 19:41:33, Nathan Thom wrote: >Hi all, I implemented a pretty quick killer move algorithm into LittleThought >last night from memory of what its supposed to do. > >A couple of questions: >1. First, i incremented the killers[from][to] array by the inverse of the >current depth (ie MAX_DEPTH - depth) when a cutoff occurred. In my code depth 0 >is the root and it increases downwards. I figured this would be better than >plain depth because cutoffs closer to the root node would be better. However i >found this to slow it down considerably. To test my theory, i changed it to >plain depth and it perfored much better than normal. So, can someone explain why >incrementing by depth was better than MAX_DEPTH - depth or this just some quirk >with my implementation/test positions? Also if someone has any ideas as to a >better incremental? > >2. What is a history move? why is it different from a killer move? > >Thanks. Hi. You are talking of the history move here and not the killer. There is a paper on the net writen by Jonathan Schaeffer that discuss this heuristics in detail. To answer your question. The orginal consept was to store 2^depth where depth are reminding depth. Because of danger for overflow I think most use depth^2 now. Also in the orgin paper he stored this on the best move to (no cutoff) but this you have to experiment with, maybe also try store the hashmove, exclude capture, clear before each iteration etc. I'm not sure how it is best to test, maybe with many games and log the moveordering (fail on first, second etc.) For the test maybe an idea is to not use hasmove etc. for moveordering. The killer move is a move you store on each ply (killer[MAX_PLY]). You will store the move score here and replace it when you find a move at this play with better score. You can also store two moves at each ply to test. This have much in common with history moves, and maybe you will find oyt that the best is to use only one of them. On Ed Schroeder's page he show another method of using the killer move, it seams that he don't use history as explained in Schaeffers paper but talking of a threat history move. Odd Gunnar
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.