This paper argues for a particular account of luck by comparing two distinct versions of the modal account of luck that have been provided by Duncan Pritchard (2005, 2014). More specifically, it argues that there are three respects in which Pritchard鈥檚 earlier modal account of luck is preferable to his later account: it accounts better for the fact that luck comes in degrees, it includes a significance condition, and it better acknowledges the subjective nature of luck. The paper then discusses two consequences of the points it makes for epistemology: an alleged pragmatic encroachment, and a particular view on the relation between knowledge, luck, and justification.