Author: Thorsten Czub
Date: 07:05:54 01/24/00
Go up one level in this thread
On January 23, 2000 at 20:53:37, Imran Hendley wrote: >>1. you let the new versions play against older versions of the same >>program , or let the same program play on fast machines against >>slower machines. thats nonsense. >>2. you don't delete the doubles ! here you even count 12x the same opening !!! >>3. you only let the new programs play against new other software >>(or - as stupid, see point 1). But you never test against a big >>variety of programs on same fast hardware. >>therefore your results are ot serious. >>die schwedische eloliste ist nicht mehr das was sie einmal war. >>the autoplayers have not only brought us advantages. it seems >>they had very negative effects on you. >>:-)) >2 & 3 can be debated, but I don't think number 1 is a mistake by the SSDF. If a >human loses two identical games, the second one isn't "deleted" so I think the >ability of a program to learn from its mistakes should and does effect its >rating. it seems you mixed up the points. 2 was the doubles, not 1. i see not much sense in wasting time playing 12 times the same game. as much affect on the rating this must have, it has nothing to do with the playing strength of the program. only with the intelligence or lazyness of the programmers or the operators. my point. you can of course have other point of views.
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.