Internet Chess club on Facebook Internet Chess club on Twitter Internet Chess club on Google+ Internet Chess club on YouTube Internet Chess club on LinkedIn Subscribe to Internet Chess Club RSS feed

Happened Today in the Chess World - May 11, 2024


Deep Blue, the supercomputer that defeated the world champion Garry Kasparov

On May 11, 1997, a computer named Deep Blue proved superior to a human for the first time, defeating the world chess champion Garry Kasparov.
Kasparov, born in 1963 in Baku, Azerbaijan, learned to play chess from his father and was already a prodigy at 6.

Deep Blue was created in 1989 at IBM and designed by Chinese computer scientist Feng-hsiung Hsu and Canadian Murray Campbell. Hsu and Campbell met at Carnegie Mellon University in Pittsburgh, where they worked on a thesis project to create a computer capable of defeating a human chess champion. IBM provided them with funding and resources to perfect their work.

Deep Blue wasn't your average electronic calculator; it was a supercomputer of what experts call very high parallelism. It could process 200 million moves per second and store thousands of games, openings, and endings. It had already challenged Kasparov a year earlier, winning the first game in Philadelphia but ultimately losing the six-game match 4-2.
Hsu and Murray didn't lose heart. They brought in chess champions (such as Joel Benjamin and Miguel Illescas) to teach Deep Blue new tactics and added more libraries of openings and endgames to its memory. They even resort to retrograde analysis, starting from the conclusion of a game and working backward through all the moves that led to it.

At first glance, it seems natural that a supercomputer capable of processing millions of data in seconds would outplay the human mind, whose processing capacity is much less efficient.
But in chess, things aren't so straightforward. Deep Blue played like an amateur, analyzing the position of the pieces and assigning predetermined values, deciding its move each time.
A champion, however, has a sense of the board, establishing a true feeling for it. They can glance at the situation and study only about a dozen possibilities before deciding. The computer used brute force, sifting through millions of combinations; the champion, on the other hand, employs sensitivity and imagination.
Deep Blue's libraries of openings and endgames made it particularly formidable in the openings and endgame but not in the middle game, where human ingenuity could prevail.

From May 3 to 11, 1997, Deep Blue challenged a more focused than ever Garry Kasparov in New York.
"I will defend the human race," he pompously declared before taking his seat at the board. In the encounter a year earlier, he had realized there was only one way to beat computers: to confound them, to do things they don't expect. Kasparov began the first game with the unconventional Barcza System, an opening that doesn't follow a precise sequence but leaves several possibilities open.

Deep Blue searched its memory for a response to this attack but found none, losing in 45 moves. In the second game, Kasparov made a crucial mistake that could have led to a draw, but he didn't capitalize on Deep Blue leaving its king somewhat exposed in the endgame.

The third, fourth, and fifth games ended in a draw, and on May 11, with the score tied at 2 ½, the decisive showdown began, followed by hundreds of journalists hoping to herald the dawn of a new era in which machines would do all the work and humans could finally rest. Kasparov once again set out to confound Deep Blue with the Caro-Kann Defense but then allowed the computer to sacrifice a knight, a move that disrupted his strategy to the point of surrendering after just 19 moves.

Deep Blue - Garry Kasparov (1997 match - Game 6)
White to move
Can you guess the next move by Deep Blue that discombobulated Kasparov's strategy?


8.Nxe6! Qe7 9.O-O fxe6 10.Bg6+ Kd8 11.Bf4 b5 12.a4 Bb7 13.Re1 Nd5 14.Bg3 Kc8 15.axb5 cxb5 16.Qd3 Bc6 17.Bf5 exf5 18.Rxe7 Bxe7 19.c4 1-0

View past articles