What Was Deep Blue?

Mary McMahon
Mary McMahon

Deep Blue was a supercomputer designed by IBM specifically for the purpose of playing chess. The computer distinguished itself in 1997 by beating Garry Kasparov, one of the top chess players in the world. Although Deep Blue was retired after this match, the match laid the groundwork for increasingly sophisticated chess computers and programs, and many chess players now utilize chess software as a learning and practice tool to keep their games sharp.

Pieces for chess, which Deep Blue played.
Pieces for chess, which Deep Blue played.

The inspiration for Deep Blue came to Feng-Hsiung Hsu in 1985. Hsu started to develop a computer which he called Chiptest, later bringing the idea to IBM when he joined to company in 1989. At first the resulting computer was code-named Deep Thought, and jokers at IBM later proposed “Deep Blue,” a melding of “Deep Thought” and “Big Blue,” IBM's nickname.

IBM's Deep Blue was a supercomputer that exhibited artificial intelligence by playing chess against human Grand Masters.
IBM's Deep Blue was a supercomputer that exhibited artificial intelligence by playing chess against human Grand Masters.

This supercomputer was specifically designed for the purpose of playing chess, with the developers taking a hardware based approach, rather than focusing on software like that used in modern chess programs. Like other supercomputers, Deep Blue was incredibly powerful for its time, with 1.4 tons of hardware to back up a chess program written in C and capable of calculating up to 200 million potential positions every second. Deep Blue could see up to 40 moves, or plies as they are known in the chess world, ahead.

The computer drew its strategy from an extensive record of chess games played by Masters and Grand Masters. Using the records of these games, Deep Blue could consider a wide range of possible moves, gambits, and strategies which would potentially allow it to dynamically respond to moves made by an opponent. Deep Blue also improved its game by playing several matches with chess Masters, with programmers learning from the computer's mistakes.

Deep Blue's first match against Kasparov took place in 1996, and Kasparov won the match. The 1997 match featured a substantially updated Deep Blue, however, and the computer integrated adaptations from its experiences in the previous match against Kasparov. Deep Blue won the match, taking two games outright to Kasparov's one and gaining another point and a half from three draws, for a total of three and a half to two and a half.

Kasparov later decried the 11 May 1997 match against Deep Blue, arguing that the computer displayed such depth of intelligence that humans must have intervened during the games to help Deep Blue win. He demanded a rematch, which IBM refused, and the issue became a topic of some controversy in the chess and computing communities. Some knowledgeable chess critics pointed out that Kasparov's strategy in the 1997 match was extremely conservative and very out of character for him, suggesting that he might have won if he had played with his usual aggressive, dynamic style.

Mary McMahon
Mary McMahon

Ever since she began contributing to the site several years ago, Mary has embraced the exciting challenge of being a EasyTechJunkie researcher and writer. Mary has a liberal arts degree from Goddard College and spends her free time reading, cooking, and exploring the great outdoors.

You might also Like

Discussion Comments


I know humans like to think they cannot be replaced by computers, but for many jobs we can. With chess, programmers can load so much information into a computer program that it is virtual impossible for the computer to lose no matter who the person it is playing against. Computers are much better chess players today than Deep Blue was when it played.


In high school, my friend and I had our own version of Deep Blue. We would play his home computer, which had a chess playing program on it. This was back before there was a personal computer in everyone's home, so this was a big deal to us.

The program had different levels. I think it had 10 levels. One was the easiest level and then 10 would have been the most difficult. I think we worked our way up to level five or so, and it was both of us against the computer. Also, we could undo our moves when we saw that we had made a mistake. I guess that was cheating, but we needed all the help we could get and, like I said, we never got past level 5 anyway.

Post your comments
Forgot password?