ABSTRACT
In this chapter, I propose to explain the way in which the frequency theory of probability has been superseded by the propensity interpretation and also by the ‘measure-theoretical approach’, as it is now often called. I shall try to show that, even though this approach supersedes the frequency theory, it also lends it a kind of post mortem justification; for the frequency theory becomes ‘almost deducible’ from the measure-theoretical approach. That is to say, within the measure-theoretical approach it can be shown that the probability is zero that we may accidentally hit upon a random sequence that violates the demands of the frequency theory of von Mises. In other words it is ‘almost certain’ that these demands will be satisfied. 1
