Nate Silver

Nate Silver @NateSilver538

The forecast has been fairly stable for a few reasons. 1. It tries to adjust for convention bounces 2. It also hedges against (often volatile) post-convention polling by using some pre-convention data, too 3. It expected the race to tighten to begin with — PolitiTweet.org

Nate Silver @NateSilver538

Our *forecast* really hasn't shifted that much, though. Trump's Electoral College chances are 32%, up from 30% before there were any post-RNC polls and 27% two weeks ago before the DNC. https://t.co/ajG88SznSA — PolitiTweet.org

Nate Silver @NateSilver538

Now our polling average starting to move a bit more aggressively, to Biden +7.1. Keep in mind that's still a fairly sizable lead, though (about Obama's margin of victory over McCain in 2008). https://t.co/cy51vc5isJ https://t.co/HelxcDRWJT — PolitiTweet.org

Nate Silver @NateSilver538

Part of it I guess is you don't look smart by pointing out the obvious, the obvious being that the candidate who has been ahead in the polls all year (Biden) is the favorite, though the question of *how much* of a favorite requires a lot of technical and domain expertise. — PolitiTweet.org

Nate Silver @NateSilver538

Can't resist making one more run at this. A chart like this is nonsense, and the analysis behind it is lacking any sort of empirical anchor and is otherwise hopelessly confused. It's amazing that they shared it with a reporter because they thought it would make them look smart. https://t.co/LKLIqDNLrX — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias @jbarro I'm just saying, the NYC suburbs may not be Vail, but there are a LOT of options that aren't cookie-cutter suburbs either. So I can see a lot of people saying "this is 85% as good as Vail, plus I can go to NYC for culture/friends/family any time I want". — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias @jbarro I don't know New Jersey very well at all, but the northern NYC suburbs are objectively pretty nice (kind of pretty, interesting architecture, small-town feel, good trains to NYC) and I can see people having lots of reasons for wanting to live there. — PolitiTweet.org

Nate Silver @NateSilver538

RT @nataliemj10: Ariel said it better: https://t.co/uwSWY8anuJ — PolitiTweet.org

Nate Silver @NateSilver538

RT @nataliemj10: Re #2: THERE IS NO EVIDENCE OF THIS. Stay in your lane, JPM. https://t.co/kfktYNssWS — PolitiTweet.org

Nate Silver @NateSilver538

By the same token, making one bet at a -10% or -20% ROI every 2-4 years ain't going to cost you very much. You'll sometimes wind up winning anyway, you'll get a fun "sweat", and you'll gain social capital among people who think betting on elections makes you seem sophisticated. — PolitiTweet.org

Nate Silver @NateSilver538

It's also the rarity of elections. Say an expert can achieve a 20% ROI on bets on presidential elections and 10% ROI on midterms. Would be impossible to achieve in nearly any other context involving markets. But can they make a career off one stochastic bet every 2-4 years? No. — PolitiTweet.org

Nate Silver @NateSilver538

Most of what makes political prediction markets dumb is that people assume they have expertise about election forecasting because they a) follow politics and b) understand "data" and "markets". Without more specific domain knowledge, though, that combo is a recipe for stupidity. — PolitiTweet.org

Nate Silver @NateSilver538

Both of these propositions are almost entirely lacking in evidence, to the point where they're more superstitious than empirical, but are an interesting window into the mindset of techbros and financebros who are buying up Trump shares on prediction markets. — PolitiTweet.org

Nate Silver @NateSilver538

@chrislhayes I think it's also a problem though that the media *assumes* it benefits him, and has largely adopted the White House's framing on that point, without citing much if any evidence. — PolitiTweet.org

Nate Silver @NateSilver538

@chrislhayes It's not at all clear that it benefits him politically and it may do the opposite. — PolitiTweet.org

Nate Silver @NateSilver538

The nearest proxy for this issue environment is June, when COVID-19 numbers were getting better and there was a lot of focus on both the protests and the looting that was occurring in some cities. Trump *lost* ground in the polls during that period. — PolitiTweet.org

Nate Silver @NateSilver538

This is smart, but I'd add one more wrinkle: I think one of the flaws in the current conventional wisdom is in the assumption that the focus on crime/protests/unrest will either benefit Trump or have no effect, when it's plausible it benefits Biden instead. — PolitiTweet.org

Nate Silver @NateSilver538

@haralabob Well then hopefully other people can read 538 and make some money by betting on our forecasts. I can't do so myself. — PolitiTweet.org

Nate Silver @NateSilver538

@haralabob Not really. My employer doesn't let me bet on them, I wouldn't think it would be ethical to do so anyway since our numbers can move markets, there's not a ton of liquidity, and there aren't that many elections so it's high-volatility. But, their prices are usually pretty dumb. — PolitiTweet.org

Nate Silver @NateSilver538

Political betting markets have become so dumb as to perhaps be a contrarian indicator at this point. — PolitiTweet.org

Nate Silver @NateSilver538

Everyone's losing their minds with this WhY WoN't BiDeN CoNdEmN ThE LoOTeRs! / Is ThE RaCe TiGhTeNinG?!?! news cycle. Go take the rest of the weekend off. Or take 2 weeks off and come back after Labor Day, which is about when we should get a better read on things. — PolitiTweet.org

Nate Silver @NateSilver538

@ForecasterEnten You should write the first take! https://t.co/qQs7p9z4Re — PolitiTweet.org

Nate Silver @NateSilver538

We're mayyyybe 3 weeks away from the first wave of "New York City is back!" takes. — PolitiTweet.org

Nate Silver @NateSilver538

Let's also be clear what "changing subj" means here. It means Trump is rooting for—perhaps even trying to incite—violent clashes between protestors, counter-protestors and law-enforcement officials, to distract attention from a pandemic that's killed 180,000 Americans. — PolitiTweet.org

Nate Silver @NateSilver538

I'm always a bit wary when people at the top news organizations—who's job in some sense is to decide what's newsworthy—talk about the news cycle as though they have no agency over what the news cycle consists of. — PolitiTweet.org

Nate Silver @NateSilver538

RT @mattyglesias: @NateSilver538 Also, something your model does a good job of clarifying is that time is a factor here — every week that p… — PolitiTweet.org

Nate Silver @NateSilver538

3. Polls show voters give Trump poor marks on protest handling; 4. Polls show voters give Trump poor marks on race relations; 5. Trump is often careless in his messaging and in crisis management. So to me, it seems hard to know how this will play out. — PolitiTweet.org

Nate Silver @NateSilver538

Given the lack of polling evidence so far, are there good priors to assume Kenosha will help Trump? I don't know about that either. Consider: 1. The unrest is happening while he is president; 2. Trump's standing in the polls WORSENED in June amid outbreaks of violence; — PolitiTweet.org

Nate Silver @NateSilver538

If we get down to Biden +4/5 or something, that's different, and about the point at which the speculation starts to be justified. But we haven't seen numbers like that yet, and I'm not sure it's particularly likely that we will in the short term. — PolitiTweet.org

Nate Silver @NateSilver538

If we go from Biden +9 pre-RNC to say Biden +7 post-RNC once we get more polling, I'm not sure that speaks favorably to Trump's re-election odds or to his message working. Bounces are typically a bit larger, in fact, and the post-RNC period should be a high-water mark for Trump. — PolitiTweet.org