Key excerpt (But it's worth reading the full thing):
But the real value-add of the model is not just in calculating who’s ahead in the polling average. Rather, it’s in understanding the uncertainties in the data: how accurate polls are in practice, and how these errors are correlated between the states. The final margins on Tuesday were actually quite close to the polling averages in the swing states, though less so in blue states, as I’ll discuss in a moment. But this was more or less a textbook illustration of the normal-sized polling error that we frequently wrote about [paid only; basically says that the polling errors could be correlated be correlated between states]. When polls miss low on Trump in one key state, they probably also will in most or all of the others.
In fact, because polling errors are highly correlated between states — and because Trump was ahead in 5 of the 7 swing states anyway — a Trump sweep of the swing states was actually our most common scenario, occurring in 20 percent of simulations. Following the same logic, the second most common outcome, happening 14 percent of the time, was a Harris swing state sweep.6
[Interactive table]
Relatedly, the final Electoral College tally will be 312 electoral votes for Trump and 226 for Harris. And Trump @ 312 was by far the most common outcome in our simulations, occurring 6 percent of the time. In fact, Trump 312/Harris 226 is the huge spike you see in our electoral vote distribution chart:
[Interactive graph]
The difference between 20 percent (the share of times Trump won all 7 swing states) and 6 percent (his getting exactly 312 electoral votes) is because sometimes, Trump winning all the swing states was part of a complete landslide where he penetrated further into blue territory. Conditional on winning all 7 swing states, for instance, Trump had a 22 percent chance of also winning New Mexico, a 21 percent chance at Minnesota, 19 percent in New Hampshire, 16 percent in Maine, 11 percent in Nebraska’s 2nd Congressional District, and 10 percent in Virginia. Trump won more than 312 electoral votes in 16 percent of our simulations.
But on Tuesday, there weren’t any upsets in the other states. So not only did Trump win with exactly 312 electoral votes, he also won with the exact map that occurred most often in our simulations, counting all 50 states, the District of Columbia and the congressional districts in Nebraska and Maine.
I don't know of an intuitive test for whether a forecast of a non-repeating event was well-reasoned (see, also, the lively debate over the performance of prediction markets), but this is Silver's initial defense of his 50-50 forecast. I'm unconvinced - if the modal outcome of the model was the actual result of the election, does that vindicate its internal correlations, indict its confidence in its output, both, neither... ? But I don't think it's irreconcilable that the model's modal outcome being real vindicates its internal correlations AND that its certainty was limited by the quality of the available data, so this hasn't lowered my opinion of Silver, either.
Jump in the discussion.
No email address required.
Notes -
This makes me think of someone stuck on a very sticky wicket, trying to justify an argument that was fundamentally wrong. Of course there are facets of any sophisticated but wrong argument that are right. You can highlight the correct facets and minimize the wrong facets. You can pre-prepare reasons for why you might be wrong to conserve credibility.
Nate has the rhetorical skills to pull it off. But it still feels very slimy. The 90 IQ twitter pleb mocking him with '60,000 simulations and all you conclude is that it's a coin flip?' may not be that numerically literate. But he has hit on a certain kind of wisdom. The election wasn't a 50/50 or a dice-roll. It was one way or another. With superior knowledge you could've called it in Trump's favour. Maybe only Bezos and various Lords of the Algorithms, French Gamblers and Masters of Unseen Powers knew or suspected - but there was knowledge to be had.
I prefer prediction models that make money before the outcome is decided, not ones that have to be justified retroactively. Nate wasn't heralding before the election that this 6% was the modal outcome, it wasn't really useful information.
Do you have a substantive disagreement with his argument?
If Nate put his money where his mouth was, he'd have lost $100,000. He talks the talk (after it's decided) but doesn't walk the walk when it actually means anything.
https://x.com/NateSilver538/status/1842211340720504895
Did the other guy send the contract?
He says so and that Nate later refused to sign.
Did the other guy provide proof that he sent the contract?
To be fair, in the absence of Nate denying it, I don't think he necessarily needs to provide proof.
That would not be fair. In the absence of Nate confirming that he refused to sign a contract, a claim of having sent the contract is just a claim absent further evidence.
My curiosity / eyebrow is raised because Ranger is raising this bet as a character failure on the part of Nate Silver, but the proffered evidence is of the conditional offer of a bet, not that the bet was accepted as offered but that Silver refused to sign it.
This leads to a couple of issues for which more information than has been provided is needed.
-Did the other person actually accept the bet, or are they just claiming so with post-election hindsight? (i.e. is he talking the talk after the election is decided?)
-Did the person try to modify the terms of the bet offered that would render the offer void? (i.e. did he refuse to walk the walk when it mattered?)
-Did the person fail to meet the conditions of the offer of bet? (i.e. did they not have their lawyer do it, but tried to make their own contract- thus invoking the payment risk issue raised?)
I've no particular strong feeling on Nate Silver one way or another, but if someone wants to make a character failure accusation with linked evidence I'd generally prefer the links to be evidence of a character failure.
I disagree. All Nate has to do is say "no you didn't, you fucking liar", and if Keith can't provide evidence of sending him the contract, he's the one that's going to suffer reputational damage. On the other hand, if Nate says that, but Keith promptly provides evidence, this will look even worse for him. Since Nate knows for a fact whether or not he received the contract, his decision on how to react to the claim tells us something about the truth value of the claim that he was sent the contract.
There are also scenarios that would explain a lack of reaction. Maybe after the spat Nate blocked Keith, and has no knowledge that he's now going around claiming that he sent the contract. So while the lack of reaction doesn't outright prove the contract was sent, I maintain that the potential reputational damage that can result from the claim is a weak form of evidence in itself, and thus it is the demand to provide hard evidence that's unfair.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link