Clinical observation indicates that self-efficacy (SE), though generally linked with maintaining smoking cessation, can be excessive. The states-of-mind (SOM) model offers a testable means of predicting when positive thinking will be detrimental. Specifically, the ''positive-monologue'' SOM (ratio of positive to positive + negative thoughts greater-than-or-equal-to .69) is seen as maladaptive. This model was tested by relating SOM ratios for SE-relevant thinking, measured in articulated thoughts shortly after smoking cessation, to later smoking status. Contrary to predictions, positive-monologue SOMs were (a) typical and (b) predictive of longer periods of abstinence than were less positive cognitive balances. Several explanations are considered for this divergence from the SOM model.