The art of uncertainty How to navigate chance, ignorance, risk and luck

Spiegelhalter (The Art of Statistics), a statistics professor emeritus at the University of Cambridge, delivers a stimulating survey of the myriad ways humans have attempted to quantify the unknown. "Probabilities are subjective judgments," Spiegelhalter contends, pointing out that calculating them requires deciding what kinds of information to include in a dataset and which situations count as positive outcomes. Exploring how people have strived to predict the future with statistics, Spiegelhalter describes how in the 1690s, English astronomer Edmond Halley reviewed data on the frequency with which people died at various ages to tabulate how much the English government should charge for annuities, and how in the late 1980s, the European Centre for Medium Range Weather Forecasts developed a prediction system that involved running weather modeling software 50 times under slightly different starting conditions to see which meteorological phenomena were most likely to occur. Elsewhere, he expounds on calculating coincidences, noting that the odds that a monkey typing at random would produce the complete works of Shakespeare is equivalent to "winning the lottery every week for 20,000 years." Spiegelhalter's explanation of Bayesian statistics--which, at its simplest level, makes contingent predictions that are updated in the face of new evidence--is among the most accessible readers are likely to find, and the case studies effectively ground the mathematical discussions. This is a sure bet. (Mar.)

Probability with a twist. Spiegelhalter, professor of statistics at Cambridge University and bestselling author ofThe Art of Statistics, emphasizes that he will deliver facts about uncertainty, avoiding matters of debate such as the best Beatle song, what to wear this evening, or the existence of God. Statistics, rules, and even equations come thick and fast, but few readers will miss his warning: Never trust words to express uncertainty. Many times this century, governments have raised the terrorism threat level, warning that an attack is "likely." The number of attacks that followed--zero. Few agree on the meaning of "unlikely,possible,likely,probable,rare, and so on," but a number is straightforward. Having set the scene and perhaps delivered more information than readers want to know, Spiegelhalter provides his expert opinion on simple subjects that turn out to be not so simple. Coincidences turn out to be common. Another long chapter discusses luck, which one expert called "the operation of chance, taken personally." Still another points out that most people who die in car accidents are wearing seat belts, yet belts remain lifesaving. This is proved by Bayes' theorem, which demonstrates how additional facts can make sense of nonsense. Random numbers are essential in many fields outside of statistics, but they're extremely difficult to achieve; most readers can't produce them. He concludes with a long discussion of making decisions and managing risks--perhaps his most practical chapters. Readers should not expect strategy for winning at Las Vegas or the office election pool. Spiegelhalter admits frankly that many concepts are tough going, and readers whose mathematics classes include a taste of statistics will have an easier time. A satisfying look at the nature of uncertainty. Copyright (c) Kirkus Reviews, used with permission.