Probability concepts permeate physics. This is obvious in statistical mechanics, in which probabilities appear explicitly. But even in cases when predictions are made with near-certainty, there is are implicit probabilistic assumptions in play, as it is assumed that molecular fluctuations can be neglected. How are we to understand these probabilistic concepts? This book offers a fresh look at these familiar topics, urging readers to see them in a new light. It argues that the traditional choices between probabilities as objective chances or degrees of belief is too limiting, and introduces a new concept, called epistemic chances, that combines physical and epistemic considerations. Thinking of probabilities in this way solves some of the puzzles associated with the use of probability and statistical mechanics. The book includes some history of discussions of probability, from the eighteenth to the twentieth century, and introductions to conceptual issues in thermodynamics and statistical mechanics. It should be of interest to philosophers interested in probability, and to physicists and philosophers of physics interested in understanding how probabilistic concepts apply to the physical world.