Author: Rolf Tueschen
Date: 13:21:11 01/20/06
Go up one level in this thread
On January 20, 2006 at 15:05:50, Uri Blass wrote: >otherwise we will never be able to prove that A is stronger than B and >saying always that rybka is stronger than other programs with more than 95% >confidence is long saying that is no necessary(I will not be surprised if it is >even more than 99% confidence if you combine all the results and analyze it >statistically. > >There are are some technical problems that one program may be lucky in playing >weak opponents but these problems are small and I never saw superiority that is >so clear like rybka's superiority in tests and it is not only the CEGT tests. > >Uri You are good in stats but not so good in applying stats to CC. Of course you cannot prove that Rybka IS stronger because the others have not even started to react on Rybka. Read what I discussed with Bob. I agree with you and e.g. Simon and Lagersh. and almost everybody, that Rybka has an impressing record at its entry. But it's not possible yet to say that it's better than everybody else like we could say that Kasparov is better than the whole rest of chessplayers (except Kramnik per favor). Simply because the surgery in CC is different from the talents in human chess brains. That is what you and all oversee. So in other words, Uri, take me by these words, they are tru and will last to be true for the next 500 years - you cannot "prove" mit mathematics what is not in the "nature" of chessmachines. There is no natural development, Uri. And alone therefore your whole application of certain methematics is just wrong. - Here I am now the one who says that I cant explain it to lays. Bob knows quite well what I'm talking about. He no lay! :)
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.