Author: Heiner Marxen
Date: 07:56:05 05/08/04
Go up one level in this thread
On May 06, 2004 at 05:10:23, Mike S. wrote: >On May 05, 2004 at 20:00:05, George Tsavdaris wrote: > >>(...) > >> As all moves lead to mate, all moves are the best ones. > >Please have some chess culture and some taste! :-)) > >Which doesn't necessarily mean that the shortest mate is always the "best" from >a human viewpoint. For example, a #6 can be much more clear - giving the >defender less options or the like - than an "inhuman" problem-like #4. People >who like chess problems will have the opposite opinion, but will accept your >statement even less. My carbon processor prefers the mate with "the smallest proof tree". Even for Chest this is a meaningful metric, although it is obviously not applied to the mate itself, but to the non-mating branches of the tree. >A statement like "Every variant which ends with mate, is equally best no matter >how long it is" is an attack against chess culture... I strongly recommend >programmers (and fans) not to underestimate the bad impression of "uglyness" a >chess program can give, when a too big compromise between good play and >practical success have been made. > >There are cases, not only nullmove-related one's, when 2600+ chess engines move >like patzers for such reasons. Computerchess can do better. Sometimes, too much >is sacrificed for a few more engine competition Elo points... I dislike that, >even when it happens only in 0.x% of games. It's wrong priority (better strength >of course, yes, but not at all cost please). > >The increasing importance of computerchess for chess in general brings an >increased responsibility for style and beauty of top-level chess, for >programmers. > >Regards, >M.Scheidl In general I agree with your points. Cheers, Heiner
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.