Covid 8/​20: A Little Progress

Link post

(Writer’s note: Due to the ut­ter train­wreck that is WordPress’ new ed­i­tor, this week’s post was writ­ten in Google Docs and then cut and pasted into WordPress. This is an ex­cel­lent ex­am­ple of the prin­ci­ple that any given thing is likely get­ting worse even as things in gen­eral im­prove. As an ex­per­i­ment, I’m open­ing the draft doc­u­ment to pub­lic com­ments – I think com­ment­ing on the side as the post goes is ac­tu­ally a bet­ter method, and I en­courage you to read it and com­ment on it there.)

The Covid news is not as ex­cel­lent as it looks at first, but it is still good.

The pos­i­tive test per­centages, and the num­ber of peo­ple test­ing pos­i­tive, are drop­ping fast. The num­ber of peo­ple hos­pi­tal­ized is drop­ping as well. The num­ber of re­ported deaths mostly con­tinues on its plateau, but that al­most cer­tainly re­flects a drop­ping num­ber of ac­tual deaths by this point.

Those pos­i­tive test per­centages look less im­pres­sive when bro­ken out by re­gion, which we do this week for the first time, but im­prove­ment is still im­prove­ment.

There’s even a rapid test freshly ap­proved by the FDA, thanks to the NBA. More on that in its sec­tion, but don’t say sports never did any­thing for you.

There’s also a lot of noise out there. There’s some new offi­cial ly­ing for our own good I need to com­plain about.

It’s still im­por­tant to cel­e­brate the good times, when the times are good. Even if those times aren’t as good as we would like or hope.

Thanks to some C# code, I am now down­load­ing and pars­ing the .csv files from CovidTrack­ingPro­ject, which makes the data far eas­ier to an­a­lyze.

Let’s run the newly ex­panded num­bers.

Pos­i­tive Test Counts

June 4-June 10 35487 24674 55731 22693
June 11-June 17 41976 22510 75787 17891
June 18-June 24 66292 26792 107221 15446
June 25-July 1 85761 34974 163472 16303
July 2-July 8 103879 40139 202863 18226
July 9-July 15 108395 53229 250072 20276
July 16-July 22 117506 57797 265221 20917
July 23-July 29 110219 67903 240667 26008
July 30-Aug 5 91002 64462 212945 23784
Aug 6-Aug 12 93042 61931 188486 21569
Aug 13-Aug 19 80887 63384 156998 20857

Here we see a hope­ful con­tinu­a­tion of pre­vi­ous trends. The South and North­east con­tinue to rapidly im­prove. The West has re­sumed im­prove­ment af­ter last week’s set­back. The Mid­west con­tinues to be stuck in place.

Death Counts

June 4-June 10 743 1297 1230 1936
June 11-June 17 778 1040 1207 1495
June 18-June 24 831 859 1204 1061
June 25-July 1 858 658 1285 818
July 2-July 8 894 559 1503 761
July 9-July 15 1380 539 2278 650
July 16-July 22 1469 674 3106 524
July 23-July 29 1707 700 4443 568
July 30-Aug 5 1831 719 4379 365
Aug 6-Aug 12 1738 663 4554 453
Aug 13-Aug 19 1576 850 4264 422

This shows a large in­crease in deaths in the Mid­west, and a de­cline in the West. I am guess­ing that some of last week’s deaths in the Mid­west got de­layed, and things are much closer there to be­ing steady. The South saw no­tice­able im­prove­ment, which makes sense, but it’s too early to know for sure. The North­east is back about where it makes sense for the North­east to be.

