Saturday, April 4, 2020

Coronavirus: Light at the End of the Tunnel


  • On March 11, the World Health Organization declared a Covid-19 pandemic
  • On March 12, Hawaii decided that all large events are canceled. Disneyland announced its closing.
  • On March 13, Pres. Trump declared a 'state of emergency' in the U.S.
Right then I started collecting data for the World and for the US. All my data were taken around noon from ncov2019.live, and rounded up to nearest hundred cases.

Here is the summary of my data and projections which suggest that there is light at the end of the tunnel.


The first three graphs show the number of cases for three weeks starting in mid-March. The second graph suggests that a change has occurred and that the rate of daily growth has started declining. The third graph suggests that daily growth has become 'permanently' under 20% and declining. The fourth graph summarizes these three weeks of 21 daily data points and shows a clear downward trend.

In this writing, the last data point was collected at noon on April 4. Where do we go from here? 

If the rate of decline is slow, as the third graph suggests, then the shape of new cases will follow the red line on the bottom graph. New cases will peak at the very end of April. Around mid-May, the new cases in the US will be the same as now at about 300,000 per day. This is a likely scenario. The likely maximum of infections in the U.S will be about 35 million, and the maximum of new daily cases of will be about 1.2 million.

If the rate of decline is fast, as the fourth graph suggests, then the shape of new cases will follow the yellow line on the bottom graph. New cases will peak around April 15 and on May 15, the US will record only 100 new cases. This is a wonderful statistical outcome but not likely to occur unless the mitigations, warm weather and other factors serendipitously conspire to put a lid on this pandemic. Even under this optimistic scenario, the number of infections in the U.S will be about 12 million.

I'll keep following the numbers but indeed there are statistical indications that there is an end to the 'Covid-19 tunnel' in late spring.

======
There is great risk in making forecasts using only a few and fairly volatile data points. Every week adds precious new data that enable better future predictions. I will leave this post alone and will post updates separately. In this way, we can see how close we got in predicting 6 weeks into the future based on 3 weeks in the past.