An update on a pre-registered result about the coronavirus

https://www.kalzumeus.com/2020/04/21/japan-coronavirus/

by balfirevic • 6 years ago

214 144 6 years ago

We project a true count of over 500,000 infections, including more than 5,000 severe cases, and a breakdown in provision of care (“overshoot”) in Nagoya, Osaka, and Tokyo, before the end of April.

The current count is ~21,000 with the number of confirmed daily cases declining for 6 days straight, from 741 to 338 yesterday. So... not only wrong on the numbers themselves, but also on the scale, trajectory and impact on society.

Japan's response seems commensurate with the problem, and this reads to me like an excessively verbose illustration of how one can spin even pre-registered falsifiable claims whichever way you want regardless of outcome. 'See? We were mostly right'.

Maybe that's too harsh.

akavi3• 6 years ago

Yeah, I'm withholding judgement till "the end of April", but given Japan's doubling time is roughly 10 days, you'd expect there to be ~2.5k severe cases right now. Given that Japan's official count is ~11k, is there reason to believe 1/4 of all official cases are "severe"?

Does anyone have up-to-date data on "excess deaths" in Japan for the past few months?

ceejayoz4• 6 years ago

Given that Japan's official count is ~11k, is there reason to believe 1/4 of all official cases are "severe"?

Yes. That fits data out thus far.

https://www.cdc.gov/mmwr/volumes/69/wr/mm6912e2.htm

Although the majority of reported COVID-19 cases in China were mild (81%)

Combine the 19% severe cases number out of China with Japan's aging population (which COVID-19 hits substantially harder), and 1/4 is entirely reasonable.

Loughla• 6 years ago

Maybe I read that wrong, but wasn't the point that Japan wasn't acting appropriately, and that white paper, along with other research, forced (for lack of a better word) them to own up to the scale of what they were seeing?

Isn't the point that, while the particulars were proved false, the overall thrust of the article that Japan was stalling action which MAY lead to those numbers was true?

Did I miss the point of this entirely?

kspacewalk24• 6 years ago

Honestly, who has 1-1.5h to read this without skimming through most of the verbiage to tease out the point? I read the "pre-registered" part. It was wrong on a bunch of points, both numbers and impact. I strongly doubt a major world economy of 126 million people with many world-class epidemiologists and doctors and generally non-disastrous political leadership was 'forced' into anything by this.

The hubris of this guy is quite off-putting.

SpicyLemonZest6• 6 years ago

Everyone with disastrous predictions about the coronavirus can and does make that claim. If my predictions aren't true, give it a few weeks and exponential growth will make them true. If they still aren't true, then they would have been true, but whatever we did a few weeks ago must have been the correct reaction to my predictions.

The very general thrust - that Japan surely had community spread - I agree was true and not widely acknowledged at the time. But I don't think that exempts the more specific projections from analysis.

ekianjo• 6 years ago

You forgot that test capacity in Japan is still close to nil.

panarky2• 6 years ago

One might argue that a "true count of infections" is not falsifiable -- people aren't getting tested, causes of deaths aren't properly attributed, officials are juking the stats.

A more falsifiable claim is "a breakdown of provision of care in Nagoya, Osaka, and Tokyo, before the end of April".

If patients are in parking lots or bodies are in refrigerated trucks and mass graves, that evidence is harder to hide.

There are 10 days left in April, but at this point, is there any evidence of "a breakdown in the provision of care"?

barry-cotter5• 6 years ago

Refusing to test, secret testing locations, over two hours on hold on the COVID hotline. Things aren’t looking great.

https://www.facebook.com/groups/524159784940511/permalink/52...

