Under the decision rule specified by Kandori, Mailath, and Rob (1993) [Kand
ori, M., Mailath, G.J., Rob, R., 1993. Learning, mutation, and long run equ
ilibria in games. Econometrica 61, 29-56], myopic adjustment can lead to su
rprising results, including coordination on strictly dominated strategies.
We show that under an alternative decision rule, convergence to Nash equili
brium is guaranteed. Moreover, if rare mutations are introduced, risk domin
ant equilibria always correspond to long run equilibria. (C) 1998 Published
by Elsevier Science S.A. All rights reserved.