Nate Silver

Nate Silver @NateSilver538

@mattyglesias LOL, missed that earlier. I appreciate it. — PolitiTweet.org

Nate Silver @NateSilver538

And FWIW, I think that's probably what we're going to wind up with once all votes are counted. (And we should wait… https://t.co/1W6eIO4MOE — PolitiTweet.org

Nate Silver @NateSilver538

Don't want to stir things up *too* much but it seems like if a forecast says that Biden is favored because he could… https://t.co/7PAsG8DwkS — PolitiTweet.org

Nate Silver @NateSilver538

It'll get figured out eventually but it's going to take 4 or 5 variables. — PolitiTweet.org

Nate Silver @NateSilver538

¯\_(ツ)_/¯ https://t.co/eehs4ElQ1U https://t.co/Kk9qF4BTmR — PolitiTweet.org

Nate Silver @NateSilver538

RT @b_schaffner: Interestingly, our estimate of the national vote margin is not too sensitive to our likely voter model (Biden's margin is… — PolitiTweet.org

Nate Silver @NateSilver538

@NickRiccardi Yep, exactly. — PolitiTweet.org

Nate Silver @NateSilver538

Much to Fivey's chagrin—or perhaps delight—Biden keeps oscillating between like 89.9% ("favored") and 90.1% ("clearly favored") in our forecast with every random poll we add. (And boy have there been some random polls today.) https://t.co/ajG88SznSA https://t.co/FdOT1ODMe7 — PolitiTweet.org

Nate Silver @NateSilver538

In other words, if you knew that Biden was +1 in GA, +2 in NC, +3 in AZ, only -1 in TX and +8 or +9 nationally, you'd probably guess that he had a slightly larger lead in Florida polls than he does. — PolitiTweet.org

Nate Silver @NateSilver538

Some debates on here about FL (the conventional wisdom is bearish on Biden relative to polls). One thing to keep in mind is Biden's actually underperforming there relative to other states. He's up 2 there when our model thinks he "should" be up 4 based on his numbers elsewhere. — PolitiTweet.org

Nate Silver @NateSilver538

@ForecasterEnten We shall know soon. Or maybe not so soon in PA. But at some point. — PolitiTweet.org

Nate Silver @NateSilver538

Pennsylvania isn't a huge mail-voting state, but Dems having returned 82% of mail ballots vs. 71% for Republicans is something that stands out a bit. Returned mail ballots are +43 Democrat but outstanding ones are just +9 D. https://t.co/sTPHw4KZsl https://t.co/nPpj6Uw9qE — PolitiTweet.org

Nate Silver @NateSilver538

@MattGrossmann Quite conceivable that expanded access to mail voting increases turnout, net of ballot spoilage, which could lead to a pro-Biden polling error. — PolitiTweet.org

Nate Silver @NateSilver538

I think our forecast has followed an intuitive path this year. It's hedging its bets more in the summer, then it starts to get more confident after Labor Day. But it's been pretty smooth, in contrast to our bouncier 2016 model (though of course the polls were way bouncer in '16). https://t.co/XKxUCiT8FN — PolitiTweet.org

Nate Silver @NateSilver538

The comment is: Having a few elections under your belt helps a *lot*. No matter how much you test things in the lab, there are some things you're going to learn only by seeing how your forecast reacts to real data in real time. (I'm sure this applies to lots of other stuff too.) — PolitiTweet.org

Nate Silver @NateSilver538

Not gonna do any model meta stuff since 1) who cares 2) everyone's forecasts are *pretty* similar. But just one stray comment on this interesting reading from the Economist team on how they think their forecast (Biden 96%) may be overconfident: https://t.co/mQGs3YuVNS — PolitiTweet.org

Nate Silver @NateSilver538

One state I am a bit curious about is Texas. Our forecast has been pretty skeptical about Biden's chances there, expecting Trump to close well. But instead Biden got a decent run of polls there this weekend and it now it has him within ~1 point. https://t.co/486ZZ385cV https://t.co/efAeDAaXaX — PolitiTweet.org

Nate Silver @NateSilver538

Nor is there anything magical that will happen when we run the final model just after midnight. And hey, pollsters: Please do get those polls to us by midnight! We expecting that to be our final run of the season. — PolitiTweet.org

Nate Silver @NateSilver538

Bottom line: our final forecast is likely to be very close to what you see now (i.e. 90/10 or thereabouts). Could it be 92/8 or 88/12? Sure. But there's so much polling in most of these states now that new polls aren't shifting things much. — PolitiTweet.org

Nate Silver @NateSilver538

Biden got a tiny further boost this morning (from 89% to 90%*) from this plus some decent polling for him. But now there are no longer any further gains to be had for Biden from the mere passage of time. * Yes, just over the "clearly favored" line in our verbiage — PolitiTweet.org

Nate Silver @NateSilver538

The factors (reducing the amount of uncertainty and reducing the weight assigned to the prior, which expected the race to tighten) had gradually been helping Biden. They're why his win % had been drifting upward even while the polls were basically flat or tightening by a hair. — PolitiTweet.org

Nate Silver @NateSilver538

2. One type of uncertainty—"drift", or how much polls change between the present day and the election—is also now gone. Of course, there's still the chance that polls could be *wrong*. But there's no longer time for polls to change (though a few more will straggle in today.) — PolitiTweet.org

Nate Silver @NateSilver538

1. The forecast is now totally polls-based; there's no longer any prior for economics/incumbency. The weight assigned to this prior had been declining anyway so that it was close to zero, but now it's actually zero. — PolitiTweet.org

Nate Silver @NateSilver538

So... this is a special morning for our forecast in that there's now no more time until the election. (The model treats Election Day and Election Eve as equivalent to one another.) This has a couple of minor effects. https://t.co/aEBz22RXEs — PolitiTweet.org

Nate Silver @NateSilver538

It is also a bit self-serving in that if you have a poll with Trump +1 in PA when the consensus has it Biden +5, "voter fraud" is a way to excuse why your poll was wrong if Biden wins. — PolitiTweet.org

Nate Silver @NateSilver538

Final Morning Consult state polls. https://t.co/GeM5vHtrTs https://t.co/zJlFfLcxrA — PolitiTweet.org

Nate Silver @NateSilver538

This is semi-interesting. In Nevada, 1.4% of mail ballots so far required a signature cure, when the voter didn't sign the ballot or the sig didn't match the one on file. Dems have cured 51% of their rejected ballots so far, vs. 35% for Republicans. https://t.co/jxCrxz1urt — PolitiTweet.org

Nate Silver @NateSilver538

This isn't "disaster porn" for Biden voters. It's trying to sort carefully and analytically through different sources of potential polling error, including debunking some I think are overrated. Not going to tweet out snippets so would encourage you to read the whole thing. — PolitiTweet.org

Nate Silver @NateSilver538

👋 I'm here to remind you that Trump can still win. https://t.co/DqVm8I5Svd — PolitiTweet.org

Nate Silver @NateSilver538

@chrislhayes Yeah... I'd love for a pollster to go back in think about whether the insanely high turnout so far changes any of their assumptions. Nobody seems to know what to do with it. — PolitiTweet.org