4/17 Friday: At noon my friend wakes up and asks if I can call her ward's 保健所 and make a PCR test appointment for her. I call them at 12:30.
It took 2 HOURS to set up a PCR test.
As I expected, the staff was INCREDIBLY resistant to offer a test, and spent 2 hours asking me 100 questions about her symptoms, what treatment she had received so far, what hospital she went to the night before, name of the doctor who recommended her a test etc. I'm put on hold for an extra hour as they sort their things out. They finally call me back and tell me all of the information for her test. But here's the kicker. It's a secret location. It's a medical facility that's currently closed, but is being used as a corona testing site on the downlow. To enter the building, she'll have to walk through the parking lot and use the staff entrance. The woman on the phone makes me promise not to tell ANYONE other than my friend the name of the place where she is being tested. Because if people knew they were doing testing there it would "cause a commotion." I really wish I was making that last part up. It's absolutely horrific how much they are trying to cover this up and keep the official numbers low. I am currently waiting for them to call again with her test results. Getting her tested required 1 ambulance ride, two visits to the hospital and 5+ hours on the phone. This is ridiculous and something needs to change."

XorNot6• 6 years ago

Exponential growth means fully half of the present population occurred in the last growth cycle (presuming R0 ~ 2).

With a mean time to symptoms of 14 days, depending on what Japan is doing, 10 days is an age (though it's worth noting that severe respiratory symptoms set in in week 3).

ForHackernews2• 6 years ago

Yeah, I don't understand at all what the purpose of publishing this hash was supposed to be. Is it just to be able to say "I told you so!" (or in this case, I didn't tell you so, but I really did think this) at some later date?

In a world where I was wrong, people who trust me professionally would likely think less of me. That was an acceptable risk.

Yeah, what a hero taking an extremely slight risk to his reputation in the middle of a pandemic.

Sorry, I normally generally like things patio11 writes on here (he's the bingo card guy, right?) but this whole thing reads like a lengthy ego-stroke.

nico_h4• 6 years ago

The point of the hash is a marker to force yourself to come to terms with your idea. You make it public and then act on it.

The author has a visa to stay in Japan, a country with strict rules on the disruption of public order. Spreading unfounded speculation on a pandemic might rule afoul of such rules.

HackOfAllTrades2• 6 years ago

"I don't understand at all what the purpose of publishing this hash.."

The author was explicit: "sending up a flare", hoping "someone would read between the lines".

whycombagator• 6 years ago

The current count is ~21,000 with the number of confirmed daily cases declining for 6 days straight

Yes, but all of those 6 days are still higher than 7 & 8 days ago. In fact, all the days you picked are also higher than they were through March 15st-April 4th too - and several days between April 4th onwards as well.

Today's new cases in Japan are many multiples higher than most days in March - but I do hope they continue their recent trend of decline.

So... not only wrong on the numbers themselves, but also on the scale, trajectory and impact on society.

Maybe, but they did use the term "true count" which appears to mean: asymptomatic and those who have it but haven't/won't get tested.

As the US has learned, once you start to ramp up testing the case count ramps up too.

Looking at Japan's testing, they are testing more per day now than in March[0] but their total number of people tested is still not very high at almost 117k[1] - given they have ~125m population.

[0] https://ourworldindata.org/grapher/full-list-covid-19-tests-...

[1] https://ourworldindata.org/grapher/full-list-total-tests-for...

barry-cotter5• 6 years ago

Looks like the breakdown in provision of care will happen before the end of April after all.

https://twitter.com/patio11/status/1252836409406152709?s=21

NHK confirms that coronavirus-appropriate hospital beds are 80%+ filled in Tokyo, Osaka, Ishikawa, Hyogo, Shiga, and Okinawa.
Experts threshold of concern for imminent impact on medical care is 30%, again per NHK.
The 28 prefectures which are above that threshold will not fit in this tweet.
NHK reports that several prefectures which were unable to sustain the prior strategy of hospitalizing all diagnosed cases have begun moving less severely affected patients to hotels/etc or asking them to self-sequester at home, temporarily freeing up bed space.
!NHK reports that while reprioritization of hospitalization strategy and urgent measures to add more beds may buy some time, the situation risks a collapse in the provision of medical care, and that experts explain we cannot be optimistic.

waffle_ss• 6 years ago

It's not too harsh. It seems pretty clear to me the blasé conclusion that essentially "things are gonna get much worse in a month" (no duh) was loosely backfilled with numbers after the conclusion was already made.

