Minimum Cost Sequential Analysis: (An unpublished manuscript located in King's College, Cambridge–comprising 6 typed pages, 7 handwritten pages and 3 figures)

Abstract

This chapter discusses about Turing's paper titled, “Minimum Cost Sequential Analysis” that is an unpublished manuscript located in King's College, Cambridge. In this paper, Turing does not give an explicit citation, but he refers to Barnard's use of straight-line boundaries for the acceptance or rejection of a batch of items. In addition, Turing uses the notation p, as did Barnard, for the proportion of defective items in the batch. On the second page of Turing's paper, he gives examples to show that it is necessary to assume a prior distribution for p. He calls it “a priori” distribution as was common 50 years ago. Turing says that the solution to the problem of minimum cost sequential analysis will not depend very greatly on the form of the a priori distribution of p. In modern terminology, he is asserting robustness or insensitivity of the procedure with respect to the prior. After describing the basic logic, Turing discusses the indigestible algebra that is appropriate for the Turing archives.