Overthinking: Binomial maximum entropy (errata) p. 312

[rxg]

I think this is a typo:

If we now set the expected value of q equal to λ, then H(q) = H(p)

I think it should be:

If we now set the expected value of q equal to λ, then H(p) = -H(q, p) ≥ H(q)

this is more in line with the strategy used for Gaussian, and Figure 10.4 makes clear that not all candidate distributions (with the same expected value) have the same entropy