In his 12-year reign as world chess champion, Garry Kasparov has
earned a reputation for both brilliance and aggressiveness. He
is widely considered the greatest player in history. But on
Sunday afternoon, after resigning the sixth and deciding game of
his match with the IBM supercomputer known as Deep Blue,
Kasparov sat slumped and glassy-eyed as he awaited questions in
a midtown Manhattan ballroom. "He looks like a DMV photo,"
cracked international master Mike Valvo.
Kasparov's capitulation shocked everyone, coming just one hour
into a game that he needed to draw in order to tie the match.
Things had gone much differently 15 months ago, when Kasparov
defeated an earlier version of Deep Blue 4-2 in Philadelphia.
But since then IBM's computer scientists had enlisted the help
of four grandmasters, and this latest teaming of technology and
human intelligence threw Kasparov some curves. In Game 5, for
example, no one anticipated that with one of Kasparov's pawns
poised to reach the last file and become a queen, Deep Blue
would simply ignore it and launch an attack with its own king.
That stunning shift of focus set up a perpetual check and forced
Kasparov to offer a draw.
"The computer will be unbeatable in five or 10 years," says
Frederick Friedel, an expert on artificial intelligence and
computer chess who served as one of Kasparov's seconds. "Garry
will understand much more about chess, but he will still lose
because he will make mistakes."
Kasparov, for one, was not ready to concede supremacy. At the
postmatch press conference he came back to fiery life, arguing
that his biggest mistake had been to accept the IBM team's
conditions, which included playing on consecutive days and being
denied access to Deep Blue's log, a record of its evaluations of
moves. "It's very important that Deep Blue enter competitive
chess and go through the scrutiny other players go through," he
said. "Then, I assure you, I will tear it to pieces."
May 18, 1997