Covid 3/​4: Declare Victory and Leave Home

Link post

Health officials look on in horror as individuals both vaccinated and unvaccinated, and state and local governments, realize life exists and people can choose to live it.

This is exactly what I was worried about back in December when I wrote We’re F***ed, It’s Over. The control system would react to the good news in time to set us up to get slammed by the new strains, and a lot of damage can get done before there is a readjustment. The baseline scenario from two months ago is playing out.

The good news, in addition to the positive test percentages continuing to drop for now, is that we have three approved vaccines rapidly scaling up and are well ahead of the vaccine schedule I anticipated, having fully recovered from last week’s dip, and it looks like the new strains are more infectious but not on the high end of the plausible range for that.

The J&J vaccine was approved this week, after a completely pointless three week delay during which no information was found and (for at least the first two-thirds of it) no distribution plan formed. Anything I put at 98%+ on a prediction website isn’t fully news, but the other 2% would have been quite terrible. Supply will initially be limited, but will expand rapidly, including with the help of Merck.

Meanwhile, now that we were provided a sufficiently urgent excuse that we were able to show that mRNA vaccines work, we’ve adopted them to create a vaccine for Malaria. Still very early but I consider this a favorite to end up working in some form within (regulatory burden) number of years. It’s plausible that the Covid-19 pandemic could end up net massively saving lives, and a lot of Effective Altruists (and anyone looking to actually help people) have some updating to do. It’s also worth saying that 409k people died of malaria in 2020 around the world, despite a lot of mitigation efforts, so can we please please please do some challenge trials and ramp up production in advance and otherwise give this the urgency it deserves? And speed up the approval process at least as much as we did for Covid? And fund the hell out of both testing this and doing research to create more mRNA vaccines? There’s also mRNA vaccines in the works for HIV, influenza and certain types of heart disease and cancer. These things having been around for a long time doesn’t make them not a crisis when we have the chance to fix them. And your periodic reminder that the same is true of health’s final boss, also known as aging.

Also, please note that I have been given the opportunity to offer Covid Micro-Grants; see the section below for details. If you can use $1k-$5k to complete a project to help with Covid-19, please don’t hesitate to apply.

Let’s run the numbers.

The Numbers


Last week: 4.9% positive test rate and an average of 2,068 deaths.

Late prediction (Friday morning): 4.5% positive test rate and an average of 1,950 deaths (excluding the California bump on 225).

Result: 4.2% positive test rate and an average of 1,827 deaths after subtracting the California bump.

Great news. I’ve found it pays to be conservative in predicting changes, so when we get the full ‘baseline scenario’ style changes like this, I’m going to undershoot. This was essentially the good scenario, and it bodes well. Deaths continue to lag behind, despite increased vaccination effects for the elderly, in ways I don’t entirely understand. The theory that it’s lag can’t explain the bulk of it because it doesn’t match the past data.


NOTE: Arkansas reported net negative deaths this week, which seems unlikely, so I set them to a plausible but low number (40) instead.

Jan 7-Jan 13628039637383475222378
Jan 14-Jan 20524933867207437020212
Jan 21-Jan 27628132178151422221871
Jan 28-Feb 3552430788071341020083
Feb 4-Feb 10493726877165342918218
Feb 11-Feb 17383722215239270013997
Feb 18-Feb 24365224334782242713294
Feb 25-Mar 3383416695610195813071

There is no plausible story where deaths in the south could be on the uptick for real, but the Arkansas adjustment goes the other way and there weren’t any other glaring mistakes. My assumption is that this is data lag after the storm and isn’t a real change, slash there’s a lot of noise in when deaths are measured in ways that still do not make sense to me but which have happened too many times to not acknowledge.

Positive Tests

Jan 21-Jan 27260,180158,737386,725219,817
Jan 28-Feb 3191,804122,259352,018174,569
Feb 4-Feb 10144,90299,451255,256149,063
Feb 11-Feb 1797,89473,713185,765125,773
Feb 18-Feb 2480,62564,857150,493110,339
Feb 25-Mar 366,15158,295151,253115,426

Test counts bounced back this week and that’s likely accounting for the bumps up in raw positive test counts in the Northeast and South. The situation is still clearly improving. Doesn’t mean I would start lifting mask mandates.

Test Counts

NOTE: This table will not be in future editions unless I can find a new data source for it that’s reasonable to use. Suggestions for a new data source are great.

