[Distributed at a CEH drought seminar for the water industry, held at Wallingford on 21 June 1996. The reproduction is authentic down to the misspelling of millennium. Remember, this was 1996 ... sort of.]
Recall those heady days in in the last millenium when, after triumph in Euro '96, Terry Venables took on the job of grooming young talent to help Glenn Hoddle retain the World Cup when it was staged in England last year. But how did he set about picking the 9 to 12-year olds who were to become the Ince, Gascoigne and McManaman of 2010? And why did it end the way that it did? Ivachek Buck of the New Guardian reports exclusively.
In 1997, the Football Association (a governing body for football in England and Wales) secretly commissioned the Institute of Hysalaries (IH) to investigate the relationship between midfield strength and physical, social and environmental factors. Terry and Glenn helped choose the factors: Glenn was keen to include IQ and religious conviction, while Terry insisted that alcohol intake was important.
The research team gained access to data on a hundred midfield players, picked to represent the different standards of the professional game in England in the 1990s. In confidence, Terry and Glenn ranked the players according to their footballing ability. Although some players agreed to undergo special tests in laboratory conditions, it did not prove possible to obtain a complete database of all the experimental factors.
The researchers got round the problem by developing many different regression models to estimate the player's ranking from the information available. Surprisingly, "Lives close to London" did not prove to be a significant factor in any of the models. Of the many statistically satisfactory models, 20 were selected that were both conceptually reasonable and provided a range of ways of estimating midfield ability, according to the information available.
After much additional research - including the development of advanced models to forecast how youngsters would respond to grooming by Blair's British Sports Scholarship scheme - Terry's assistant, Ellen Saccharin, was finally in a position to run through the electronic applications of more than 10000 young hopefuls. The results she presented were staggering: no fewer than 50 candidates were predicted to reach rank 1 (Gazza) level. Terry swiftly bought more shares in Football-came-home (FCH), realizing that another English triumph in 2010 was inevitable.
So how did things go so badly wrong? It can at last be revealed that Ellen Saccharin was then in the pay of one of Venables' arch enemies. Having come across an obscure report on over-estimates of drought severity in the UK water industry in the mid 1990s, Ellen had secretly applied each of the 20 regression models for which she had the relevant information. Typically, she had been able to use at least five of the equations, and often as many as 12. The estimated rankings she reported to Terry had in each case been the highest found. Perhaps because the bias introduced was favourable to FCH, no one in the company had thought to question the estimates too closely. Presumably the regulatory authority was caught off-foot.
So, as we remember Scotland's famous triumph at the National Sports Stadium last year, let us lay the blame where it really belongs. Why didn't IH's Duncan Reed explain the drought severity assessment problem in layman's terms?