The Signal and the Noise

Nate Silver, who built a system for forecasting baseball performance and later founded FiveThirtyEight, asks why most predictions fail and what sets apart the few that succeed. His answer is that people misjudge probability and uncertainty, and that experts and amateurs alike confuse confidence with accuracy. Silver visits forecasters in fields from weather and earthquakes to poker, elections, the stock market and basketball, and uses case studies including the 2008 financial crisis, climate change and weather forecasting, the last of which introduces the idea of calibration. He makes the case for Bayesian reasoning, in which probabilities are revised step by step as new data arrives. The 2012 book received the Phi Beta Kappa Society award in science the following year.

Bill Gates put it on a year-end reading list, noting that its subject reaches well beyond politics and that it came out before the election Silver forecast so accurately.