Sample text for A new era : how Garry Kasparov changed the world of chess / Michael Khodarkovsky and Leonid Shamkovich.
Bibliographic record and links to related information available from the Library of Congress catalog
Copyrighted sample text provided by the publisher and used with permission. May be incomplete or contain other coding.
Game Two: Sunday, May 4, 1997
Deep Blue--Kasparov
Before Game Two, Garry had decided to let Deep Blue play the Spanish Game
(Ruy Lopez) against him, and immediately before the game he decided to
play a very "left wing" variation that the computer would not be able to
find in its database of chess openings.
The game started out with Garry's playing the Spanish Game with 5...Be7--an obvious choice given his anticomputer strategy. The opening was
played with the standard Spanish Game moves. Deep Blue obviously had
access to Spanish Game theory and was playing according to firmly
established chess principles.
It was soon becoming apparent to everyone that something different was
happening in Deep Blue's play. Its ability to develop a rational plan, its
patience in realizing plans, its sophisticated maneuvers--all this made
Deep Blue seem quite human in its play. It was eerie to watch--Garry was
being outplayed positionally by a machine! Computers aren't supposed to be
able to do this and yet it was happening before our eyes. Apparently, the
IBM Deep Blue team had modified some of the parameters of Deep Blue's
scoring functions between rounds one and two which would account for the
machine's ability to limit its opponent's counterplay to practically
nothing. The implications were staggering. We could now never be sure
which Deep Blue to play against! Garry was getting squeezed off the board.
With no attacking possibilities of his own, all he could do was play a
patient, waiting defense, hoping to fend off Deep Blue's eventual attack.
Deep Blue gradually improved its position and with a highly prized human
virtue--infinite patience--prepared a decisive breakthrough into Garry's
position. With 36. axb5, the breakthrough was achieved. The route into
Garry's position would be the a-file. The only question was, could Garry's
defense hold?
Deep Blue played the breakthrough masterfully while preventing any chances
of counterplay by Black with its 37. Be4!, a move which surprised many
observers since Qb6, netting some extra pawns, seemed more in keeping with
a computer's materialistic bent. With no counterchances, Garry just waited
with the rest of us for the end. After Deep Blue played 46. Ra6, Garry,
having seen enough, resigned. The audience, the commentators, the
pressroom, and perhaps the world were stunned not so much by a machine's
having defeated a World Champion, but by the manner in which it had done
it. The great subconscious fear of machines conquering their makers--and
their makers being powerless to stop it just as Garry seemed powerless
against Deep Blue--seemed to be in everyone's conscious thoughts.
Game Two--The First Questions
We entered several game positions into our computers for evaluation and
waited to see if they would select the same moves...
While the computers were working, and as the shock of what had happened
began to wear off, something started to bother us--the inconsistencies in
Deep Blue's behavior. For example, in Game One and until move 35 of Game
Two, Deep Blue spent three to four minutes on every move. Obviously it had
been programmed to make a move in that amount of time. But before its move
35, Deep Blue took fifteen minutes to make sure that Bxd6 was a good move
for White.
Another inconsistency was the way Deep Blue had played between Game One
and Game Two. In Game One, it played the way we were accustomed to see a
computer play--materialistic and tactical but unable to comprehend
higher-level things like long-term positional weaknesses and piece
coordination. In Game Two, material was not important to the machine and
it played with a strategic and positional competence achieved by very few
grandmasters.
These inconsistencies bothered us and Garry; we started to feel
suspicious. However, we had nothing solid to go on. There was also the
possibility that the Deep Blue team had set us up by having Deep Blue play
in its original style in Game One, only to surprise us with its newer
capabilities in Game Two. All this bothered Garry and it became difficult
for him to concentrate completely on his preparations for Game Three. With
these thoughts dominating our minds, we went to sleep. Our dreams could
not possibly match the fantastic events that were to occur.
Library of Congress subject headings for this publication: Kasparov, G, K, (Garri Kimovich)Deep Blue (Computer)Chess Tournaments, Chess Collections of games