This paper considers learning rules for environments in which little prior and feedback information is available to the decision-maker. Two properties of such learning rules, absolute expediency and monotonicity, are studied. The paper provides some necessary and some sufficient conditions for...