The attraction of this book is its discussion of the philosophical foundations of game theory. Those parts are useful and worth a read (particularly Section 1, and to some extent Sections 6 and 7). However, the book is plagued by quality problems, and I was disappointed by it. Exercises were added in the Second Edition to enable use as a general text in game theory, but the exercises are an afterthought -- few and inadequate. Also, the quality of the writing and explanations is generally serviceable, but often poor. For example, subgame perfection and backward induction (Section 3) are simple, straightforward concepts when explained properly, but the explanations in this book are completely botched and incomprehehensible. If you're learning game theory, my advice would be to learn the fundamentals from a standard text.
Also, there is a major disconnect between the book's grandiose description of game theory and the reality. The description: "[Game theory's] practitioners have great designs for it, claiming that it offers an opportunity to unify the social sciences and that it is the natural foundation of a rational theory of society." That's a laughable characterization of the powers of game theory. Game theory, at best, is a collection of interesting, but extremely skeletal, mathematical ideas that say almost nothing about the social sciences or human rationality. This is a case of Taleb's "Ludic Fallacy" gone wild -- "the misuse of games to model real-life situations."