Nate Silver

Nate Silver @NateSilver538

@gelliottmorris @DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 You've chosen an aggressive model specification that basically says "highly volatile elections and/or large polling errors are no longer very likely because of polarization", a hypothesis that is really only supported by a few elections ('04, '12, maybe '08 and maybe *not* '16). — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris @DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 This is a sample size of 18, or whatever. You have to be very very careful about how you specify the model *especially* around the fundamentals where the number of defensible model specifications far exceeds the sample size. — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris @DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 You're saying the very high volatility is elections circa 1976-1992 is no longer germane (b/c of polarization) and giving the last few elections extremely high leverage in your model. You may not *realize* that's what you're doing, but that's the upshot of what you're doing. — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris @DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 If you basically (i) throw out a bunch of older data because of some speculative theory about polarization AND (ii) don't make any attempt to accommodate for the fact that we're having an election in the midst of a pandemic, you're going to wind up with an overconfident model. — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris @DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 If you have Biden at 99% to win the popular vote in July then I don't know that you've avoided it. That is very, very high. Even 90% (1 in 10) or 95% (1 in 20) would be quite radically different than 1 in 100. — PolitiTweet.org

Nate Silver @NateSilver538

@DanRosenheck @Nate_Cohn @collprof @Michael_Cobb68 A potential issue here is that "fundamentals" are subject to a large degree of overfitting on the post-WWII data so you probably either want to expand the sample or correct for that somehow. — PolitiTweet.org

Nate Silver @NateSilver538

Plausible hypotheses: 1) New cases are among younger people 2) New cases are on average less severe (masks = lower viral loads?) 3) Cases are actually decreasing too, and the flatness is an artifact of the data (false positives?) 4) Reflects long-term patients being discharged — PolitiTweet.org

Nate Silver @NateSilver538

Slightly curious about the New York COVID data. For the past 6 weeks or so, cases and the positive test rate have been ~flat (with modest fluctuations from day to day). But hospitalizations have fallen almost in half from 1358 on June 18 to 743 today. https://t.co/1ZJ2qHZOiO https://t.co/5jW58hOnf0 — PolitiTweet.org

Nate Silver @NateSilver538

RT @trvrb: @NateSilver538 I'd agree with this framing. Long fuse on infections becoming reported cases and even longer fuse on resulting de… — PolitiTweet.org

Nate Silver @NateSilver538

We've seen that states and people *will* change their behavior when cases are noticeably increasing (say R >= 1.2 or so). But, there's a lag before that improves the situation, during which time things can get pretty out of hand. — PolitiTweet.org

Nate Silver @NateSilver538

Thread. The US never really decided between suppression (keep R<1) and mitigation (take limited measures to e.g. keep hospitals from overflowing; basically, Sweden) and these regionally-driven flare-ups are the somewhat inevitable result. — PolitiTweet.org

Nate Silver @NateSilver538

In general, there's more evidence of a plateau in new cases Arizona than in Florida, and more in Florida than in Texas. But these states aren't the only sources of concern. Other parts of the Southeast are problems, for instance. — PolitiTweet.org

Nate Silver @NateSilver538

Second day in a row with a record number of cases. We're also seeing a lot of tests reported, especially in some of the worst-affected states, and Texas also reported quite a few (5K) backlogged cases today. So this is very very bad, though maybe slightly less bad than it looks. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 951 Yesterday: 977 One week ago (7/10): 854 Newly reported cases T: 77K* Y: 71K 7/10: 67K Newly reported tests T: 852K* Y: 831K 7/10: 823K Positive test rate T: 9.1% Y: 8.6% 7/10: 8.1% * record — PolitiTweet.org

Nate Silver @NateSilver538

That's partly because there are likely some unobserved characteristics (maybe the voter in Florida is Cuban-American and the one in California is Mexican-American) but also because states have different political cultures and are subject to different levels of campaign activity. — PolitiTweet.org

Nate Silver @NateSilver538

Part of the issue is that seemingly demographically similar voters in different states may not actually vote all that similarly. e.g. a 46-year-old Hispanic man in California probably has different views on average than a 46-year-old Hispanic man in Idaho or Florida... — PolitiTweet.org

Nate Silver @NateSilver538

It's a decent stand-in when you have few polls and weak priors. Our models do use it a little bit of it. But generally speaking, looking directly at the polls in a state is quite a bit more accurate. And often, so are simpler "fundamentals" methods (e.g. national polls + PVI). — PolitiTweet.org

Nate Silver @NateSilver538

At the risk of restarting the MRP wars: For the last 3 models I've designed (midterms, primaries, now revisiting stuff for the general) trying to impute how a state will vote based on its demographics & polls of voters in other states is only a mediocrely accurate method. — PolitiTweet.org

Nate Silver @NateSilver538

@conorsen Yeah, I have no way to assess those odds other than that people whom I might expect to be skeptical seem increasingly optimistic. I do think it'll take some time to get everyone vaccinated and that the first vaccine might not be the most effective one. — PolitiTweet.org

Nate Silver @NateSilver538

I do think it's worth pausing to note that COVID-19 is only ~7 months old and yet there are already a lot of very promising developments in terms of treatments and vaccines and just general knowledge about the disease. https://t.co/xO1DCnHa4i — PolitiTweet.org

Nate Silver @NateSilver538

RT @galendruke: 🎧 New pod 🎧 White Democrats express a desire for racial equality and inclusiveness but don't always support policies aimed… — PolitiTweet.org

Nate Silver @NateSilver538

@_schem I'm The Raptor obviously — PolitiTweet.org

Nate Silver @NateSilver538

RT @_schem: tag yourself i'm KC Wolf https://t.co/n0Uc1l6KGD — PolitiTweet.org

Nate Silver @NateSilver538

RT @_schem: had a ball of a time working on this @JPlanos piece, featuring incredible art by Clara Kirkpatrick (https://t.co/jNXQNyKg6k) an… — PolitiTweet.org

Nate Silver @NateSilver538

And Trump has an advantage in a standard-fare partisan election because of the Electoral College. — PolitiTweet.org

Nate Silver @NateSilver538

Still, I can imagine that—with a low bar for success and people fatigued by months of lockdowns—he'd have a more tenable position with competing claims ("worst is behind us" vs. "still really bad") that got treated by voters/media a fairly standard-fare partisan argument. — PolitiTweet.org

Nate Silver @NateSilver538

But there probably is a narrow window where by Nov., there have been ~2 months of real improvement. Would that be enough to reelect Trump? I don't know. I tend to think the public would still give him bad marks for COVID, and keep in mind he was behind before any of this started. — PolitiTweet.org

Nate Silver @NateSilver538

What we can say for sure is that the US has *already been hit really hard*. And even if the numbers are improving by the fall—which, to be clear, may take some luck—we'd probably be looking at spring before there's a robust economic recovery and things feel 85% "normal" again. — PolitiTweet.org

Nate Silver @NateSilver538

THREAD: In forecasting Trump's fate, one needs to be careful as there is a fair amount of uncertainty as to what the trajectory of the pandemic will look like by November. Overly precise predictions of COVID numbers in X country/state by Y date have generally not fared well. — PolitiTweet.org

Nate Silver @NateSilver538

One "fun" thought: the period from, oh, say, Nov. 3 through Jan. 20 could be critical to building out the US's vaccine infrastructure, and yet that's a period where you could have a lame-duck president and one can imagine Trump being less than helpful under those circumstances. — PolitiTweet.org