This paper adopts an aggressive stance on short-term and long-term weight losses produced by treatment of obesity. For many years the behavior therapy field has settled for small weight losses. This has resulted in part from a focus on "maintenance," which implies something worth maintaining. Complacency has occurred because of the contest mentality underlying the design of parametric studies which compare behavior therapy to various alternatives and find behavior therapy to be the "treatment of choice." It is not surprising that behavior therapy emerges the victor in these contests, given that the comparison groups are pseudotreatments such as nondirective group therapy rather than the widely used forms of dieting (e.g., Weight Watchers, Overeaters Anonymous, self-imposed diets). There is renewed cause for hope, however. The most recent generation of behavioral programs can produce substantial weight losses, apparently because programs are both longer and better. This paper proposes specific methods for testing the limits of treatment for obesity, in hopes that greater initial losses will be accompanied by more effective approaches to maintenance. In addition, we propose that more attention be focused on matching individuals to treatments, so that professionals serve a dual role of providing their treatment to those who are best suited and referring others elsewhere.