Among the top election quants, Nate Silver reigns supreme
Not only did FiveThirtyEight's Silver pick all 50 state winners in the presidential race, he also beat out his polling aggregator rivals for sheer margin of accuracy.
While there's already been whole swimming pools of ink devoted to the Election Day prediction performance of polling aggregators like FiveThirtyEight blogger Nate Silver, CNET is ready to hand out one more round of kudos to the king of the quants.
By now, anyone following the presidential election knows that Silver successfully predicted the winner in the race between President Barack Obama and former Massachusetts Gov. Mitt Romney in all 50 states., earning him worldwide admiration and validating a polling aggregation model that had drawn mockery and ire from many pundits.
But Silver wasn't the only one to do exceptionally well in the prediction department. In fact, each of the five aggregators thatyesterday -- FiveThirtyEight, TPM PollTracker, HuffPost Pollster, the RealClearPolitics Average, and the Princeton Election Consortium -- successfully called the election for Obama, and save for TPM PollTracker and RealClearPolitics handing Florida to Romney, the aggregators were spot on across the board when it came to picking swing state victors.
Still, even within the club of those who used computational analysis of thousands of national and statewide polls to peg the outcome of the election, someone had to be the most accurate. And a CNET examination of each aggregator's performance reveals a single winner -- unsurprisingly, Silver.
In addition to picking the winner in all 50 states -- besting his 49 out of 50 slate in 2008 -- Silver was also the closest among the aggregators to picking the two candidates' popular vote percentages. All told, he missed Obama's total of 50.8 percent by just four-tenths of a percentage point (50.4) and Romney's 48 percent by just three-tenths of a point (48.3) for an average miss of just 0.35 percentage points. HuffPo Pollster and RealClearPolitics tied for second with an average miss of 0.85 points.
In preparing to make these comparisons, CNET surveyed 11 swing states. In the end, Silver was closest to the final margins among the candidates in seven of them and also had the best overall record, missing by an average of just 1.46 points in the 11 states. TPM PollTracker was second with the closest predicted margins in three states, and the second-best average margin, 1.80 points.
It is worth noting that while Silver's final pre-election calculations showed a tie vote in Florida, he still predicted a 50.3 chance that the president would prevail in the Sunshine State.
The performances by Silver and his fellow polling aggregators should be sounding alarm bells in the halls of long-venerated pollsters like Gallup -- which, by the way, predicted that Romney would win the national popular vote by a point.
For the Nate-haters, here's the 538 prediction and actual results side by side twitter.com/cosentino/stat...-- Michael Cosentino (@cosentino) November 7, 2012
But the situation does create a bit of a chicken-and-egg problem. While the results of pollsters like Gallup and many of its competitors are looking increasingly suspect in the face of more accurate predictions by Silver and other aggregators, and while some may begin choosing to ignore those traditional polls, the aggregators could not do what they do without the standard polling systems.
In the meantime, even Silver may need to tip his cap to someone who seems to have done an even better job at prognosticating the final presidential election results. In a blog post today, dailyKos founder Markos Moulitsas noted that he had predicted exactly the final Electoral College vote totals and reported an average margin in the swing states that was less than Silver's or that of any aggregator. Moulitsas' methodology? A savvy but seemingly manual reading of last-minute pre-election polls. Maybe the quants don't have all the answers after all.
Correction, November 8, 4 p.m. PT: The above chart originally misreported the number of points by which TPM PollTracker was off in Florida. The correct figure is 1.7 points.