I normally get through poker books as quickly as I can, but this one's been sitting on my shelf for a while. It looked a bit foreboding, and I though the maths involved might be difficult. This suspicion was enhanced by discussion of Bayesian analysis and maximum likelihood methods in the introductory section, which is certainly more proper maths than I'd seen in a poker book before.
My fears were pretty much unfounded. After the introduction, there wasn't much more of that kind of thing. More maths than your average poker book, but nothing that'd challenge even a first-year undergrad.
The first couple of sections are introductory - the first being a maths warm-up, the second a kind of wishy-washy fake game theory. The third section is proper game theory. A variety of toy games are extensively analysed. It gets a bit monotonous, and some of the cases in a maths game theory book would have been just left as exercises, but it does build up good insight into the real game. Perhaps the best chapter of the book is the final one of this section, which covers the application of the game theoretical ideas to real poker play.
The fourth section covers risk management - risk of ruin, Kelly criterion, etc. It's a bit of a gear change, and pretty straightforward, but it's good to see it, pulling in ideas more regularly seen in financial maths. The final section covers tournament play, game theory with more than two players, and a final summary.
It's a good book. The authors clearly know their maths, and are trying to present it as clearly as possible to a less mathematical audience, but at the same time are quite willing to go into a lot of depth. The presentation is a real let down, though - a small font and tiny margins crams maximal amount of text onto each page, at the expense of legibility, and the maths typos are too numerous to mention, so you'll need to basically re-derive the equations yourself to make sense of what's written! In short, the authors claim their approach is the future of poker, and they make a convincing case.