Bayesian Inference & Occam's Razor

In chapter 28 of his book Information Theory, Inference and Learning Algorithms, David MackKay uses an example of a box behind a tree as a demonstration of the embodiment of Occam's razor in Bayesian inference.

What's behind the tree? One box or two? What is the basis for the belief?

I really enjoyed the discussion and his example. The book is available online as a PDF, and the example begins on page 343.

link