My findings were simple, hidden in plain sight: the early outbreak of SARS-CoV-2 was faster growing than many epidemiologists estimated. One thing we don’t screw around with in finance is estimating exponential growth rates – that’s how you estimate returns, and returns are the bread and butter of finance. According to my estimates, case growth rates were much faster than all the conventional models estimated at the time, and faster growth rates with the same start date implied more cases, a larger subclinical iceberg of infections, lower odds of successful containment, and lower severity.

With faster growth rates, estimates of the severity of the pandemic become extremely sensitive to the start date; 2-day doubling time starting 20 days ago will generate 1,000 infections, but 2-day doubling times starting 60 days ago will generate 1 billion infections. Every 2 days we err in our highly uncertain estimate of start dates, our estimate of pandemic size, and burden changes by 2x.

I shared a write-up on these findings privately and most people saw rubbish. One professor from Oxford even said directly to me that if Harvard is saying one thing and Alex from Bozeman is saying another, he will believe what Harvard says. As I tried to warn people of an oncoming bullet train of a pandemic in February 2020, I received stern emails from distinguished professors high up the totem pole claiming that if I shared my findings openly, it could sow complacency and “disrupt the public health message.”

In emergency situations, I’m accustomed to flattening hierarchies, trusting information as sincere, and ensuring communication flows. In the early days of Covid, however, our information was not shared widely, the tall totem pole of academic hierarchy did not flatten, and my statistics were assumed to be rubbish.

Jay Bhattacharya, meanwhile, found the ruby.