Author: Ratko V Tomic
Date: 16:55:19 05/23/00
Go up one level in this thread
> >> While I agree that SSDF could use recalibration to human >> ELO, the problem is not that the SSDF ELO is always too high. >> For the top programs their rating is too high, for the bottom >> ones it is too low when compared to human players. So the >> problem is that it exaggerates the ELO difference between >> programs (relative to human players). > Everyone that has followed this understand the reason for this. I am glad SSDF recognizes that the problem is in exaggerated rating differences (which on high end gives overrating and on low end underrating). Since obviously the list would be more useful (to humans) if it matched closer the computer vs human strengths, does this acknowledgment mean SSDF will try to compensate for the problem? (Or create secondary ranking more useful to human players.) >Today almost every player plays and analyzes with computers all day long. >Today people understand the way to play them. Simple as this. >The program that probably suffers most from the above facts is Fritz because >almost every chessplayer use Fritz, not necessarily because it's better than >other programs but it is very strong tactically and you have chessbase > excellent database functions. One can look at the list spread from various angles. In my note I had explicitly separated the anticomputer strategies and drastically new kind of programs from the flexibility (or dimensionality of variation) component. The latter can be fixed in uniform and automated way by choosing a smaller ELO factor (maybe 300 instead of 400) for the existent database of comp-comp results. The same method would be much harder for human-computer results, trying to separate which humans played anticomputer (or even anti-specific program) and which played generic chess. One would need to look individual games and separate them in order to use a customized ELO factor for each of the two sub-types. In any case, whatever causes one sees as dominant, the acknowledgment of the problem is the first step toward solving it.
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.