Author: blass uri
Date: 01:46:02 06/19/98
Go up one level in this thread
On June 18, 1998 at 20:54:36, Bela Andrew Evans wrote: >I don't think you'd want to take people's saved games to rate >computers -- you could get all kind of skewed data that way. After >all, people only have to save the games they want to save -- which >might be the ones they win, or which makes a certain program look >good. I can beat my computer programs a lot -- if I take back a >couple of tactical blunders I make each game :). I would hate to >have that kind of data end up in some rating list. > >No, the only way to do it would be to have people play the computer >programs under somewhat tournament-like conditions. In addition, I >think it would only be fair if people played the computer programs >they DID NOT own or were familiar with -- you want to know the >programs general strength, not how it gets beat up by somebody who >has played hundreds of games against it in his pet openings and >knows every strategic error the program makes in those openings. I agree it would be fair if people played the computer programs they DID NOT own or were familiar with but we cannot force it because people want to win. unpopular programs and programs we cannot buy will have better results against humans because of this reason. Uri > >Bela Evans
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.