That fundamental ideas in applied mathematics would be developed in a brewery sounds sufficiently improbable, but the story is true and intriguing. The statistical technique most often used to study events of low probability was discovered by a Polish mathematician and an employee of the Guinness brewery.
It is one year since the tsunami struck south-east Asia. And it is the way of economists to interpret tragedy in terms of statistics and to describe human behaviour by abstract models. That was how at this time last year my mind turned to the ways we measure and analyse such things – the mathematical analysis of catastrophic events of low probability.
I still have a sense of wonderment that statistical methods can be applied so effectively to these issues. Processes that are immensely complex and inadequately understood – such as the eruption of volcanoes and the shifting of the earth – can be well described, though not predicted, by abstract mathematical relationships. The classical approach to extreme events such as the tsunami uses a technique called the Poisson distribution which describes the incidence of very low probability events. This was simultaneously discovered by two people about a century ago. One was a Polish mathematician who used it to explain the incidence of deaths from horse kicks among the Prussian military. He did so from his university office, without the slightest knowledge of German army training.
But the activities of the other inventor of this brand of statistics were as mundane as those of Professor Ladislaus Bortkiewicz were rarefied. The Poisson distribution was also discovered by an employee of the Guinness brewery in Dublin. When I noted this in the column prompted by the mathematics of the tsunamis, the editor found the idea that fundamental ideas in applied mathematics would be developed in a brewery sufficiently improbable that he asked me to check. The story is both true and intriguing.
This pioneering account of the Poisson distribution was published under the pseudonym “Student” and the anonymous author was one of the most important contributors to modern statistical theory. The t-statistic, probably the most widely used test of statistical significance, was for many years known as “Student’s t” after the man who first described it.
We now know that “Student” was William Gossett, who spent his entire working career with Guinness, eventually becoming managing director of its UK operations. Some accounts suggest that Gossett used a nom de plume in order to conceal his academic activities from the controlling Guinness family. They had prohibited their employees from publishing any material about the company’s affairs after a worker, his tongue loosened by the company’s products, revealed details of their brewing process.
But in fact, Gossett’s identity was known not only to fellow statisticians but to his employer – the company +insisted on the pseudonym so that it could turn a blind eye to the breach of its rules. And Gossett’s ideas were not the idle musings of his leisure hours. The Poisson distribution was discovered through observation of the process of fermentation in beer. The t-statistic is a measure of how much confidence can be placed in judgments made from small samples. Gossett invented that measure to enable the quality of brews to be monitored in a cost- effective manner.
Scientific brewing began in Europe in the 19th century. The great French biologist Louis Pasteur published a volume of studies on beer: Copenhagen and Munich were world centres in brewing technology. But Ireland’s family-controlled brewing industry was slow to follow. Claude Guinness was, however, an innovator. Educated at Winchester and New College, Oxford, Claude set out to recruit the best science graduates from Oxford and Cambridge. It is difficult to conceive how radical it then was for a manufacturing company to hire the best scholars from the oldest universities. Not until the 1960s did such businesses, which routinely stressed practical skills over theoretical knowledge, undertake the milk round of the campuses.
By the outbreak of the first world war, Guinness had become the leading company in the British Isles in the application of biochemistry and statistics to industrial processes. Between the wars, the men who had graduated from ancient seats of learning to the brewery graduated further, to the executive suite: Oxbridge men hired in Claude’s time came to dominate senior management roles in the company. One of the first to be recruited, T. B. Case, became managing director of the whole business. Gossett led the company’s expansion into the mainland and its Park Royal brewery opened in London in 1935.
Guinness became the first multinational business in what is an industry still largely organised on national lines, although an unquenchable thirst for the stout of the homeland among the Irish diaspora was as important as the ambitions of the company’s management in promoting global expansion.
The marketing campaign the company conducted using the unprovable, and today illegal, slogan “Guinness is Good for You” is still a landmark in brand advertising. Gossett devised econometric techniques to assess the effectiveness of their spending.
Illness forced Claude’s early retirement from the family business. His recruitment strategy was a phenomenon of one generation only. After 1945 and the retirement of Claude’s men, the company went into decline. Only in the 1980s did the Guinness family again look outside for fresh talent. The man they found, Ernest Saunders, did not match Gossett in either intellect or ethics: he merged the eponymous brewer with Distillers, into what is now Diageo, before resigning ahead of trial and imprisonment for illegal share support.
How much did Guinness customers and shareholders gain from Claude Guinness’s search for the best minds? Claude’s people were not only intellectually distinguished but effective. Although the business may have been better at developing these applications than using them, Guinness was for a generation probably the best-run firm in its industry. The Guinness family and their company certainly incurred more benefit than cost from indulging the academic leanings of the man still best known to statisticians as “Student”.