Author: Andrea Griffini
Date: 04:53:30 06/19/03
Go up one level in this thread
On June 19, 2003 at 00:18:30, Peter Skinner wrote: >the programmer says it doesn't use tablebases, yet >kibitzes mate in 40 scores... I find that odd. It wouldn't be hard to actually code a quick mating algorithm for special positions (i.e. R+K/K) even if having it being optimal may be is a bit harder. Claiming a mate in 40 doesn't mean there can't be a mate in less... but just that the program knows how to mate in 40, and for that a special algorithm is enough. Also actually computing a mate using reverse analysis is not that hard for simple cases, but I don't think this is as useful as a precomputed table because during the search it's important to know *immediately* that the position is say a mate in 20. Also "building" on the fly the complete mating table for simple cases in RAM should require just a few seconds... I wonder if this can be classified as using tablebases or not. I was thinking to add reverse analysis for just K+pawns endings... even if enumerating the winning positions is probably harder than for checkmates. Andrea
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.