This years Winter Olympics in Vancouver is finished. Norway had a very good year, and performed as expected. Did the nations perform as predicted by professor Daniel Johnson, and his formula?

About two weeks ago I wrote an article about Professor Daniel K.N. JohnsonDaniel K.N. Johnson and his formula for success. Together with a student he has analyzed the results from all the modern Olympic Games, and found a formula that tells us how many gold medals and total medals a nation is expected to win. This is based on social parameters such as GDP and population.

The formula predicted that Canada would win the most medals, while Russia would win the most gold medals. To make it interesting I placed a small bet on both these nations. Unfortunately, I lost my money. Let's take a look at the results:

These results can be interperted in many ways, but we see that as expected there is a positive tendency that the results are correct, with some that are clearly wrong. I got an email from Professor Johnson yesterday, as a respons to the article I wrote about this. It was a short text that he probably sends to everyone that has shown interest in his work.

You were kind enough to express an interest in my economic predictions of the Vancouver Olympics earlier this year, so I thought I would give you the first chance to critique them after the fact as well.

It was a fascinating year to watch and to be at the Olympic Games. Our model was 87% accurate on total medal counts, 81% on gold medals in particular. While that sounds pretty good, I’m pretty critical of the model’s performance this year. When you consider that it has averaged 94% accuracy over the past 5 Games (87% for gold medals), this year was clearly different. Data on past accuracy and predictions are online at http://faculty1.coloradocollege.edu/~djohnson.

On one hand, the predictions for the host nation were spot on (predicting 27 medals, actual count 26), as well as for Norway (predicted 26, actual 23), China (predicted 12, actual 11) and the Netherlands (predicted 10, actual 8).

On the other hand, if you consider some of the big surprises among the top 14 teams, the model is less impressive this year than it usually is. The US and Germany both outperformed the benchmarks by huge margins (by 11 and 10 medals respectively), while Russia and Finland fell far short (at 8 and 9 medals lower respectively). Why? We suspect that the US and Germany put disproportionate funding and effort into medal-winning success at these Games, but we’ll need sports analysts to discover the real answer. All that the economics model can say is that these four results were completely out of line with historical performance relative to socioeconomic resources.

In terms of gold medals, the model accurately predicted the counts for Austria, Sweden, Italy, Switzerland, Korea the Netherlands, all within 2 of the actual results. That’s of course not counting the 62 participating nations for whom the model correctly predicted zero gold medals.

However, Canada’s gold rush was completely unanticipated by the model. Having a history of not winning gold when they host the Games (a performance in the Summer Games of 1976, repeated in the Winter Games of 1988), Canada now holds the single-year record for most Winter gold medals by any nation, let alone a host nation, in the entire history of the Winter Games. There’s simply no way that a model could predict that sort of spectacular turnaround.

Who were the big winners? France was not predicted to win more than a few medals, and ended up with 11, two of them gold. Poland, the Czech Republic and Japan all ended up with five or more medals, representing really striking performances by those nations given the model. Australia, a Summer Games powerhouse, had its best Winter showing ever, with 3 medals. Kudos to athletes from those nations in particular, who clearly helped their nations to ‘outperform the resources available to them’.

If we consider rankings alone, the model is only 75% correct for the rankings of all participant nations, and only 67% for rankings among the top 12. So it appears that the model may need a little re-calibration for the 2014 Games in Sochi, and perhaps even before the 2012 Summer Games in London.

See you all there! Like me, I hope that you are celebrating these Games for the athletic excellence on display, while you celebrate the nations who surprise us with their ability to support that excellence via the funding and infrastructure that the model clearly shows to be important.

---Dan.

So there you have it. A summery from the man behind the formula for success. I would like to thank Professor Johnson for sending me this email, and for answering questions I have sent him.