Author: Russell Reagan
Date: 15:38:15 07/12/03
Go up one level in this thread
On July 12, 2003 at 14:28:45, Dana Turnmire wrote: >The main question would be is there a gold standard of chess suites that all >problems have been confirmed to be sound? There are so many out there so which >one would be best as far as testing a program? Personally, I think that the best test positions for most engines are positions in which we can clearly show that one move or variation is winning and superior to all other moves and variations. It might take a lot of work for us humans to show that the particular move or variation is THE superior one, so don't mistake "clear" for "simple". The purpose of this for most engines should be to find bugs or holes in the engines playing strength. An example of questionable test positions are some of the Nolot positions where it is unclear if the "best" move is really superior to all others. Also note that I said that this applies for "most engines". By "most engines" I mean engines that are not the strongest (like Fritz, Shredder, Junior, and so on). Most engines don't need to be worrying about ridiculously complex test positions. Most engines would benefit more from patching up some of their "holes" and being able to find the best move in positions where there is a clear best move. I'm not even sure if I would include the top amateurs in this category. For instance, Crafty is very well tested, and it probably has very few major holes. I'm sure a GM could find some smaller holes to take advantage of, but not many major ones.
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.