Author: Terry McCracken
Date: 21:07:49 12/30/02
Go up one level in this thread
On December 30, 2002 at 04:17:27, Uri Blass wrote: >On December 30, 2002 at 04:03:58, Terry McCracken wrote: > >>On December 30, 2002 at 00:18:30, robert flesher wrote: >> >>>To denude a program of it's opening book and play fischer random ,shuffle, or >>>whatever... will serve only to accomplish that we no long have chess. By this i >>>mean that opening preparation plays a very critical role and engines are even >>>tuned to understand specific chess positions. Take away Kasparov's opening >>>knowledge and he would never have become World Champion. FACT! is Kasparov still >>>very good at position judgment??....yes.....is he still good at >>>tactic's??....yes. Think of the last time you were on a chess server and lost to >>>someone who was considerably lower rated than you....OK tell me you have never >>>lost because someone nailed you in your favourite opening because they received >>>the lastest informant or NCO,MCO,ECO before you(and they did homework :). In the >>>end the opening is as important and the middlegame, or endgame. IMHO ...Cheers~ >> >>If you want to know _true_ engine power, you need to turn off the books, >>thinking on opponents time etc. >> >>Alan Tomalty (Computer Chess Expert of Komputer Korner) presented this idea back >>in 1987! >> >>He suggested that "thinking" on opponants time to be disabled also. >> >>The problem I see is, that these games need to be operated manually, due to the >>absence of books, to avoid repeated positions. >> >>The SSDF doesn't have the luxury of testing engines this way, too time >>consuming. >> >>The idea certainly has merit, but it boils down to how much time you lose in >>such testing, as it can't be done by auto-play. >> >>Yes the opening is a vital aspect of chess, and chess can't be complete without >>it, but this kind of testing is to rate engine strength in the middlegame, >>mostly and endgame, without EGTB's. >> >>Terry > >endgame without tablebases? > >I think that it is ridicilious when some authors simply removed knowledge >after adding tablebases. > >I read that nimzo cannot mate in KQ vs K positions without tablebases(these >problem is new and nimzo could mate before the programmer added tablebases). > >Uri Well that's a problem now isn't it? I've seen programmes checkmate 14 years ago withe BN + K without EGTBS, let alone KQ + K!! That is preposterous that a programme can't find mate with QK + K without EGTBS! Terry
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.