Author: Reinhard Scharnagl
Date: 05:28:00 09/13/04
Go up one level in this thread
On September 13, 2004 at 06:11:28, Uri Blass wrote: >On September 13, 2004 at 05:53:39, Reinhard Scharnagl wrote: > >>On September 13, 2004 at 05:46:08, Kurt Utzinger wrote: >> >>> Hi Eduard >>> You are right. All the computer matches played so far >>> vs Kasparov, Kramnik and some others are statistically >>> spoken rather meaningless. The best thing to find out >>> real strength of computer programs would be to let >>> participate them at mixed tournaments where programs >>> are not allowed to change parameters and opening book >>> lines every day and where they will have to fight many >>> different opponents. >>> Kurt >> >>Hi Kurt, >> >>that would be nearly the right way. But because of the limited brain of >>a human being and the rule not to be allowed to use any scripts or book >>material during a game, the computer programs should also be limited in >>its usage of persistent data. I have written on that and often mentioned. >> >>Regards, Reinhard. > >I remember your proposal > >I think that it is better to have also tournaments when you limit the size of >the data and the hardware but do not prevent programmers to change the program >during the tournament. > >There should be basically 4 types of torunaments of computers against humans so >everybody is going to be happy: > >1)Tournaments with no limitations about hardware and no limitation about >changing software. > >2)Tournaments with no limitation about hardware but programmers are not allowed >to change the software during the tournament. > >3)Tournaments with limitations about hardware(including size of the program) but >no limitation about changing the software during the tournament. > >4)Tournaments with limitations about hardware(including size of the program) >when programmers are not allowed to change the software during the tournament. > >Uri Hi Uri, I do not at all want to shrink possible kinds of tournaments. But the question has been, when tournament results should be ELO relevant and then being compared to human ELOs. Reinhard.
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.