Level of detail is a method that involves optimizing the amount of detail that is simulated for some entity. We introduce an event generation method to optimize the level of detail of upcoming events in a simulation. Our method implements a cognitive model, which uses an estimate of the player’s knowledge to estimate their interest in different aspects of the world. Our method predicts the salience of upcoming events, and uses this salience value to define the level of detail of potential new events. An evaluation of our method’s predictive capacity shows generally higher accuracy than a baseline predictor.