Author: Uri Blass
Date: 11:03:36 11/05/00
Go up one level in this thread
On November 05, 2000 at 13:32:13, pavel wrote: >On November 05, 2000 at 13:18:03, Thorsten Czub wrote: > >>On November 05, 2000 at 12:40:17, pavel wrote: >>>1.Rhg1 Nf6 2.h3 Bxh3 3.Kb1 Nbd7 4.Bh6 Nh5 5.Qc2 Ndf6 >>> ² (0.51) Depth: 9/29 00:07:55 14671kN >>> >>>(mahamood, MyTown 06.11.2000) >>>--------------------------------------------------------------------- >> >>gandalf4.32f says: >> >>depth 11 1/41 +0.35 Rhg1 Nf6 Bh6 ... >> >>>I think there is always one "best" (or good) moves that one program finds (or >>>sees) that others dont. Doesnt make "that one" better or "untouchables" than >>>others. >> >>of course. but i do not only present ONE position and thats all. >>The beta-testers of gambit-tiger came with lots of data, games, >>matches, examples. >> >>if gambit-tiger would only "solve" this position, >>without all other results, i would say: pah - ONE position. >> >>but this is not my idea, not my point. >> >>Gandalf, CSTal, Gambit-Tiger, maybe others too, are on the right way. > >right way? tunning evaluations? >all (top) program has reached a point that you can't do much improvement on the >program; other than doing fancy tunning in eval codes, better PC (speed), >learning. > >these are (AFAIK) the only ways to make your program better at this stage of >"chess programing". >(this is my assumption, but i believe in it) There are a lot of ways to do programs better. It takes time because humans are stupid and need a lot of time to find the right ways. Programs of today are clearly better than programs of previous years and not only because of better evaluation. Programs of the next years will be better than the programs of today. Uri
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.