Nate Silver Admits He’s a Lagging Indicator, Declines a Bet on Senate Outcome

The Case For A Republican Sweep On Election Night

FiveThirtyEight founder Nate Silvers presents The Case For A Republican Sweep On Election Night. 

The discussion is between Nate Silver and Nathan Redd. The latter is an “alter ego” who tries to convince Silver that the Republican outlook is even better than the 538 model shows. 

Here is part of the discussion that I pieced together. 

Redd: This is the easiest election call I’ve ever seen.

Silver: Well, we’re not really making “calls” at FiveThirtyEight, exactly. You know how the model works. It’s probabilistic. And the model says the Senate is about as close to 50-50 as it gets.

Redd: You think it’s 50-50? OK, let’s bet then. Steak. Sushi. Rangers tickets. You name it. To me there are two scenarios: Republicans win big or Republicans win small. Democrats barely even have a majority to begin with and there’s a backlash every midterm election.

Silver: I get that, but it’s not so straightforward. Voters may be unhappy, but they’re agnostic about which party they prefer. I don’t see how you can be so confident about all of this.

Redd: I’m not confident! I just told you, I don’t know if we’re going to win small or big. But lately I’ve been thinking big. You want to know why?

Silver: I’m sure you’re going to tell me.

Redd: I’ll tell you a detail. Your polling average still has Democratic John Fetterman up in Pennsylvania’s Senate race. He’s not going to win after that debate.

Silver: Ahead? I mean, just barely. And it’s still not entirely clear how much the debate changed things. But I agree with you in principle: if there’s news that hasn’t yet been fully reflected in the polls, that’s going to make our model a lagging indicator.

Redd: So, we got a bet?  On the Senate.

Silver: I’m going to pass. Our model has it at 50-50, it’s been trending Republican, and you make a couple of semi-persuasive points, like about Pennsylvania.

Redd: Semi-persuasive! That’s the nicest thing you’ve ever said to me, Silver! And good luck on Tuesday, but you won’t need it. It’s going to be an early night.

Mish Comments

The quotes are all in order but pieced together skipping a Redd prediction about a New York governor race.

The dinner bet with Silver’s alter ego is fictional, except for the fact that Silver would have declined it. 

I agree with Silver that his odds are not a “call”. 

But here’s the key point, that Silver admits: “If there’s news that hasn’t yet been fully reflected in the polls, that’s going to make our model a lagging indicator.“

In this regard, Silver is way lagging. He will rely on very stale polls, even placing placing high odds on them.

Perhaps Silver needs a +- confidence indicator. On some of these races it would be +-20 percentage points. 

In retrospect, Silver does have a confidence indicator. 

When it was 55-45 in favor of Democrats he called it a “dead heat”. Now that it’s 55-45 in favor of Republicans it’s still a dead heat. 

Q: Did the odds change from 52-48 to 45-55 in three days? 
A: No but Silver’s perception of the odds did.

+- five points is a dead heat to accommodate perceptions and lags while discounting obvious momentum.

This post originated at MishTalk.Com.

Thanks for Tuning In!

Please Subscribe to MishTalk Email Alerts.

Subscribers get an email alert of each post as they happen. Read the ones you like and you can unsubscribe at any time.

If you have subscribed and do not get email alerts, please check your spam folder.

Mish