Nate Silver

Nate Silver @NateSilver538

I am very willing to believe that China is manipulating its numbers. But these stories on "intelligence officials say China numbers unreliable" don't seem to contain any actual intelligence and mostly seem to cite fact patterns common in all countries. https://t.co/OPcBbXR4dQ https://t.co/nlKPit2qil — PolitiTweet.org

Nate Silver @NateSilver538

I don't know if this will hold, say, 3 weeks from now (to say nothing of 3 months from now, etc.) But so far the degree of fairly level-headed, calm-and-collected adherence to social distancing rules and norms has been higher than you might have thought going into this. — PolitiTweet.org

Nate Silver @NateSilver538

A bit more hopeful on balance than the bleak numbers of the past two days. 28K new cases today vs. 26K yesterday. But testing increased quite a bit, to 118K today from 101K yesterday. — PolitiTweet.org

Nate Silver @NateSilver538

So, the UK had tested 65,000 people as of March 19. Not great, but a lot of countries got off to slow starts. But now they're only up to 163,000 tests. That's only about 7,000 tests per day over the past two weeks.😬Quite pathetic. https://t.co/cU257eNuYq — PolitiTweet.org

Nate Silver @NateSilver538

RT @kat__stafford: NEW: Michigan is now reporting the racial breakdown of COVID-19 cases. So far, 10,791 positive cases have been reported… — PolitiTweet.org

Nate Silver @NateSilver538

Meanwhile, actual Democratic voters are drifting *more* toward Biden as the coronavirus crisis unfolds. The behavior of real people is completely divorced from the narrative on here, a lesson folks probably already ought to have learned re: Biden. https://t.co/9EKwbItFG9 — PolitiTweet.org

Nate Silver @NateSilver538

The delegates are technically free agents (even on the first ballot, *pledged* delegates aren't legally *bound*) but the delegates they are *chosen by the candidates* so you can expect loyalty. The Biden delegates are gonna want Biden unless Biden doesn't want the nomination. — PolitiTweet.org

Nate Silver @NateSilver538

Here's how Andrew Cuomo could be nominated. Joe Biden drops out. That *could* happen. He could get sick. There could be some huge scandal. If Democrats need a last-minute replacement for Biden, then Cuomo might be the choice you'd bet on. Short of that though, very very unlikely. — PolitiTweet.org

Nate Silver @NateSilver538

I hate this line of thinking (AnYtHiNg CaN HaPpEn NoW BeCaUsE TrUmP! 2016! 2020!) — PolitiTweet.org

Nate Silver @NateSilver538

The "good news", I suppose—I am making gigantic air-quote motions as I type this—is that the outlook hasn't changed that much over the past few weeks. These same experts were predicting 200K US deaths two weeks ago and 260K now so not actually that different given the wide range. — PolitiTweet.org

Nate Silver @NateSilver538

2) Some models are only seeking to forecast the current wave of infections. Conversely, the experts in this survey are trying to forecast through the end of the year, which could include a second wave. — PolitiTweet.org

Nate Silver @NateSilver538

Two differences to be mindful of: 1) Some models *assume* that social distancing will remain in place until an appropriate time. The experts the survey do not necessarily assume that, however; they are trying to account for real-world uncertainties. — PolitiTweet.org

Nate Silver @NateSilver538

Want a more realistic range for COVID-19 deaths in the US given all the uncertainties involved? Experts in a survey we're publishing say it's somewhere between 70,000 and 1.7 million. The mean guess is 260K, but the range is wide, obviously. https://t.co/ij2siBJuDl — PolitiTweet.org

Nate Silver @NateSilver538

RT @MattGrossmann: Bernie Sanders’s support declined in counties where a covid-19 case appeared before the primary, driven by a political f… — PolitiTweet.org

Nate Silver @NateSilver538

Italy is back up to some high test volumes (by their standards)—40k tests reported today after having fallen to ~25k/day over the weekend—so the slowdown in their numbers is likely a bit sharper than it looks. — PolitiTweet.org

Nate Silver @NateSilver538

Meanwhile, reducing travel by ~67% in places *without* stay-at-home orders isn't awful. But it could well be—although everyone's guessing here until we look at the data 2 weeks from now—that 67% isn't enough to get R0 <1, whereas 84% is. — PolitiTweet.org

Nate Silver @NateSilver538

Not the headline here, but people reducing travel by ~84% in states with stay-at-home orders seems good. That's a lot. Enough that we might hope to get R0 under 1 in those states (i.e. the number of new infections will start to shrink). https://t.co/7jcq7bpeXZ https://t.co/9kjSPGK3Fd — PolitiTweet.org

Nate Silver @NateSilver538

RT @jazzmyth: This is an excellent story on the 🆕⬅️. The graphics are integral to communicating the results—and they're beautiful! So many… — PolitiTweet.org

Nate Silver @NateSilver538

@PatrickRuffini I also can't imagine that there would be any shortage of demand for tests if they were more widely available. People with fairly severe symptoms often can't get tested. To say nothing of people with mild symptoms or just want to be careful (say, visiting an older family member). — PolitiTweet.org

Nate Silver @NateSilver538

@PatrickRuffini When there's lots of reporting that there's a bottleneck: https://t.co/eUp61h34hu — PolitiTweet.org

Nate Silver @NateSilver538

This is a case where having 5-10 different models to average out would be better. — PolitiTweet.org

Nate Silver @NateSilver538

It's no problem if models change from day to day. If anything people usually err by sticking to their priors too much in the face of new evidence. But I'd like to know, for instance, how many deaths they projected to happen in NY *today* vs. how many actually happened. — PolitiTweet.org

Nate Silver @NateSilver538

Does @IHME_UW keep an archive of what its previous predictions were? It's really important to do so and it's something we try to make easy for people whenever we're running some sort of sports or elections model at 538. — PolitiTweet.org

Nate Silver @NateSilver538

@SethS_D Seems more likely that rich people have an easier time getting tested, thus the denominator is larger in rich neighborhoods. — PolitiTweet.org

Nate Silver @NateSilver538

Frankly, whenever the media reports on the number of new detected cases *without* context on how many tests were conducted, it creates disincentives for governments to do more testing since doing fewer tests may make things look superficially better (see also: China). — PolitiTweet.org

Nate Silver @NateSilver538

But we'd much rather have the number of tests still increasing! (Even though it would lead to the number of detected cases increasing.) We're gonna need a **lot** more tests, and tests with a shorter turnaround time, to detect new hotspots, and then to get back on our feet. — PolitiTweet.org

Nate Silver @NateSilver538

This is the flip-side of what I was talking about a week or two ago when the number of tests was increasing very rapidly. Now that we have a more constant baseline, the nominal rate of increase in detected US cases is not nearly as high. — PolitiTweet.org

Nate Silver @NateSilver538

Pretty clear we've hit a testing bottleneck now. A lot of tests are stuck for processing at private labs. Other states may be limiting tests to very sick/hospitalized people. — PolitiTweet.org

Nate Silver @NateSilver538

Some of the countries with socialized/government-run medical systems seem to get off to much faster and better-organized starts re: testing, but then don't ramp up capacity as rapidly as the more decentralized / privatized countries (like the US) do. — PolitiTweet.org

Nate Silver @NateSilver538

Despite not making the obvious eff-ups of some other countries, the situation in France is not looking great. They'd only conducted 100K tests as of a week ago. Case fatality rate is looking similar to Italy and Spain. https://t.co/9fFIJJjqhz — PolitiTweet.org