Game AI Planning Analytics: The Case of Three First-Person Shooters

Abstract

We present a general framework for Game Artificial Intelligence Planning (AIP) Analytics. The objective is to provide analytic tools to study and improve AIP components and their use in video-games. Extraction and formatting of AI data is first described and discussed. Then AIP metrics are listed with examples and illustrations from three popular First-Person Shooters: F.E.A.R. (2005), KillZone 3 (2011) and Transformers 3: Fall of Cybertron (2012). The patterns we discovered in our study clearly show the AIP component is called more often by the game over the years.