It all started a few days ago when I made the mistake of commenting on a Twitter thread. And a Twitter thread about how policy affects death and doing the usual cherry-picking about how some favored policy in State A caused it to have lower death rates than that trogolodyte science-denying State B. Of course, you know the Team colors for A and B. But… let’s not put things in the Memory Hole, policies and recommendations have been all over the place- characterized as “evolving,” but looking more like a random walk than any actual data-driven convergence. Media and government have found a symbiotic relationship: Panic Porn gets clicks, views, and ad revenue, and government can (without even any pretense of constitutionality and respect for fundamental civil liberties) use the resulting hysteria to massively increase its control over our lives and choices, and completely distract from mundane things like two-decade-long wars. “Science” is commonly invoked.

Now a disclaimer: I am a working research scientist with a rather broad set of experiences and a LOT of experience dealing with environmental health research funded by the NIH and NSF, but I am fundamentally a physical chemist. So I’m not the guy to talk to about protein structure of virus coats and mechanisms of immunological pathways at other than an educated layman level. But handling and interpreting data, that’s something I do. Ditto, understanding the dynamics of research funding, well enough to have euchred many millions of your tax dollars for research programs. So I know how that game is played, and am numerate enough to recognize bullshit and spin, and the thing that set me off was a doozy.

Here’s a chart that was trotted out and used extensively by various howler monkeys:

Statistic: Death rates from coronavirus (COVID-19) in the United States as of November 18, 2020, by state (per 100,000 people) | Statista
Taken from Statista

Man, it’s easy to see the effect of the fantastically enlightened policies in the states near the bottom! Oh wait, their policies are RANDOM and changing weekly as well. This observation did not seem to calm the howler monkeys in the least- one even claimed a major difference between two states in the middle that were only 2 deaths/100k apart, assuming that no one could look at the chart and catch him in the lie.

Looking at the chart (and assuming the data are more or less correct), something struck me. So let’s try a really goofy, nutty, out there hypothesis: that other than really brutally stupid policies like New York government’s deliberate placement of COVID patients into nursing homes, the various policies and decrees have essentially NO influence on outcomes. They neither help nor hurt (as far as COVID). Now of course, there’s demographic differences between states (ages, ethnicities…), and some correlation doubtless can be found there, but just for shits and giggles, let’s assume spherical cows and hypothesize that things like closures, curfews, talismasks, and the like have no influence, whether mandated or not, and that the demographic differences are second order.

If this hypothesis were true, what would we expect to see? I think that if we sorted the data into a histogram with bins representing death rates and bar height representing the number of states with death rates in that bin, we’d see something like a normal distribution; it won’t be a perfect Gaussian because you can’t have negative death rates since Lazarus, so it will have a Poisson distribution character.

Let’s do it. Statista has a great feature allowing the download of their data into Excel. This is delightful for geeks like us who can slice it, dice it, or visualize it.


Well, whaddaya know? It DOES seem to distribute normally (I didn’t get a chance to fit it to a Poisson, but eyeballing, I’ll opine that λ = 5 or so will follow it tightly). There’s a little tail thickness from NY and NJ, but otherwise, this is pretty fucking close to exactly what we predicted: political policy has almost no effect. It’s almost as if viruses spread and mutate without really caring much whether bars are open or closed. Like, I dunno, the common cold?

Now clearly there could conceivably be effective policies focused on actual vulnerable populations instead of kids and healthy young adults. Conceivably. But no politician is going to conceive of that- welcome to your new chains. And as politicians are guided by “science” (meaning political appointees and government bureaucrats), the chains will only get heavier. Count on the media to not present the data in this format, where the futility of relying on politicians is far too evident.

So, let’s say it again: public policy of the sorts we’ve been pursuing (universal restriction of basic civil liberties and compulsory obeisance to symbolism of dubious efficacy) has almost no effect, positive or negative, on a virus that is a powerful replicating machine presenting relatively low risks to the majority of the population. It’s random.