Writing in a circuitous manner may have been necessary due to the author's immigration status but it's also a very convenient smokescreen that makes his group's results look more dramatic.

For example, the timeline for March 24th, the day before this epic whitepaper was released: "a sporting event is postponed." This "sporting event" was the Tokyo OLYMPICS! But showing that Japan was already starting to take COVID-19 a lot more seriously than some of their public pronouncements may lead you to believe before the whitepaper release date might undercut the article's implication that his whitepaper changed the public narrative and thus the author would be less deserving of asspats and clout.

I've heard people say that this pandemic highlights the (unfortunate) severe lag time of real-life discourse to Twitter's discourse (what is "known" on Twitter versus public narrative).[^1] From that angle this report could be a good thing to try to short-circuit that gap. My opinion is the attempt failed (more accurately their conclusion was already stale), and this writeup brings it right back around to the worst parts of Twitter's discourse - personal brand marketing and clout-chasing.

[^1]: While that's undoubtedly true, Twitter also generates a massive amount of garbage takes which are conveniently forgotten through severe selection bias (looking back with hindsight to see who was right is easy). Online takes also have zero cost compared to reality; you can throw infinite virtual darts at a dartboard but in reality you can steer your meat machine down only one path. Thus there will always be a lag between the two.

simonebrunozzi4• 6 years ago

Oh my. The Japanese government should feel ashamed to have hidden all of this from the public (edit: I mean the real data and the real gravity of the situation, of course). I, for one, fled the US to come to Japan for a few weeks because I mostly believed these numbers.

I hate this. I really hate that politicians (all over the world, in this specific case in Japan) can afford to treat us like this, and usually not face any significant consequence.

kspacewalk24• 6 years ago

I, for one, fled the US to come to Japan for a few weeks because I mostly believed these numbers.

SARS-CoV-2 thanks you for your efforts to help it find new populations to replicate in.

SketchySeaBeast• 6 years ago

Yeah, considering that you can be asymptomatic for quite a few days that was an extraordinarily selfish act. Fear made them a vector. Feels like the guy in a zombie movie who tries to break the door down to get into the safe room.

simonebrunozzi4• 6 years ago

The prior 3 weeks I was in a remote area with essentially no contact with anyone. I came back to SF, stayed at home for one night, and left.

Also, if I feel endangered, I have every right to decide where to go. You might see it as selfish, I see why. At least I am paying my useless and expensive health insurance in the US (out of my own pocket, because you know what? I'm unemployed right now) to support the system, even if I am not taking advantage of it.

Sometimes things are a bit more complicated than "SARS-CoV-2 thanks you for your efforts...".

smcl2• 6 years ago

Please tell me this is a joke comment

9nGQluzmnq3M• 6 years ago

I usually like patio11's writing style, but this is super verbose even by his standards and oddly evasive/defensive to boot. There's a number of important messages here, but they make up perhaps 10% of the content, and I still don't really fully understand the rationale behind the cloak and dagger "hash of report" business.

simonebrunozzi4• 6 years ago

Perhaps that's on purpose. Perhaps he wants to avoid being sued, or similar things. Patrick has a ton of credibility and I would give him the benefit of the doubt here.

joebot123• 6 years ago

The purpose of the overall evasiveness could be to prevent personal retaliation of some kind - that would be understandable. But why avoid mentioning the word Olympics and instead call it "the sporting event" repeatedly? Especially now when your connection to the white paper is public.

OJFord• 6 years ago

This is a very different essay than my typical work. It may be judged to very different standards in very different quarters than usual. Please excuse my need to write more reservedly and participate less in subsequent commentary than I usually would. Some truths have social consequences, acknowledged or not.

dahdum• 6 years ago

I thought the same, like he's walking on eggshells? Maybe it's his assocation with Stripe or due to someone else in the working group? The memo doesn't read as particularly controversial nor did it recommend much concrete action.

I left Japan in early March, by personal anecdote masks seemed pretty ubiquitious (and all public facing employees were wearing them), hotels were screening guest temperatures, warning signs everywhere, etc.. Time will tell, but I still think I'd have been better of staying there than returning to the US.

nico_h4• 6 years ago

The memo doesn't read as particularly controversial nor did it recommend much concrete action.

