< Forecasting Tuesday


Friday, November 02, 2012

BROOKE GLADSTONE:  It’s the week before the election and there seems to be two prevailing views about what will happen on Tuesday, but they don't break down along party lines. One camp asserts that President Obama will most likely win re-election. That camp includes FiveThirtyEight’s Nate Silver, who told us Thursday afternoon that there's a 79% chance of an Obama win, a number that he has since revised upward.

Also, in Silver’s camp, right-leaning journalists like the Washington Examiner's Timothy Carney, who wrote a piece Thursday titled, “Why I think Obama Will Win, in Detail.” But there’s another camp, poll skeptics who still think it's a tossup. Joe Klein of Time Magazine is one. He wrote a column simply titled, “I Don’t Know.” The left-leaning blog, Truth Out, is also skeptical, as is former Republican congressman, now MSNBC host, Joe Scarborough.

JOE SCARBOROUGH:  Nate Silvers says this is a 73.6% chance that the President’s going to win. Nobody in that campaign thinks they have a 73 – they think they have a 50.1% chance of winning. Anybody that thinks that this race is anything but a tossup right now is such an ideologue, they should be kept away from typewriters, computers, laptops and microphones for the next ten days, ‘cause they’re jokes.

BROOKE GLADSTONE:  Nate Silver, you are so confident in your predictions you challenged Joe Scarborough to a $1,000 bet that Obama would win.

NATE SILVER:  It’s pretty simple. Obama leads in the polls in states like Ohio, Iowa, Wisconsin that would suffice to give them 270 electoral votes. Therefore, he's the favorite. If you're a pundit and you're someone who’s paid to render your opinion and you care about being accurate, I think the argument is stronger, both in an intuitive way and based on this complex formula that we have for Obama, based on the swing state polls. But I think Joe is ducking the question, which kind of lets you know that he’s more of an entertainer than, than someone who’s really doing any kind of analysis or reporting –


NATE SILVER:  Yeah [LAUGHS]. Look at all the cases in the past since 1968, where candidates two or three points ahead in a swing state like Ohio, as Obama is now, in the average of polls, and how often did he win? So you count those as wins. How often did he lose? You count those as losses.  And then you have a percentage score. That's really all we’re doing.

BROOKE GLADSTONE:  Okay. Truth Out points out that few people answer telephone polls - only 9% of people do - and that those weirdos [LAUGHS] certainly aren't representative –


- of the average voter.

NATE SILVER:  Oh sure, and look, there – there are two contradictory kind of trends here. On the one hand, if you just look at how well have the polls actually done, they’ve been getting better. So 2004, they nailed the election, 2008. On the other hand, if you look at how many people actually answering these calls, then it’s fewer and fewer people. So which of those will, which will prevail on Tuesday, we're not sure. We assume the polls could miss, but if they’re biased, it could be either way.

BROOKE GLADSTONE:  How does it feel to be a legend in your own time?

NATE SILVER:  [LAUGHS] It's been very odd to have FiveThirtyEight itself be a story, and I think that itself kind of tells you a little bit about maybe where people's priorities are misplaced. You know, people sometimes say, oh, you know, people focus too much on the polls, to begin with, right? But when you then focus on someone who builds statistical models based on the polls, you’re one level further removed away from anything that really affects the average American in any obvious way.

BROOKE GLADSTONE:  I don’t know, I think that you yourself, if you weren't affecting this false modesty right now –


- would concede that people following your meta poll would be closer than people who were trying to follow individual polls.

NATE SILVER:  Well, I’ll tell you what I’m less modest about though, I think we have a better way of doing things, and the reason I got into this is because I found the horserace coverage extremely vapid, right, where a lot of it was just kind of recycling quotes from the campaigns and, and talking in clichés but didn't really hold up to scrutiny. And I had seen in baseball, a field I was in before covering politics, how you had kind of the “Moneyball” revolution seeped its way into kind of journalism first and then inside the game.

Politics, is kind of the, the reverse almost, where campaigns themselves are quite data driven but I think the people who cover them are not as much. They’re always more allured by what they think is the, quote, unquote, narrative, what - what I think, frankly, is the – is the spin.

BROOKE GLADSTONE:  You’re kinda like the Jonah Hill character –


- of politics.

NATE SILVER:  I've gotten that comparison more than the Brad Pitt comparison,


 - unfortunately.

BROOKE GLADSTONE:  [LAUGHS] I’m sorry, At some point, you’re going to make a high certainty prediction about something high profile, like maybe a presidential election, and it’ll break the other way. It won’t even necessarily mean that you were wrong. In this election, for instance, by your numbers, if you run a scenario ten times, Romney would win at least twice.

NATE SILVER:  I will stand here, or sit here right now, and guarantee you we’re gonna be wrong [LAUGHS] sooner or later. What’s different is that we tell you how likely we are to be wrong and how often. I mean, the irony here is that in my book, “The Signal and the Noise,” I talk a lot about how we think about uncertainty and how do we measure uncertainty, but that's something that even people who are, who are versed in statistics aren’t very good at doing, and certainly I think pundits and reporters aren't very good at doing, where everything has to be either a 50-50, we can't weigh in or, oh, this is certain. And usually the truth is somewhere between zero and one hundred.

BROOKE GLADSTONE:  But if Tuesday is one of those times where Romney wins, is your goose cooked? Will you ever work in this town again?

NATE SILVER:  [LAUGHS] I mean, not if some of the bigger hacks in the media establishment get their way.

BROOKE GLADSTONE:  But those hacks have an even lower record of probability, we’ve noted on this program, than –


BROOKE GLADSTONE:  - than you have, by far.

NATE SILVER:  Well, so – so look, I think people are competitive, right? And a lot of what I do implicitly critiques the horserace coverage. We’ve  had, I should say, a very steady forecast. We had all 50 states the same when we launched the model in June. We know which states have more Republicans than Democrats, we know it’s gonna be a close election based on the economy, and it’s, it’s maybe not all that complicated.

But people want to pretend that someone wins the day and there are all these ups and downs and momentum and the roller coaster and game change. And all these terms that are used are basically BS, and when someone calls them on that and asks them to provide their facts and to defend their analysis, and they can't do it, they become very, very defensive and they want to shoot the messenger.

BROOKE GLADSTONE:  Don’t hit me, Nate. [LAUGHS]

NATE SILVER:  [LAUGHS] Not you! Not you. You have – this show,  this program has self-awareness.

BROOKE GLADSTONE:  Nate Silver blogs at FiveThirtyEight, and you can find him on the New York Times website. Thank you so much.

NATE SILVER:  Thank you.



Nate Silver

Hosted by:

Brooke Gladstone