Author: Pete R.
Date: 10:05:57 07/26/00
Go up one level in this thread
On July 26, 2000 at 12:42:41, blass uri wrote: >I do not see the point of comparing the evaluation with other top programs. It might be worthwhile if you have an amateur program that is 800 points or so weaker than Fritz or whatever. Of course if you succeed then you will be duplicating Fritz' weaknesses, though you will be much better than you were before. Ultimately it is much more desirable to have very accurate test suites, but this is not an easy thing. I think Dann has shown a great many errors in common test suites for example. What is the most accurate test suite for positional chess? I don't know. How can a test suite be created where we have a high confidence that white is slightly better, or much better, black is slightly worse, etc.?
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.