WUHAN VIRUS DAILY UPDATE FOR MARCH 20
Even when you're used to building projections, it feels kinda strange when your next data point goes precisely how you said it would from the prior data point. Well that's what we have today. Don't worry - that's as much of a fluke as anything - we still don't have a clear indication as to when this thing will peak from the data itself and we expect more surprises before it's all done. But there's some comfort when things go exactly as expected even though one data point is nothing at all like a trend. So let's see...
Today there were 17,038 total incidents at a growth rate of 45.34% and 219 deaths at a growth rate of 36.88%. Both down by 6% from the prior day which is a positive development although way too early to call a trend. This is, however, exactly what we predicted because the good news is that this comes from 138,521 total tests which is 34,576 new tests - another 50% bump in new tests since the day before. This increase in tests is absolutely critical and is how places like Singapore are keeping this thing suppressed. It's what allows us to direct our efforts where they will matter without having to shut down our economies which the idiot governors of California and New York are now doing. If they had widespread testing they could be using a scalpel instead of a bulldozer.
At current rates - general likelihood for death will drop below 1% within 4 days. Initial projections, before the new testing capacity came online, was that couldn't happen until the 28th. A positive development.
What do we know? This is not yet demonstrably contained - but the ratio of incidents to deaths is behaving in the manner we would hope it does as new testing comes online. The risk right now is that the time from detection to death might be longer in USA than other countries due to superior health care access so there could be a bump in the rate of death in the next week or two. If that doesn't happen then, based on behaviors in Italy and China, we should hit the peak of this within two weeks.
How will we know? When the daily rates of growth for both incidents and deaths drops below the moving averages for both (currently at 44.02% and 34.73% respectively) - that will be a strong indication. When they drop below the average of the moving averages then I'll be able to reasonably pick the "end date" of the pandemic.
My bet right now is that this will happen in April and the end date will be in June. But that's a gut call and not based on any proper projections for which there is certainly inadequate data to provide any credibility for.