Flatten the Curve of Armchair Epidemiology

That dismissal seems a bit lazy and flippant too. Plus attribution of motive. Like it or not, there is a grain of truth to the idea that techies (young and old) are particularly prone to this kind of thing. We're rather notorious for it. It's a natural and predictable consequence of having often been the smartest person in the room, able to leap ahead of others who have studied a subject (only slightly) longer. Equating income with intelligence doesn't help either. It takes a while to learn that "opine first, study later" doesn't work so well when dealing with fields equal or superior to our own in technical depth and rigor requiring years of effort to gain expertise, or that it's actively dangerous to indulge such impulses when the stakes are far higher than all but a handful of computer bugs.

> Notice that it is the highlighted part they want you to tweet out.

You mean the highlight that medium applies automatically based on user behavior and the author has no part in controlling?

They implicitly are, as they're arguing that these men aren't experts. Judging this would require one of two things:

1. The author can tell what is expertise based on the words spoken, the quality of the material produced. But then they'd need to be an expert themselves.

2. The authors defines expertise purely in terms of credentialism. Someone works at university X and not university Y, therefore they are experts and anyone else isn't.

As they mock people who engage in "armchair epidemiology" it's not (1), so it must be (2). But it's that absence of critical thinking that deserves to be flattened. It's an attitude of "don't think just obey" - the exact thing that a university education supposedly trains out of people. Supposedly. It's how Google ends up deleting videos of doctors because they aren't "experts" in health or disease, whilst leaving the pronouncements of people who can write R up.

But you're overlooking the most likely outcome - that nobody is qualified in this field. However software developers can admit this and epidemiologists can't.

Bearing in mind for a moment that epidemiologists are amateur programmers, their work output is literally papers+code, they have a major handicap that they can't stop and say "maybe this doesn't work". What else would they do?

I've found the comments by Caswell Bligh very insightful on this topic. Bligh is one of the awful "men in tech" programmers the article laments, who has been writing his own models based on epidemiological papers. Perhaps even from an armchair. Everything he's concluded is worth reading but I'll quote some extracts:

https://lockdownsceptics.org/2020/04/26/latest-news-12/#comm...

Do ‘R0’ or the other ‘R’ values even mean anything when ‘infection’ versus ‘exposure’ can’t even be defined? If social distancing reduces the level of exposures but not the number, does this mean the ‘R’ number is the same, or reducing? Maybe the lockdown means that the population is being quietly, in the background, inoculated to a certain level of resistance (which is what I think is probably happening) but that it doesn’t show up in antibody tests.

Many of these things will never be known. They are probably not even predictable with viruses that are not ‘novel’. Virologists and immunologists may know some of it, and they also know what they don’t know. The modellers don’t know any of it.

As such, models will never be able to give you absolute values for the number of infections, deaths and so on. We may as well not bother with the units on the axes and simply use the model to illustrate the different shapes that are possible

In this interview there's a useful quote:

https://fivethirtyeight.com/videos/how-one-modeler-is-trying...

At about 8.00: “The nature of SEIR models is they do tend to always show this huge increase up to the point of saturation where most people get infected and very often that has not historically happened”.

https://lockdownsceptics.org/how-reliable-is-imperial-colleg...

@Hugh Osmond. What you’ve articulated – and it’s something that’s been bothering me – is that ‘R0’ and the other ‘R’ variants are not a real characteristic of the virus. Snapshots during the epidemic may give you observed values for them, but that doesn’t mean that they represent something fundamental. Attempting to both derive, and then use values for ‘R’ in models, has led to this disaster. It needs to be shouted from the rooftops.

‘R0’ is not a characteristic of the virus. And saying “‘R0’ depends on the virus but also age/population density/pollution etc. etc.” is just another way of saying the same thing. If you know ‘R0’ in advance, you’ve already anticipated what the simple model is going to tell you – there’s almost no point running the model.

In the more sophisticated, more realistic model we shouldn’t even try to accommodate an ‘R0’ input at all, and nor should we try to modify an ‘R’ as time goes on. That’s the wrong way round.

