.@Cernovich Because you got a lot of things about the coronavirus right, allow me to jump in with a few things about 'the models' that I think you're not seeing quite right: 1) The 'doomsday' 2.2 million deaths prediction was always a 'do nothing/no mitigation' scenario in the...
...Imperial College simulation (imperial.ac.uk/media/imperial…). Clearly, we aren't doing nothing. Complementary, they also consider a hard suppression scenario with school shutdowns and a lot of social distancing; US doesn't run out of ICU then and is predicted <100,000 deaths.
What is a bit confusing is that the Imperial model predicts an ugly (and deadly) peak later this year, even with the current suppression. Why? Well, they assume at some point, we'll lift the lockdown. But immunity is still low and we don't have a vaccine. So what they're saying.. 
...is we're postponing the bulk of the problem for now. Hopefully, we'll come up with better drugs etc., but without that, the model is right not to create a false sense of security. 2) The UW IHME model that is now receiving a lot of attention in the US is a different kind of...
...model in the sense it doesn't seem to simulate at the micro-level, but perform curve fitting (statmodeling.stat.columbia.edu/2020/03/29/the…). Where do they find curves of COVID-19 deaths under lockdown? A few weeks ago, only in Wuhan data, but now also in Italian, Spanish, etc. data, which they...
...actually document quite nicely, also how the better data affects their death estimates for the US (healthdata.org/sites/default/… and healthdata.org/covid/faqs). Also when we started social distancing, we didn't know how important which element of it was - now we're learning about it..
...Just saying, the models will continue to get some things quite wrong, some things less wrong, and some things even right, but I think it's important to criticize them for the right kind of reasons, so that we as a society can take well-informed decisions on what to do next.
