The setpoint model of human metabolic adaption to diet-induced weight loss has been introduced and examined considering a constant energy density in the body weight lost. Body composition analysis indicates that the ratio of fat and nonfat lost is constant for an individual but is dependent on the initial percent body fat, producing an energy density varying from individual to individual. In this study, body composition and the setpoint model are used to examine weight loss dynamics. Comparison is made to a Harris-Benedict (HB) based model using a nominal energy density (3500 calories/lb). Conclusions: the weight dynamics for a given reduction in calorie intake are completely determined by the initial percentage of body fat. The HB model is predicted to accurately track the weight loss for the very obese who experience little metabolic adaption and whose energy density is close to the nominal value. However, for subjects with decreased initial body rat percentage, the HB model first underestimates then overestimates the weight loss. The crossover time and the maximum overestimate is dependent on the initial body fat percent and the percent dietary calorie reduction. A characteristic time and an energy density ratio are defined and used to calculate a maximum rate of weight loss for a given calorie reduction.