The Physics of Wall Street: A Brief History of Predicting the Unpredictable

Weatherall, a philosopher of physics, traces the movement of physical and mathematical ideas into finance through a series of biographical chapters. He starts with Louis Bachelier's 1900 thesis modelling prices as random walks, then covers Maury Osborne applying statistical mechanics to log returns, Benoit Mandelbrot arguing that price changes follow fat-tailed distributions rather than normal ones, Ed Thorp moving from blackjack to convertible arbitrage, Fischer Black and the option pricing formula, Doyne Farmer and Norman Packard's Prediction Company, and Didier Sornette's work on log-periodic precursors to crashes.

The argument the book makes is a defence. Weatherall holds that mathematical models did not cause the 2008 crisis, and that the failures came from using models outside their stated assumptions, from institutional incentives, and from ignoring what the modellers themselves said about limits. He points out that Mandelbrot and Black both wrote explicitly about when their models would fail. His constructive proposal is that economics needs the physicist habit of treating a model as an approximation valid in a specified regime, and he ends by calling for a large coordinated research effort.

The history is accurate and readable and the biographical detail is good, particularly on Bachelier's neglect and on Thorp.

The defence of the modellers is contested. Critics including Mandelbrot's own followers and Nassim Taleb argue that the problem is not misuse but that Gaussian-based risk models are structurally wrong for financial data, and that the profession's continued reliance on value at risk demonstrates it. The book also gives little space to the political economy of why banks preferred models that understated risk.