I recently caught up with the old documentary Game Over: Kasparov and The Machine, about the famous match where the IBM chess super-computer Deep Blue defeated the world champion Gary Kasparov. I mainly watched this because I enjoyed following the World Championship Candidates Tournament recently, the reporting of which featured copious comparison of the players moves against modern chess engines. The documentary can be watched for free on YouTube here.

This documentary was not nearly as bad as I was led to believe. Its technique, although clumsy at times, was basically sound – use of flash-back footage was fairly good, pacing was fairly good, the set-up was not too bad, the interviews were well-conceived and edited,  the music was JFK-like but serviceable, and Kasparov is a pretty fun guy to watch and listen to, as he is kind of a character. The central argument of the documentary is that IBM cheated in game 2. In a complicated mid-game position, Kasparov tried to trick the machine by making a bad move on purpose that offered the machine a pawn to take for free. Deep Blue responded instead with a fairly sophisticated positional move that wound up dooming Kasparov to lose the game. He simply could not believe any machine was capable of making a move like that, and insisted that some hidden, human chess player must have intervened at that moment. The question of why IBM would cheat (one might honestly wonder, after all,) is answered by reference to the PR and revenue windfall reaped by IBM after emerging victorious from the match. Kasparov further claims that Deep Blue played differently at that moment than at any other moment in any other game, and he feels that he went on to lose the match (without the necessity of further human intervention by IBM) because he could not recover psychologically from the “obvious” instance of cheating in that early game. It did not help matters that IBM kept Deep Blue in a separate, hidden room, they refused to release the game logs when Kasparov requested them after game 2 (they did release them later,) and they forever dismantled Deep Blue immediately after the match.

Of course now this all seems pretty silly, with modern PC-based engines like Rybka and Houdini playing so far above the top grand-masters it’s mind-boggling, and which when tasked with reviewing Deep Blue’s actual moves against Kasparov show how comparatively weak Deep Blue would be now as a player. Clearly Kasparov was wrong – computers are capable of making moves like that. The only question is were they then? But there has to be a first time for everything – maybe that was simply the first time a chess engine was powerful enough to prioritize long-term considerations over falling for a simple, obvious trap like Kasparov’s.

Facility at the game of chess has always been more or less equated with superior “intelligence.” My feeling about Kasparov’s defeat is that it showed what a limited, perhaps even neurotic, intellectual endeavor chess actually is. It’s a very dreary, machine-like kind of skill, one in which machines have now totally eclipsed us. The recent Candidates Tournament made this abundantly clear. There was something strangely post-modern and science-fiction-like about the constant comparison of the moves of these ultra elite chess players to a PC program that would have absolutely wiped the floor with any of these guys, and which discovered moves and strategies so profound no human would ever see them in a million years. All the drama and interest from that tournament came from human frailty, capriciousness, and happenstance. Almost all serious errors leading to interesting outcomes (that is, losses) occurred while people were rushing as a result of completely arbitrary time rules. That guy “Chucky” was clearly unstable, so from one game to the next you never knew if he was going to play brilliantly or self-destruct, and in the end his emotional issues had a staggering effect on the tournament’s outcome! The tournament even brought out how much we all miss the Cold War, only to disappoint us when in the end the Russians clearly weren’t conspiring against the Westerner Carlsen (or at least not all of them were!)

Before Deep Blue, we could pretend that chess masters were doing something really deep and significant in their play, something machines could never do, much the way machines could never paint the Sistine Chapel or compose a Beethoven string quartet. I think IBM did a real service to society in dispelling this notion, and in the end I felt kind of bad for the IBM tech nerds who invented Deep Blue, that they never really got to enjoy their technical achievement because of Kasparov’s very public and paranoid accusations and denunciations. I guess rich IBM stock holders got to enjoy their achievement, however!