One of the challenges in anyone’s life, especially in one’s job, is figuring out what’s really going on. For me, and anyone involved in the financial sector, this is especially complicated because there is an infinite amount of data, interpretations, arguments, and conclusions that will tell you up, down, right, left, buy, and sell—all at the same time.
The physical sciences have been dealing with this problem for a long time and refer to it as isolating the signal—the meaningful information you need—from the noise, the random events that can mask the signal.
Nate Silver is the statistician behind the popular blog www.fivethirtyeight.com, which was one of the best-performing political prediction engines during the past election cycle. His expertise extends well beyond politics, though, to include baseball, where he developed a pioneering database, and online poker, where he made his living for several years.
Silver’s real interest, which knits these seemingly disparate career paths together, is how to characterize uncertainty and how to isolate usable information—signal—from noise. He has done so successfully in the three domains mentioned, and he wrote The Signal and the Noise to describe his process and how it has been used in different fields.
I was hoping for insights into how Silver applies his methods to politics, a notoriously noisy data environment. I had no expectation that this would be a textbook, but I did look forward to some details that might translate to my own career.
In that regard, I was a bit disappointed, but I still enjoyed the book and found it worth reading. Silver does a very good job of walking us through the history and current state of several different realms: weather forecasting, politics, baseball, earthquakes, and economics. He outlines how previous forecasters developed models, how they failed, and how they’ve been improved. He also provides intellectual context in the form of a good explanation of Bayes’ theorem, the increasingly accepted statistical model that underlies his own work.
Readers will finish the book with a solid conceptual understanding of how these methods are used in various fields. What you won’t take away is detailed knowledge of how the actual prediction process works.
An in-depth explanation of how Silver’s predictions for the last presidential race evolved in real time, for instance, and how his models changed at each stage of the race, would have moved the book from good to great. By including real data and processes, he could have given the reader a much better understanding of how to put these ideas into practice.
For anyone interested in the big picture, this is a worthwhile read—very good popular science, with an emphasis on the popular. For those hoping to peer into the head of a justly celebrated analyst, you won’t find what you’re looking for.
Verdict: Recommended, but understand what you’re getting.