A lot of what you are reading about how bad corona is going to be are from modeling and computer simulations. They are useful tools and I honestly believe most people running them are diligent and are not aware of how bad the WHO is but the assumptions underlying them 1/n

Can yield radically different outcomes with relatively small tweaks. Let me give you an example. This is a GitHub model which is the basic type used by the epidemiologists you read about. You can play with the different assumptions and inputs to 2/n https://gabgoh.github.io/COVID/index.html

See what happens. It is a useful exercise and you can see how easy it is with a few small tweaks we can end up in radically different places. I want to highlight arguably the most fundamental variable of R0 and what happens when you use the WHO mean and empirical researched 3/n

numbers from doctors, epidemiologists, and mathematicians around the world. This is the rough US population with the WHO R0 mean of 2.26. (It's actually 2.25 but I can't input that). We have not yet actually hit the peak and expected 40k dead. But what happens IF.... 4/n

We move the R0 to a number researchers are finding quite common between 3-4? This seems a more accurate spread number from what researchers say BUT it also gives us crazy other numbers. For instance, while we may be past the peak it also says we have had 6m deaths from corona 5/n

and 37 million hospitalizations from corona. These are of course absurd numbers which we can easily observe are not true. So out of this very simplistic exercise there are a couple of key points. First, what you are reading in the headlines about deaths and things like that 6/n

need to be scrutinized. Second, play with this website. It's cool and you learn a lot. Third, lots of variable at play. Fourth, a model or simulation is not empirical research. Do not confuse the two. Fifth, my own instinct is that the R0 from empirical research is higher 7/n

than the WHO guidelines widely used by many people but that the broad severity is lower. Otherwise we would have 6m dead. Revisit your priors, be willing to change your assumptions, ask question, wide variance forecast.