Abstract
Search procedures generally discover bounds as partial results before a final value is obtained. In particular, chess programs typically store such bounds in a hash table, along with any exact results, for every position processed. These values save computing time if the position is encountered again. For chess programs, upper bounds, lower bounds or exact values may occur. This paper shows how upper bounds, lower bounds and exact values can conveniently be represented using a single numeric scale, which enables program code to be slightly simpler, and avoids the necessity of a separate data item to distinguish bounds from exact values.
Get full access to this article
View all access options for this article.