Test Counts (Old Chart)

Here’s the chart I make from the spread­sheet each week:

Date USA tests Pos­i­tive % NY tests Pos­i­tive % Cu­mu­la­tive Pos­i­tives
June 11-June 17 3,446,858 4.6% 442,951 1.1% 0.66%
June 18-June 24 3,638,024 6.0% 440,833 1.0% 0.72%
June 25-July 1 4,331,352 7.1% 419,696 1.2% 0.82%
July 2-July 8 4,461,980 8.2% 429,804 1.1% 0.93%
July 9-July 15 5,196,179 8.4% 447,073 1.1% 1.06%
July 16-July 22 5,481,861 8.5% 450,115 1.1% 1.20%
July 17-July 29 5,757,290 7.8% 448,182 1.1% 1.34%
July 30-Aug 5 5,079,828 7.9% 479,613 1.0% 1.46%
Aug 6-Aug 12 5,014,573 7.5% 502,046 0.9% 1.58%
Aug 13-Aug 19 5,229,319 6.3% 543,922 0.8% 1.68%

A small bounce back in ca­pac­ity is very good to see given the al­ter­na­tive. The drop in pos­i­tive rate is large and highly wel­come. It’s the best look­ing statis­tic we’ve seen in months. But see the next sec­tion.

New York’s pos­i­tive test per­centages con­tinue to drop. I worry that this effect is mostly ‘we have more test­ing ca­pac­ity and marginal tests are all com­ing back nega­tive’ rather than a drop in in­fec­tions. Of course, in im­por­tant ways that is ex­cel­lent. It means test­ing truly is ad­e­quate, and the true in­fec­tion rate might not be much higher than the pos­i­tive test rate.

Pos­i­tive Test Per­centages and Hospi­tal­iza­tions By Region

Cur­rent Hospi­tal­iza­tions (av­er­age daily value over pe­riod):

Date North­east Mid­west South West
326 to 41 10977 116 1524 1625
42 to 48 24021 1716 3937 3835
49 to 415 33757 9068 6160 6466
416 to 422 34173 9753 6639 7408
423 to 429 31065 10467 7383 7609
430 to 56 26378 11058 8579 7174
57 to 513 21709 10505 8376 7250
514 to 520 17801 9462 7975 7178
521 to 527 14333 8718 8015 7073
528 to 63 11859 7826 8053 6950
64 to 610 9077 6740 8160 7232
611 to 617 6688 5661 8642 7286
618 to 624 5308 5032 10203 8266
625 to 71 4446 4920 13462 10324
72 to 78 3949 5075 17846 12427
79 to 715 3899 5630 28559 13892
716 to 722 3904 6127 33428 14523
723 to 729 3875 6594 33706 14255
730 to 85 3657 6672 30839 12863
86 to 813 3412 6703 27897 11379
813 to 820 3141 6639 24275 10214

This is the first time I’ve pre­sented the hos­pi­tal­iza­tion data, so I figured it made sense to go back a long way.

I’ve been skep­ti­cal of us­ing the hos­pi­tal­iza­tion data due to all the ways that data can be warped, but at this point it seems like a long view over time offers enough in­sight to be worth­while. It is an­other per­spec­tive, and it sug­gests strongly that the cor­ner has been turned.

Pos­i­tive Test Per­centages (Note: Wash­ing­ton has been given 100k nega­tive tests per week be­cause it stopped re­port­ing nega­tive tests over two weeks ago)

For in­di­vi­d­ual states, see the chart in the Pos­i­tive Test Per­centages tab on the spread­sheet.

Per­centages North­east Mid­west South West
326 to 41 32.17% 16.70% 11.55% 11.84%
42 to 48 36.20% 16.16% 12.97% 8.11%
49 to 415 35.03% 16.49% 11.27% 9.45%
416 to 422 30.92% 17.28% 10.66% 5.55%
423 to 429 22.07% 13.78% 7.41% 7.56%
430 to 56 17.88% 12.44% 6.80% 6.29%
57 to 513 9.92% 9.40% 5.88% 5.01%
514 to 520 8.49% 7.19% 4.71% 4.33%
521 to 527 5.41% 6.72% 5.77% 4.20%
528 to 63 4.24% 4.61% 6.18% 5.04%
64 to 610 2.65% 3.96% 6.20% 5.08%
611 to 617 2.04% 3.32% 6.97% 5.35%
618 to 624 1.90% 3.82% 9.49% 7.24%
625 to 71 1.78% 4.08% 11.61% 8.32%
72 to 78 2.20% 4.71% 13.15% 8.59%
79 to 715 2.11% 5.12% 13.77% 8.53%
716 to 722 2.49% 5.13% 13.29% 8.56%
723 to 729 2.54% 5.51% 12.32% 7.99%
730 to 85 2.58% 7.26% 12.35% 6.68%
86 to 813 2.30% 5.67% 14.67% 6.98%
813 to 820 2.06% 5.62% 9.41% 6.47%

