Author: Serge Desmarais
Date: 22:12:04 09/10/98
Go up one level in this thread
On September 11, 1998 at 00:43:19, Bruce Moreland wrote: > >On September 10, 1998 at 21:21:29, Serge Desmarais wrote: > >> Your program was the Champion of the World at blitz against other >>programs/computers, but I think you were not completely satisfied about how it >>ranked in longer time tournament? Despite that it did very well, if I remember. > >In the 1995 WMCCC, with average hardware, it finished tied for third (highest >bukholz points amongst the three programs tied for third), and beat both Genius >and MChess, who finished first and second. > >In 1996 it got 6 wins and 5 draws to finish 1/2 point out of first, in clear 2nd >place, on nice hardware (people either brought 200 mhz pentium pro's or used the >supplied 133 mhz pentiums, mainly, and I brought a pentium pro 200). In this >tournament Shredder won with one loss and two draws, but the loss was to me. > >In 1997, it had wonderful hardware (600 mhz alpha overclocked to 767 mhz), but >it lost twice and drew a few times to finish tied for I think fourth, fifth on >bukholz points. > >People can say whatever they want. They can say I got lucky in 1995. They can >say that the field was poor in 1996. They can say that I should have done >better on nice hardware in 1997. > >Perhaps all of this is true. I don't know. Of course I am unsatisfied, I >didn't win any of these events. > >But I don't think that there is any evidence that my program is particularly >deficient at longer time controls. > >bruce I never said it was "deficient". I only said that your program seems to have "something" that makes it play better than the others at faster time controls. And I would not say that a program that can play at an IM level or more is "deficient". An interesting idea could be to try entering your program in an event as the KK Cup? I am not a programmer of chess programs myself, but I figure that when your program is getting performances over 2400 or 2500, it must be tough to improve it (mush tougher than to improve one rated at 1200!), because there it must be only fine-tuning. As for humans : what makes the difference between a GM rated 2600 and one rated 2700? Despite the numbers, I do think it is only tiny things. Serge Desmarais
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.