A New Hashing Method with Application for Game Playing¹

Article type: Research Article

Authors: Zobrist, Albert L.^{; *}

Affiliations: The University of Wisconsin, Madison, Wisconsin

Correspondence: [*] Present affiliation: The Rand Corp., Santa Monica, CA. Correspondence address: see inside back cover.

Note: [1] This article is a reprint from: Technical Report #88, Computer Sciences Dept., The University of Wisconsin, Madison, Wisconsin 53706, April 1970. The Editors gratefully acknowledge permission from Dr. A.L. Zobrist to reprint this report often referred to in publications while being inaccessible in many libraries normally thought of as well-stocked. Dr. Zobrist also informed us with hindsight on his work. It has been refined in 1971 for subsets of the chess-board for the Zobrist-Carlson-Kalme chess program (as published in the 1973 Scientific American article; it has been published with more details in the 1977 CACM article), then it has apparently been used in the OPS5 work at CMU some time thereafter.

Abstract: A general method of hash coding is described together with an application for programs which play board games such as checkers, chess, and Go. An auxiliary method which detects retrieval errors is proposed. The error rate can be precisely controlled depending upon how much space in the hash table is devoted to the auxiliary method.

DOI: 10.3233/ICG-1990-13203

Journal: ICGA Journal, vol. 13, no. 2, pp. 69-73, 1990