<p>Adam Kolber argues that retributivists should insist not just that factual guilt is established beyond a reasonable doubt before they endorse punishment, they should condition punishment on proving that the defendant deserves to be punished. Perhaps they need not prove desert beyond a reasonable doubt, but they should still withhold punishment if the overall likelihood of desert is not very high. And he argues that there are so many reasons to doubt that defendants deserve punishment that retributivists cannot conclude that the likelihood of desert is high enough to punish. I argue that his argument is doubly mistaken. First, the reason to use the proof beyond a reasonable doubt standard for factual guilt rests on assuming the truth of retributivism. It does not call for applying that standard to the preconditions for accepting the truth of retributivism. Second, a different sort of inquiry is called for if one doubts the moral framework of retributivism itself. One needs to ask: How likely is it that punishment would wrong the defendant if some other moral theory is true? If the prime alternatives are consequentialism and abolitionism—as Kolber’s argument presupposes—then the relevant question is: Should the possibility that one of these alternative theories is true give a retributivist pause? I argue that the possible truth of consequentialism is very unlikely to provide a reason to doubt that punishment that would be just if retributivism were true would wrong a defendant if consequentialism were true. As long as a retributivist is fairly confident that abolitionism is wrong, Kolber’s arguments do not present a reason to hesitate to punish in accord with retributivism.</p>