This week’s dra­matic drop was driven by the South, which went from an out­lier high num­ber that likely re­flected data er­rors, to an out­lier low num­ber that prob­a­bly also re­flects data er­rors. My guess is that some nega­tive tests got time-shifted from last week into this week. If you in­te­grate the last two weeks, you get a pos­i­tive rate for the South of 11.7%, which is a small but no­tice­able im­prove­ment over what came be­fore that. We also see the North­east and West reach their low­est lev­els in a while.

Note that these num­bers came from us­ing C# code to parse the CovidTrack­ingPro­ject’s .csv files, so there is a larger than nor­mal chance of mean­ingful er­rors creep­ing in. We man­u­ally fixed Wash­ing­ton. If you find an im­por­tant er­ror, please do let me know.

Put­ting all these per­spec­tives to­gether yields a con­sis­tent pic­ture.

Put­ting the Num­bers Together

Things are slowly im­prov­ing in the North­east and West, and im­prov­ing some­what more rapidly in the South, but not rapidly enough that we can be con­fi­dent putting kids back in school or colder weather won’t re­verse those trends. In the Mid­west, things are mostly hold­ing steady.

My pre­sump­tion is that the Mid­west is not im­prov­ing be­cause it is not get­ting suffi­cient benefits from herd im­mu­nity.

Go­ing state by state, we see dra­matic im­prove­ments con­tinue in Ari­zona, where the pos­i­tive rate was over 20% from 618 to 730 and is now down un­der 10% and fal­ling fast. Hospi­tal­iza­tions are also down by more than half. Alabama and South Carolina and Arkansas are also re­cov­er­ing fast, so fast that the drops in pos­i­tive test rates look sus­pi­ciously like data er­rors. Florida is re­cov­er­ing far slower, still with a 15% pos­i­tive rate, but down sub­stan­tially from peak and fal­ling. Texas has too many data er­rors to tell for sure.

We see sub­stan­tial im­prove­ment in Cal­ifor­nia, al­though that could largely be them fix­ing data er­rors.

The only state where things are clearly rapidly get­ting worse is Hawaii. From mid-April un­til mid-July they held pos­i­tive rates un­der 2% but now they’re headed up steadily and are over 10%. Hawaii has all the signs of an un­con­trol­led wave that’s go­ing to look like what re­cently hap­pened in Ari­zona, Texas and Florida.

Look­ing for­ward, we should ex­pect the death rates to slowly drop, and the hos­pi­tal­iza­tion rates to also slowly drop. Pos­i­tive test num­bers are likely to re­bound up­ward a lit­tle bit, be­cause this week’s rate was likely ar­tifi­cially low, but the real pos­i­tive rate is likely to con­tinue to drop.

Alas, none of this progress looks to be fast. Un­til we get tech­nolog­i­cal in­no­va­tion, that seems un­likely to change for the bet­ter. Which means that for all prac­ti­cal pur­poses, you’ll be stuck in this same limbo for many months.

The pos­si­ble ex­cep­tions are the heart of the North­east re­gion plus Ari­zona.

Ari­zona may have overkil­led the herd im­mu­nity thresh­old suffi­ciently that it can squash things within a few months. It would be un­sur­pris­ing if that’s the safest place in the coun­try by the end of the year, and the place most con­fi­dent that it’s go­ing to stay that way.

The core North­east, mean­ing New York, New Jersey and New England but not Penn­syl­va­nia or Delaware, has strong test­ing in place, strong ex­ist­ing herd im­mu­nity lev­els and low cur­rent in­fec­tion lev­els. New York City is, as Nate Silver noted this morn­ing, ar­guably the safest place in the coun­try.

