I have been unsatisfied with the estimates of prevalence and incidence of COVID-19 around the world as well as the many doomsday scenarios. The epidemiologists and modelers also appear to be making things too complicated and including too many ill-defined variables. My attempt is to make forecasts based on the most solid data we have and that is, deaths. I’ve tapped into the World Health Organization (WHO) and and Johns Hopkins University (JHU) databases and a couple of other data aggregators.
The main premise of my model is that by following the daily death statistics and observing their trends, we can forecast important parameters, e.g., number of cases, deaths, and how they vary with time. Death statistics, however, are a lagging indicator because they represent infections 2-3 weeks ago. So forecasting comes with its caveats. However, by observing the daily/weekly changes one can deduce if we are still accelerating or getting a handle on the problem. For example, if R0 is ≤ 5 and we can reduce social contact by ≥ 80% then we should be able to contain the virus. China and S. Korea tamed it. Italy and Spain now have contained the virus as well as other previous hotbed countries. The U.S. and U.K are dead last not surprisingly since they were the last to declare a national emergency and whereas the U.K at least socially isolated as a national policy, the U.S. administration pretty much did nothing and left things to the states, which most have responded admirably, but not all.
I do caution that events change daily and are very unpredictable. Even days old blogs can become obsolete quickly. That is also a reason to be very diligent and follow events daily.
Please feel free to pass this link on to anyone who may be interested.
A little about me can be found on:
And a huge thanks to Vasiliy Loskutov for launching the blog site.