Francis Galton

Francis Galton is presented as the nineteenth-century thinker who first made regression to the mean intelligible. A half cousin of Charles Darwin and a renowned polymath, he recognized that unusually extreme results tend, on a subsequent measurement, to move closer to the average—not necessarily because anything has caused them to improve or deteriorate, but because outcomes combine persistent factors with luck or other variable influences. His work supplies the historical foundation for a central warning about judgment: people readily mistake statistical regression for evidence of punishment, reward, treatment, or talent.

Identity and Discovery

Galton discovered and named regression to the mean late in the nineteenth century, roughly two hundred years after the theory of gravitation and differential calculus had already been developed. The delay underscores how unnatural the phenomenon is to ordinary intuition. Regression is common wherever measurements are imperfectly related, yet it tends to disappear into plausible stories about causes and intentions.

His discovery began with hereditary measurements. In his 1886 article “Regression towards Mediocrity in Hereditary Stature,” Galton described experiments on successive generations of seeds and comparisons between the heights of children and their parents. Large parent seeds tended to have smaller offspring, while very small parent seeds tended to have larger offspring. He reported that the mean filial regression toward mediocrity was directly proportional to the parents’ deviation from the mean. What first appeared to be a striking biological observation gradually became, through years of work and with help from the leading statisticians of his period, a general account of imperfect correlation.

The Statistical Principle

Galton’s decisive insight was that correlation and regression are not separate concepts but different perspectives on the same relationship. Whenever the correlation between two scores is less than perfect, regression to the mean will occur. The direction is predictable even when the individual outcome is not: an extreme score on the first measurement tends to be followed by a less extreme score on the second.

The principle becomes clearer when an outcome is divided into durable and temporary components. In a golf tournament whose average score is 72, a player who shoots 66 on the first day has demonstrated both above-average talent and better-than-average luck. A player who shoots 77 may be relatively weak and may also have had an unlucky day, although either inference can be wrong in an individual case. When predicting the second day, the player’s talent is expected to persist, but the luck attached to the first day cannot be predicted; the rational forecast therefore moves toward average performance rather than simply repeating the first score.

The same logic explains why a later jump after an unusually poor jump may be better without any effective instruction, and why a later jump after an unusually good jump may be worse without any damaging intervention. The change does not require a causal explanation. It follows mathematically from the fact that luck contributed to the initial result. Galton’s importance lies in making this impersonal mechanism visible where observers instinctively search for a human cause.

Influence on Judgment and Prediction

Galton’s principle illuminates the misleading experience of training instructors in the Israeli Air Force. Instructors observed that cadets often performed worse after being praised for an especially clean maneuver and better after being reprimanded for a bad one. They concluded that punishment worked and reward failed. Regression offers a different explanation: praise and punishment followed unusually good or bad performances, which were partly produced by fluctuating factors. The next performance was therefore likely to move toward the average regardless of the instructor’s response.

The same corrective is applied to predictions about reading ability and college grades. A child’s early reading age and later GPA share factors such as genetically influenced aptitude and family support for academic interests, but each outcome also contains factors specific to that outcome. A precocious reader may later receive lower grades because of an unhappy love affair, an adolescent skiing accident, or another event unrelated to early reading. Prediction should therefore preserve the shared component while reducing the influence of outcome-specific and uncertain factors.

Galton’s lesson also bears on expert forecasting. In Philip Tetlock’s twenty-year study of 284 professional commentators and advisers, more than 80,000 predictions were collected about political and economic events. The experts performed worse than an equal-probability baseline, and even in regions they knew best they were not significantly better than nonspecialists. Knowledge improved forecasting only slightly, while the most knowledgeable and famous forecasters often became less reliable because knowledge strengthened their confidence more rapidly than it strengthened their predictive accuracy.

Intellectual and Thematic Significance

Galton’s work changes the meaning of apparent improvement and decline. It warns against crediting an intervention merely because a bad result is followed by a better one, or blaming a punishment merely because a good result is followed by a worse one. The statistical movement toward the mean can imitate a story of learning, discipline, treatment, or decline while remaining entirely independent of those explanations.

His discovery also exposes the broader human tendency to prefer causal narratives to mathematical inevitability. Extreme outcomes invite explanations, and observers commonly infer stable talent, stable weakness, or deliberate influence from single performances. Galton showed that the less-than-perfect relationship between measurements is itself enough to generate the pattern. Regression therefore functions in the book not only as a statistical concept but as a discipline of intellectual humility: it separates what an outcome reveals about enduring ability from what it owes to temporary luck, and it makes the limits of prediction impossible to ignore.