Updated: Mar 26

The human mind doesn't comprehend exponential processes very well. We tend to project exponential processes as linear in the short term, but underestimate the long term effects of exponential growth rates. And the COVID-19 viral outbreak falls into the latter category.

So I decided to put together a rough model to estimate various scenarios based around 4 primary variables:

1. Daily growth rate of outstanding cases (first derivative of # outstanding cases, currently running around ~50% growth from last week)

2. Weekly growth rate decay (second order derivative, # outstanding cases) - this is the most important variable

3. Starting # of present cases 

4. Death rate % of total cases

Here is my Coronavirus Scenario Model that is based on the above variables that will be updated daily.

Keep in mind that the growth rate and decay rate are both assumed to remain static (they don't change) over time, so this is a large assumption in the model. However, if the growth rate actually increases, these scenarios would be mild compared to reality. If the decay rate tends to increase over time, these scenarios may turn out to be too aggressive. But this is the best data we have at this point.

Bottom line, this is serious, and it would be better to overreact to ensure that the second derivative (decay of the growth rate) is as negative as possible.

Said another way, we need to take coordinated action as a society so that the number of infections will grow by less and less each week.

Some believe that this is a massive media/crowd overreaction. But based on simple exponential math, it is not hard to see how this outbreak becomes truly unmanageable in size if not appropriately handled.

I am not an alarmist. I am a realist, driven by data, not opinions. This too shall pass, as long as we remain calm and take appropriate measures.

More data on COVID-19:

Johns Hopkin's real time database

Andrzej Leszkiewicz's COVID-19 database

The math behind exponential growth and infection rates of COVID-19


©2018 by Impressions. Proudly created with