It has become common, in works on the foundations of probability, to recognize that the word “probability” has been used in two senses: credence, having to do with the degrees of belief of a (possible idealized) agent, and objective chance, which is thought of as being a matter of physics. On the assumption that the fundamental laws of physics of physics are deterministic, a number of writers, from Bernoulli onwards, have rejected objective chance and have adopted a conception of probability on which all probability is epistemic. In spite of this, it seems that, even if the laws of physics are deterministic, it makes sense to ask whether a coin toss or a roulette wheel is fair or biased. In this talk, Myrvold will argue that this is not an illusion, and that there is a need for a notion of probability that goes beyond the familiar dichotomy of credence and chance, a hybrid notion that combines epistemic and physical considerations. Myrvold will make a case that this notion is the one appropriate for both statistical mechanics and traditional games of chance.