The reformulated balanced states of mind (BSOM) model (Schwartz, 1997) proposed new cognitive-affective set-point ratios based on a mathematical model of consciousness (Lefebvre, 1990) to differentiate among pathological, normal and optimal balances. Using data derived from the Affects Balance Scale (Derogatis, 1975), the reformulated set-points were empirically evaluated by tracking changes in affect balance SOM (ratio of positive to total affect) in 66 depressed male outpatients undergoing cognitive-behavior therapy (n = 45) or pharmacotherapy (n = 21). Confidence interval estimations indicated that across treatments both remitted (SOM = .35) and unremitted (SOM = .35) patients had pathological pretreatment affect balances near the predicted set-point (.38). At post-treatment, affect balance for remitted patients (SOM = .74) progressed to a normal dialogue near the predicted set-point (.72), whereas unremitted patients maintained a negative balance (SOM = .41). Using Hamilton and Global Assessment Scale ratings, remitted patients were classified into average and optimal responders. At post-treatment, average responders achieved an affect balance (SOM = .70) near the normal dialogue set-point (.72), whereas optimal responders progressed to an affect balance (SOM = .81) at the optimal dialogue set-point (.81), supporting the theoretical distinction between normal and optimal balance. A mathematically generated measure that distinguishes normality and optimality permits increased quantitative precision in comparative psychotherapy outcome research.