Author: Uri Blass
Date: 10:57:46 05/02/03
Go up one level in this thread
On May 02, 2003 at 13:01:01, Drexel,Michael wrote: >On May 02, 2003 at 12:02:29, Uri Blass wrote: > >>On May 02, 2003 at 11:33:39, Tony Hedlund wrote: >> >>>On May 02, 2003 at 09:47:43, Uri Blass wrote: >>> >>>>On May 02, 2003 at 09:44:37, Mogens Larsen wrote: >>>> >>>>>On May 02, 2003 at 09:08:08, Uri Blass wrote: >>>>> >>>>>>Does Ruffian has some other kind of learning that is not book learning to >>>>>>prevent losing another game in the same way? >>>>> >>>>>No, I'm quite sure it doesn't. >>>>> >>>>>Regards, >>>>>Mogens >>>> >>>>In this case then I see no reason for the ssdf to test it. >>>>I do not see what is interesting in seeing a program losing the same game in the >>>>same way. >>> >>>I've never seen Ruffian doing that. >>> >>>>The ssdf should check if a program can lose the same game twice and refuse to >>>>test programs that can do it. >>> >>>Such a program must be so weak that it's never considered to be tested. >>> >>>Tony >> >>I do not think that the problem is of playing strength. >>I understood from Mogens's post that ruffian can lose the same game >>twice even if it does not happen often. >> >>Uri > >What do you think would happen if Leo in his tournaments would reject >all programs which can lose the same game twice because this could change the >results? > >Michael I think that it is his tournament and it is his right to do it. This is also a different case because Leo does not have limited number of programs to test because he does not play hundreds of games with every program. The ssdf has limited time and it is better if they use it for programs that learn. The point is that I do not think that the rating of programs that do not learn is very interesting. Uri Uri Leo does not give rating to programs
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.