Well it was pointing out that the current plan was B.S. and had already failed, and that the testing plan in place resisted falsification by preventing testing of unconnected / asymptomatic people. no biggie.

The point of the memo was to gather and surface disparate information already public separately un-worrisome but that together might be able to be paid enough attention to for a (positive) governmental reaction.

It was to force agencies to connect dots that the authors believe were not being connected. (Maybe it just wasn’t anybody’s job to connect them)

pnathan• 6 years ago

That is his m.o. when critiquing Japan.

If you read between the lines to what he is really saying, I would think you would revise your opinion substantially.

philwelch3• 6 years ago

What appears “evasive/defensive” to one culture may appear merely polite to another culture. As always I would presume good faith here.

davidw2• 6 years ago

Agreed - but it did leave me a bit ... curious, if nothing else.

SkyMarshal4• 6 years ago

I still don't really fully understand the rationale behind the cloak and dagger "hash of report" business.

I didn’t read every word of it but got the sense there were political and/or Japanese cultural sensitivities why he didn’t initially feel comfortable with a full public, attributed release.

SpicyLemonZest6• 6 years ago

I think you're being a bit too accommodating. McKenzie does have a flair for the dramatic; he did the same thing with Tether where there really was no reason he couldn't just say what he thought when he thought it.

At the same time, yeah, it seems very plausible that a white guy in Japan announcing the Japanese pandemic response is run by fools would face blowback for it.

glofish• 6 years ago

I found the post to be lengthy, lacking focus and specificity. More compelling case could be made in a fraction of its length. Hardly worth reading.

The tedious timeline, tweeting hashes as if that meant anything other than: "do you see how I was right all along".

The whole concept of tweeting a hash of a whitepaper as form of endorsing it is kind of absurd.

akavi3• 6 years ago

I don't think the set of people whose opinions patio11 cares about would forget about him tweeting that hash (especially given the emphasis he placed on it).

The hash concept depends on social norms, and it's a fairly strongly enforced in his corner of the internet.

simonebrunozzi4• 6 years ago

Over a roughly 2 week period, we would expect approximately 64,000 passengers to fly from Japan to Singapore. 3 infections in that population is a rate of approximately 47 basis points, which is 5X the 9 basis points rate of infection in Japan. If one believes the government, the rate with surveilled clusters backed out be a tiny fraction of 9 basis points.

This argument is a bit flawed. People that travel from Japan to Singapore are not a perfect sample of Japan's population - they might be more likely to be infected, as they might travel more than the average Japanese person.

lolc• 6 years ago

Three is way too few samples to draw any statistics from anyway. It does suggest untracked transmission in Japan. And that's about it.

When I read that part I was disappointed to have been lured into reading another iteration of "I multiply numbers without tracking associated uncertainty". Sigh. That stuff runs rampant at the moment.

scythe• 6 years ago

Must the conclusion be all one way or the other?

Maybe cluster containment was a bad strategy, but not horrible? Maybe masks and voluntary distancing had a little effect, but not enough? Maybe cases were undercounted, but not to the tune of 90-99%?

grosales• 6 years ago

This is just my 2 cents but according to https://www.worldometers.info/coronavirus/#countries Japan has only done 923 tests for 1M of their population. That's much lower than any developed nation. One thing the article claims is the spread of the virus by asymptomatic infected people. Given this low level of testing - it's hard to gauge the true level of infection in the country - although the continuous decrease on daily cases is good, it doesn't mean it cannot spike back up. We just have to wait and see what happens.

SyneRyder• 6 years ago

To give an indication of how low that testing is, on that site Japan is ranked #119th country in the world in terms of testing.

m_mueller• 6 years ago

What I find confusing though is Japan's low death numbers. I was assuming that spikes in mortality couldn't be hidden, but recent comments from inside sources in Japanese hospitals are starting to make me doubt this assumption. Still, there is not a crisis on Lombardia-level going on, which one would expect given the relative inaction of Japanese government. I'm guessing the truth is somewhere in the middle - Japan's reported numbers are underestimating the issue, but at the same time Japanese society is as a whole more resistant than at least Southern Europe.

