Author: Shaun Graham
Date: 06:12:02 07/15/98
Go up one level in this thread
On July 14, 1998 at 22:57:29, Robert Hyatt wrote: >On July 14, 1998 at 00:28:52, Shaun Graham wrote: > >>On July 13, >>> >>> >>>based on experience I don't think this is true. The best chance comes early, >>>not later, because Fritz has serious holes in it, as does every program I can >>>think of, and GM's are going to find 'em, and drive trucks through 'em. >> >>First of all human players have their own holes to deal with. I'm not going to >>argue against that if GM's play anti-computer chess that they will win more >>games against a computer of course they will. If you are playing a competitive >>game against a computer program and use anti-computer techniques, the computer >>WILL NOT perform like a GM, because they can't adapt and punish these methods as >>a human GM would. However if you are training with a computer and play your >>normal game(regular GM chess) against the computer just as if it were an unknown >>grandmaster, then it will perform like a Grandmaster. So the question is, (1) >>is Fritz a GM strength against typical non anti-computer chess? I would say >>certainly. (2) Is it a GM against non typical (anti computer) chess? I would >>say definitely not,for the reason that the programs are currently unable to >>adapt to such situations. > > > >Please take a deep breath, and reflect on what you are saying for a minute, >and you'll realize how flawed your logic is. You are saying that when a human >plays "anti-computer" he can beat the computer but when he doesn't he won't. You take a deep breath, and reflect on what you are claiming i said. I never said that a human not playing ant-computer chess wouldn't win, but indeed, the likelyhood is considerably reduced depending on the player(which is most players) > >When playing humans in chess, in my first chess-club, I learned who liked to >play wildly and was a good calculator, and who liked to play quietly and get >into a complex endgame. And in each case, I did my very best to give my >opponent exactly what he didn't want. In a swiss system you don't have games and games of experience to decide this when you don't know the opponent. Thus this analogy is inaplicable. > >The point? If I have to play fritz, I'd probably get beat by it. But I >wouldn't get beat in a wild tactical game, because I would at least exert enough >control in the game to avoid such. In a swiss system you don't often know your opponent, you are going to play what you think is best based on the experience you have. In this case of testing you don't have experience of fritz, you are playing the player you got paired with. >>> >>>And giving them more tournaments to find these openings is suicidal. So >>>the key questions to me are: (1) is Fritz the best commercial program? >>>answer seems to be yes, if you are talking about playing computer vs >>>computer matches. It's proven that on the SSDF it seems. (2) Is fritz or >>>any other program a GM? not even close yet. Good chess players find and >>>exploit holes in every program running. I think that many programs might >>>fit right in the middle of the IM class, although they will not be nearly >>>as consistent as IM players, because they will beat an IM on tactics, then >>>get crushed in an elementary ending. A PC GM is a ways away, IMHO, still. >>> >>>Getting closer, but not here. *yet*. But given enough time... >> >> >>Regardless of it as actually DONE OR NOT! the point is that if Fritz were placed >> in a tournament setting even just a single tournament,(masquerading as a human) >>would it be able to perform at a level equal to that of other GMs in the >>tournament. I would say most definitely yes. If it performed at a level equal >>to that of GMs then it can be said it performed at GM strength. The result is, >>that when you or i play a game against the program using normal play, will it >>take advantage of our mistakes to the degree of a GM even a low rated GM. >>Arthur Bisguier is old he is a GM, he is 2300+ USCF, Fritz would destroy him in >>a match probably even if he did know it was fritz he was playing. So the title >>is absolutely inconsequential. The ultimate question is ELO rating. If Fritz is >>2450-2500 ELO that's all that's of any importance. > > >I don't think so. I believe that after 2 rounds, everyone would "know". I >can pretty well identify a computer after looking at two games using a computer >to help me. GM's can do it too. And once they find out, it's all over, all you Ridiculous, in a european swiss tournament with 50+ games going on, you don't actually see every game that is played, and a guy sitting with 2 points after two rounds is not going to be thought to be using a computer. >get from then on is "anti-computer".
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.