In the field of personnel economics, there are few opportunities to convincingly test for salary returns to specialization as against versatility. This paper performs such a test by modeling returns to performance measures associated with two different skills practiced by running backs in the National Football League. We find gains to both specialization and free agency with substantial predicted differences in returns for alternative skills. These differences vary across the salary distribution. In the top half of the salary distribution, model simulations show that specialists in either particular skill generate higher marginal returns than do versatile players.