So while things will likely re­main mostly static, in those ar­eas, that’s good. Static works. If I knew things would re­main static, I’d be look­ing for a new place in Man­hat­tan to re­turn to. In­stead, I’m wait­ing a few more months to avoid the un­cer­tainty of the win­ter and school, as well as the elec­tion, be­fore tak­ing that step.

C.D.C. Once Again Reap­plies to De­lenda Est Club

I am go­ing to re­ject their ap­pli­ca­tion once again. They are al­most cer­tainly un­der ex­treme pres­sure from the White House that has crip­pled their abil­ity to re­spond, and have mostly done (as far as I can tell) ex­cel­lent work in the past. On this par­tic­u­lar is­sue, they’re say­ing some­thing rather than say­ing noth­ing, and I don’t want to make the perfect the en­emy even of the lousy, let alone the good. They have my sym­pa­thies. I want to cut them as much slack as pos­si­ble.

But no more than that. Please don’t push it.

This week’s ap­pli­ca­tion is in the form of an ad­vi­sory on those who have had Covid-19 and re­cov­ered.

In that ad­vi­sory, they say that for three months, such peo­ple do not need to quaran­tine, but they do need to con­tinue mask wear­ing and other so­cial dis­tanc­ing mea­sures.

This, of course, ended up be­ing used pri­mar­ily as an­other way for the Very Se­ri­ous Peo­ple to warn that im­mu­nity is go­ing to fade Real Soon Now and can’t be re­lied upon, so we need to keep sac­ri­fic­ing to the Gods and pun­ish­ing any­one who doesn’t, lest we in­vite the Gods’ wrath.

If you say that some­one is safe for three months, but not for four months, you are con­spicu­ously say­ing that they are not safe in month four. So that be­came the head­line.

As I some­how have to keep say­ing: Four months from March would be July, so we know that im­mu­nity not last­ing four months is ut­ter non­sense. At a bare min­i­mum it al­most always lasts five.

Then there’s the mat­ter of still re­quiring such peo­ple to con­tinue so­cially dis­tanc­ing. Quaran­tin­ing when called upon is far more effi­cient and im­por­tant than so­cially dis­tanc­ing. If you don’t need to quaran­tine, you don’t have to so­cially dis­tance.

Un­less so­cial dis­tanc­ing is about sac­ri­fic­ing to the Gods, rather than about not spread­ing the virus. And the CDC is say­ing you have to put your life on hold be­cause oth­er­wise peo­ple might see you not putting your life on hold.

If that’s what you mean, come out and say it, you ly­ing cow­ards. See if peo­ple agree with you.

If that’s not what you mean, then I am here to in­form you this is not how this works, nor is it how any of this works.

This blog’s offi­cial po­si­tion (not med­i­cal ad­vice!) is that if you are sure you have pre­vi­ously had Covid-19 and you are sure you have re­cov­ered, then you can pre­sume you are effec­tively in­definitely im­mune. You can act as if you can nei­ther be in­fected nor in­fect oth­ers, un­til such time as we get con­firmed me­dia re­ports of large num­bers of re­in­fec­tions. When and if that hap­pens, you can do the math on timing and figure out how long you have. Un­til then, your ac­tions are al­most purely sym­bolic.

Which mat­ters! You’re still set­ting an ex­am­ple. Con­sider set­ting a good one. But do it on pur­pose. Eyes open.

Tyler Co­hen echoes this with his re­minder about T-Cell im­mu­nity, which he calls the dog that didn’t bark. As Tyler says, such re­sponses are com­mon and well-known to those in­vested in im­munol­ogy and pub­lic health. Yet we effec­tively didn’t hear any­thing about them un­til Au­gust. If there is po­ten­tially good news, the Very Se­ri­ous Peo­ple see it as threat­en­ing to un­der­mine peo­ple’s will­ing­ness to sac­ri­fice, and thus the news is sup­pressed. It’s not sup­pressed by some sinister con­spir­acy, but rather by mak­ing sure that it is in no one’s per­sonal in­ter­est to talk about it, and let­ting na­ture (and Na­ture) take its course.

