On optimality of the LR tests in the sense of exact slopes. II. Application to individual distributions