How to predict Olympic results before the games even start

From The Washington Post:

Just two weeks before the opening of the 2016 Rio Olympics, athletes are training furiously, tourists are packing their bags, and researchers are fine-tuning their predictions.

Predicting the number of medals that each country will bag at the Olympic Games has long been a past-time of betting markets. But researchers at the Tuck School of Business in Dartmouth have also been known to get in on the game. Earlier this week, Camila Gonzales, a graduate of the school, released predictions for how many gold medals and total medals each country will win in the 2016 Olympics, based on a prediction method developed by Tuck professor Andrew Bernard and another professor more than 10 years ago.

Their method of forecasting a country’s Olympic success is incredibly simple, but the researchers say it’s also the most accurate model out there. It predicted the total medal count for the London Olympics with 98 percent accuracy, they say.

For 2016, their model predicts that the U.S. will come out on top, winning 105 medals overall, ahead of China with 89, the United Kingdom with 67, the Russian Federation with 62 and Germany with 48. They also predict that the U.S. will lead the gold medal count with 48, followed by China with 38.

What’s fascinating about their method is that it doesn’t rely on any knowledge about the sports or athletes themselves. Instead, it mostly comes down to economics.

They forecast medal totals based on only four factors: A country’s population, its comparative level of wealth, its performance in previous Olympic Games, and whether it is hosting that year’s Olympic Games.

A country’s population is important because the larger it is, the bigger the potential pool of talented athletes that country has to draw on. However, population alone isn’t the key — it’s not as if …

Continue Reading