Anecdote: When I first read Feynman's _Statistical Mechanics_
book, I was mystified by the discussion starting on page 2,
i.e. the derivation of the Boltzmann factor. I just figured
that was because I wasn't smart enough to follow the argument.
I suspected that the result was true, but the argument was
missing some steps.
After many years, I came to realize that not only is the
argument full of holes, the conclusion is not entirely
true.
The discussion in Kittel & Kroemer (at the beginning of
chapter 3) analyzes the same Gedankenexperiment. The
discussion is better, but still leaves much to be desired.
There are serious conceptual issues that are not even hinted
at ... including questions about what we mean by "temperature"
and "thermal equilibrium".
I found I understood it better after working out some
numerical examples.
Beware: This topic involves some deep conceptual issues
*and* a host of devilish details. I wouldn't go anywhere
near this topic if it weren't so very important.