Could You De­tect Covid-19 Via Voice Changes?

That’s a pre­limi­nary the­ory com­ing out of MIT. It sounds like their sam­ple sizes are still su­per low and putting any value on this is still very pre­ma­ture, but I figured I’d share it any­way be­cause it’s a pretty cool con­cept.

Col­lege Without Foot­ball Poses Risks To Stu­dent Athletes

Univer­sity of Ok­la­homa head foot­ball coach Lin­coln Riley an­nounced Satur­day that nine of his play­ers tested pos­i­tive for Covid-19. The re­sults came back af­ter the team re­turned from a one-week break from pre­sea­son prac­tices.

Said Riley: “You know when you give play­ers time, there is risk in that. This isn’t the NBA, we don’t have a bub­ble. We all have to con­tinue to work to do a bet­ter job by all ac­counts. We’re still con­fi­dent in the plan we have.”

In similar news, UNC Chapel Hill can­cel­led its in-per­son classes af­ter find­ing mul­ti­ple Covid clusters and run­ning a pos­i­tive test rate of 13.9%. Which to me seems like an over­re­ac­tion, and an­other ex­am­ple of re­act­ing to the in­evitable re­sult of a difficult trade-off with apolo­gies for one’s lack of sac­ri­fices and a cho­rus of cries that “I’ve made a huge mis­take.” A com­mon theme here is that if what you are do­ing is seen as reck­less, then a mostly or­di­nary rate of pos­i­tive tests will be seen as highly blame­wor­thy, whereas if what you are do­ing is not seen as reck­less, it’s con­sid­ered fine.

Coach Riley’s com­ments re­mind us what is at stake. It’s not safe for stu­dents to be on cam­pus not play­ing foot­ball. They’re go­ing to break the rules, and they’re go­ing to get in­fected. What they need are some semi-pro­fes­sional sports that al­low them to fol­low proper dis­tanc­ing rules and get the fre­quent test­ing they need to stay safe. The good news is that a large por­tion of those who never played sports are prob­a­bly huge nerds, who both fol­low rules and are already so­cially dis­tant to be­gin with. So to­gether the jocks and nerds can work to­gether and beat this thing. Unity 2020.

Or if that’s not good enough, send ev­ery­one home en­tirely. Be­cause then, when they get in­fected re­gard­less, at least it won’t legally be in any way your fault.

World’s Most Im­por­tant Scien­tific Re­search Performed by NBA

A col­lab­o­ra­tion be­tween the NBA and Yale, funded by a grant from Emer­gent Ven­tures at the Mer­ca­tus Cen­ter at Ge­orge Ma­son Univer­sity, has led to the cre­ation of a sal­iva-based test. That test has now been ap­proved for emer­gency use by the FDA.

The test can be done in two hours, from sal­iva, with­out spe­cial­ized equip­ment. Its method is be­ing open sourced, so any lab can perform it. They an­ti­ci­pate it will cost about $10, and it is roughly as ac­cu­rate as a PCR test.

Over time, this could al­low us to test ev­ery­one pe­ri­od­i­cally, which would in turn be suffi­cient to squash the pan­demic en­tirely. We could then re­turn to our old lives even with­out a vac­cine.

The next sec­tion will talk about the role of Emer­gent Ven­tures here. The role of Yale was as a place that is still ca­pa­ble of do­ing at least some ac­tual sci­ence when con­di­tions are suffi­ciently good and the fund­ing is available.

The key thing to un­der­stand here is that this could not have been done with­out the NBA and its bub­ble.

The NBA bub­ble had three nec­es­sary el­e­ments.

First, the NBA bub­ble cre­ated a pop­u­la­tion that could be phys­i­cally con­strained, mon­i­tored in­tensely and tested fre­quently, via mul­ti­ple meth­ods, with­out the need to con­sult with ‘med­i­cal ethi­cists’ and go through an IRB. It was a nat­u­ral ex­per­i­ment. This al­lowed tests to be com­pared and eval­u­ated in a way they couldn’t be el­se­where.