EE84M3i• 6 years ago

One theory is that people are dying but it hasn't been discovered yet because of people staying inside. see https://en.wikipedia.org/wiki/Kodokushi

Having just ridden the Yamanote today though, it's unreal how empty it is.

will4274• 6 years ago

I mean, look at the comorbidity data - heart disease, hypertension, diabetes - aren't the Japanese one of the healthiest people in the world?

btilly7• 6 years ago

Japan is a country where at the best of times, a lot of people die at home and are not discovered for some time. If you remove social contacts thanks to lockdown, this is going to become more common.

See https://en.wikipedia.org/wiki/Kodokushi for more.

twic2• 6 years ago

We often conflate asymptomatic and presymptomatic people: my understanding is that asymptomatic people aren't very infectious, but presymptomatic people are. I can't find a solid citation for this, though.

The mechanistic basis for that is simply viral load: asymptomatic people have some virus, but not enough to ever make them really ill, whereas presymptomatic people have loads of virus, but it hasn't made them really ill yet.

garmaine2• 6 years ago

With this disease it appears asymptomatic people are equally infectious.

timerol2• 6 years ago

You are correct that official coronavirus deaths are low. But official influenza deaths (blue diamonds) were almost double their baseline value (green line) and above a threshold value (purple line, but I don't know enough Japanese to know what threshold) back in week 8 of this year (late Feb, early March, depending on how you count it.)

https://www.niid.go.jp/niid/ja/flu-m/2112-idsc/jinsoku/1852-...

I wonder if there are other, non-coronavirus causes of death that are significantly above their normal values?

sneak2• 6 years ago

It would be exceptionally difficult—if not outright impossible—for me to write something like this and not once make even a passing attempt to imply blame or culpability for a core dishonesty that may well end up causing thousands of deaths.

Patrick, I admire your professionalism and restraint in service of your goals.

sdffdsfdsfsddsf• 6 years ago

Assuming the ratio of cases with symptoms to cases without symptoms usually remains the same, I must admit I don't immediately understand why it would be important to determine the amount of asymptomatic cases to determine the risk. It seems to me the growth rate of (symptomatic) cases would be the major indicator for that. (Obviously it would be of general interest, but I mean specifically with respect to gauging the risk).

It all seems to hinge on the assumption that with the implemented "strategy A", nobody with pneumonia would be tested for Covid19 unless they have a known connection to a symptomatic case. That of course would be a mistake. But were people really that complacent that it would never occur to them to test for Covid19, even as the rest of the world is talking about nothing else? (Availability of tests is of course a factor here, but should only be a temporary impediment).

I am interested in it, because here in Germany there also is currently a discussion of the virus spreading silently, only to come back in full force in fall if we become complacent. I must admit I don't fully understand the logic behind that, as surely when it spreads silently, it also has a proportion of symptomatic cases in its wake, so that in the end, it doesn't spread silently after all.

vikramkr• 6 years ago

The asymptomatic cases can also spread the disease, and a single individual is enough to trigger another outbreak if they slip through the cracks, so for contact tracing and everything you need to know how many asymptomatic cases there might be and where they might be. And the wake might not be spread evenly throughout the population, you could end up with younger individuals being asymptomatic and creating a breeding ground for the virus to spread and mutate in a college town before, even though it seems to be under control, it spreads back to vulnerable populations where everyone is symptomatic. That's one extreme of this but we don't know without random testing how bad it is towards that end

PeterisP• 6 years ago

It's not important to determine the amount of asymptomatic cases, but it's important to detect asymtomatic individuals so that their contacts can be traced and tested as well.

This was not done, and "strategy A" failed because of that.

seemslegit2• 6 years ago

There is no way Japan's faux-success is due solely or even dominantly to asymptomatic cases, they had to be turning a blind eye to symptomatic ones as well and attributing them to pneumonia, flu etc. The fact that the reported rates began rising only after the Tokyo Olympics were announced cancelled is almost a USSR-era farce.

