The paper argues that there is such a thing as luck in acquisition of candidate apriori beliefs and knowledge, and that the possibility of luck in this “armchair” domain shows that definitions of believing by luck thatp offered in literature are inadequate, since they mostly rely on the possibility of it being the case that not-p. Whenp is necessary, such a definition should be supplemented by one pointing to variation in belief, not in the fact believed. Thus the paper suggests a focus upon the agent and her epistemic virtue in the account of epistemic luck in general.