Author: Dave Gomboc
Date: 09:01:01 06/02/99
Go up one level in this thread
On June 02, 1999 at 06:26:29, blass uri wrote: > >On June 02, 1999 at 04:53:34, Dave Gomboc wrote: > >>On June 02, 1999 at 04:38:38, Abhishek Roy wrote: >> >>>How the heck does a mass market program get up there at no.2 breathing down >>>the neck of Fritz 5.32 and above Hiarcs 7. That too, before the much awaited >>>new King engine (3.something if I am not mistaken) is released. >>>Does this mean that despite all the utterly useless bells and whistles in >>>CM6000 it still has the strongest engine? >>>Hard to believe. >> >>My guess is that the lack of ability to autoplay it means that competing >>programs aren't tuned against it as well as they are against each other. >> >>Dave > >I believe that most programs(if not all of them) are not tuned against other >programs. > >If programs were tuned against other programs then I could see many cases when >program has a winning advantage out of opening book but it is rare. > >I also do not see results of >=90% against programs like Fritz3(pentium90) or >Rebel8(pentium90) and I believe that it is easy to get >=90% >You need only to autoplay against them and >remember the games that you won to be part of the opening book. > >The fact that old Fritz5 had stupid repetition lines in the opening book is also >an evidence that Fritz's programmers did not autoplay because if they did it >they could see these lines and delete them from the opening book of Fritz5. > >I do not understand the reason for these short draws by repetition against >weaker opponents and it is clearly a bug in Fritz5. > >Uri I didn't mean book tuning. I meant engine tuning. The people make some changes, see how well it works on some test suite, if that is okay, do some autoplaying... but the autoplaying is never done against CM6K. Dave
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.