S-Curves and the Coronavirus

“It’s tough to make predictions, especially about the future.” Baseball player and philosopher Yogi Berra

The term “Black Swan” is often used to refer to the realization of an unexpected risk that has a big impact. It is sometimes overused to refer to any unwanted occurrence, but it definitely applies to the coronavirus. No one saw it coming and it is having major ramifications across the globe. When will it end? That is hard to predict. The life cycle of a variety of phenomena that have occurred in the past can often be described by an S-curve, illustrated below.

These graphs have been shown to accurately approximate the total number of compositions by musicians over the course of their careers, the total number of publications by professors, and cumulative cases of diseases, among many other phenomena. Can it be used to predict when the spread of the coronavirus will start to subside? These S-curves start with a period of acceleration, then hit an inflection point and start to slow. But is the slowing reliable in real time? I believe the S-curve may be better at explaining the past than predicting it. As a case in point, a week or so ago I noticed someone on LinkedIn had posted a graph that seemed to show that the number of cases was starting to slow down, like in the image below.

However, just a few days later, the trend picked back up again. As of March 9th, the trend is shown below.

In hindsight, one likely explanation is that testing was lagging. The slow down reflected the relative lack of testing, not the spread of the disease. But now that testing is on the rise, will a slow down be a reliable indicator that the spread is over? Maybe. It is always easier to explain the past than predict the future. As Yogi Berra also used to say “It ain’t over ’til it’s over.”