Some philosophers believe that we can, in theory, justifiably pre-punish people – that is, punish them for a crime before they have committed that crime. In particular, it has been claimed that retributivists ought (in principle) to accept pre-punishment. The question of whether pre-punishment can be justified has sparked an interesting and growing philosophical debate. In this paper I look at a slightly different question: whether retributivists who accept that pre-punishment can be justified should prefer (ordinary) post-punishment or pre-punishment, or see them (in principle) as on a par. The answer is complex: asking this question brings to light unrecognised distinctions within both retributivism and pre-punishment, giving us four different answers to the question, depending on what kind of retributivism and what kind of pre-punishment are combined. Surprisingly, given that it is usually presented as a second best, to be pursued only when post-punishment is unavailable, some combinations will find pre-punishment preferable.