Author: Uri Blass
Date: 03:16:40 10/06/01
Go up one level in this thread
On October 06, 2001 at 05:37:22, Thorsten Czub wrote: >On October 05, 2001 at 13:14:54, Bruce Moreland wrote: >>>I am interested in writing a program which uses this "human" approach.If you >>>want to join me,you are welcome. >>>Regards, >>>Ashwin. >> >>How is it going to work? >> >>bruce > >in the movie A.I. , first minutes, they make fun about chess programmers who >thought it would be AI to teach a chess program chess. the makers of AI are IMO >right. what chess programs of today do is AS (artificial stupidity). > >Even the mobility function of touring was more modern than todays chess >programs. > >one major reason is IMO the material values. > >they are a kind of plague. >the rook gets 5 points, no matter if it can move or not. It is not truth. I know that the top chess programs use mobility in their evaluation. If the evaluation is not high enough is another question but they use mobility. >its there although it is NOT there. > >programs often play as if it would be enough to add the material values and feel >confident. Not truth. programs have often positional scores that are bigger than a pawn even in the middle game. Junior has even bigger positional scores than most of the other programs and it even plays sometime bad sacrifices because of it. My Junior7 see positive evaluation for itself in the last game of your tournament inspite of the fact that it is losing. The scores are wrong but the problem is not positional scores that are not big enough. The opposite Junior overevaluate the black king's problem and it believes that the black king is unsafe. [D]Junior 7 - Gambit-Tiger2 6k1/pbq2p1p/1p2n1pB/1B2p2P/P2rP3/4Q3/2r2PPK/R2R4 w - - 0 1 Analysis by Junior 7: 29.Rxd4 exd4+ 30.Qg3 Qxg3+ 31.Kxg3 Bxe4 ³ (-0.62) Depth: 3 00:00:00 29.a5 Bxe4 = (-0.10) Depth: 3 00:00:00 29.f3 gxh5 = (-0.05) Depth: 3 00:00:00 1kN 29.hxg6 hxg6 30.Rxd4 exd4+ = (0.18) Depth: 3 00:00:00 1kN 29.Bd3 Rc3 30.hxg6 hxg6 ² (0.62) Depth: 3 00:00:00 2kN 29.Bd3 Rc3 30.Rac1 Rxa4 31.Rxc3 Qxc3 ² (0.55) Depth: 6 00:00:00 6kN 29.Bd3 Rb2 30.a5 bxa5 31.Rab1 Rxb1 32.Rxb1 gxh5 33.Qg3+ Kh8 = (0.17) Depth: 9 00:00:03 555kN 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.