This paper investigates a user model for a system which automatically generates interactive stories. These stories are focused on dilemmas in order to create dramatic tension. The system is provided with knowledge of generic story actions and dilemmas based on those cliches encountered in many of today’s soap operas. These dilemmas and story actions are instantiated for the given storyworld and a story planner creates sequences of actions that each lead to a dilemma for a character (who can be the user). The user interacts with the story by making decisions on relevant dilemmas and by freely choosing their own actions. In this paper we propose a method of generating a model to predict user decisions. This model is used to optimise the user’s experience — by selecting the next dilemma to be presented according to what will be the most interesting and by looking ahead when planning.