An implicit assumption of prior literature on strategic behavior of auditors is that auditors know the preferences of their colleagues. In this study, we conduct an experiment to investigate the validity of this assumption. In our experiment, we match a manager with a pair of top and mediocre audit seniors, as assessed by their firm. Each auditor predicts the choices that will be made by other auditors on two tasks that differ in their level of ambiguity. Our results show no difference in the accuracy among managers, top seniors, and mediocre seniors when they predict the choices made by specific individual auditors for both tasks. When predicting the number of managers and seniors who will choose a specific option on the high‐ambiguity task, managers outperform top seniors, who in turn outperform mediocre seniors. For the low‐ambiguity task, we find no difference among managers, top seniors, and mediocre seniors. Our results provide some limited support for models of strategic auditor behavior, and indicate that the ability to predict the choices of others is a dimension of an auditor’s expertise.