Yes, in this paper (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2571012) I also consider the case where people make mistakes. You can study equilibrium even in models where people are not perfectly rational.