Creating an Emotionally Adaptive Game

To optimize a player's experience, an emotionally adaptive game continuously adapts its mechanics to the player's emotional state, measured in terms of emotion-data. This paper presents the first of two studies that aim to realize an emotionally adaptive game. It investigates the relations between game mechanics, a player's emotional state and his/her emotion-data. In an experiment, one game mechanic (speed) was manipulated. Emotional state was self-reported in terms of valence, arousal and boredom-frustration-enjoyment. In addition, a number of (mainly physiology-based) emotion-data features were measured. Correlations were found between the valence/arousal reports and the emotion-data features. In addition, seven emotion-data features were found to distinguish between a boring, frustrating and enjoying game mode. Taken together, these features convey sufficient data to create a first version of an emotionally adaptive game.