Computer Chess Programs In A Glance

Over the years, chess has caught the attention of programmers and scientists. Through the passage of time, computer chess served as the yardstick for analyzing the capabilities of a computer in comparison with a human. Solving the game turned out to be more difficult than they imagined.

The pioneer chess program was designed by Konrad Zuse, a German inventor. However, the program was not used and unpublished until 1972. Thus, it was American Claude Shannon who was officially recognized as the "Father of Computer chess." He introduced what is known as "minimax algorithm," which involves selection of the most probable move on the premise that the other player would respond with their best move as well.

The earliest chess programs utilized brute force. This is a method wherein a player extensively finds the best position in a tree of moves with the help of heuristic evaluation function. As the player takes advantage of their position based on the evaluation function, the opponent tries to reduce that advantage.

The first fully-pledged computer chess was designed by Alex Bernstein and company for an IBM 704 in 1957. It investigated all possible four plies (an attempt by black or white is called a ply) in approximately 8 minutes and could beat amateur players.

A minor program that played chess on a 6 x 6 board had been in use a year in 1956 on the MANIAC I, constructed at the Los Alamos National Laboratories. However, the pioneer program that competed in an actual tournament was MacHack VI, designed by Richard Greenblatt for a PDP-6 at Massachusetts Institute of Technology. Eventually, it ran on all PDP-10 computers, which are used in universities and research centers.

While Herbert Simon and Allen Newell predicted that a computer program would emerge world champion in a decade, it took more than that before a computer emerged victorious against the world champion. This is because good human players have the ability to distinguish patterns in the chess board which is why they inspected only few possible moves.

On the other hand, chess programs play a decent game because they inspect a wide range of possibilities in a span of one second. Simon and Newell was credited for modifying the minimax solution and utilized "alfa-beta pruning", a strategy that determines negative branches in the move tree and eliminates them early.

These are just some of the evidences showing man's attempts to beat complicated games such as chess.