Monday, July 16, 2012

Predicting Gold

Let's see if statistical modeling pans out.

The 2012 Summer Olympics won’t end until August 12, but the results are already in. Team USA will be the overall medal winner, although China will win the most gold. Russia will take third place and host country Great Britain will be fourth.

Statistical modeling. It’s that accurate.

The Tuck School of Business at Dartmouth has perfected this predictor, basing their model on surprisingly few factors: population, GDP, and home-field advantage. The most successful teams come from countries with a population base large enough to produce cadres of athletes, and with economies big enough to support them. More amorphous, but historically proven, is the host-country effect. Olympians are spurred on, apparently, when performing for their own countrymen. That’s what gave Greece an outsized 16 medals in the 2004 Athens games, and Australia 58 medals in Sydney, 2000.

It’s a simple concept with some powerful math plugged in. Tuck has fine-tuned the weighting for each variable, resulting in an accuracy ratio well into the 90th percentile for the last several Olympiads. For Beijing in 2008, their numbers were 95% correct.

Those are betting odds, and they have us thinking about a lot more than Olympic gold. Statistical modeling is used throughout the business world, from insurance to financial services to macroeconomics. It’s a way of quantifying a world of possibilities, and giving startlingly precise glimpses of the future.

Chances are your business is already relying on statistics in ways you’re not even aware of. Chances are you can (and should) be using them a lot more. Statistical modeling is complex, difficult to master, and decidedly unsexy. It doesn’t even exactly deliver the gold. But as the Tuck School so ably demonstrates, it can tell you exactly where the gold is going to be.

The C4:

  1. The Tuck School of Business at Dartmouth has created a statistical model that predicts, with up to 95% accuracy, who will win the most gold, silver and bronze at the Olympics.
  2. That’s the Tuck School of BUSINESS. Clearly they’re not overly concerned with international athletic competition. They’re making a powerful point about the versatility and reliability of statistical modeling.
  3. It’s a point you should heed. Your insurers' actuaries do. So does your broker, your banker, and probably your baker. Statistical modeling is how modern businesses can see the future and control their own fates. Are you using it to its fullest potential? Statistically, probably not.
  4. OK, thanks to Tuck we already know how London will end. But Michael Phelps will still be in the water, and Kerri Walsh will still be in the sand. We won’t miss this for the world, no matter what Tuck predicts. Go Team USA!