Kasparov Vs Deep Blue

Are you serious? Am I actually supposed to reply to this nonsense?

If you are not going to contribute to the discussion, then do us all a favor and keep your bad jokes to yourself.

To those who are interested (OODA_Loop, at least): I will be in possession of the book I have been citing in a couple weeks. It is in storage. When I have dug it out, I will present the information I have been trumping up during this discussion and I would be happy to talk about it.

Okay. The summary of the information gleaned from "A New Era..." by Michaerl Khodarkovsky and Leonid Shamkovich:

Kasparov won the first game of the 1997 rematch after his 45th move. In game one, the machine played "the way we were accustomed to see a computer play - materialistic and tactical..." (p. 199)

In game two, however, "material was not important to the machine, and it played with a strategic and positional competence achieved by very few grandmasters." (p. 199) After the game, Kasparov's team entered a particular position into their computers that stemmed from the move 37.Be4. It was expected that Deep Blue would play 37.Qb6, which would win at least two and maybe three pawns - the kind of thing a computer never passes up (p. 198). Though their computer programs (Fritz 4 and Hiarcs) searched for 2 days and 2 nights straight, neither engine ever found 37.Be4 (p.198-99). In addition, the computer gave Kasparov a chance to draw by perpetual check on its 45th move, but the champ missed it and resigned instead. After being alerted to this fact, and with suspicions already stemming from the drastic change in style of play, "Garry requested explanatory printouts for three moves, 35.Bxd6, 37.Be4, and 45.Ra6." (p. 202) The IBM team agreed to furnish the printouts To Ken Thompson, a computer chess expert who held no personal stake in the match results.

According to "A New Era," the printouts were never furnished (p. 203). Kasparov addressed the audience and the IBM team at the post game press conference for game three (a draw), and essentially called his opponents out, since they went back on their word to furnish the printouts (p.209). The head of the IBM team, CJ Tan, said that "after the rematch they would show Garry printouts of all the games." (p. 209)

As you probably guessed already, this promise never even came close to being delivered. When a printout was finally delivered to Ken Thompson, it was for the move NOT played by the machine, 37.Qb6 (p. 211). The printout consisted of all the logical reasoning that the computer used in order to 'know better' than to play 37.Qb6, yet there was no information at all about the move actually played: 37.Be4.

For games five and six, the final two games of the match, Kasparov requested through the Appeals Board that full printouts of both games be sealed and locked away immediately after their completion (p. 217). CJ Tan said this time that it was not possible to do so (p. 217). After game five was over, however, Tan suddenly changed his tune again and did produce a printout (p.224). The obvious implication is that he needed to ensure that the printout would be 'clean' before turning it over, and this of course is why he would need to wait until after the conclusion of the match to furnish all printouts. Conceivably, he would have the chance to doctor them before turning them over if he chose to do so.

Still no requested printouts had been produced from game two - the ones that were requested. At the time of the book's printing, no printouts had ever been received (p. 232).

By the time game six had come around, Kasparov had lost his fighting spirit (p. 225), and what the audience was witnessing was, "...a very tired man pitted against relentless electronic calculation aided by psychological warfare." (p. 228)

Notably, "To his credit, Garry accepted all the responsibility for his loss." (p. 232)

After Deep Blue had won the rematch, Kasparov (while on national TV) challenged the Deep Blue team to a third ten-game rematch (p. 233). He had several well-intended regulations that IBM refused to agree to, and the third match never happened. One request was that the rematch should be organized by an independent party with no stake in the results - someone other than himself or IBM (p. 233). Another was that he be able to study ten printouts of Deep Blue's past games, as would be available in a match between two human players (p. 233). What is wrong with these simple requests, I don't know, and no explanation was given as to why they were refused.

So the stink of it is that the computer played like a human GM in game two, including the missed perpetual check variation at the end of the game, and didn't do so at any other time. Then, mysteriously, no proof could be provided of the computer's thought process for any of the moves it made in game two - only a move it decided against playing.

Makes me wonder... If it doesn't do the same for other people out there, then I guess I'm out of touch. It just seems to me that the IBM team should have no problem turning over any and all printouts to anyone who requests them and that there is no need for the cat-and-mouse B.S. that took place here unless they have something to hide.

My ego has nothing to do with the topic. I have no vested interest in Kasparov's dominance or lack of dominance over a computer chess playing program.

In my "false claim" that "any car can beat a racehorse," I did include in my post the very important caveat that it would have to be "under the right conditions," which you convieniently left out of your quote when you got back to me here. Hmm...

Anyway, the "nonsense" I spoke of was not just the reference to 19th century cars vs. racehorses. Cars during the 19th century were certainly not capable of going 40+ mph. But I don't think the person who originally posted that remark was referencing 19th century cars... If they were, they failed to point out that very important detail.

Moreover, the real "nonsense" I spoke of was your comment regarding Deep Blue's lack of hands. Do you really think that since "the laws require the player to move the pieces with 'one hand'," Deep Blue should have lost by forefit because it failed to comply with those laws??? Come on... That's absurd. There are separate laws for computer vs. human games, as most chess afficionados already know.

By the way, if you have any comments regarding the real meaning of this discussion, rather than critical comments about one of its contributors, then please share them.

A.) Nobody, including Garry himself, ever said Kasparov was way better than all the other GMs in the world. If you have made that assumption, then that is your opinion. It may be shared by others, but that is not the focus of this discussion.

B.) Even if he is better on an individual basis, that means nothing when other GMs are allowed to collaborate (rather than just go one-on-one against him) with each other, and with a machine that can calculate two hundred million moves per second.

And the reason that it would be an issue if the GMs suggested moves for 'the thing' is that THEY WERE NOT PERMITTED TO! It is a violation of the rules for the match! Simple as that. What if Kasparov wanted to move twice in a row? Would that be allowed? No, because it is not permitted by the rules of the match.

Wow. Seriously?

I remember GM Ian Rogers saying that IBM later released all the computer's evaluations for each move. Dunno if its true or not...

Never heard this, but it may be true.

Either way, I've never seen any real explanation for all the runaround B.S. that went on during the match.

Really? 2775 compared to 2750 is "way better" to you???

Wow. Standards have gotten lower.

With a rating difference of only 25 points, even at the lowest levels of all competition, the "superior" player can only be expected to win about 52% of games. That's "way" better?!? Seems like a very slight difference to me.

Also, a bunch of players "weaker than yourself" does not even come close to being the equivalent of a team of GM's, which IBM hired for the match. Let alone the fact that they had a super-machine to crunch any variations they wanted it to. There is a reason why using an engine on chess.com gets you banned. I would think that the same logic would be in place during a match such as this one.

All of a sudden you are a Super GM? And these "weaker" players (GM Joel Benjamin, GM Nick DeFirmian, and GM John Fedorowicz) cannot hope to measure up to you because they are so weak?

Please...

And I do firmly believe that Kasparov, had he won, would still have pointed out the indiscretions of the IBM team - as he did immediately after game 2, game 3, and game 4 at the post-game press conferences, before the results of the match had even been decided.

Church!

Have there been any other chess computers