Japan has a leading hospital-bed-per-capita capacity in the OECD which could help them conceal this but I believe if occupancy of those beds was graphed for the last two months we'd see a curve indicating an epidemic comparable with US and Europe.

twblalock• 6 years ago

Then where are the overwhelmed hospitals? Where are the bodies?

seemslegit2• 6 years ago

They're not overwhelmed yet because they have capacity (NYC hospitals are struggling but are not over capacity either), deaths happen all the time especially in a country with aging population like Japan and are just not counted as covid19, like the author says it will probably be visible in excess deaths statistics once those are published.

jeffreyrogers2• 6 years ago

This seems a bit over-the-top. Sure, maybe Japan isn't testing and is just treating pneumonia cases as pneumonia cases instead of the cornavirus cases they actually are, but if that is allowing society to function mostly normally and isn't changing death rates all that much I don't see the harm. Arguably that is a better response than shutting down your economy and creating massive debt problems that will need to be resolved later.

XorNot6• 6 years ago

The death rate from COVID-19 is a constant: about ~1% of all cases, so far as is currently known.

If the R0 of the virus is about 2 (which from evidence is about the case for uncontained spread, possibly higher) then half of your entire case load is discovered in the last increment.

That period is (roughly) 14 days from infection to the onset of symptoms. We cannot build a hospital, nor train doctors and nurses, in 14 days.

But here's the thing: the hospitalization rate - which is what keeps the death rate at 1% - is around 10 - 15%. And both those numbers are constants, multiplied against an exponential factor. And hospital stays are around 3 to 4 weeks long.

By the time your hospital beds are 50% full, you're already out of capacity - the next wave is going to be being denied treatment. And in fact, anyone else who has an otherwise treatable but potentially life-threatening condition is also going to be denied treatment. At which point a lot of your deaths aren't going to be directly due to COVID-19...but that's really not going to matter.

jeffreyrogers2• 6 years ago

Okay, but this appears to not be happening in Japan. Why not? If your answer is, it's going to happen, then why hasn't it happened already if things are truly as dire as the original essay is saying?

Edit: It also appears to not be happening in Sweden. I think a lot of the modeling of this is flawed.

porker• 6 years ago

Near the top patio11 writes: "Our choice to be anonymous was a considered one and is discussed in more detail below.".

I cannot find where it's discussed in more detail below. Can someone point me to it?

weinzierl• 6 years ago

I love Patrick's posts but this is long, even for his standards. Can someone summarize the main points?

HackOfAllTrades2• 6 years ago

No. It's well worth your time to read the whole piece.

rmccue2• 6 years ago

The clearest communication about the use of hashes rather than direct publication is laid out quite simply in patio11's original memo:

I have concerns about saying this loudly because I have an immigration status and because Japan has previously shot the messengers in times of national emergency, including in the wake of the 2011 disaster. As someone who has spent their entire adult life in Japan, who lives here with my wife and children, and who has family on the front lines, I pray daily that I am wrong.

sneak2• 6 years ago

At which URL did you find that text? I can't locate it on his site.

I can confirm, though, as an expat, that there is a special kind of fear/restraint involved in needing an opt-in, discretionary permission from a government to be permitted to physically return to your home (which you already paid for) and your family. It causes a lot of security-conscious people to err far on the side of caution, which, indeed, may not even be an error at all.

Immigration policies generally affect a very small minority of people in a country, so strengthening the rights around them simply isn't a priority in most societies, and risks/chilling effects like this fall out as a result.

rmccue2• 6 years ago

http://media.kalzumeus.com/covid-19/japan-covid19-memo-20200... - linked from “I did not publish the memo at the time.” just before the “ Why did you publish the hash and not the memo?” title.

BerislavLopac2• 6 years ago

There are few conspiracies in the world. There are many systems with complicated decision-making processes, internal data flows, and incentive structures for actors within them.

