Understanding statistical mechanics for a non-physicist I study statistics and not physics, but I am interested in the role probabilities have within physics, e.g. within classical mechanics. I wonder whether it is possible to understand classical mechanics (or any other physical field that is based on probabilities) with basically no pre-knowledge of physics? And if so how would you recommend me to approach this topic. My main interest is the philosophical interpretation of probabilities (are probabilities objetive, etc.) and I am currently reading the book "Probabilities in physics" (2011) which is sometimes hard to follow without an understanding of quantum mechanics, statistical mechanics, etc.. What I hope for is that an understanding of statistical mechanics will improve my understanding of probabilities. 
So my question is basically which books would you recommend to achieve this goal? Or which topics would you reccoomend me to try to understand?
 A: The problem is that an undergraduate statistical mechanics course will typically meld with a lot of thermodynamics and thus touch on very conventional physicsy analysis -- things like the kinetic theory of gases, reversible engines, defining the entropy as $\int dQ/T$, which might not suit the path that you're trying to go down. Advanced statistical mechanics courses are liable to be more mathematical but they also require a more high-level view of physics which may be more accessible but maybe not. For example, one needs to introduce the Lagrangian and Hamiltonian pictures of physics in order to derive a key result called Liouville's theorem.
With that said I do happen to know a nice set of lecture notes on advanced statistical mechanics from a professor I had at Delft named Jos Thijssen; right now apparently a copy is available at this link, that might be up your alley. 
A: I would say statistical mechanics and quantum mechanics are the two big fields in which probability plays a big role in physics. They also exemplify two opposing situations: in statistical mechanics, probabilities are basically the same as in a game of cards, born out of chaotic (but deterministic) shuffling of a complex system, except the numbers involved are so big that probabilistic laws are effectively fully accurate descriptions of the system. In other words, if you tossed one billion billions coins, it would be pretty natural for almost exactly half of them to come out as head. Quantum mechanics, on the other hand, seems to feature probabilities as an ontological property of nature. Not just a measure of our ignorance, but a genuine feature of the laws of the universe - though the jury's still out on that one.
I'm not an expert in good books for laymen to approach these concepts. I've read multiple ones in which they were discussed, mostly quantum mechanics at least, but I can't say any one struck me as better. If you're willing to go a bit into the formalism as well - which can only help you get a deeper understanding of the ideas - then I'd suggest you check out The Feynman Lectures on Physics, which are an amazing introduction to almost all core topics of physics.
