Following the conclusion of the Republican and Democratic national conventions, we’re mercifully facing down the final days of this election season. With the race heating up, policy wonks and voters are turning their attention to the other breakout star from 2008’s Presidential race: statistician and blogger Nate Silver, whose predictive models during that race proved more accurate than the projections of expert pollsters and the pundits.
Silver’s star has grown considerably since his days crunching performance data of ballplayers for the sabermetrics-obsessed readers of Baseball Prospectus using PECOTA, or Player Empirical Comparison and Optimization Test Algorithm, a system he developed and sold to the company. In addition to forecasting the results of this year’s Presidential Election — he’s projecting an Obama win on November 6th with 314.9 Electoral Votes versus Romney’s 223.1 as of this morning — Silver releases his first book on September 27th.
The Signal and the Noise: Why Most Predictions Fail – But Some Don’t explores his methodology, and aims to serve as a primer on predictive analytics for a mass audience. In the book, he argues that overconfidence is the greatest threat to predictive accuracy, causing a “prediction paradox” that requires data be approached with objectivity and humility.
Silver spoke of the need for researchers to “embrace uncertainty” during a keynote address at Greenplum’s Data Science Summit 2012. “Despite all this information in our world,” Silver said, “there are all kinds of places where we’re not making much progress in making predictions or forecasts.” He pointed to stock market analysis and disaster preparedness as two notable examples of existing models coming up short in practice, to catastrophic effect. To find out why, check out the full video of Silver’s Data Science Summit 2012 keynote address.
About the Author
