Affiliation:
1. NextMove Software Limited, Cambridge, U.K.
Abstract
In a defining event for the field of Artificial Intelligence (AI), the first game of chess skill between a human and computer took place in 1956 (Chess Review (1957) 13–17; The Machine Plays Chess? (1978) Pergamon Press). In this match, Dr Martin Kruskal from Princeton University played White against the MANIAC I computer at Los Alamos Scientific Laboratory in New Mexico, programmed by Paul Stein and Mark Wells. Due to the very limited capacity of computers at the time, which couldn’t handle a full 8 × 8 chess board, the competitors played “Los Alamos Chess”, a minichess variant using a 6 × 6 board without bishops. For this game, White played without a queen, opened with P-K3 and ultimately won against the machine opponent in 38 moves. Here we show that Black can force a win in 21 moves.
Reference15 articles.
1. Bell, A.G. (1978). The Machine Plays Chess? Pergamon Press.
2. The B∗ tree search algorithm: A best first proof procedure;Berliner;Artificial Intelligence,1979
3. Hitech wins North American Computer-Chess Championship;Berliner;International Computer Games Association (ICGA) Journal,1985
4. Breuker, D., Allis, V. & van den Herik, J. (1994). Mate in 38: Applying proof-number search. In J. van den Herik, B. Herschberg and J. Uiterwijk (Eds.), Advances in Computer Chess (Vol. 7).
5. A comparison of minimax tree search algorithms;Campbell;Artificial Intelligence,1983