Endgame: Challenging the Masters

Narrator: This was suspense beyond anything the world of chess had ever known. Garry Kasparov, 34 years old, a cheerful and confident player throughout a dominating career, hovered over his pieces in the deciding game of a match with an implacable challenger: Deep Blue - a computer. Kasparov won the first game in a breeze, but the next day he got the shock of his life.

To Kasparov, Deep Blue played the second game much more strongly, differently, "unlike a computer," he said. Kasparov conceded, and the loss wore on his mind for the rest of the match.

In each of the next three games, the world's pre-eminent chess players, man and machine, battled each other to a draw. Kasparov's early scorn for Deep Blue's abilities had proven a great mistake.

In the final game, the computer led with the white pieces, and soon Kasparov's fans had to admit that the once unthinkable might actually be happening. How had it come to this?

It was the work of a generation of impassioned computer engineers and programmers that led to Deep Blue. Human chess champion Garry Kasparov was being challenged, really, by other humans.

They were, arguably, as brilliant as Kasparov, and no less disciplined and determined. They passionately believed they could teach their machines to surpass the performance of the human brain.

The programming pioneers of the late 1940s knew and respected chess; it was a smart person's game, and a worthy challenge. But after all what a Grand Master did was just look ahead and analyze various possibilities, wasn't it? And computers could be taught to do that. But not easily; and for decades, not very well.

Still, by the mid-1980s, computer chess programs began challenging and occasionally beating Grand Masters. It remained unclear whether they could ever defeat the world's best.

In 1996, after seven years of research and development, a team at IBM believed they were ready to challenge Garry Kasparov.

In Game One of six, Deep Blue eked out its first win against the Champion. Kasparov learned fast and won the match decisively. His confident proposal for a rematch in 1997 was immediately accepted by the IBM team. They were learning, too.

The afternoon of May 11, 1997. Game Six, the deciding game. Deep Blue challenges Kasparov's brain with an array of 256 processors that can examine 200 million possible moves, every second. They call it the "brute force" approach. Kasparov doesn't try, he can't try, to consider every possible move.

He knows from experience what's important, and he relies on the human mind's extraordinary ability to recognize and remember patterns. The patterns of chess experience lived in Deep Blue's software, too. Grand Masters had coached IBM's programmers to deepen Deep Blue's "book" - its library of what humans knew "from experience" - about how to win. Deep Blue could now play as well as a Grand Master, with all the benefits of hindsight.

In the early moves of Game 6, playing by its book of openings, then in the mid-game analyzing millions of possible scenarios by sheer brute force, Deep Blue has led Kasparov into making a poor move. Kasparov is rattled; he defends what he can, but it's clear that the computer will reliably do what he himself would do and he recognizes that he has already lost.

On Deep Blue's 19th move, the champion resigns. For people watching around the world, it is a slow-motion moment.

Match Commentator: "The IBM program Deep Blue has defeated World Champion Garry Kasparov... in an absolutely stunning..."

Narrator: IBM retired Deep Blue, and it never competed again. But the improvement of chess software did not stop in 1997. Today, inexpensive computers - with their large memories and sophisticated software can play chess as well as Deep Blue did with its massively parallel processors.

Yet Deep Blue vs. Kasparov was justly called historic. The programmers had proved it was possible to build a chess-playing machine that could defeat the best human opponent. Nevertheless, the machine they had built did not play chess by "thinking" in the same way a human does. A machine that can think remains the dream and it's still many years and quite a few startling breakthroughs away.