Some Necessary Conditions for a Master Chess Program

Since 1967 there has again been great interest in chess programming. This paper demonstrates that the s t r u c t u r e of today's most successful programs cannot be ex tended to play Master level chess. Certain basic requ i rements of a Master player's performance are s h o w n to be outside the performance limits to which a p rog ram of this t ype could be extended. The paper also examines a basic weakness in the tree-searching model approach when applied to situations that cannot be searched to completion. This is the Horizon Ef fect , which causes unpredictable evaluation errors due to an interact ion between the static evaluation func t i on and the rules for search termination. The ou t l ine of a model of chess playing that avoids the Hor izon Effect and appears extendable to play Master leve l chess is presented, together wi th some results a l ready achieved.