Deep Blue – Legendary IBM chess computer

On May 11, 1997, IBM’s Deep Blue supercomputer bombed the Russian chess general Garry Kimovich Kasparov.

Kasparov is considered to be the strongest player in history with 2851 ELO rating, which was ranked the highest in about 20 years from 1985 to 2005 and was the undefeated world chess champion from 1985 to 1993.

Previously in 1996 the Deep Blue supercomputer was tested with Kasparov. The computer won the first game, but then Kasparov won the game. By 1997, IBM along with a new Deep Blue upgrade, was determined to defeat Kasparov again. They unveiled a new version of the software that has been enhanced with multiple super hardware running in parallel. Joel Benjamin – American chess expert was analyst and consultant for Deep Blue.

The determination of IBM in Deep Blue was designed to “fight Kasparov”, all the opening steps that Kasaprov has used are installed in the machine dictionary. And this time they have succeeded. In the six games, Deep Blue beat Kasparov 3.5-2.5. At the final game, Deep Blue spotted Kasparov’s mistake from the start of the match and decided to try to break Kasparov’s game from the start of the game so Kasparov finally surrendered.

Kasparov lost to Deep Blue in 1997.

The 1997 game was live broadcast to millions of people. Under the terms of the deal, winning Kasparov, the founder of Deep Blue was received $ 700,000, while Kasparov was received $ 400,000. IBM alone sought profit through advertising reached $ 50 million.

The way of chess playing of human and computer is completely different. Man chooses his move after evaluating a minimum number of options that he thinks is the most appropriate.

And the computer does not have that ability to intuition, it in turn mechanically reviews all possible alternatives on the chessboard and then by the exclusion method, chooses the scheme it considers the best. The number of alternatives is enormous.

Before a chessboard, for example, with 30 steps able to move legally, man choose by experience and intuition some of the options he thinks are most reasonable, calculating what happens after 5-6 steps or longer depending on the level, then decide to move.

The computer, on the other hand, it in turn reviews the possible evolution of all 30 alternatives, including the most “silly” ones, until the end of the game. For example, after each step there may be 30 steps returned by the opponent, the computer must calculate the number of moves after a step of 30×30 = 900.

If each option is pre-calculated with a length of only 5 steps, the number of steps is 590,490,000,000,000. Which each of the options to determine the win of the game can be an average of 20 to 40 steps (!). So this is going to be a clash between the miraculous intuition of human and the speed of computer computing (millions of steps per second).

Leave a Reply

Your email address will not be published. Required fields are marked *