We develop a multi-agent framework based on probabilistic cellular automata theory to describe off-equilibrium dynamics in the context of the economic problem of price adjustment in different strategic situations as investigated experimentally by Fehr and Tyran (2001) and (2008). It is found that the main experimental findings, namely suboptimal aggregate behavior in terms of sluggish adjustment after a fully anticipated money shock, can be reproduced and largely explained by the interaction of sophisticated and naive agents. Furthermore, a range of conceptual issues as e.g. the source of endogenous beliefs on the other players rationality is addressed within our multi-agent framework. We find that, if costs/payoffs act as driver of rational behavior, then endogenous beliefs and consequential aggregate behavior are driven by the particular off-equilibrium time-dependent payoff/cost profile rather than by total off- equilibrium payoffs/costs that naive agents face in the respective strategic situation.