Measuring strategic depth in games using hierarchical knowledge bases

This paper presents a measure intended to quantify the relative strategic depth of games as experienced by human players. The measure is based on the complexity (number and specificity of rules) of a hierarchical knowledge base that is extracted from playtraces. As a proof-of-concept, we compute the proposed measure for three arcade-style games and compare the results to the strategic depth reportedly perceived by human players in a survey. We find that our measure is practicable and able to capture the differences in strategic depth very accurately.