The Drunkard's Walk

The Drunkard's Walk: How Randomness Rules Our Lives illustrates the difficulties of writing a popular book on statistics: before getting to one's point, one must spend several chapters introducing concepts in successive layers, to build a foundation for the reader to understand the eventual point. Leonard Mlodinow does this here, with the first nine of the ten chapters building the foundation for the tenth. Although such a narrative structure could prove awkward and threaten to capsize, each of the ten chapters here are engaging, filled with historical anecdotes, and clearly explain the concepts and tenets of statistics and probability. Unfortunately, the final, eponymous chapter, while the realization of all the concepts discussed previously, feels almost truncated, because since the author spent nine chapters getting us to this point, it feels like its assertions could be fleshed out over several more chapters, rather than being terminated where it is.

This is certainly a good starting point as a popular introduction to statistics -- although many of the concepts introduced are counterintuitive, they're explained and fleshed out relatively quickly and easily -- I particularly enjoyed the chapter on Bayesian probability in this regard. While not nearly as mathy as it could have been -- the mathematics of probability are glossed over (none of those hideous upside-down 'u' symbols) -- The Drunkard's Walk does feature paragraphs liberally sprinkled with numbers. The only issue here is that any math is free form, and some pages read like giant word problems. Rarely is everything tied together neatly with equations, which can make for necessary re-reading. However, as a popular primer, one would be hard-pressed to do better. (A slightly more in depth look at similar themes is available in Chances Are . . . by Michael and Ellen Kaplan.)