Friday, November 16, 2012

Was The 2012 US Presidential Election Results Mathematically Predicted?

Did a 34 year old statistician and “poker number expert” used tried and true mathematics to predict the outcome of the 2012 US Presidential Elections with better than 90 percent accuracy?

By: Ringo Bones

Despite attracting widespread ridicule from traditional political pundits in what was regarded as the closest US Presidential Race in years, a 34 year old statistician and poker numbers expert named Nate Silver gave incumbent US President Barack Obama a 90.9 % chance of success by using tried and true mathematical techniques way before the final voting results came through. Nate Silver used an elaborate series of tried and true calculations to correctly call the outcome in all 50 states by running thousands of computer calculations based on innumerable factors such as polling results and voting outcomes of previous elections. His achievement opens the door for a more statistics-based approach to polling says experts. Given Silver’s mathematically predictive success, will the outcome of the next US Presidential Election – in 2016 – be predicted with better than 90 percent accuracy using his method?

The highly mathematical approach used by Nate Silver – which has echoes of the book and a Hollywood film that stars Brad Pitt about baseball statistics titled “Moneyball” – involves running hundreds of mathematical calculations for the polling data of each state based on a myriad of factors, including election results from the past and more recent polling data. Even though Silver is an avowed Obama supporter, his “magic mathematical formula” calculated that there was a 90.0 % likelihood of an Obama win and the President would win 332 Electoral College seats compared to 206 for Republican challenger Mitt Romney – correct if the Democrat wins Florida which, at the time, was yet to declare when Silver made his prediction.

Years before gaining fame for his mathematical predictions for the 2012 US Presidential Race, Nate Silver’s mathematical skills were already well known in the American journalistic world because back in 2010 his existing blog at was bought by the New York Times for a princely sum. And then Silver revealed to them his own proprietary mathematical models for predicting election results based on a combination of electoral history, demographics and polling results. I wonder if Silver’s mathematical method works in the actuary world of determining cost competitive premium rates for insurance policies with “sketchy” statistical data. 

No comments: