There has been a great deal of discussion in the recent literature regarding the supposed phenomenon of "epistemic luck." This is the putative situation in which an agent gains knowledge even though that knowledge has come about in a way that has, in some sense to be specified, involved luck in some significant measure. Unfortunately, very little of the literature that deals with epistemic luck has offered an account of it that is anything more than suggestive. The aim of this paper is to offer a more nuanced elucidation of what is involved in different types of epistemic luck. More specifically, an account of luck is proposed and several varieties of epistemic luck are shown to be compatible with knowledge possession, in contrast to two other varieties whoe status is much more problematic. It is argued that by being clear about what is involved in epistemic luck one can gain an insight into several central debates in epistemology, including the "Gettier" counterexamples, the problem of radical scepticism and the co-called "metaepistemological" challenge to externalist theoris of knowledge.