Signal detection methods were used to identify predictors of successful weight loss in 177 mildly to moderately overweight women and men assigned to one of two weight-loss programs. Predictors included initial demographic, physiological, behavioral, and psychosocial characteristics, and program type (e.g. diet-only and diet-plus-exercise). Successful weight loss was defined as a loss of at least two units of body mass index at one year. Four subgroups were identified. Participants in the diet-plus-exercise program who were initially more satisfied with their bodies and did not have a history of repeated weight loss were most likely to succeed (63% succeeded). In contrast, participants assigned to the diet-plus-exercise program who were either extremely dissatisfied with their bodies or who had a history of repeated weight loss were at similar risk for failure as participants in the diet-only program (only 26% to 35% succeeded). The results underscore the potential utility of exploring these subgroups further to inform the development of new treatment strategies to increase the likelihood of success.