Author: Marcus Kaestner
Date: 04:54:10 11/10/00
Go up one level in this thread
>I can add that there is a reason to doubt all the rating of chess programs >because of statistical error but when Fritz6 is in part of the cases fritz6a >there is a also more noise and not only with fritz but also with the opponents >of fritz. > >The opponents that were lucky to play Fritz6 and not fritz6a are going to get >better rating. you are right. but this fades away, because now we are playing only with f6a. also there is a number of maybe 80 games with f6 and under more than 2200 the influence is not so big. but yue are right, there is an error. but this error will not be more than maybe 10 elo for a specified program. marcus
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.