I picked up The Drunkard’s Walk due to some oblique recommendation—it may have been Amazon’s—since its likely treatment of probability and randomness promised to be both difficult as well as interesting. It wasn’t until I read the jacket flap and saw that Mlodinow had written for both MacGyver and Star Trek: The Next Generation that I decided I had to read the book1.
A drunkard’s walk (or random walk) is a playful scientific term for the phenomenon behind Brownian motion; that is, the random motion of molecules. More generally, it refers to any path which is the product of random movement, including, for instance, the finances of a gambler.
This book is Mlodinow’s attempt to make the relatively opaque world of probability more translucent to us uninitiated and doctorate-less schmucks. You may remember studying probability in middle school (I do); you may also remember being largely baffled by it: the only thing I took away from my class was how to calculate the number of unique possibilities in a string. For instance, a 5 character alphanumeric string consists of a combination of 26 letters plus 10 numerals. Therefore _ _ _ _ _ can be filled in 36×36×36×36×36, or 60,466,176 possible combinations.
Real probability is much more complicated than that textbook example, in large part because factors are rarely independent of one another, and the complex relationships between causal factors is hard to quantify. In fact, when Mlodinow illustrates some of the historical abuses of probability/statistics, turning a complicated real-world situation into an oversimplified textbook problem is usually the root cause.
Mlodinow, who writes books with Stephen Hawking, knows that you the reader aren’t smart enough to write books with Stephen Hawking; neither would you be smart enough to understand the book if he filled it with complicated equations or large charts of numbers. In fact, The Drunkard’s Walk will not teach you how to calculate probability (except perhaps the sort of simple example listed above); neither will it make you an expert on chi squares or p values. What it will do is destroy some of your preconceptions about, say, “lucky streaks,” which are neither lucky, nor streaks. Mlodinow shows, for instances, that “hot” sports team can have their seasonal performance replicated by a series of coin tosses; so many of the things to which we ascribe special significance are in fact a series of events which follow the normal distribution (the “bell curve”) for random events, and in which outliers (like lucky streaks) inevitably regress to the mean.
The book’s other focus is notable names in the history of statistics and probability, and in this way Mlodinow occasionally meanders into the pleasantly anecdotal, which is usually a welcomed break. Like Bill Bryson’s in his treatment of science history, Mlodinow intersperses these biographical snippets carefully and expertly; they don’t feel bolted on like Malcolm Gladwell’s attempt at the same in Blink; they feel like natural, narrative arms to his larger expository goal.
By the end, I was feeling a little disoriented, trying to mentally reconcile relationships I knew to be causal with Mlodinow’s deconstruction of so many of those same phenomenon. It’s perhaps a testament to the extraordinarily complicated nature of probability2 that even after reading the book, the subject’s opacity diminished only ever so slightly. But the book itself was an entertaining and informative read, and a good primer for anyone interested in the subject.