Two
Carnegie Mellon students,
Feng-hsiung Hsu and
Thomas Anantharaman, joined efforts in
1985 to produce
Chiptest, the first step towards the creation of the chess
supercomputer Deep Blue. With the help of
Murray Campbell and
Andreas Nowatzyk they completed
Deep Thought in 1988 which was capable of computing 720 000 positions per second and had a Grand Master level rating.
Garry Kasparov played Deep Thought in
1989 and beat it easily.
IBM understood that searching through vast
databases, identifying relationships in the information and coming up with innovative solutions had value beyond the
chess world and hired Hsu, Campbell and Anantharaman and by
1991 Deep Thought 2 was capable of searching over 6 million moves per second. In
1993 Deep Thought 2 is renamed Deep Blue. In
1997, in its last match with Kasparov, Deep Blue was capable of analyzing 100 million positions a second
1.
Deep Blue is most widely known for the games versus Garry Kasparov and for the fact that the computer managed to beat the World Chess Champion. Before and after the two main competitions Kasparov had expressed frustration over IBM not being forthcoming with the information he felt was vital for him to compete.
This is what Garry Kasparov said about being beaten by Deep Blue in 1997.
" It is as simple as this: I just didn’t understand that the match against Deep Blue was actually going to be a match against IBM, a heartless corporation which to me bears some resemblance to the former CC CPSS (Central Committee of the Communist Party of the Soviet Union), with which I had many struggles during the ’80s. This match bore no resemblance to chess, because its result was never, so to say, certified: IBM refused to submit print copies of the chess lines (something I insisted on, as it was impossible to make the analysis without them). Any world record, to be approved, must in its own way survive some "drug tests". For a computer these "drug tests" are the print copies of the chess lines, which show the way the machine made this or that move, this or that decision. This was never done, the computer being completely disassembled so that no one might be tempted to try to verify the results. So the matter was closed… I believe today we are close to reviving the experiment, but with the condition that the machine must be under the control of independent experts. That is under the control of an arbiter supervising the game as normally in chess."2
In the same speech Kasparov hinted that IBM was using a low level Grand Master to correct obvious mistakes a computer might have made and that IBM refuses to prove this wrong and that the computer has since be disassembled.
1 the timeline and positions per second calculations are all from IBM website at research.ibm.com/deepblue
2This was three years later, responding to a question during a speech before the students of
The Humanitarian University in
St. Petersburg