Critique of: “Ranking Procedures and Subjective Probability Distributions”

Professor Smith's article “Ranking Procedures and Subjective Probability Distribution” (Smith, L. H. 1967. Ranking procedures and subjective probability distributions. Management Sci.14 (4, December) B-236–B-249.) is a worthwhile attempt at making the Bayesian decision theory concepts more operational; however, he has fallen far short of this objective. The example he chose (obtaining a subjective probability distribution on market share) is in many ways inappropriate for Smith's technique, and although his procedure is more “mathematical” than some existing methods, it is not as relevant. There are some relatively minor points in Smith's article on which one might quibble, but there are also enough major, substantive matters that we will direct our critique to these issues.