You know why? Because GIGO.
The actual models tend to be some mix of random diffusion models, from fluid dynamics, and mix and match markov chains. With the more sophisticated trying their hand at stochastic differential equations. Which has it own serious problems for the unwary.
Really good post jmc - an excellent summary of the difficulties.
I haven't looked at the papers yet. I know how I'd go about trying to model all this (at least as an initial stab at things) - but I'd be under no illusion that any model I could generate would be an accurate and faithful representation of reality - even if I did get the numerical analysis and subsequent coding correct. I would hope that I'd be able to get some useful insights from such a model - but there would be a zillion caveats I'd be writing in.
If they're using Markov chains they're already making some assumptions about the stochastic behaviour they're modelling - are those assumptions justified? Maybe - but as you point out, one has to be very careful.
I think you drawing an analogy with economic modelling is excellent - and spot on. It's probably worthwhile making more of this when discussing these epidemiological models with the science sycophants who seem to think that just because a paper passes peer review it's like the 10 commandments carved in stone.
The cure should never outweigh the disease..well it clearly has hasn't it..
A very depressing litany Mia 😥
I've been trying to process what we've been doing to ourselves for months now - and still haven't quite fathomed how we've managed to go from where we were to where we are - all for a virus that isn't that deadly except for older and more vulnerable people.
It's no wonder SAGE contains so many behavioural scientists - I think 2020 has demonstrated (yet again) just how powerful state-sponsored propaganda can be.
Was drastic remodelling even possible previously. I think not.
. . . .
I'm sure there are a million other things that make lockdown feasible.
I fear you are right. I'm OK diddling about with equations and pen and paper - I'm much less confident when it comes to understanding people and societies.
I'm still having trouble getting my head round how we've gone from coping OK with respiratory infections like flu to completely losing our marbles over covid. And how fast this has occurred.
I suspect a lot of the points on your list hit their respective nails on their heads 🙁
Neil Ferguson, tried all that, Rudolph Rigger.
😆
Like I said, these models are notoriously difficult to get fully right.
You don't need to know when they are right, you just need to know when they are wrong. The Ap Priori assumption is that they are always wrong as it is with all software. The onus must be placed on the manufacturer of the epidemiological model to do (let's call them) phase 1,2 and 3 trials for models, and follow up monitoring. We need to put software under the Medicines and Healthcare products Regulatory Agency. Due to rank scientific incompetence, ICL introduced a medical software without proper medical trial, the effect has been worse than thalidomide. Why do scientists never learn. The same happened with radiation therapy in the early days:
Modellers should always bear in mind the first rule of modelling: all models are wrong, only occasionally are some less wrong than others. If this simple maxim were recognised a bit more by their protagonists perhaps their advice to our politicians would be rather more measured.
