The optimal matching model of support suggests that supportive behaviors are effective when they match recipients' needs or goals. In 2 studies, we used quasi-signal detection to test this model. Specifically, we simultaneously modeled the associations of affective and relational outcomes with matching ("hits"), underprovision ("misses"), and overprovision ("false alarms") regarding emotional or practical support. In both studies (N = 44 and 38 couples, respectively), both partners in committed relationships reported daily receipt of and desire for support, as well as moods and relational outcomes for 21 days. Emotional (but not practical) matching was associated with favorable relational outcomes only; overprovision had a minimal effect. In contrast, both emotional and practical support underprovision were associated with adverse outcomes, both affective and relational in nature. Study 2 ruled out the possibility that stress is the sole mechanism responsible for the adverse effects of underprovision, and found evidence for another, relational mechanism; specifically, the results documented the role of perceived partner responsiveness as a mechanism mediating the deleterious effect of underprovision.