Talk:Game Over: Kasparov and the Machine

Strong PV section removed

I removed a "summary of the film" section that had been added by 72.137.244.109 since it was strong POV. There is nothing wrong with enlarging the description of the film, but it must be done in a way that makes it clear that what is being described are the opinions expressed in the film and not facts. (See NPOV) E.g. it is wrong to say that the chess community suspected foul play on the part of IBM as that is not true (of the chess community in general). Any summary needs to be expressed in terms like: "The film expressed the opinion that ...". And it needs to be said that these opinions are highly controversial - they are not generally agreed upon. And I haven't seen the film so I don't know whether the film says this or not, but it is definitely factually untrue that IBM did not release the game logs. They were made available to Kasparov right after the match and were put up on IBM's web site soon after. See the other wikipedia articles on Deep Blue. - Hayne 15:08, 20 March 2006 (UTC)

I think there is some miscommunication here: Kasparov asked for the release of the game logs, in which he meant a summary of everything that Deep Blue was thinking during each move. IBM released the moves that it actually made, not the moves it was thinking of making (which was what Kasparov actually wanted, because he suspected that they were cheating!) Sasha 05:47, 30 July 2006 (UTC)

No - I don't think there is any miscommunication here. The logs that were released right after the match ended and which are still available on IMB's web site [1] are the full info on what Deep Blue was "thinking". The logs show the evaluations of the "principal variation" (the move that was considered best at that stage in the "thinking" and the expected continuation). Sometimes the logs show that one move was initially considered best but then more analysis made it change to consider a different move as better. - Hayne 00:51, 1 August 2006 (UTC)

Deep Blue Controversy

Although IBM's Deep Blue officially defeated Garry Kasparov in their 1997 confrontation, there is much controversy surrounding that match. Primarily, Kasparov and all of the Chess Community felt that IBM was cheating during their match by having humans intervene in Deep Blue's thought process. It is universally accepted that a human and a computer together is much more powerful a force than a human and a computer individually.

These suspicions were made particularly clear in Game 2 of Deep Blue’s 1997 match with Kasparov. Kasparov forced Deep Blue in to a closed game, during which computers are tend to have the most trouble in. He also offered Deep Blue two pawn sacrifices, which would have given Deep Blue a short-term advantage, with possible consequences in the future. Computers tend to snatch at the free material while paying less attention to king safety. However, alarms went off when Deep Blue did not take the pawns. Instead, it performed a move that was far more representative of former World Champion Karpov. Eventually, Kasparov resigned the game, saying that it was hopeless at the end. He was genuinely suspicious that there was human intervention in that game, due to the fact that Deep Blue made too many "human" moves, and did not play like a computer.

The suspicions raised in Game 2 are not merely the work of conspiracy theorists: when Kasparov resigned the game, he had missed a simple perpetual check, which would have easily drawn the match. Kasparov was horrified when he realized this, and the Chess Community was deeply concerned: how could a computer play so well, but miss something as basic as a perpetual check? This is unfathomable to think about.

Kasparov and the rest of the Chess Community were seriously suspecting human intervention in Game 2 due to this fact. It is very possible that there was a relatively strong human (not necessarily a grandmaster) watching the screen, and Deep Blue needed permission from this person to make a move. If the human noted that Deep Blue's contemplated move would place its king in danger, the human would drop the main log, and revert to the second-best, safer move. It is effectively impossible for a computer, especially one like Deep Blue, to miss a perpetual check, like Deep Blue did in Game 2. However, it is rather plausible that a human might, especially when playing under pressure (as was Kasparov).

In fact, there is much evidence of unethical behaviour and human intervention during the match. During one of the games, Deep Blue crashed, and IBM mechanics were sent over to repair it. In all fairness, this should have resulted in a resignation by Deep Blue and a victory by Kasparov. However, being the good sport that he is, Kasparov merely chuckled and started chatting merrily with his mother, when the referee of the game shot him down and stated that he was not to communicate with anyone. "Why not?", replied Kasparov, "they're talking to their guy all the time!". He was referring to the IBM mechanics, who were, indeed, constantly modifying Deep Blue, during and in between matches.

