No. If all the states fall within the range he predicted, and Romney wins, Silver’s reputation should not suffer.
George Will predicts Romney will win 321 electoral votes. I suspect he’ll 1) end up being drastically wrong and 2) being drastically wrong will have zero effect on his career.
But in the public’s eye, it will. Math and numbers are not intuitive for a lot of people. Even though Nate Silver has said repeatedly that he WILL be wrong at some point, that he MUST be wrong, statistically speaking, the average person is not going to hear that.
You mean Michael Barone won’t be living under a bridge and eating cat food after this election?
It’s almost as though he was pulling numbers directly out of his ass, and then, with hours to go and realizing how remarkably stupid those numbers were, he adjusted them based on nothing; as if there never really was a mathematical model in the first place.
But, that would be impossible to believe considering the credence he was given in conservative circles!
You left out his best one!.
The unskewed polls thing is pretty dumb. But the polls in general have been much more favourable to Obama over the last week then when the site started, with even Rasmussen and Gallup giving Romney just a 1 point advantage. So even if he kept his methodology consistent, his prediction should’ve moved significantly in Obama’s favour.
Do you understand how math works? A 90.9% chance means that outcome is by far the most likely. It also means there’s a 9.1% chance it doesn’t happen. Silver doesn’t guarantee anything, he’s just saying if things go the way they appear they are already going, Obama’s re-election is extremely likely. Not guaranteed.
Now, the unskewedpolls folks, aren’t they the ones giving Romney a 106.8% chance of victory? That’s staking your career on the outcome. Methinks unskewedpolls.com will be a very lonely website for the next couple of years.
Then they’ll be back.
Here in Philly, there’s a lot of voters. More than '08, in the precincts I’ve seen.
Also a lot of confusion resulting from moved and consolidated polling places and suppression tactics (robopolls with false info, etc.)
Hoping the first overcomes the second.
Nate is in the proverbial cat-bird seat. It would take years and years to actually prove that his predictions are bunk (if they are, which I’m not saying).
Dude, not 100%. TV stations in battleground states are making bank.
Mathematically, yes - but all it will take is one election to “prove” his model is bunk as far as the public is concerned.
My god, that would be sweet. And if he could take Fred Barnes with him I’d even chip in so they could get the good cat food.
Yep, and I will always think that that supreme court Citizen’s United decision was a giveaway to big media companies.
I thought that was not really the case, since the TV stations have to sell to the candidates at their lowest rates. (Don’t know about PACs.) They’d rather have free spots to sell to corporate advertisers at a higher, normal market rate.
I don’t think that’s true. He has predictions for ten or so close states. That ought to be enough to tell if there’s something wrong with his model.
I’m not sure the focus on Silver is healthy though. The accuracy of his model is pretty closely tied to the accuracy of the polls. If he’s wrong (or right), I think we’d be better off having a discussion on “what was wrong (right) with the polls” rather then “what was wrong (right) with Nate Silver”
If the polls are reasonably accurate, Obama wins. It’s as simple as that. There is a possibility that the polls are systematically wrong in a way they haven’t been since the start of regular state polling but how they will be wrong necessarily falls in the realms of conjecture.
So here is my conjecture:
- Momentum: the idea that undecided voters will break towards whichever candidate is winning. Romney of course has been desperately trying to create the appearance of momentum for weeks but it’s pretty clear that the in the last few days it’s Obama who has it, albeit fairly modest.
2)Ground game: Anecdotal evidence suggests that Obama has invested massive amounts of resources building a ground game and his team has the huge advantage of learning from their 2008 experience. Reports I have read suggest that his ground game is especially strong in Iowa, Nevada and Ohio, and if he wins those three he will almost certainly win. Polling indicates he is well ahead in early voting and the numbers on the party registration of early votes, which is a noisy variable especially in Ohio, suggest he is ahead though not as much as 2008. Certainly there is no reason to believe his ground game is worse than Romney.
-
Narrative: Always the most difficult to assess but it’s hard to see how Romney has any advantage here in the last week. The biggest story by far is Sandy and clearly Obama’s handling of it has been well received by most people. Other stories include the jobs numbers, the Powell/Bloomberg endorsements and the controversy over the jeep ads. I am struggling to think of a big story which has helped Romney significantly in the last week.
-
Pennsylvania: There has been a debate about whether this is a sign of confidence or desperation and I think it’s the latter. Let me put this way: Suppose polls were tied in Ohio, would you really spend time in PA where even Rasmussen says Obama is clearly ahead. Wouldn’t it make much more sense for Romney to double his efforts in Ohio and Virginia since he will likely win if he wins both states. OTOH suppose as public polls indicate, Romney is behind by about 4 in OH and about 5 in PA. Then yeah it would kind of make sense to take a fling on PA. A long shot but then so is Ohio and PA hasn’t really been exposed to his message as much so why not see if it works. In a Bayesian sense I think the swing towards PA is evidence that Romney’s internal polls are pretty similar to public polls.
So all the above favors Obama. What about Romney? Well there seems to be a lot of enthusiasm in his rallies. And he still seems to poll pretty well on the economy so perhaps some undecideds will break towards him for that. But it’s pretty thin stuff IMO.
So bottom line: I expect the polls to be pretty accurate but if they are wrong I wouldn’t be surprised if they underestimate Obama’s numbers in key states by a point or two.
We should find out soon and I will certainly have been wrong if Romney wins or even makes it very close like say losing by half a point in Ohio.
Only the candidates get the lower rates. Super PACs are charged whatever the going rate is at the time they place their buy.
What’s sad is that if Nate is “wrong” he will be eviscerated. While when George Will, Dick Morris and all those knuckleheads are wrong they probably will get even more writing gigs.
He shifted by 84 electoral votes in three days. The polls haven’t been THAT favorable!