The second sentence here is absolutely correct, but it directly contradicts the first one. Because a "conspiracy" is nothing more than having a group of those actors, with aligned incentives, coordinating to work together while hiding that coordination with other actors whose incentives are in opposition.

robocat• 6 years ago

I disagree. The second sentence is about uncoordinated actors making decisions, without anyone conspiring together. The emergent behaviour may look like a conspiracy, which is what he is pointing out.

BerislavLopac2• 6 years ago

What is the difference between "looking like a conspiracy" and "being a conspiracy"?

ohazi• 6 years ago

Most of this seemed blindingly obvious to me and my peers (in the US) at the time that Patrick tweeted his hashes. Several of the replies to his tweets correctly guessed the topic, and most came to similar conclusions. Anybody with half a brain knew what was going on.

Maybe Japanese culture really is "different" and things really did need to be done this way, but I find myself agreeing with other comments here that this whole effort seems puzzlingly labrynthine with relatively little to show for it, other than providing an opportunity to tweet hashes.

davidw2• 6 years ago

Yeah, I think it didn't take much to put two and two together, even without much familiarity with Japan:

https://twitter.com/davidnwelton/status/1241556680833298434

joe_the_user• 6 years ago

Yeah, making a secret prediction seems only justified when your position is in the extreme minority and when it contains fairly surprising and/or specific claims. When your claim is just supporting a "strong minority viewpoint", doing so secretly ridiculous.

wglb• 6 years ago

I think the article points out the justification as a counter point to the likely official statements everywhere that "Nobody could have known".

mherdeg• 6 years ago

Cool. Did you trade on it?

AnonC• 6 years ago

I felt, while my subjective confidence was 90%, that I was likely miscalibrated. The likelihood that one non-expert, doing casual sleuthing in his spare time, had scooped not just any expert but almost all the experts and almost all the parties in formal authority felt infinitesimally small.

So...was it just a coincidence that the results seem vindicated? Or is the above statement just too much modesty? Or both?

AgentME• 6 years ago

There's a whole section at the end with his theories about why you could expect someone like him to get this better than the experts (the experts were too busy managing the intricacies of Plan A instead of constantly rigorously checking that its assumptions still held and whether it was time to give up and switch to Plan B, etc), and he makes a point against having too much modesty like he did if you have good data.

gregmac• 6 years ago

A long time ago I did a bit of work on disaster alert systems. They’re not dissimilar to fire alarms. The engineered purpose of a fire alarm is not merely to let people know there is a fire. Many will have already perceived the fire. The alarm, buttressing training delivered far before the alarm rings, gives you unquestionable and immediate permission to evacuate. We know that otherwise some people, smelling smoke and feeling uneasy, would look around the room, see other people not moving, and conclude “Who am I, to disrupt everything going on by shouting ‘Fire! Fire! Fire!’?” The history of humanity has seen far too many rooms where no one shouted ‘Fire!’ early enough.

This alone is a pretty interesting point. Can apply to many other systems that are much less significant than literally "life or death."

throwawayiouyt• 6 years ago

This novella was worth reading just for this gem. It is equally applicable in "boiling frog" scenarios as in the "fire" cases.

timerol2• 6 years ago

I really appreciate this post. It's hard to read, but this seems very intentional. There are a number of salient points that are very specifically not stated. Some questions deliberately not answered:

Who was on the Working Group, other than Patrick? Dr. Tyler Cowen spread the results, but no one is mentioned as working in the group other than Patrick. This seems designed to limit potential social fallout to other parties. The whole "Someone read between the lines" section would be much easier to read with the relevant person named. Also consider that Patrick may be limiting fallout to himself here, with the bold statement early on that "I instigated the Working Group and I was the primary author of its white paper."

Did the NYT article writer correspond with the Working Group? The white paper was published with plenty of time before the article was written, and the explicit goal of the Working Group was to "leverage a news organization." Also consider that the NYT article was published on Thursday. And when discussing the goals of the working group, "This implied a publication deadline of Thursday. We assumed media organizations would need our work by Wednesday to check it and produce reporting informed by it." Note that Dr. Cowen's quick post was also published on Thursday, possibly to increase deniability when referring to that goal date.

