How A Sabermetrics Nerd Predicted Last Night’s Election And Changed Politics Forever

The date was October 14th, 1984. The Tigers led the Padres 3 wins to 1 loss in Game 5 of the World Series. Behind 5-4 in the top of the 9th inning, the Padres handed the ball to their best reliever, Goose Gossage. After spending the last 9 seasons as arguably the best reliever in the game, Gossage would end up in 5th place for the Cy Young vote and 13th place for the MVP vote in ’84. After walking the Tiger’s number 9 hitter, and then allowing the next hitter to reach safely on a sacrifice bunt, Goose held the Padres’ World Series hopes on his shoulders. Alan Trammell came to the plate and dropped a successful sacrifice bunt to set up a second and third with 1 out situation for the team’s best hitter, Kirk Gibson.

Dick Williams, the Padres’ pitching coach, jogged up to the mound to discuss how they’d approach the situation. Catcher Terry Kennedy had already thrown up four fingers to call an intentional walk, but Williams convinced Gossage that he could strikeout the batter. It just so happened that he couldn’t, and as Gibson rounded the bases with his second homerun of the game, the 1984 Detroit Tigers were engraved atop the Commissioner’s Trophy.

28 year later, October 14th, 2012, Governor Mitt Romney, a Detroit native, sat atop his biggest momentum of the Presidential election. It was following a successful first debate with President Barack Obama that propelled his chances of winning over the prior two weeks. Most of the media became so absorbed in the aftermath of the debate, that Romney was at times considered a favorite. Over the next few weeks leading up to last night’s election, Obama gained some momentum in a couple of debates, as well as some last minute campaigning, but the election was generally still considered too close to call in a handful of swing states.

As I type this out, the election results have mostly been released, and there’s no doubt that Obama has won a second term as President. What has me more intrigued, is what came from a blogger over at Five Thirty Eight. Not only did Nate Silver predict the winner of the election, but he has thus far correctly projected the winner of all 50 states, as well as the projected Senate winners thus far. This after a 2008 election where he correctly projected 49 of 50 states, as well as each Senate election, and I’m starting to think that Silver is a witch.

Back in 1984,  a 6 year old Silver watched his hometown team win the World Series at the hands of Kirk Gibson’s 9th inning home run. Not only would the Tigers be engraved on that trophy, but baseball, the averages, the walks, the homeruns, it would all be engraved in his mind at a young age.

After graduating the University of Chicago and majoring in economics, Silver went on to consult at KPMG in 2000. At the time, Moneyball had risen in Oakland, and teams were slowly beginning to accept the objective meaning behind Sabermetrics. With the rise of the computer, and now the Internet, baseball data was easier to obtain than ever. The movement strived to better understand the value of players tools, and then to better project a player’s future worth.

Tired of a job he didn’t enjoy, Silver followed his love for baseball and numbers, and committed his free time to Sabermetrics. In 2003, he sold his work under the name PECOTA to Baseball Prospectus. The system used a number of forecasting methods, ranging from production, usage, position, and personal attributes, to predict what a player would produce in the future. With the sale, Silver left KPMG to explore work at BP, where he would advance PECOTA, as well as manage and write.

During his 6 years at Baseball Prospectus, Silver made a name for himself in the Sabermetric community. Not only were his PECOTA forecastings incredibly accurate, but his work writing led to many new ideas within the community.

For me, this story truly starts back in 2008. I studied Political Science in college, and with my interest in baseball, I found the concept of poll aggregation dishearteningly similar to what happened in baseball during the early stages of Sabermetrics. During that time, men like Bill James were hated among the majority of the professional scouting community for their extreme ideas. Although the movement started in the late 70′s, Bill James was still hated amongst a large portion of baseball in the early 2000′s. Simply put, people who spent their lives learning how to scout men playing a game, didn’t believe numbers could predict the human element of player’s. After all, the game is played by humans and not computers.

The beliefs in politics about election polls weren’t far off from the same beliefs that scouts had when denying Bill James. Although polls were generally understood to have some significance, no one believed that you could predict the human element of voters. After all, votes are made by humans and not computers.

Poll aggregate FiveThirtyEight.com launched on March 7th, 2008, and Silver revealed himself the writer in late May. His early methodology was consistent with the way Baseball statistics were being created and further interpreted, and I suppose the Baseball Prospectus link on the website also had me intrigued. It’s fair to say I was addicted during that election year, but little did I know that Silver would go on to be so accurate.

From that point, Silver has dedicated himself to better his forecasting models. Although the results were there, pundits wouldn’t stop criticizing his methods. You don’t have to look very farm back to see professional pundit’s doubts and hecklings.

Last night’s results proved the old ways wrong, and the new ways right. As it stands, with 95% of the vote in, Obama has 50.2% of the popular vote, and Romney with 48.3%. Silvers predictions? 50.8% Obama and 48.3% Romney. From a political science perspective, this changes the game, and anyone who doesn’t adopt these methods will be left in the dust.

As a political scientist, I don’t see last night as a win for Obama or a win for Silver, it’s a win for numbers. This is something that I and other amateur analysts try to emphasize. Sure, we can put faith in Silver, FiveThirtyEight, PECOTA, Sabermetrics, and all this methodology, but what truly rings true is that the numbers mean something. Polls can and do predict elections.

But this goes further than forecasting baseball or politics though, last night, Nate Silver disproved that there is a human element haunting the future. For now, he’s changed politics and campaigns, but I doubt it stops here.

Follow Me On Twitter