This research examines preinvestment forecasting processes in escalation of commitment, considering two questions: whether individuals are able to accurately predict their behavior and affect in escalation situations and how forecasting processes may be linked to actual escalation. Three experiments demonstrated that individuals underpredicted their escalation and overpredicted their postescalation regret. Two of the experiments also indicated that the less individuals predicted being entrapped, the more they escalated. Counter to expectations, anticipated regret did not predict escalation. The discussion focuses on the theoretical and practical importance of forecasting on escalation and on the importance of understanding both behavioral and affective forecasting effects simultaneously.