Really, I think you’ve articulated something extremely important here. It’s been driving me mad for weeks: the illness is being defined in terms of a standard, simplistic, mathematical model that was easy to compute before the invention of computers. But even now, if you want to make a more sophisticated model, you are still forced to talk about its results in terms of the simple model. When I told a biologist friend of mine I was creating a model, the first thing she asked me was “What value of R0 are you using?”. This is how we have arrived at this disaster.

>when neither have faced this before

>The fact that mistakes have been made doesn't mean there's nothing to their field and anyone can do it

I was not making that claim. The premise is: should every non epidemiologist keep their mouth shout and trust the domain experts blindly? Don't they have the right in a democracy to question the science and therefor the legitimicy of the suggested policies?

That's a hard sell, when the field of experts themselves have little consensus.

I might be seeing this from a different bubble, over here in the Netherlands. We have had our equivalent of the CDC show and explain their models and policy choices in detail. We got powerpoints and everything.

Now before you make assumptions: https://www.weforum.org/agenda/2019/11/countries-preparednes... .. the Netherlands ranked 3rd in preparedness.

Yet, it turned into a shitshow. It was already going wrong in Italy, when they claimed their models suggested we have it fully under control if only we wash our hands, and they then let one of the biggest social events happen (think mardi gras). Even though we already had our first cases.

Italy's expert initially said they had it under control. They suggested cultural differences in health and healthcare made it such a bigger problem for China than for them. Then it went to shit in Italy. Then the dutch experts said, after a few cases, they had it completely under control and the reason it went wrong in Italy was cultural differences in health and healthcare. Yeah, racist ignorance.

And all their preparations didn't include supply chain management or the human factor. For example: the relationship (worldwide) between the lockdowns and the available of PPE -vs- the availability of PPE and the spread of the virus. Because none of those things are part of the known models. So they just exclude it entirely.

So, given all of that, yeah, I would like to see their homework in more detail, and we should all put in the intellectual effort, to sort this stuff out.

>Are there no software bugs?

More than any other engineering discipline. So software engineers do actually have more experience debugging collapsing systems build on the wrong assumptions.

>Can just anyone come in and fix those?

Yes, mostly. Its called a pull request.

Astrophysics does have pretty considerable on-going problems as a result of its inability to do real experiments. Even regular physics is heading the same way. Dark matter is basically a giant admission that our models of the universe don't match reality and we have no idea why or how to fix them. String theory does trigger big debates about whether it's actually science or not.

Astrophysics is saved as a science largely by the universe's close relationship with more testable on-earth physics - the motion of planets is Newtonian motion which can be experimented on locally, the physics inside a star is something that can be replicated in a Tokamak, and so on. Sometimes observations are all we have to go on, like with the light bending observation that cemented Einstein as right, but that's rare. Mostly physics yields to experiments. When it doesn't you do see decades without progress.

those models are constantly checked against new findings

I think I'll have to dispute that. My research into epidemiology and its history suggests that this part never seems to happen except by outsiders. Models produce totally failed predictions, in fact often reality falls outside even the uncertainty bounds that are already very generous.

Scientists would at this point stop and say, hmm, there seems to be a problem with germ theory. Our predictions are wrong so we need a refined theory. They'd then design experiments to figure out what that problem is.

Epidemiologists never do this, perhaps because they often don't have a medical background. They can't simply switch contexts and start doing lab experiments. So they simply declare themselves to be highly successful despite all the evidence they weren't, and wait for the next outbreak. Then they jump up and present often the same or very similar models to policymakers as "science", and inform them if they don't immediately implement the epidemiologists recommendations (which are always extreme) millions will die.

COVID is a case in point. Imperial presented a model that was in the words of its author based on "thousands of lines of undocumented C written over 13 years ago to model flu pandemics":

https://twitter.com/neil_ferguson/status/1241835454707699713...

It's rather ironic that we've seen endless attacks on people who "claim it's just like the flu" when one of the world's leading epidemiologist literally used a flu model to predict COVID.

But this statement implies epidemiological theories and models didn't really change in 13 years. And when you look at these models in depth, that isn't a surprise. They're astonishingly simple. There hasn't been any breakthrough in germ theory from these people, and their models always predict massive epidemics that don't actually happen. To explain this they'd need to do medicine, or at least honestly admit that current theory is inadequate to explain observed reality. They never do this, which is why in my eyes they aren't scientists.