We have had many reminders of the limits and risks of statistical predictions about human behavior. When I ventured last year to offer a formula describing Supreme Court action in the state “right to counsel” cases over a period of years, I was mainly concerned to show that, contrary to accepted judicial doctrine about decisions in unique cases, consistencies in the Court's collective judgments could be demonstrated in this area, and stated in quantitative terms. I did not suppose that the formula was more than a method of approximation. I did not offer it as the perfect formula, or the only one; and in particular I did not claim that it was capable of indefinite extension to other areas of constitutional law. An exploratory effort begins with what is available. But I did take into account, so far as the available data permitted, not only the Court's decisions, but also the votes of individual justices on each of the pivotal factors on which their positions could be identified or imputed. I did not claim more for the product than that, so far, it works. With this general statement, let me turn to Fisher's article.