Se­cond, the NBA bub­ble cre­ated a large di­rect fi­nan­cial in­cen­tive, jus­tify­ing and pay­ing for most of the ex­penses in­volved, and thus al­low­ing the amount of sci­en­tific funds re­quired to finish the job to be rel­a­tively small.

Third, it cre­ated enough of a sense of need and ur­gency from oth­ers that the Very Se­ri­ous Peo­ple and med­i­cal ethi­cists, who un­der nor­mal cir­cum­stances never let a good vol­un­tary co­op­er­a­tion to­wards the com­mon good hap­pen on their watch with­out go­ing through all the proper chan­nels, didn’t seek to shut the whole thing down in or­der to show how very se­ri­ous they are. There was lit­tle or no talk about how ‘waste­ful’ this was, or how ‘in­equal­ity’ was in­volved, or any other non­sense like that.

This is how we are al­lowed to figure things out in 2020. We can use a spe­cial case like the NBA bub­ble to cre­ate a world-sav­ing test as a side effect, and then maybe be gra­ciously al­lowed to share it with the rest of the world.

If there is a fully nat­u­ral ex­per­i­ment go­ing on where all the work hap­pens to be done any­way, then maybe it’s eth­i­cal to an­a­lyze it and figure some­thing out. This is from ESPN’s ar­ti­cle on the test.

“My goal is not to test ath­letes,” Grubaugh said. “That’s not my tar­get pop­u­la­tion. My tar­get pop­u­la­tion is ev­ery­body. There were con­cerns about part­ner­ing with the NBA when all these other peo­ple need test­ing. But the sim­ple an­swer ended up be­ing the NBA was go­ing to do all this test­ing any­way, so why not part­ner with them and try to cre­ate some­thing for ev­ery­one?”

Good thing the NBA was do­ing all this deeply awful test­ing any­way, thus al­low­ing sci­ence to take place. Ever since we banned that highly un­eth­i­cal thing called ‘sci­en­tific ex­per­i­ments’, it’s been rough out there.

World’s Most Im­por­tant Scien­tific Re­search Funded By Blog­ger’s Per­sonal Fund

I wrote this post two years ago urg­ing my read­ers to ap­ply for grants from Emer­gent Ven­tures. In it, I pre­dicted that Emer­gent Ven­tures was likely to have a big­ger im­pact than the two billion dol­lars that had re­cently been given away by Jeff Be­zos (which, as I noted then, was a pre­ma­ture mis­di­rec­tion of funds from the world’s great­est char­ity, which as we all know is Ama­, but was definitely di­rected to­wards a good if in­effi­cient cause).

I pre­dicted this be­cause Tyler Cowen would be dis­tribut­ing the money per­son­ally, based on quick turnaround de­ci­sions to find those who could use the money to study or do things with po­ten­tially high im­pacts, with­out a bu­reau­cracy or wor­ries about how things looked weigh­ing him down. Also with­out the need to prove effort or hours worked or any­thing else weigh­ing down those who got granted money. The ‘se­cret sauce’ was to make sure not to ‘keep the crazy off his desk.’

Two years later, the fund has ex­ceeded any rea­son­able ex­pec­ta­tion with this one grant alone, play­ing a key role in a po­ten­tial dra­matic ad­vance­ment of Covid-19 test­ing. It is likely that Emer­gent Ven­tures will, by the end of 2021, have had a mean­ingful im­pact on the world’s over­all econ­omy, health and well-be­ing.

If you are in the busi­ness of us­ing money to at­tempt to make the world a bet­ter place, or even if you are in the busi­ness of pre­tend­ing to do that in or­der to feel bet­ter about your­self or look bet­ter to oth­ers, or to help oth­ers in do­ing any of that, please don’t ig­nore this data point!

For those whose char­i­ta­ble fo­cus is any­thing other than ex­po­nen­tial risk, and who don’t want to take di­rect ac­tion them­selves or fund in­di­vi­d­ual pro­jects di­rectly, if your money’s first stop is not en­sur­ing that Emer­gent Ven­tures is fully funded, that seems like a mis­take.

