CHESS

THE GAME in this space last week, between Mr Garry Kasparov and a computer, evoked vivid memories for me of a splendid New Year's Eve spent in the company of my good friend Alan Turing as 1945 was turning into 1946.

Our work at Bletchley, which as all now know played the crucial role in defeating the Hun, was now over, and Turing had invited me over to ask my advice about what he should do next.

"Have you thought," I asked him, "about building a machine that could play a mighty fine game of chess?"

"But that's impossible," he blurted out, "unless...". He sank into deep thought. When he continued, his voice was breaking with excitement. "You mean an algorithmic process to play the perfect chess game. I've always said it, Colonel, you're a genius." (He always called me Colonel at times of great intimacy.) Then he rushed to his room and I didn't see him for eight days.

When he emerged, he was carrying a heavy contraption of wheels and levers and was shouting "Forward and starboard. Forward and starboard. It's perfect! Watch this."

He put the machine down and clicked a chess board on to its lid. The gears whirred and the following moves appeared as if by magic on the board:

1.h4 h5 2.g4 hxg4 3.h5 Rxh5 4.Rxh5

"I've tried other algorithms, but this is the one!" he exclaimed. "Each side moves to the furthest forward square as possible, and as far towards White's right-hand side. And you always capture if you can."

4...g3 5.fxg3 g5 6.Rxg5 f5 7.Rxg8

The machine thought a bit about that one, evidently preferring it to Rxf5 because g8 is further forward.

7...f4 8.Rxf8+ Kxf8 9.gxf4 e5 10.fxe5

"Do you see?" screamed Turing. "Now as far forward as possible and..."

10...Qh4 mate (see diagram).

"... the perfect game is over."

As far as I know, that was the last time Turing ever touched a chessman. Had he done so, I doubt that Kasparov would have overcome his artificial intelligence.