We study the steady states of a system in which players learn about th
e strategies their opponents are playing by updating their Bayesian pr
iors in light of their observations. Players are matched at random to
play a fixed extensive-form game, and each player observes the realize
d actions in his own matches, but not the intended off-path play of hi
s opponents or the realized actions in other matches. Because players
are assumed to live finite lives, there are steady states in which lea
rning continually takes place. If lifetimes are long and players are v
ery patient, the steady state distribution of actions approximates tha
t of a Nash equilibrium.