Author: stuart taylor
Date: 08:28:56 11/14/00
Go up one level in this thread
On November 14, 2000 at 05:49:19, Thorsten Czub wrote: >On November 13, 2000 at 21:09:49, stuart taylor wrote: >>I think I understand now. > >thats always very good :-))) > >>Well! that almost makes me want to cry! (together with an emotional Gambit >>Tiger). > >?! > >>GT does not look anything special at all according to these results. > >50% against Junior6a. >So how will Junior want to beat GT in Elo-points ? > > >>You might say, "but it's not enough games to indicate very much"! >> But I can tell you, that I anyway am always upset to see a program which I >>want to get for analysis, and main use, lose even one game, esp. to a previous >>generation program, since that indicates some inadequacy somewhere, compared to >>something else which is available. > > >you are upset? why ? > >>I would prefer if they were many draws, even >>if only the occasional win. That would look more reliably better. But now, it >>looks like in many ways, the old Junior6a is stronger than GT. > >junior6a gets - if it is lucky, 50% against GT, Rebel-Tiger gets better >results against Junior. But - IMO GT is the better program. > >i don't see why you are upset. its very normal that programs play different >against different opponents. Because if I analyse a very deep position with GT, or whatever I think is the best, I will not feel sure until I set up another program and wait many hours to get the same position analysed again, just in case the first program missed something which its "fear-enemy" will not miss. I will never have a program I can feel I can trust.(as showing as much as a program can show). >you have always ONE program that is a kind of fear-enemy. If that's its worsed enemy from last and new generations, then that's ok! >thats normal. Normal, but not what I'm after. > >>S.Taylor
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.