There is much more evidence of unethical behaviour on the behalf of IBM: they had promised to release Deep Blue's main logs for all the games after the match had been completed, but they backed out on that deal. When Kasparov defeated Deep Blue by a considerable margin in 1996, he agreed to a re-match. However, when Deep Blue defeated him in the 1997 rematch, IBM refused another showdown, so Kasparov could never redeem himself. Finally, IBM was indeed modifying Deep Blue in between the matches, and playing with Kasparov’s mind in the process: Deep Blue effectively gave Game 1 to Kasparov, furthering the mental implications that Kasparov’s Game 2 defeat had on him.

Kasparov's loss in Game 2 was effectively a loss of the entire match. Chess is famous for psychological warfare between players, but one can obviously not conduct such warfare against a computer. Kasparov collapsed psychologically: Games 3, 4 and 5 were drawn partially due to the fact that Kasparov could not recover after Game 2: he was constantly thinking "something is up here", and he could not quite place his finger on it. IBM furthered Kasparov's suspicious and deliberately fed his paranoia: they were playing with his mind. In the end, Kasparov simply wanted to end the match: he gave Deep Blue the 6th and final game, without even putting up a fight.

Furthermore, chess matches are seldom executed without both players conducting extensive research on their opponent, including their strongest / weakest openings, play style, etc. Deep Blue had a myriad of information on Kasparov (of course: it was built for the sole purpose of defeating him), but Kasparov was not given any information whatsoever about Deep Blue before the match. This would be incredible barrier facing players at the Grandmaster Level: one can not expect a fair match when one player knows everything about his opponent, while the other knows nothing at all.

However, any Chess player could tell that Kasparov was a stronger player than Deep Blue: Kasparov played an emotional game against something with no emotions, and the vast controversy surrounding the matches, as well as IBM's unethical toying with Kasparov's mind, had detrimental effects on Kasparov's game. Perhaps this is why IBM refused another rematch to serve as a tiebreaker: now that Kasparov knew what he was up against, a greedy corporation playing a death-match against him rather than the friendly experiment they had originally claimed that it was, he may have been able to sufficiently prepare himself once he was prepared for all the psychological tricks and had reviewed its past games. In the end, it was not Deep Blue that defeated Kasparov: Kasparov was defeated by his own emotions in the face of unethical play. Sasha 05:51, 30 July 2006 (UTC)

I feel compelled to respond to some of the comments by user "Sasha" above.

Sasha said "It is effectively impossible for a computer, especially one like Deep Blue, to miss a perpetual check" but this betrays an ignorance of how computer chess programs work. They calculate move by move and always have a limit (usually referred to as a "horizon") beyond which they can't "see" what is going to happen. In the case of checks and other forcing moves, the limit is usually extended to be larger number of moves than usual but there is still a limit. if the situation is sufficiently complex, then computer chess programs often do not "see" perpetual checks until they get closer.

Sasha said "IBM mechanics, who were, indeed, constantly modifying Deep Blue, during and in between matches" but in fact there has been no evidence whatsoever that "mechanics" or programmers modified Deep Blue while a game was in progress. Modifications of the program parameters between games was not disallowed by the rules of the match and was done. But note that this is analogous to Kasparov reading a chess book inbetween games and being influenced to change his gameplay because of something in the book - nothing unusual or wrong with that!

Sasha said "they backed out on that deal" but as I have pointed out above, IBM did in fact release the full logs as promised and they are still available on the IBM web site. - Hayne 01:07, 1 August 2006 (UTC)

• To expand on above, the perpetual check theory has been disproven by modern chess engines such as Rybka. They show white has a way of stopping the perpetual check and coming out better and in all likelihood deep blue was advanced enough to see the variation since it chose the line. Since Kasparov resigned, the world may never know. You can see an illustrated video of what i'm referring to in Kingscrusher(on youtube)'s video here http://www.youtube.com/watch?v=FvA6zKPUwqE .

Removal of "needs infobox" tag

This article has had its infobox tag removed by a cleanup using AWB. Any concerns please leave me a message at my talk page. RWardy 19:55, 10 September 2007 (UTC)