Forecasting the Olympics
Published 20 August 2012
Professor Robert Fildes is Director of the Lancaster Centre for Forecasting and ex-President of the International Institute of Forecasters. In this opinion piece he asks how well the pre-Olympics forecasts fared.
So who won and who lost – at least in the various forecasting competitions seen at the London Olympics? For one, G4S lost – they needed to make a number of forecasts to get the security staff in place, and this they failed to do in a dramatic fashion. A poor forecast of those who would turn up (or was it just insufficient ‘stock’ of people backing up?) – definitely a booby prize. I wonder did they do any forecasting at all? The London Organising Committee got it right by having a vast over-capacity of volunteers, but that's an easier problem to solve.
The second area was who would turn up in those allocated seats – and again, the estimates were widely off the mark. And embarrassing to the London committee, their only serious mistake, although the various predictions of hotel occupancy were also much too high. This in turn led to poor pricing and the low numbers of casual tourists turning up.
The only area of success – though a mixed record – was the high-profile forecasts of the medal table. But what was being forecast? The total medal count? The number of golds? Or the position in the table? Every forecaster could perhaps claim some success. The medal prediction errors averaged around 10% with China being the big problem.
And the winning forecasters? The Wall Street Journal makes that claim and a quick check supports this. Their forecasters did it by using a disaggregate approach and included expert opinion. Other more aggregate econometric models that used GDP, population, previous performance, investment in sport and home advantage – amongst other variables – did less well. And predicting golds is obviously tough when so many favourites are pipped at the post.
Professor Robert Fildes
Robert Fildes is Distinguished Professor of Management Science at Lancaster University Management School and Director of the Lancaster Centre for Forecasting. He was co-founder in 1981 of the Journal of Forecasting and in 1985 of the International Journal of Forecasting. For ten years from 1988 he was Editor-in-Chief of the IJF and remains an associate editor. He was president of the International Institute of Forecasters between 2000 and 2004. In 1976 he wrote one of the earliest business forecasting textbooks. Though long out-of-print, many of its core ideas have survived the test of time. Robert is a Fellow of the International Institute of Forecasters and of the Operational Research Society.