Author: Dieter Buerssner
Date: 15:20:07 05/08/02
Go up one level in this thread
On May 08, 2002 at 16:32:37, Steffen Basting wrote: >I've just written my first chess program. It searches about 6 plies ahead >(standard alpha - beta with iterative deepening + aspiration windows) plus >quiescent search. As it is quite frustrating that it is often victim to traps >(especially when men are pinned) i wonder how to improve the playing strength. >It would be nice to hear some suggestions - i am a beginner to chess >programming... If you get 6 plies, the other comments may be most worthwhile - especially move ordering. killer moves and history heuristics are not difficult to implement. I think SEE is not easy, and for a new program MVV/LVA should be enough, and time spent on other ideas may be more efficient. You will get easily deeper, when you implement null move - this can give a real boost. Hash tables will help to cut the tree size down (due to move ordering and due to avoiding to search whole subtrees). Another boost (and perhaps the most important one) I think can be expected by search extensions. Everybody will do this different in detail, but some common ideas seem to be in many engines - analyze lines with check(s) and/or recaptures longer. The easiest way to start may be - just don't count checks and recaptures as plies (don't decrement remaining depth). Regards, Dieter
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.