This is how you give your char­ity the chance to save the world. Not with a whim­per or a com­mit­tee. With a rene­gade who plays by his own rules. With a bang.

Take Me Out to the Ballgame

The St. Louis Car­di­nals, in the wake of their Covid-19 out­break, had 41 peo­ple trav­el­ing with the team from St. Louis to Chicago.

They rented 41 cars and had ev­ery­one drive them­selves there.

If you want to be safe, be safe.

Noth­ing is im­pos­si­ble if you have heart.

Tac­ti­cal Vaccination

A pa­per that came out this week showed a re­sult that seems su­perfi­cially strange to many, but which struck me as in­tu­itive.

Sup­pose we have a vac­cine. Should we give it to old peo­ple first, be­cause they are more vuln­er­a­ble? Or should we give it to young peo­ple first, be­cause they spread the virus more?

Their con­clu­sion was that if we had a highly effec­tive vac­cine (>50% effec­tive) we should give it to young peo­ple, be­cause it can stop the spread of the virus. Whereas if it’s not as effec­tive, we can’t stop the virus that way, so we should in­stead vac­ci­nate the el­derly so they don’t die.

That seems… ob­vi­ously di­rec­tion­ally right?

With any pan­demic, there are es­sen­tially ‘suc­cess’ wor­lds and ‘failure’ wor­lds. Failed wor­lds rely on herd im­mu­nity and let things run their course, re­gard­less of how fast or slow that out­come is reached. Suc­cess­ful wor­lds con­tain the virus be­fore that hap­pens. If you have a chance to move from failed to suc­cess­ful, or make your­self suc­cess­ful faster with many less in­fec­tions, that is far more im­por­tant than other marginal strate­gic im­prove­ments.

How­ever, if you’re go­ing to fail ei­ther way, then pro­tect­ing the most vuln­er­a­ble and oth­er­wise miti­gat­ing the dam­age be­comes the pri­or­ity.

Shifts be­tween jus­tifi­ca­tion via ‘this is for miti­ga­tion’ and ‘this is to ac­tu­ally suc­ceed’ have been a big prob­lem.

If High School Must Ex­ist, Teach Statis­tics In High School

Here is this week’s ‘what do the school num­bers ac­tu­ally look like’ ex­er­cise. This time it comes from Mis­sis­sippi.

Ash­ton Pittman gives us the num­bers on Mis­sis­sippi schools: https://​​twit­​​ash­ton­pittman/​​sta­tus/​​1295462052757221378

He meant it as an ob­vi­ous scan­dal. Look at all these out­breaks!

Google says: In 2013 Mis­sis­sippi had 493,650 stu­dents en­rol­led in a to­tal of 1,063 schools in 162 school dis­tricts (ver­sus 82 coun­ties).

The first thing to note is that there are 71 school dis­tricts with at least one in­fec­tion, and the to­tal in­fec­tion count is 444. The av­er­age ‘cluster’ is 6 peo­ple, if that cluster counts ev­ery­one in the dis­trict. Which does not seem re­motely rea­son­able. So it doesn’t look like we’re see­ing huge out­breaks.

There are 493,650 stu­dents. 199 of them have tested pos­i­tive. That’s 0.04%. Hard to be con­cerned about that, or think it means kids are catch­ing Covid-19 at school at an alarm­ing rate.

There are 32,311 pub­lic school teach­ers. 245 of them have tested pos­i­tive. That’s 0.76%, which is far more con­cern­ing.

That’s a While ‘dozens’ of schools just had their first day of class (which, of course, means that at those schools the in­fec­tions we’ve found didn’t hap­pen at school, yet some­how are always used as a fur­ther con­dem­na­tion of the safety of schools!), there are a thou­sand schools in the state, so let’s say that on av­er­age they’ve been open for two weeks.

So let’s pre­sume these statis­tics come from the past two weeks, and see how it com­pares to the pop­u­la­tion.

Dur­ing that pe­riod, the state had 11,008 pos­i­tive tests out of a pop­u­la­tion of 2.976 mil­lion, which is 0.37%. Our source for es­ti­mates sug­gests that fully 3% of Mis­sis­sippi caught the virus dur­ing that time.

