Author: Thomas Mayer
Date: 08:34:59 11/10/01
Go up one level in this thread
Hi Christophe, > Let me repeat it for the third time, and maybe you'll get it this time: > > by letting very different players (in term of strength) play each other YOU > GET A RATING WITH A HUGE ERROR BAR. I get you already... but I ignore it somehow... :) But I ask you: give me some engines with lower strength on Athlon 1333... Maybe Fritz 1 on Athlon 1300 ? It is maybe around 2400 on this machine... we can calculate with the formula +70 ELOs by each hardware doubling... Any idea how ChessTiger on Palm would score then ? Or any other dedicated computer ? Or maybe try the Tasc R.30 against some of the top scorers on SSDF. The rating difference wouldn't be that huge... But I doubt that it would make a rating that is near to the rating it has on the SSDF list. >For example if you let Chess Tiger for Palm play Athlon 1333 top programs only, >you are probably going to get an estimated SSDF elo of 1000 for Tiger, PLUS OR >MINUS 1500 points. Elo calculation would not give such a big error bar. And it is simply unfair to compare engines with that speed difference of processors, even if the SSDF list says they have about the same rating... DID YOU UNDERSTAND ME NOW ??? > They are useless because after playing them you have collected almost no > information, mathematically speaking. I get games ! What's wrong with calculating a rating out of that ? The calculation would not show such a big error bar, even with those 9 games posted in the winboard forum. Try it with EloStat, if you want... :) But to repeat it, I hope you understand me now: THIS IS NOT SOMETHING AGAINST CHESSTIGER ON PALM ! It's just a statement that we should be very careful with comparission with such huge speed differences in CPU power... The speed difference will influent the result much more then the chess strength of the contestants. >On the other hand I agree with you on the fact that subtracting 100 elo points >to the bottom of the SSDF list was unfair, because accurate ratings had already >been established for those slow computers. well, we have at least one point, where we agree... :) Greets, Thomas
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.