The Signal and the Noise: Why So Many Predictions Fail — but Some Don't

Silver surveys forecasting across domains — baseball, weather, politics, earthquakes, economics, chess — and catalogues the systematic ways predictions fail. The book's strongest chapter, on Bayesian reasoning, is a crash course in the mental move that separates useful forecasters from confident ones: updating beliefs with evidence instead of defending them. For product direction it pairs well with Duke's Thinking in Bets and Saffo's forecasting rules — three different angles on the same problem of acting under uncertainty. Silver is a clear writer with a data journalist's instinct for the telling example. More than a decade old and aged mostly well; the chapters on economics and epidemiology are the most dated, the chapter on Bayes still the most useful.