Albert L. Zobrist

The first ever program to play whole Go games was that of Albert L. Zobrist (USA). It was written as part of a PhD thesis on pattern? recognition.

The program divided the game into four phases: fuseki, extension towards tengen, defence and attack of loosely staked out territory, and yose. Each phase had a separate set of match patterns to find the most appropriate moves.

Zobrist's program played a game against George Cowan (and won by 7 points) on the 17th of November 1968.

Zobrist introduced the idea of using an influence function to segment the board into black and white territory, too.

Bibliography:

• Zobrist, A. L. A model of visual organisation for the game Go. In Proceedings of the Spring Joint Computer Conference, volume 34, pages 103 - 112, 1969.

• Zobrist, A. L. Feature Extractions and Representation for Pattern Recognition and the Game of Go. Phd, Graduate School of the University of Wisconsin, August 1970.

• Zobrist, A. L. A new hashing method with application for game playing, April 1970