Author: Alessandro Damiani
Date: 16:44:52 11/22/02
Go up one level in this thread
On November 22, 2002 at 19:08:27, Steffan Westcott wrote: >On November 22, 2002 at 12:52:54, Alessandro Damiani wrote: > >>Steffan? You wrote someting you name SME? Does it include a SEE? > >Yes, SME (Static Move Evaluation) is a generalisation of SEE where the first >move in the 'exchange search tree' is not necessarily a capture. The exchange >logic does have some gross simplifications (to constrain the large number of >possible exchange patterns) but the results work reasonably well. > >SME is there when the other heuristics have dried up, so its not used >everywhere. Its designed to discourage 'obvious' daft moves like moving a piece >to a square attacked by an unpinned enemy pawn, or moving pieces away from a >co-ordinated attack, etc, and encourage 'obvious' sensible moves like developing >a piece, defending own king under attack, etc. SME is heavy, but its only there >to generate the first few 'plausible' moves, its not used throughout the entire >movelist, its not used on every search node, and much of its calculation (which >is a subset of a full position evaluation) is performed at most once per search >node, with much lighter processing per subsequent move suggestion. > >Forgive me if I don't entertain any more questions on SME, I've probably said >enough for now and I have to have some secrets! > No problem, I wasn't asking about SME to use it myself. I was wondering if it is of the complexity of a SEE and if your move generation idea allowed a more efficient static evaluation like SEE. I use my own method anyway, which is simpler than a SEE but with enough accuracy. When do you think you will have your engine ready for playing? I still need some weeks to finish rewriting my engine. Alessandro
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.