At 9:47 p.m. Eastern Time on Tuesday, my mother texted me.
“Nate Silver is a rock star,” she said.
“Not yet,” I replied.
President Barack Obama had not yet sewn up his re-election in the campaign against Mitt Romney and the contest was a narrowly run affair. Virginia, a state that I called home for many years, was still up in the air and Romney held the advantage, if ever so slim. Florida, as is its wont, was too close to call. Ohio, the lynchpin of Obama’s “firewall” hung in the balance. There was still a chance that Silver, the New York Times statistician and author of FiveThirtyEight, that predicted Obama had a 90.5% chance of winning the election on Tuesday morning, could be wrong.
At 11:26 p.m., my mother texted me again; “Once again, Nate Silver = rock star.”
This time, I held no reticence. “Indeed,” I said. “I’ve known that since 2003.”
Silver’s victory was a triumph of logic. A vindication of simple math and data analysis over emotional ideologues and punditry.
“The wizards and alchemists are being exposed and done away with,” ReadWrite Editor-in-Chief Dan Lyons said of Silver. “It’s the new Age of Enlightenment.”
When it was all said and done, Silver’s win was total and complete. The degree of accuracy with which Silver’s model predicted the final outcomes of races across the country was almost stunning. Many people and polls predicted that states like New Hampshire, Colorado, Iowa, Minnesota and Wisconsin would fall to Romney. There was corollary evidence to suggest that it might end up that way, with Romney picking up endorsement from large local newspapers in both Iowa and Wisconsin in the days heading up to the election.
As the night went along, Silver’s prediction for the winner of just about every “battleground” state on the board was not only correct, it was stunning in its degree of accuracy of the final vote count.
Take a look at the map at the top of the article for the actual election results. Now, look at Silver’s prediction map below. No coincidence, they are very similar.
Let’s take a look some of Silver’s predictions to the actual results in battleground states.
- Silver’s prediction: Romney 50.6% — Obama 48.9%
- Actual results: Romney 50.6% — Obama 48.4%
- Silver’s prediction: 50.7% Obama – 48.7% Romney
- Actual results: 50.8% Obama – 47.8% Romney
- Silver’s prediction: 51.1% Obama – 47.9% Romney
- Actual results: 52.1% Obama – 46.5% Romney
- Silver’s prediction: 49.8% Obama – 49.8% Romney
- Actual results (pending): 49.8% Obama – 49.3% Romney
As of the time of this article, Florida is the only state that has not declared a winner in the presidential election. Silver not only predicted the correct winner in just about every swing state to a high degree of accuracy, his model predicted the tie that became Florida.
After Silver made his now-infamous bet with MSNBC morning news anchor “Morning” Joe Scarborough, the political pundits of the country came out of the woodwork to decry Silver as some type of fraud, a stooge of the liberal-leaning New York Times and in the pocket of the Obama campaign. None of that was ever true, no matter how much you wanted to believe the talking heads on Fox News or CNN. Silver’s approach has nothing to do with punditry. There is no emotion in his model. It is the logical approach to predictive data analysis taken to the nth degree.
“Those of us in the pundit-ocracy make our bread and butter by telling people what the truth is as we see it from our gut,” comedian and faux-pundit Stephen Colbert told Silver in an interview on The Colbert Report on Monday.
“I am not very pro-pundit, I have to say,” Silver said. “If pundits were on the ballot against, I don’t know [Ebola], I would probably vote Ebola. Or third party.”
In the end, Silver gets the last laugh. He correctly picked the winner in all 50 states, many times hitting the exact percentage of votes for each candidate. After proving his predictive statistical model in baseball, he has now refined it for politics. And made a whole lot of people look like fools along the way.
Read How Statistician Nate Silver Threw A Wrench Into Traditional Election Metrics.