DateUSA testsPositive %NY testsPositive %Cumulative Positives
Jan 7-Jan 1313,911,52912.2%1,697,0346.6%6.97%
Jan 14-Jan 2014,005,7209.7%1,721,4405.9%7.39%
Jan 21-Jan 2712,801,2718.8%1,679,3995.3%7.73%
Jan 28-Feb 312,257,1237.7%1,557,5504.6%8.02%
Feb 4-Feb 1011,376,5416.4%1,473,4544.1%8.25%
Feb 11-Feb 1710,404,5045.2%1,552,5553.5%8.41%
Feb 18-Feb 249,640,1094.9%1,502,7413.2%8.55%
Feb 25-Mar 310,610,0924.2%1,701,8293.1%8.69%

The bounceback in test counts helps explain how positive test percentages fell so much week over week, and makes trends in New York look troubling. I’m going to be in the city this coming week, and it might be that I got in exactly in time given I’m not yet vaccinated.


Our progress here suddenly looks great. I expected a surge to happen in March and am pleasantly surprised to see it happen this large and this quickly. The one concern is if a bunch of this is catch-up efforts after the snowstorms cleared, in which case we might effectively be back on our old pace for a few more weeks.

The future numbers are even more promising, if you can wait a few months:

I’m quite happy about this of course, and do expect the vaccines to arrive, but in an important sense it’s important to realize this is literally Fake News. What’s fake is the claim that this is news, that something has changed. Nothing changed. Biden has been pursuing a hyper-aggressive policy of under-promise and over-deliver to the point of absurdity, in order to claim maximum credit. This is the natural result. I do understand the motivation, but in addition to the continuing damage to his credibility and government credibility in general (which is bad for vaccines in particular, but in general represents a truth-tracking update) it is of course highly unhelpful. If you want people to hold the line, telling them the end is in sight is exactly what you should be doing. Especially if it’s true.

The question is whether we can count on this pattern to continue. I don’t mean that in a judgemental way, I mean that in a truth seeking way. If we can assume that what is said is designed to make the end result look as impressive as possible, then we can properly evaluate the claims coming from the new administration. We’d get to have Pravda which always lies (in the same directions), instead of the New York Times which keeps you guessing by sometimes telling the truth. It would be especially nice if this pattern extends beyond the pandemic. Presumably at some point there will be a time to claim to have delivered the goods, which complicates matters.

Could it be? Vaccinating people overnight?

We finally are going to vaccinate at night, it seems, in order to make it clear who is getting which vaccine. Or, alternatively, we can think of this as offering the rent-controlled good-but-hard-to-get thing during the day (Moderna/​Pfizer vaccine at a time you want to be awake) versus the market rent good-enough thing at night (J&J vaccine, which you bid on by willing to make a trip in the middle of the night at increasingly terrible hours). It’s a really bizarre way to do a little bit of an obviously correct thing, but at this point we’ll take whatever we can get.

Meanwhile, in North Carolina they have open vaccinations except for those who refuse to lie to government officials, who go to the back of the line:

How much is vaccine capacity worth, and how much are we underinvesting in it even now? About this much.

How good is our vaccine prioritization? About this good:

How much are we gonna have how fast? Hopefully this much, and hopefully faster:

Faster wouldn’t actually surprise me, since we have an authority systematically under promising.


It is Italy’s turn to worry as cases trend upwards. Mostly it seems like Europe is doing what it takes to stabilize things while it suffers several months of extra pain thanks to their collective decision to be penny pinchers with regard to vaccines. That decision seems like the essence of the European project at this point, emphasizing things seeming fair and polite and making sure everything abides by all the rules and regulations, whether or not that is compatible with life. One must not underestimate the value of keeping the peace, but these trends likely keep accelerating, and I doubt it ends well.

Farewell, Covid Tracking Project

On March 7, the Covid Tracking Project will stop collecting data. There are many other data sources out there, but I still don’t have one I’m fully happy with. I primarily want easy access in table form of the number of tests, positive tests, hospitalizations and deaths, on a daily basis, including a full history. This needs to be available for the nation and if at all possible for individual states; more granularity beyond that is a bonus, as is any additional data.

John Hopkins has been suggested as an alternative data source. The data itself seems excellent, but like most places they seem obsessed with giving it to us in graph form rather than table form, which is useful at a glance but super frustrating when I’m trying to create spreadsheets and my own graphs and charts. Also, they list their data source as… the Covid Tracking Project. So they have the same problem I do, and we’ll see if they still have good data next week.

Anyway, once again opening the floor for any suggestions.

The wikipedia data on deaths and positive tests is great, but as far as I can tell it doesn’t include the number of tests, so it doesn’t tell me the denominator (the total number of tests).

Announcing Covid Microgrants

Thanks to a donor who wished to remain anonymous, I am able to offer Covid microgrants. These will be grants of $1000 to $5000 each, for those who have a Covid project which they could finish given this small amount of additional funding. If you’re interested, fill out this Google form. Applications close on 3/​12/​21, and decisions will be quick and based only on my own judgment. I am very curious to see the quantity and quality of applications that come in, and if things go well this could happen again. Please don’t hesitate to apply, or to encourage others to apply.

