One of the notions running around right now is that the COVID virus — they're now calling the virus something like SARS.CoV.2019 now, and COVID-19 is a name for the illness, so the terminology has gotten away from me again — but in any case, one of the notions running around now is that actually there are millions of undetected cases out there already.
Okay, so this is maybe barely possible. R0 for the virus would have to be WAY bigger than estimated. That Gardner paper, using an R0 of 2.2, estimated on 31 January that the number of unidentified cases was around 10 times the actual count. So, if that estimate is good, then as of today there will be something like one million cases. Other people have estimated as much as 20 times, which gives two million.
Aside: If you haven't been following this, R0, often read "R-naught" is a measure of how infectious a disease is. You can think of it as the number of follow-on cases you can expect from an individual case, and R0 = 2.2 means you evect on average 2.2 more cases from every case identified. For comparisons, flu is usually 3-5, chickenpox is around 10, and measles is between 12 and 18.
But that's a pretty small number for a massive global pandemic, with panicky millions of undetected cases, so let's multiply that by 10 — and get 20 million.
R0 is going to need to be much more than the estimated 2.2 — closer to measles, higher than the flu or chickenpox. But this has some interesting consequences that we'll get to in a minute.
In the meantime, let's look at the panic-press recently on the actual disease.
It seems there are three groups pushing panic.
- Some Qanon people, including the ones who are identifying a patent for a vaccine for a different virus, that causes pig diarrhea, as a "coronavirus patent" that proves it's a Chinese bioweapon.
- Some venture capitalists and skeevy finance websites pushing estimates of hundreds of thousands of deaths in months
- And some politically-biased news sites — and yes, CNN and MSNBC, I amlooking at you — with commenters who are practically orgasmic that THIS, FINALLY has the walls closing in on Trump.
Here's the problem — as I've pointed out repeatedly, every one of these asymptomatic or mild cases increases the denominator on the case fatality rate. (For any journalists reading this, that's the number on the bottom of a fraction.) As of 11:13 Eastern time today, there are 116,166 confirmed or presumed cases, and 4088 deaths, giving a case fatality rate of 4088/116166 or 3.5 percent. But make the denominator 10 times that, and it's 0.35 percent. Make it 20x and it's 0.17 percent.
If, in fact, there are 20 million or more, mostly occult undetected cases, then that's 0.02 percent or about two per ten thousand.
Honestly, at this point we should be doing one thing anytime anyone makes a prediction of doom and catastrophe and catastrophic doom, and that's asking them to show their work.
Just saying "it's exponential" doesn't count. First of all, some of these estimates require big exponents, and the actual data don't support that. We could compute an estimate of what the exponent would be (hint: (1+n)^r) but we can use actual data more quickly by simply looking at a semi-log plot of the statistics, which the estimable JHU COVID-19 dashboard conveniently provides. Here's the key insight: if the function is exponential, the slope of the log plot is the exponent. So the closer to a horizontal line the semi-log plot gives, the lower the exponent and the slower the growth —even if it's an exponential function.
I forgot to include a "math warning" there, but the log-and-short of it is that the catastrophe predictions require that semi-log to have a substantial upward slope, and it's just not there.
As well as looking for people to show their work, then, I think there's another question to ask, one that is as important now as it was in 80 BC: cui bono? "Who benefits?"
Frankly, I don't understand the motivations of conspiracy theorists. Maybe you all can explain what Qanon's expected benefit might be in the comments. The Monolithic Legacy Media likes it because it's another chance to dunk on Trump, and while it probably is too late to impeach again, at least there's the chance it will be Trump's Katrina — or Chernobyl.
The financial people? To understand that, you need to remember that traders aren't like most people. They don't profit from the gains their stocks make — they profit from their ability to buy low and sell high. In other words, they profit when the market is really volatile.
So, among those who benefitted from a panic over the weekend are the people who saw a good chance there would be a big drop on Monday. In other words, panic selling, because if you aren't panicking, panic selling is a buy signal, as well as making short sales pay off.
Do I know that's what happened, and those mathematically-improbable stories of hundreds of thousands of deaths in the next few months were intended to hurt Trump and panic markets?
No.
But I still ask cui bono?
No comments:
Post a Comment