Predicting The Winter Olympics

For most of the world, the Winter Olympics is a spectacle of strange sports (curling) and even stranger names (luge). For the host Canadians, it's a rare and glorious shot at world domination through sacred winter sports like hockey.

So when economist and Olympic prognosticator Daniel Johnson predicted Canada is poised to be top dog, you might expect the country to be thumping its collective shoulder pads in celebration. But hold off on the celebratory poutine. Medal predictions are tricky, perhaps more of an art than a science, and it's not clear that economists or anyone else using objective data are best suited for the endeavor.

Does home field advantage matter? Do you take individual sports and athletes into account, or look at broad demographic trends?

The most important number, by far, is a country's past medal count. That is the benchmark of all Olympic medal predictions, and it's often hard to beat. (Here's a handy interactive graphic for medal winners at all Summer games since its inception in 1896, and here's a less glitzy database for both Games.) Putting aside boycott years like 1980 and 1984, not much changes, and when it does, it's instantly noticeable. China had a massive jump in their medal haul from 2004 to 2008, mostly due to enormous infrastructure improvements and vast amounts of resources that the country pumped into athlete developments for the Beijing Games.

Daniel Johnson, who is himself a Canadian with a reported love for curling, doesn't factor individual sports or athletes into his calculations. Johnson uses per capita income, population, political structure, climate, and home field advantage to construct his model. (Yes, the latter is statistically significant: the host country earns 1.8% more medals than it would otherwise.) Johnson's predicted results? Twenty seven medals for Canada, followed by 26 each for the United States and Norway.

His previous accuracy has been highlighted by the press, but it's a bit misleading. For example, Johnson claims he predicted the total number medals by country in the 2008 Beijing games with 93 percent accuracy, but simply using the Athens totals as a control predictor will beat Johnson's predictions for eight of the top ten countries. And of those other two countries, his U.S. prediction (103) beat the Athens number by one measly medal, and foreseeing China's vast improvement hardly merited a distinguished prophet award.

There are various ways to valuate Olympic success. Some use the total number of gold through bronze, and others only the gold medals. Given the competitive spirit of the games, it's best to stick to bragging rights: who beats whom in the rankings of total medals won. For instance, even if the US wins fewer medals than they did in Turin, but are still ahead of Canada, Germany and the others, it will be hailed as a victory. By this measure, Johnson's work on the Beijing games is even less impressive: he only nailed the top three - US, China, and Russia, which are as close to no-brainers as you can get.

To measure the boldness in Johnson's Vancouver predictions, an instructive rubric would be the major departures from performances in Turin in 2006. That means his calls on Sweden, Norway, and Germany (changes of +7, +11, and -9, respectively) are probably more significant than his forecast of a three-medal Canadian improvement. Still, Canada has never been the top medal-earner so he deserves some credit for boldly predicting a win for the home team.

If you can't trust an economist, who can you trust? The market, obviously. Betfair, a British prediction market, has Germany as the 5/4 favorite for the most medals, Canada second at 12/5, and the US is third, with odds of 9/2. Hubdub, a prediction market that uses virtual money, says Canada has a 52 percent chance to finish with the most medals, followed by Germany (29%), and US (18%).

There are other methods to predict Olympic medal counts. Individual sport by sport analysis is the more traditional approach. USA Today predicts Canada will come out on top with a whopping 34 medals, which would be one of the greatest Olympic improvements in modern times. The Associated Press made its own forecasts for individual sports, and has Canada on top with 30 medals, followed by Germany and the US with 27 apiece.

The oddsmakers aren't shying away from making predictions, either. Ladbrokes has Germany on top, just ahead of Canada, with the US a distant third.

Canadians should be cautiously optimistic with these Olympic predictions. But, as a Canadian, I can tell you we would gladly forgo the most medals or golds for the only thing that really matters: hockey supremacy.