In some hypothetical situation (the specifics of which aren't terribly important) I know that two possible reasons for choosing near-game theory optimal strategy as opposed to near-optimal exploitive strategy are for playing against unknowns and for playing against someone who adjusts better than you. Are there others?