Many observers consider that Kasparov's accusation of cheating was the reaction of a sore loser, but we can't dismiss the charges too easily. The former World Champion was there on the scene and knew how to conduct a chess game better than any other person on the planet.

Let's look at the public record of the cheating charges.

Played by a computer?

There is one well known position where Kasparov was certain that no computer could possibly select Deep Blue's move. It's from game two and is shown in the following diagram. Kasparov, who is in serious trouble, expected 37.Qb6. In 'Game Over' his side claimed that White wins a Pawn or two and that no machine could resist snatching the material.

Deep Blue - Kasparov
New York 1997 (game 2)

After 36...a6-b5(xP)

Deep Blue continued 37.Be4. The IBM logs (see the link box again) show that Deep Blue calculated

• 37.Qb6 Rxa2 38.Rxa2 Bc7 39.Qe6+ Kh8 40.Be4 and now
  • 40...Rb8 41.Ra6 Qd8 42.d6 Bb6+
  • 40...Ra8 41.Kh2 Rb8 42.g3 Qf8
  • 40...Qf8 41.Kh1 Bd6 42.Ra6 Rd8 43.Ra7
• 37.Be4 Rcb8 38.g3 Qd8 39.Kg2 (39. Ra6 Rxa6 40. Rxa6 Bc7 41. Rxf6) 39...Rxa2 40.Rxa2 Bc7 41.Qa7 Bb6 42.Qa6 Qd7

The game continued 37.Be4 Rxa2 38.Qxa2 Qd7 39.Qa7 Rc7 40.Qb6 Rb7 41.Ra8+ Kf7 42.Qa6 Qc7 43.Qc6 Qb6+ 44.Kf1 Rb8 45.Ra6, and Kasparov resigned. The final position was later shown to be drawn by perpetual check after 45...Qe3.

Commenting on the game (see link box again) GM Yasser Seirawan said, 'I considered 37.Qb6! Rxa2 38.Rxa2 Bc7 39.Qe6+ Qxe6 40.dxe6 as leading to a winning ending by force.'

Seirawan flagged other unusual moves played by Deep Blue:-

• '23.Rec1! This move was the most revealing of the game. It is a move of extraordinarily refined sophistication.'
• '24.Ra3! Another excellent move by Deep Blue. Although it is an accepted stratagem to try and grab an open file by doubling Rooks, it is impressive how Deep Blue has used so many of the well worn positional motifs of this particular opening. An opening that Deep Blue’s programmers could not have reasonably expected.'
• '26.f4! Again, simply superb.'
• '28.Qf1?! Funnily enough this inferior move looks very much like a human misstep! '
• '31.R3a2!? An awkward move to explain'
• '33.Nf5! A powerful and somewhat paradoxical move.'

The comments ('a move of extraordinarily refined sophistication', 'this inferior move looks very much like a human misstep', 'an awkward move to explain') are Seirawan's. These moves are worth further investigation in the IBM logs.

Why no rematch?

Here we have little real information. Everything is open to speculation. The IBM announcement of September 1997 (link box again) was terse and worded awkwardly: 'Our scientists said they really did want to move on to other grand challenges and the company wants to'.

Supporters of the conspiracy theory claimed that IBM gained huge advantages from having won the match: millions of dollars of free publicity and an unusual jump in the stock price (though only a fraction of the 15% claimed in the film). Given these advantages, any level-headed business person (and IBM has thousands of employees fitting this description) would have agreed to a rematch. Even a loss to Kasparov would have been worth millions of dollars of publicity.

We would have asked IBM senior management two questions. Who? made the decisions -- (1) to deny a rematch? and (2) to retire Deep Blue? -- Why? It is easy to imagine scenarios other than a conspiracy theory that would lead to these same decisions. Only IBM insiders know for sure.

Despite having access to IBM's facilities and scientists, the crew behind 'Game Over' gives no sign that they asked or tried to answer these key questions. They could have also asked why it took IBM three years to release Deep Blue's logs. Their failure to ask the right questions means that everything is still open to speculation.