Noisy Introspection in the 11–20 Game

Abstract

Previous experiments based on the 11–20 game have produced evidence for the level†k model with observed levels of strategic thinking consistently ranging from 0 to 3. Our baseline treatment uses the 11–20 game and replicates previous results. We apply four models of strategic thinking to the baseline†treatment data and use these to predict behaviour and beliefs in five other treatments that employ games with a very similar structure. The best predictive performance is achieved by models that incorporate ‘common knowledge of noise’. A model of noisy introspection, which does so, predicts behaviour remarkably well.