Author: Mogens Larsen
Date: 05:10:06 05/07/00
Go up one level in this thread
On May 07, 2000 at 07:33:35, blass uri wrote: >I do not think that the fact that she found problems increase the uncertainty. >If nobody knows about problems it does not reduce the problems. Will the appearence of faulty games reduce the credibility of the test? Of course it will, especially since it could have been avoided and because not all uncertainty can be traced in the games. A minor thing could easily be an advantage to one of the programs. The attention to detail was appalling. >The fact that she searched to see if there are problems reduce the uncertainty. No, because there's no reason to assume that all problems have been discovered or that the testing procedure has been corrected, and it suggests that there's something inherently wrong with the test. The errors and factors of uncertainty were piling up, while the tester refused to engage in anything resembling a fruitful discussion. You're not afraid of discussions, which is why I bother explaining my opinions. >She was convinced that Fritz will win after the matches and not before the >matches. Maybe you ought to look at all the posts again. I could be wrong about it, but I doubt that very much. The term "as I expected" should be possible to find. >I agree but the result in the ssdf method(2 computers with autoplayer) and the >results in one computer(ponder off) give similiar results. That doesn't matter, since they're not comparable due to different setup. >In the 4 games that I found in the ssdf games the strain was significant for all >the games. > >There was a 5th game when Junior was slowed down only for the first move out of >book by a factor of 2 or 3 and in this case the tester did not want to repeat >the game. That sounds like problems, but I'm unable to comment on it, because I can't check it myself or have access to such proof. I hope you understand. >I do not think the method of the ssdf is better. >They have their own problems. > >For example I found that Fritz5 won the same game against Rebel8 5 times because >Rebel has no learning function. > >Fritz5 could have better rating if it played more games against Rebel8 so the >rating of programs is dependent on the number of games between 2 programs and >not only on the strength of the engines. > >If they decide to let Fritz5(p200) to play 10000 games against Rebel8(p90) they >can push Fritz5(p200) to be number 1 above all the programs with better >hardware. I said that the SSDF test is the best available, not that it was perfect. The learning function is a problem, no doubt about it, but it's a part of the program. Would Kasparov play the same opening again and again against Kramnik if he knew that it would be a success? I think that he would, however unlikely the scenario might be. The SSDF don't decide to play 10000 such games, so the problem isn't big, but they're more capable of responding to that. I think I've expressed my opinions concerning this issue sufficiently in this forum and I'm starting to repeat myself. If you have further questions regarding this subject, then you know my e-mail address. Thank you for exposing your views. This and related threads could have been a lot shorter if everyone involved were equally candid. Sincerely, Mogens Chr. Larsen http://home1.stofanet.dk/Moq/ "If virtue can't be mine alone, at least my faults can be my own."
This page took 0.02 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.