Inferring player engagement in a pervasive experience

We investigate the prediction of player engagement to address temporal issues arising from the long-term character of pervasive experiences such as interruptibility, mutual player state awareness, disengagement and synchronization on re-engagement. We introduce a model that operationalizes engagement in terms of the elapsed and response time in game messages. We designed and conducted an experiment based on the experience-sampling method to evaluate our model on the basis of a long-term SMS-based game called Day of the Figurines. Statistical analysis supports the hypothesis that player engagement can be predicted by the continuous data properties elapsed time and response time. Our findings point towards further research towards the adaptation of pervasive experiences to the player's temporal context.