In this paper, we study decision errors caused by the omission of part of the legal candidate moves and the inaccuracy of static evaluation in a selective minimax game tree search. Error upper bounds are presented in Section 2. A simple game tree model, which captures some basic characteristics of the Go game tree, is introduced in Section 3 for a decision error simulation study. Section 4 presents the result of this simulation study, which shows that a global selective search can be effective for game trees similar to this model. The result also reveals the existence of pathology in selective minimax game tree search. (C) 2001 Elsevier Science Inc. All rights reserved. [References: 5]
展开▼