Member-only story
A revised model for predicting the Wuhan virus.
“Those who have knowledge, don’t predict. Those who predict, don’t have knowledge. “
— Lao Tzu, 6th Century BC Chinese Poet
In reality, Lao Tzu is quite wrong — knowledge and prediction are inseparable. Those with the most knowledge are, in fact, the best at prediction. But his admonition of overly confident forecasters (like myself) is duly noted.
A week ago, I predicted the growth of the Wuhan virus (2019-nCoV) would peak at around 17,000 confirmed cases based on a simple application of the Ratkowsky Sigmoidal Growth Model (RSGM) which has been used to model the 2013 Severe Acute Respiratory Syndrome (SARS) outbreak.
It took the Wuhan virus about two days to blow past my prediction and its spread has continued seemingly unabated. As of today (7 February), over 34,800 people have contracted the Wuhan virus and 723 have died.
What went wrong with my forecast?
Pretty much everything.
First, the RSGM may be too simplistic to generate forecasts in real-time (i.e., during a crisis). The RSGM is good for retrospectively modeling the cumulative distribution of virus cases; however, using it for forecasting a pathogen like the Wuhan virus…
