99 YEARS OLD AND STILL SPORTINGLY STATISTICAL...

THANK goodness the Olympic Games are over at last. All that running, swimming, throwing, cycling and the rest of it left me quite exhausted. Now that it has finished, however, I have been able to get on with the important business of submitting the results of a proper statistical analysis which demonstrates conclusively that all the other analysts have got it completely wrong.

America did not come out best, as the medal table suggests but were, in fact, one of the worst performing nations. Allow me to explain.

As we all know, Olympic success demands a similar commitment by both the individual athletes and the nations they represent as success in any other field. It came as no surprise that the leading gold medal winners were the US, UK, China, Russia and Germany, for those are the countries that tend to come out on top in everything. But what, I asked myself, are the primary predictors of Olympic success?

My first thought was chocolate. As you probably know, The New England Journal Of Medicine in 2012 published a paper showing a strong correlation between a nation’s chocolate consumption and the number of Nobel Prizes it has won. I was disappointed to find, however, that chocolate consumption was only a very weak predictor of Olympic success.

On the other hand, Nobel Prizes do correlate quite strongly with gold medals. After all, the two top nations in the medal table, the US and UK, are also the two countries that have won most Nobel Prizes.

So if chocolate correlates with Nobels and Nobels correlate with gold medals, why doesn’t chocolate correlate with gold medals? There had to be another factor involved and I was determined to find out what it was.

I tried comparing gold medals with population size and per capita GDP (a good indication of average earnings) but obtained only a weak correlation in either case. Then I tried total GDP, which produced a much stronger result. So I set about finding the linear combination of GDP and Nobel Prizes that gave the best prediction of gold medal success which resulted in the following formula:

G = 3.0 + .05N +1.6M which predicts G (the number of gold medals won), from N (total number of Nobel Prizes won by the country) and M (the GDP in trillions of dollars).

For reasons of pure laziness, I limited the analysis to the leading 40 nations in Rio, plus a few selected for their Nobel or GDP success (which is why the formula always predicts at least three gold medals).

The formula gives predictions, for example, of 22 golds for China, 11 for Japan, 10 for France and 7 for Italy, which are very close to the actual figures of 26, 12, 10 and 7.

But America fell five short of its expected figure of 51 golds, which under the formula is worse than any country other than India, which won none against a prediction of seven.

And the country that did best of all, was Team GB, with a total of 27 golds compared with a prediction of 14. Forget America, we won the Olympics!