What Japanese organizations failed to respond to this crisis? The most damning prose in the entire article is "May all judgments be just and merciful." Patrick does not identify any Japanese organization or government by name, but says that the Working Group spoke to many of them. Also, both memos are listed in the timeline immediately before a announcement by Governor Koike. Patrick even convolutes his prose to the point where he never mentions the Olympics. I wouldn't be surprised if he had another sporting event in mind that also matched the specifics of what he wrote.

How much influence did the Working Group have in Japan's response? The NYT article is definitely one potential point of influence, but it seems that the memo itself was shared widely. Consider the difference between this humble ending to the article: "I am a responsible professional. I have no relevant expertise or authority. I have avoided unproductive criticism. I might have made some guesses, during a year when many people were guessing on many topics. I quietly told some people about them. Some guesses may, perhaps by happenstance, align with official guidance and credible published reporting." and his tweet confirming the memo (option 1): "I am materially wrong about the most consequential thing I've had to have a view on in 15 years. You should probably degrade your estimate of my ability to think through complex problems."

Interpreting things through an aggressive lens, I read this as a scathing diatribe against the Japanese government, where a non-expert foreigner living in Japan has to get the Western media involved before the government would admit to there being a problem. Props to the governor of Tokyo for being more on top of this than anyone else (perhaps with the Working Group's involvement), but the national government completely dragged its feet until it was forced by international pressure to start asking real questions.

But I'm probably reading too much into it. Patrick's too humble of a guy to say all of that.

philwelch3• 6 years ago

Who was on the Working Group, other than Patrick?

An interesting question—but if they wanted to be publicly known, they would be. I’d rather leave that to them.

Patrick even convolutes his prose to the point where he never mentions the Olympics. I wouldn't be surprised if he had another sporting event in mind that also matched the specifics of what he wrote.

The NPB season was originally scheduled to start March 20 but has been postponed, with the original postponement decision announced March 9. The Olympics were postponed on March 24; the hash dates from March 25th.

glandium• 6 years ago

Random comments in this frame of context:

- I've seen/heard some doctors saying publicly, on Japanese TV, mid-March, "I think there are 10 times more cases than reported"

- There's an average of 270 deaths per day from pneumonia on a normal year in Japan (100k).

- The largest number of deaths announced from COVID-19 in a single day was, I think, yesterday, and it was 20.

- Deaths from pneumonia that aren't tested could very well appear to be in the noise of normal pneumonia deaths, easily hiding the real death toll. It doesn't take malicious intent for this to happen, just people following the rules (see below).

- Generally speaking, Japanese institutions have a very difficult time moving outside the box. If they have a written plan that is to be followed, that is what happens.

- The plan was to test people who had 4 consecutive days of fever (2 for older people) and close contact with people already diagnosed, and that's what happened.

- It took prefecture governors starting to do things on their own for the central government to start moving. That happened for school closures (Hokkaido leading the way, then Osaka, IIRC), and sending mild cases to hotels or making them stay at home, rather than occupy beds in hospitals (started in Osaka, IIRC)

- Hospitals have a shortage of equipment, and beds. Some prefectures have ramped up the number of available beds for infectious diseases. The government finally changed the rules to allow mild cases to be isolated out of hospitals (I think about a week ago, but it's hard to keep track of time these days).

- Some hospitals have closed their doors completely. The reason? Clusters of COVID-19 in staff and patients. Not because they were treating COVID-19 patients as COVID-19 patients. But because they were treating patients that turned out to have COVID-19. Or people with COVID-19 going there for consultations. Which is not all that surprising when you know people are reportedly going from hospital to hospital to try to be tested for COVID-19. If anything, I'm surprised there aren't more clusters in hospitals.

- There have been reports of ambulances having to try a large number of hospitals before finding an ER able to attend to their non-COVID-19 patient.

I wrote this on April 4: "the political response to covid-19 in Japan is pissing me off now. It now feels like watching a 100 km/h car running into a wall in slow motion, and the car is already in contact with the wall and absorbing kinetic energy."

Almost 3 weeks later, I think we're seeing the engine entering the passenger cabin.

An update on a pre-registered result about the coronavirus

Related Stories