Key points are not available for this paper at this time.
Abstract The present paper deals with the following general type of problem. Assume that, for some decision purpose, at set of random experiments M is potentially available. From out of this set one is allowed to choose one or more experiments and perform them independently any number of times, with the only restriction that the total number n of actual experiments be given in advance. For a given decision purpose, then, the question is what experiments to choose and how many times to perform them.
G Elfving (1954) studied this question.