Insert Mission Accomplished Banner

This happened:

This kind of thing continues to happen, here’s where we were on February 25:

And here’s where they were three days later:

Then the next day, in Texas:

Even San Francisco:

The English Strain

Why do people keep making this mistake over and over again and I don’t mean Greg Abbott:

This is showing up in the case numbers! It’s showing up as a 20%-30% increase in cases!

Very few people who got infected by a B.1.1.7 strain would have otherwise gotten infected by the old strain during this same time period. Very few people who got infected by a B.1.1.7 strain would have been infected if the initial people to have B.1.1.7 had the old strain instead, because its additional infectiousness has grown its share of infections by several orders of magnitude.

Thus, if you have 80 infections with the old strain and 20 with the new, and no one’s had time to change their behaviors in response yet, this is showing up in the case numbers as about 20 new cases. It’s at least 19.

That’s how to track the impact of the new strain: All cases of the new strain should be considered ‘extra’ cases due to the new strain, until there’s enough time that the control system has adjusted behavior to account for the new infections. Period.

The switch to primarily B.1.1.7 infections seems to be poised to happen in early to mid March, which is later than I feared but clearly in the middle of the expected range.

Johnson, Johnson & Merck

In excellent news, pharmaceutical giant Merck, whose Covid-19 vaccine candidate didn’t work out, is going to help make the Johnson & Johnson vaccine (WaPo). Wonderful, and exactly how it should go. There’s available capacity (not necessarily fully free capacity, but this is a priority), everyone makes a deal, profits, looks good and does good doing it, presto.

That’s great news, and can make us even more confident we will have enough vaccine supply in the medium term, and more confident we’ll be able to help vaccinate the whole world soon after.

What this highlights is how bad the delay in approval of J&J’s vaccine was. J&J was already making doses using its own capacity, so there was a story one could tell that while this delayed some doses being delivered by a few weeks, it didn’t destroy capacity or change the long term trajectory. If days after approval, they’re finally getting to a deal to get Merck to step up, it seems very likely this deal had to wait on approval, so this pushed back half or more of J&J’s long term capacity by three weeks. That’s going to kill a lot of people.

Stockholm Syndrome

This is quite the graph, showing weekly Covid levels in the Stockholm wastewater:

(I assume Week 1 here means 2021 Jan 1-7, and so on.)

There is clearly a lot of measurement error here. There aren’t worlds in which week 4’s levels should be more than double both week 3 and week 5’s levels, nor does the jump from 42 to 43 or 34 to 35 make any sense. The last measurement is plausibly a pure data error. My best guess is that the sample isn’t effectively being taken from distinct enough locations and is effectively measuring something too local, and caught a local outbreak? Regardless of the right explanation, there’s still something being measured here, and this is the definition of off the charts. Seems worth noticing.

Noticing this, I checked in with Boston wastewater as well:

There was an upward move, but things seem to have come back and now are below the previous low point this year, so it seems like things are indeed continuing to improve. It does provide an additional suggestion that there was some sort of brief mini-surge corresponding to the uptick in numbers, but I have actual zero idea what could have caused that at that time.

Vaccines Still Work

Vaccines still work, Pfizer single-dose preventing infection edition.

Vaccines still work, Moderna single-dose preventing infection edition. More lowballing.

Vaccines still work, AstraZeneca and Pfizer single-dose edition (paper).

Vaccines still work, take essentially any vaccine you can get edition (MR). Chinese vaccine is the only plausible exception.

Vaccines still work, second doses still wasteful and J&J approval exposes this once again edition (MR).

Vaccines still work, but keep not getting approved, so here’s the rich Germans will fly to Russia, get vaccinated and leave without ever entering the country edition.

Vaccines still work, they all are awesome, but some are better than others and while you should mostly take whatever is available, you should care a nonzero amount about getting the best one you can edition, a Jason Furman Twitter thread.

Vaccines still work, we fully knew this back in July and everyone who stalled things further should be judged accordingly edition.

In Other News

We can all agree Andrew Cuomo is the worst, it seems, due to claims of sexual harassment. We were going to let the causing of and then covering up of thousands of deaths slide – I mean what politician hasn’t done that sort of thing this past year – but we have a zero tolerance policy for sexual harassment that reaches a threshold level of social media prominence. This calls for an independent investigation immediately. I’d summarize my reaction to all this as: I’m not saying Al Capone wasn’t guilty of tax evasion, and also I’m shocked, shocked to find gambling in this establishment.

It appears Operation Warp Speed had to be funded by raiding other sources because Congress couldn’t be bothered to fund it. As MR points out, this is a scandal because it was necessary, rather than because it was done. It’s scary, because it implies that under a different administration Operation Warp Speed could easily have not happened at all.

Catholic Church tells members to avoid J&J vaccine if they can, over concerns about abortion, despite Pope explicilty saying those concerns don’t apply. Divine authority, you had one job!

