Nate Silver

Nate Silver @NateSilver538

@ASDem In person hockey but maybe not on TV. — PolitiTweet.org

Nate Silver @NateSilver538

It's better to think of it this way: there are several theories for why Trump could win that very probably aren't true. But if you add up a lot of longshot possibilities, it's not that hard to get to a 10-15% chance, which is something worth taking seriously. 2/2 — PolitiTweet.org

Nate Silver @NateSilver538

I've said this before, but while it's certainly possible that Trump could win, most theories of why he can win rely on dubious evidence and don't hold up well to scrutiny. 1/2 — PolitiTweet.org

Nate Silver @NateSilver538

RT @Redistrict: Yeah, this is a poor take. Biden is not about to run 10 points behind Clinton in Philly proper. Almost all PA district-leve… — PolitiTweet.org

Nate Silver @NateSilver538

I suppose I'd note here that our *forecast* still has Trump favored in Texas, in part because it has strict voting laws (one of the few states without no-excuse absentee voting) but Biden's chances there (38%) are the highest they've been all cycle. https://t.co/486ZZ385cV — PolitiTweet.org

Nate Silver @NateSilver538

Welp. https://t.co/nXOH1SNdLP https://t.co/GKVAzZUWlG — PolitiTweet.org

Nate Silver @NateSilver538

I am not sure what's happening, but now the links to these Trafalgar polls with the dubious crosstabs have gone dead. — PolitiTweet.org

Nate Silver @NateSilver538

A broader issue here is with polling firms that behave like partisan shops even though they're nominally nonpartisan. And unlike pollsters who work for campaigns, who are mostly pretty good, a lot of them don't know what they're doing, as @Nate_Cohn says: https://t.co/Yz4fwJqrDN https://t.co/TPqTEe9q67 — PolitiTweet.org

Nate Silver @NateSilver538

I'm not a Let's Delve Into The Crosstabs guy, but some of the shit here is just crazy. Trump is not going to win 30% of the Democratic vote in Michigan. Biden is not going to win 25% of Republicans. Trump is not going to win independents by 32 points. https://t.co/YJJUHUxeE0 https://t.co/lOjynoRJLZ — PolitiTweet.org

Nate Silver @NateSilver538

I don't know exactly what they're doing, but it's not a good sign that I always know what a Trafalgar Group poll is going to say without having to open the link. — PolitiTweet.org

Nate Silver @NateSilver538

RT @ABCPolitics: .@NateSilver538 “I buy that we’re going to know quite a bit on election night … we could even get an election night call.… — PolitiTweet.org

Nate Silver @NateSilver538

Deleted my tweets that incorrectly said mail is the only way to vote in Colorado. Lots of ways to vote in Colorado! — PolitiTweet.org

Nate Silver @NateSilver538

We have AOC forecasted to win by 57 points. https://t.co/HRnLqxx3Aq — PolitiTweet.org

Nate Silver @NateSilver538

RT @jdickerson: @NateSilver538 After a season of @FiveThirtyEight podcast listening I feel like that question was bait just for you. — PolitiTweet.org

Nate Silver @NateSilver538

@injuryexpert I don't know if they'll be biased but there are a lot of ways they could just be wrong. — PolitiTweet.org

Nate Silver @NateSilver538

I'd be so careful with the early exits this year in fact that if I somehow had the option to guarantee that I would never learn about them, I'd think that would improve my election night analysis. Unfortunately, I probably won't be able to avoid finding out what they say. — PolitiTweet.org

Nate Silver @NateSilver538

(Also, the early exit polls are often decently predictive, FWIW, though I'd be very very very very careful with them this year when you have to blend in-person + mail voting.) — PolitiTweet.org

Nate Silver @NateSilver538

The early vote. Because those hot takes last for a few weeks and not just a few hours. — PolitiTweet.org

Nate Silver @NateSilver538

Georgia's moved up to a ~3% chance of being the tipping-point state. That isn't high but it's the same as say NV. And if the Biden internals have a point or two closer to the tipping point than the public polls do, then a visit starts to become compelling. https://t.co/ajG88SznSA https://t.co/UvveKaINhS — PolitiTweet.org

Nate Silver @NateSilver538

What we can say from the early voting data is that Democratic turnout is likely to be very high. Will Republican turnout be very high also? Yes, I *think* so, very likely. But it's easier to be certain that Democratic turnout will be very high because of the early voting numbers. — PolitiTweet.org

Nate Silver @NateSilver538

@davidshor @Nate_Cohn I do think it's important to look at one's edge cases! But the Economist guys tend to bring up stuff that's more debatable than wrong, and which I'm pretty sure is directionally the right approach in terms of our model's takeaways, even if you can quibble with the implementation. — PolitiTweet.org

Nate Silver @NateSilver538

@davidshor @Nate_Cohn If, as a result of that, there can be a negative correlation in certain edge cases (e.g. MS and WA) ... I'm not sure that's right but I'm not sure it's wrong either, but I'll certainly take that if it means we can handle a 2016-style regional/correlated polling error better. — PolitiTweet.org

Nate Silver @NateSilver538

@davidshor @Nate_Cohn Wasn't criticizing you, to be clear! It's a hard problem and our model leans heavily into assuming that polling errors are *demographically and geographically correlated* across states. — PolitiTweet.org

Nate Silver @NateSilver538

As I wrote earlier, our model will also use the debate as an excuse to lop off some older polling data. That could help Trump in states where the most recent polling has been a bit better for him (say, WI) and Biden were the reverse is true (say IA). But should be pretty minor. — PolitiTweet.org

Nate Silver @NateSilver538

That Gravis result is a good one for Biden, as IVR polls haven't tended to be friendly to him, although Gravis hasn't polled much recently so it's a bit hard to put it into context. But I wouldn't make too much of one IVR poll and the prior should be that the race is ~unchanged. — PolitiTweet.org

Nate Silver @NateSilver538

In terms of post-debate polling: We've got one day's worth of interviews from the 3 daily tracking polls. Biden moved up slightly in IBD/TIPP, down slightly in USC, and SurveyMonkey was unchanged. Plus a Gravis Marketing poll in PA, fully post-debate (Biden +7). — PolitiTweet.org

Nate Silver @NateSilver538

@davidshor @Nate_Cohn Our correlations actually are based on microdata. The Economist guys continually make weird assumptions about our model that they might realize were incorrect if they bothered to read the methodology. — PolitiTweet.org

Nate Silver @NateSilver538

I think we need to bail out indoor businesses like bars & restaurants to give them the option to stay closed, but there's a weird mantra liberal circles that the only reason they're open is because of a lack of a bailout. It's also because there's some degree of demand for them. — PolitiTweet.org

Nate Silver @NateSilver538

@NickRiccardi Yeah.... I just think the turnout patterns look like they're going to be different this year. Biden voters (whether D or I) seem to have really internalized the message that they need to send their ballots in early. — PolitiTweet.org

Nate Silver @NateSilver538

Or maybe some of them won't vote at all, in which case polls might be underestimating the D turnout advantage and Republicans could get some very bad results. — PolitiTweet.org