Estimating bounded rationality and pricing performance uncertainty

This paper estimates bounded rationality in two outcome, strictly competitive, extensive form games of perfect information without chance moves. In the baseline sessions, the average subject can reason effectively to about rank 6. Twelve percent of these subjects have a perfect bound on their ability to achieve the substantively rational outcome: they win all simpler games and they lose all games more complicated than the estimated rationality bound. Most subjects do not have accurate expectations about their performance against a procedurally rational opponent. Usually, they value the games at more than they are likely to earn if they actually play them.