Another reason you might want to pay money for the things you want:

Shed a tear for maybe it would also have been even more helpful to make the vaccine profitable back when it could have helped increase supply but also take whatever we can get, wherever we can get it.

Doctor Fauci’s defense against First Doses First is a combination of pure FUD and… that it would be a messaging problem?

Also that we’ve already missed the window where this would have helped much, thanks to people like him dragging their feet on this and continuing to drag their feet, so no point in worrying about it now, might as well acknowledge that the foot dragging worked:

At least the ‘this would further blow our credibility’ argument is honest and has content. It’s true that reversing these policies, when the need for first doses first is getting less rather than more urgent, would make those involved look like lying liars and/​or bumbling idiots, who mostly aren’t optimizing for outcomes, and for various reasons they’d prefer a less accurate perspective to retain its popularity.

Fauci’s new position is that ‘there are risks to both approaches’ and to continue to use variations on ‘no evidence’ and to emphasize that the second dose offers an individual additional protection, as if that was in any way in dispute. The concept of a cost/​benefit analysis, or the idea that one might shut up and multiply, let alone form a detailed model full of gears, is clearly not within his range.

At least Canada is increasingly doing First Doses First. Their statement is bold and excellent.

Zeynep post and open thread on pandemic lessons for the future.

Zeynep article in The Atlantic about how our public health messaging has been a disaster.

Post is excellent, and does a great job driving home the central things that went massively wrong with public health messaging. My only quibble is that harms from terrible regulation are treated as beyond scope and not discussed, which is reasonable in context but also feels like ignoring the elephant. Also, if you’ve been following events via my posts, Zeynep’s post is largely a case of You Should Know This Already.

In particular, Zeynep points to five key mistakes: Fear of risk compensation, telling people to use rules instead of mechanisms or intuitions, scolding and shaming especially for outdoor activities (which is a lot of why parks/​beaches were closed while indoor gyms were permitted in many places), failure to support or give people tools for harm reduction while making impossible asks (e.g. no socializing for a year), and sitting on the line of ‘no evidence’ or ‘no clear evidence’ over and over and over again.

And yes, she points out, still doing it:

We did it with masks, with transmission methods and modes of prevention, and now again and also with vaccines.

That’s all an excellent summary of the biggest failures, but I am not convinced it is fair to call them ‘mistakes.’

All of this also isn’t new, this isn’t Covid but seems highly on point (OP has lots more and is great):

Then of course because don’t be absurd and I’d be boggled to find a different answer:

Canada authorizes AstraZeneca.

Dr. Fauci graciously says it’s all right for two vaccinated individuals to have dinner together, citing “common sense” and that the risk is “extremely small.” The implication that all people involved must be vaccinated is clear, so this is a retreat from one insane position to a slightly less insane position.

Update on the White House supercluster of infections, which happened exactly the way one would expect, so no real need to click.

We shouldn’t expect anything less. CDC guidelines for citizen behavior have always been at best aspirational (you could also use the word ‘crazy’) and mostly ignored. This never seemed wise to me, since once one realizes one is not going to do what the authority demands, one often ends up doing little or nothing.

The danger is that we may have entered a new mode where people might actually listen to the CDC guidelines and make serious attempts to get people to follow them, perhaps indefinitely. “Infectious disease specialists” are like any other ‘specialist’, and think everyone should pay dearly to solve the particular problems they think about all day regardless of whether the cost/​benefit analysis would make any sense if someone ever did one. If you didn’t ignore most such ‘specialists’ you’d do nothing else all day and feel bad about falling short anyway.

Is Biden ‘following the science’ (MR) as promised? Tyler Cowen says no and presents his case. The administration allowed the CDC to issue nonsensical guidance that is similar to its usual nonsensical guidance except it’s often going to actually get followed, which is preventing the reopening of many child prisons. AstraZeneca and other vaccines remain unapproved and J&J took three weeks to approve. There is no new head of the FDA and no talk of FDA reforms of any kind. He doesn’t mention vaccine prioritization, which was also massively botched by every metric one might plausibly care about. Post also mentions some non-Covid decisions

I think Cowen’s interpretation here is wrong, and Biden is indeed Following The Science exactly the way he promised. He’s not following the science, in the sense in which science is the collective methods by which people know things, via such actions as doing experiments, gathering data, modeling the world and figuring out what causes and actions might have what effects so as to choose better causes and get better effects. He’s (Following Science

using the Proper Procedures advocated by the Very Serious People and ‘experts.’ Should we have expected anything else? Did we think we were promised anything else?

Not Covid, but Eliezer Yudkowsky science fiction ethos recommendations seem worth sharing.

Administrative Note

This week I will be in New York City. This will be awesome, and I look forward to my permanent return soon. It also means I will have limited resources and time in which to work on the post next week. It may be relatively abridged, and there is some chance it will come out on Friday instead.