When the machine beat man at the game of kings: 29 years since Garry Kasparov’s defeat by Deep Blue

On May 11, 1997, in New York, Garry Kasparov sat down at a chessboard knowing that he was defending more than just a point on the scoreboard. Facing him was Deep Blue, the IBM supercomputer designed specifically to play chess at a speed unattainable by any human being. The match was tied going into the sixth and final game, with 2.5 points for each side. Whoever won that afternoon would win the match. And what happened was as brief as it was historic.

Deep Blue won in just 19 moves. For readers unfamiliar with chess: it was not a long, strategic, or drawn-out game, but rather a quick defeat. Kasparov, then the world champion and considered by many to be the greatest player in history, lost after finding himself trapped in an awkward position almost from the opening—the initial phase of the game in which both players develop their pieces and try to control the center of the board.

Kasparov played Black, which already puts him at a slight disadvantage, since White moves first. Deep Blue opened with the king’s pawn, one of the most classic moves in chess. Kasparov responded with the Caro-Kann Defense, a solid opening known for avoiding excessive risks. In simple terms, he chose a cautious structure, perhaps seeking a controlled game in which the machine could not launch into an overwhelming tactical calculation.

The problem arose very early on. On the seventh move, Kasparov advanced a pawn to attack one of Deep Blue’s knights. The machine responded by sacrificing that knight—that is, giving up a valuable piece in exchange for opening up lines against the black king. To the casual observer, this might seem crazy: deliberately losing a piece is usually dangerous. In chess, however, a sacrifice can be devastating if it forces the opponent to leave their king exposed or paralyzes their pieces. That was exactly what happened.

Deep Blue wasn’t “improvising” the way an inspired human would. Its strength lay in calculating a vast number of possible positions and assessing which ones led to a specific advantage. IBM explained that the system could evaluate around 200 million positions per second. That capability, combined with specific preparation for the match, turned that sequence into a trap of surgical precision.

After the sacrifice, Kasparov’s king was left in the center, unable to castling safely. Castling involves moving the king to a more protected area and connecting the rooks. Losing that opportunity in the opening stages is usually a warning sign. Deep Blue then began to position its pieces according to a logic that was simple to understand but difficult to execute: putting pressure on the king, preventing the black pieces from coordinating, and not giving Kasparov a moment’s respite.

The position quickly became untenable. Kasparov wasn’t facing an immediate checkmate, but he was facing a succession of threats that were narrowing his options. Every defensive move left another problem unresolved. His queen and rooks couldn’t mount an effective defense, his king remained vulnerable, and the white pieces occupied active squares. After Deep Blue’s 19th move, Kasparov conceded.

The final score was 3.5 to 2.5 in favor of Deep Blue. It was the first time a computer had defeated a reigning world champion in a full match under tournament conditions. Kasparov had already beaten an earlier version of Deep Blue in 1996, but the 1997 match changed everything. Until then, the machine might have seemed like an extraordinary tool, or even a one-off threat. From that day on, it began to be seen as a rival capable of surpassing humans in one of their oldest symbolic domains: calculation, strategy, and anticipation.

Kasparov, who years later revisited that match in ‘Deep Thinking’, always maintained that the sixth game cannot be understood solely on the basis of its 19 moves. For him, the defeat also stemmed from psychological factors: he was still reeling from the second game, in which a move by Deep Blue seemed so un-’mechanical’ to him that it fueled his suspicions about IBM, and he arrived on the final day exhausted, uncomfortable, and lacking his usual clarity. Over time, his assessment became more nuanced: Deep Blue did not “think” like a human champion, but it combined a brutal capacity for calculation, highly specific opening preparation, and the work of a team that had studied his habits. He also believes that if he had had a day off between the fifth and sixth games, the outcome might have been different.

Deep Blue didn’t think the way we do, nor did it understand chess the way a grandmaster does. But it calculated better, wasn’t affected by the pressure of the moment, and capitalized on the mistake sooner. Kasparov’s defeat wasn’t the end of human chess. Rather, it was the beginning of a new coexistence with machines—one from which there was no turning back.

Follow MeriStation USA on X (formerly known as Twitter). Your video game and entertainment website for all the news, updates, and breaking news from the world of video games, movies, series, manga, and anime. Previews, reviews, interviews, trailers, gameplay, podcasts and more! Follow us now!