This paper examines the implications of a 'strong' version of bounded rationality popular within computational and evolutionary game theory in which agents are represented by finite automata. It is argued that this view has unrealistic features in that agents of this kind are unable to...