Player Strategy Modeling in Classic Roleplaying Game Battle Environments

Modern game developers have acknowledged the necessity of a system which adjusts the gameplay experience for gamers. Part of this system of adjustments is called dynamic difficulty adjustment which, as its name suggests, adjusts the difficulty of the game depending on information collected during gameplay. There are many approaches to accomplish this task, amongst which are to change the behavior of the computer-controlled characters according to the player’s patterns of behavior detected from a survey of past actions. This paper introduces a method to collect and process information regarding player action selections to produce an estimated model of the player’s strategy. The estimated player’s model is then used to determine the computer characters’ strategy to keep the game not too easy and not too hard.