Thus, as is usu­ally the case in such things, the 0.76% rate would be scary if teach­ers are be­ing tested as of­ten as oth­ers in the state, but not scary at all if they are be­ing tested much more ag­gres­sively. Given how im­por­tant it is to de­tect such cases, both in re­al­ity and in terms of per­cep­tion, I would be very sur­prised if more test­ing did not ex­plain this en­tire differ­ence.

I’m still not op­ti­mistic about open­ing up the schools, but when­ever I get data like this, it seems to make me less wor­ried rather than more wor­ried. We’ll see what hap­pens af­ter a few weeks. I’ll keep do­ing similar calcu­la­tions in the fu­ture when I see clean data to work with.

If you have a good source, es­pe­cially a good source that ac­tu­ally shows dan­ger, please share in the com­ments. As a re­minder, com­ments are open on the Google doc as well, as an ex­per­i­ment.

New Biggest Idiot

You may think you know who the biggest idiot prize should go to, but I be­lieve that this week we have found a new cham­pion. Be­hold, Swe­den’s An­ders Teg­nell:

“It is very dan­ger­ous to be­lieve face masks would change the game when it comes to COVID-19,” said Ten­gell, who is con­sid­ered the coun­try’s equiv­a­lent of Dr. An­thony Fauci from the White House COVID-19 task force.

He noted that coun­tries with wide­spread mask com­pli­ance, such as Belgium and Spain, were still see­ing ris­ing virus rates.

“Face masks can be a com­ple­ment to other things when other things are safely in place,” he said. “But to start with hav­ing face masks and then think you can crowd your buses or your shop­ping malls — that’s definitely a mis­take.”

He com­pletely brushed off the prospect of wear­ing masks last month, say­ing, “With num­bers diminish­ing very quickly in Swe­den, we see no point in wear­ing a face mask in Swe­den, not even on pub­lic trans­port.”

Teg­nell has ar­gued that ev­i­dence about the effec­tive­ness of face mask use was “as­ton­ish­ingly weak.”

Swe­den has of­ten been called the con­trol group. There has been great value in see­ing how things played out there with­out a lock­down. It’s still not en­tirely ob­vi­ous to me it was the wrong call.

But this is ridicu­lous.

This is tel­ling us it is dan­ger­ous and wrong to tell any­one to wear a seat belt, be­cause they might think they then can’t be hurt in an ac­ci­dent.

No, that’s not quite right.

This is tel­ling us it is dan­ger­ous and wrong to tell any­one to wear a shirt, be­cause they might not re­al­ize they also have to wear pants, and we wouldn’t want any­one walk­ing around with­out any pants.

No, that’s not quite right ei­ther.

Ah. I got it.

He’s say­ing that we have no ev­i­dence parachutes pre­vent im­pact trauma re­lated to grav­i­ta­tional challenge. We shouldn’t tell peo­ple to wear parachutes when jump­ing out of planes, be­cause we don’t have ev­i­dence that the parachutes help, and tel­ling them to wear a parachute en­courages more peo­ple to jump out of planes.

So con­grat­u­la­tions to An­ders Teg­nell, top health offi­cial of Swe­den, win­ner and new cham­pion of the Biggest Idiot award for Covid-19.


Mitch Daniels wins this week’s award for ac­tu­ally think­ing about the phys­i­cal world and do­ing things that might help, via the pro­ce­dures im­ple­mented at Pur­due Univer­sity. Via the ex­cel­lent PoliMath on Twit­ter, I learned that they are iden­ti­fy­ing the most vuln­er­a­ble stu­dents and giv­ing them room­mates that have already had Covid-19. Then, af­ter Thanks­giv­ing break, they’re go­ing re­mote for the rest of the semester be­cause ev­ery­one go­ing home to their fam­i­lies is go­ing to cre­ate out­breaks, so they need to give peo­ple time to ride out Thanks­giv­ing and Christ­mas.

If we could all think like this, by which I mean ac­tu­ally think, and then act upon it, we could beat this thing.

So bravo, Mitch. Bravo.