Gambling self-exclusion agreements enable a person to have themselves prevented from gambling for some future period. In light of evidence of their effectiveness in helping problem gamblers manage their addiction, these agreements enjoy growing popularity. In particular, several jurisdictions now oblige gambling operators to offer self-exclusion to their clientele. If self-exclusion has a unique value that is distinct from paternalistic measures, such as forced exclusion, it is surely because it prizes the gambler’s autonomy. In this article, however, I will argue that self-exclusion’s theoretical basis cannot, in fact, be found in a procedural theory of autonomy that only regards agents’ own values and decisions. Rather, I will contend that if agents may bind their future selves in only some ways—for example, by preventing themselves from gambling but not preventing themselves from self-excluding or selling themselves into slavery—it can only be because of a normative, substantive claim.