Author: Bob Durrett
Date: 18:51:30 10/30/02
Go up one level in this thread
On October 30, 2002 at 21:21:41, stuart taylor wrote: >On October 30, 2002 at 16:05:54, Dann Corbit wrote: > >>On October 30, 2002 at 15:31:38, stuart taylor wrote: >> >>>The very highest scorer if the ssdf had tested all the above, seems to have to >>>be one of these three Deep Fritz 7, or Chess Tiger 15, or CM9K Utzinger >>>settings. Does that make sense? And if so, that in any case, these would take >>>the top 3 places? >> >>Nobody knows. >> >>You'll get plenty of guesses, though. >> >>Just for fun, let's consider just these entries from the current SSDF list: >> Rating + - Games Won Av.opp >>1 Deep Fritz 7.0 256MB Athlon 1200 MHz 2763 34 -31 535 73% 2587 >>2 Fritz 7.0 256MB Athlon 1200 MHz 2742 30 -29 574 64% 2637 >>3 Shredder 6.0 Paderb 256MB Athlon 1200 2736 25 -24 831 66% 2623 >>5 Chess Tiger 14.0 CB 256MB Athlon 1200 2717 30 -30 557 61% 2638 >> >>A glance will show you that given the error bars, any of them could be the >>strongest. > >I can see that. The minus of the highest, and the plus of the lowest, change >them around. > >> >>Now, CT 15 or CM9K are unknown. They might be stronger than all of the above. >>They might be weaker than all of the above. They might be somewhere in the >>middle. >> >>By the time we could know an answer, the next iteration of all the products will >>be out. >> >>In any case, any of them are "strong enough" that it will make no difference >>which one of them we are playing against (as humans) for 99.9999% of us. > >It is good to have the feeling that you are getting the most educated advice >when trying to analyze a game, and if you are very hopeful that a sacrifice you >think is ingenious, should work, you would want the program most likely to see >the truth about it, as possible. >And when playing, you would rather be beaten by wisdom than by your human >frailties being exposed (which any good program can do). I mean, the greater the >program (nowdays) the more likelihood of its wisdom. And in that way, you are >slightly more likely to learn things from your lost games. >That's why I keep wanting the "best" in playing strength. Not because the others >are too easy. >S.Taylor Wouldn't it be nice if you had a real, human, top GM at your beck and call anytime day and night, but could turn him off when you were not in the mood? Wouldn't it be nice if you could decide what you would discuss together and not have to listen politely to some unsolicited and unwanted lecture? If the "human" turned out to be non-human, you might be upset for awhile but if you really couldn't tell the difference anyway, then maybe you could learn to live with it. : ) Bob D.
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.