Author: Robert Hyatt
Date: 19:00:36 05/04/01
Go up one level in this thread
On May 04, 2001 at 15:24:26, Dan Andersson wrote: >Succinctly put :) Its good to know that many good ideas are constantly >reinvented. But history teaches us that bad ideas are at least as >indestructible. Back to chess, the concept of 'Loose Pieces Drops Off' or LPDO >for short is another good one. A penalty for each unguarded piece, and maybe a >higher order term for coupling the penalties. The concept is based on the >observation that manywins in chess are based on the gradual disorganisation of >the opponents pieces and then finished off by a relatively short combination >winning material. Some testing makes me feel that deep searchers profit more >from this factor. An added bonus is that it is possible to choose a tricky move >even in drawn positions i.e. Tablebase accesses. > >Regards Dan Andersson I do your "LPDO" idea but change P to "Pawns" for me. IE I don't just sum isolated or weak pawns, I exponentially sum them, so that two are more than twice as bad as one, etc... Ditto for my blocked pawn and lever scoring... I don't try to evaluate "hung pieces". I depend on the search to find and remove those before letting Evaluate() have a shot at scoring the position...
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.