I finished reading this back in mid-January but have only now found the time to write it up. I've forgotten why I picked it up, probably on the strength of a blog review or something.
This is a very Americentric take on probability and decision theory, with a smattering of public choice, game theory and random other things. Math is almost absent, so there is almost no support for all the "trust me"s the author throws in. A bibliography would have ameliorated this. There are some good pop-sci treatments of various things, but it ends up being a ramble with too much opinion and not enough evidence. The ultimate advice is formulate-then-compute, and stick to it.
The bit I enjoyed most was about Lanchester's laws, which model how two opposed army-like things inflict damage on each other as a function of time. There is a good writeup of the math here.
His take on Arrow's Theorem is a bit naive and uninteresting; a more insightful approach would have furnished some perspective through the later work of Amartya Sen (and many others) or perhaps May's Theorem, but clearly there's more table-thumping to be had in banging on about how impossible voting is.