Author: Howard Exner
Date: 10:27:59 08/15/98
Go up one level in this thread
On August 15, 1998 at 12:52:25, Komputer Korner wrote: >On August 12, 1998 at 23:00:58, Ed Schröder wrote: > >>>Posted by Francesco Di Tolla on August 12, 1998 at 00:49:28: >> >>>This is the point! >>>May be I'm wrong, so I would apologize from now already for starting the >>>thread, but what do you get from the followin sentence from the commentary >>> (by Jeroen Noomen) on game 7 of the match Rebel-Anand after 1.d4 d5 2.c4 Nc6: >> >>>"I chose this opening because I wanted an unbalanced >>>game for Rebel, relying on the surprise value. Playing a Queen's Indian >>>or a Queen's gambit instead, was probably exactly Anand was hoping for. >>>He knows too much about this opening, so lets play unorthodox!" >> >>>and at move 3 >> >>>"A small success: Anand avoids the main theoretical lines, starting with >>>3 cxd5, 3 Nc3 or 3 Nf3." >> >>>Am I misunderstanding it? >> >>Hi Franz, >> >>Yes you misunderstand, let me explain.... >> >>In Man versus Machine games (like Rebel-Anand or the former Aegon >>tournament) it is *not* allowed that the operator may influence the >>game in any way. >> >>What Jeroen is talking about (and what is allowed) that you can load >>another "opening book" BEFORE the game starts. Of course it is >>not allowed to switch opening books during the game by the operator. >> >>An example, say you want to open with 1.e4 because you have lost >>the previous game with 1.d4 then simply load another opening book >>which has 1.e4 set to active and the rest of the moves on "non-active". >> >>In this way in his preparation for the match Jeroen had made 5 opening >>books for the Rebel white repertoire and 5 for the black repertoire. >> >>A good example was the Tschigorin defence in game-7 (1.d4 d5 2.c4 Nc6) >> >>In this opening book (loaded before Anand played 1.d4) the only reply to >>1.d4 for black is 1..d5 just hoping Anand would play 2.c4 and so he did. >>Then the only active move was 2..Nc6 and so we had the Tschigorin >>defence which we considered good for Rebel. >> >>So that's the way it goes in human versus computer events to escape >>from human preparations against Rebel. Since Rebel's opening book is >>generally available Rebel is an easy victim for opening preparation. >> >>To participate in human versus computer events you simply must have >>*another* opening repertoire if you want to have a good result. >> >>Hope this explains a little, feel free to ask if something is unclear yet. >> >>- Ed - >> >>>regards >>>Franz > >There is another issue here that I never tire of pointing out. A computer >program prepares it's opening by having it's human opening coder desperately >trying to stay up on the latest theory. Jeroen does a great job on this in Ed's >opinion. We won't argue with this opinion. Okay so far. However a super GM like >Anand prepares openings by investigating far beyond the latest opening theory. >There is a BIG difference. I don't know how computers will ever be able to get >rid of this disadvantage. Yes I agree this is a disadvantage. Would commercial programmers ever consider giving their customers updated opening books prior to the release of of the next program version? I realize this would not address the problemm of originating opening novelties as GM's do but it would be usefull as a quick update to the opening book. ie: to quickly incorporate the findings of others as well as ones own discoveries. Another related theme is the drawback that computers do not know there opponents and so cannot choose an opening that is suited towards gaining some advantage. As in avoiding the black side of a Ruy Lopez against GM Short for example. On the plus side for computers is their encyclopeadic recall of a wide range of openings that go quite deep. Another plus is that the odds of a computer following a specific prepared line designed to trap it must be small.(is the stonewall still an exception here?). One last point is that when forced out of book, computers are no longer getting thrashed (as you pointed out in GK vs DB). The GM may be giving up more than they are receiving in this strategy of renouncing the very openings that have provided them with known good positions in the past.
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.