This is an article related to the presidential election, which is a stupid idea. Mostly because you -- the person reading this right now -- have horrible, poorly developed political beliefs, and both you and your parents should be deeply, deeply ashamed of them. But, if possible, forget about the politics for now. If you're a Whig, forget about the Tories. If you're a Tory, ignore the Whigs.
This is about numbers, not politics.
There's a pretty good chance that, regardless of your political affiliation, you were pretty bemused by the Nate Silver frenzy this election cycle. You were bemused because Silver used to be one of us -- a guy who woke up in the morning and said, "Say, would a theoretical team of Chris Youngs beat a theoretical team of Adam Eatons? LET'S FIND OUT." He was a baseball pundit, and his numbers would catch a lot of guff when they would leak out into the regular world. But even if the numbers weren't always right, the process was respectable. (The numbers were often right, though.)
He moved on to grownup things, focusing his statistical acumen toward politics. He dug through polling histories from different pollsters, looking to see who predicted elections with more precision. He examined their methods to see if there was a reason certain polls fell prey to consistent biases. He weighted them accordingly and published probabilities.
Nothing up there should surprise you. That's sort of how the Internet baseball world works, too. When a .300/.400/.500 guy hits .200/.300/.400 over a full season, you look into it and hope the answer you find makes you a smarter baseball fan. Or, at least, a smugger one.
The anti-Silver columns and screeds that came out before the election were amazing. I'm not going to pretend there's a one-size-fits-all psychological profile for both Republicans and Democrats, but there would have been similar quips and quotes from whichever side was behind in Silver's model. If Silver were predicting a Romney romp, there would have been appeals to evidence we couldn't measure from hopeful Democrats, so forget about the politics as much as you can.
This was from the Wall Street Journal before the election:
We begin with the three words everyone writing about the election must say: Nobody knows anything. Everyone’s guessing.
They don't play the games on paper.
Who knows what to make of the weighting of the polls and the assumptions as to who will vote? Who knows the depth and breadth of each party’s turnout efforts?
There are just some things that statistics can't tell you.
Romney’s crowds are building—28,000 in Morrisville, Pa., last night; 30,000 in West Chester, Ohio, Friday. It isn’t only a triumph of advance planning: People came, they got through security and waited for hours in the cold. His rallies look like rallies now, not enactments. In some new way he’s caught his stride. He looks happy and grateful.
The clubhouse chemistry is off the charts. The manager has them all on the same page.
There is no denying the Republicans have the passion now, the enthusiasm. The Democrats do not.
This team has a focus. They just know how to win.
Is it possible this whole thing is playing out before our eyes and we’re not really noticing because we’re too busy looking at data on paper instead of what’s in front of us?
GET YOUR NOSE OUT OF THE SPREADSHEETS.
I mean, that last excerpted sentence could have been lifted verbatim from a Bruce Jenkins or Murray Chass column. But it turns out that Silver was spot-on with his predictions.
One last time, forget the politics. Here's what Nate Silver was pointing out:
There are numbers that represent what happened in the past. They can help predict the future with some degree of accuracy. These over here are the numbers that tricked us, so we have to be a little wary of them. These over here were more reliable than those other ones, so we should pay more attention to them. Let's weight them accordingly, and in the end, things can always get goofy, so let's not pretend that we know exactly what's going to happen. We can just have a good idea of what could happen.
And you know what? Silver's numbers were better than just about anyone else's, but they were still imperfect numbers. He doesn't think of himself as a wizard. He knows there were things that could have could have messed up the model, which is why he always left a little room for doubt. He probably already has some ideas on how to rejigger his formulas based on Tuesday's results.
The data we're parsing in baseball is more confusing, of course. It's not data collected by statisticians specifically to predict an outcome; it's data that's constantly being shot out of a cannon at us whether we like it or not. Don't take this as an argument that because an ex-sabermetrician did well predicting the Electoral College, Mike Trout should win the MVP.
It's just a note that right now the rest of the world is saying, "Say, those numbers might not predict everything, but they certainly mean something." I'd like to think that will happen more and more in every subject the public is interested in.
In conclusion, as I was writing this, Dave Cameron over at FanGraphs published something similar, but his article didn't have a GIF of a cat eating a lollipop.