# Here Is Why Those Covid-19 Predictions Are All Over the Place | Opinion

For those of us who believe in numbers, this is a trying time. Estimates of the number of people who might die from coronavirus are all over the place. There is the projection from a team composed of researchers from London's Imperial College and the WHO Collaborating Center for Infectious Disease Modeling that says up to 40 million people could die this year from coronavirus. 2.2 million of those would be in the U.S. Two articles published on Medium predicting up to ten million deaths in the U.S. by Tomas Pueyo have been viewed over forty million times and cited by Tom Friedman in the New York Times. (Pueyo is vice president of growth at a San Francisco start-up.)

Those estimates sound really bad, and are, although it's hard to deal with numbers that big. Here's one way to look at it. British anthropologist Robin Dunbar has calculated the number of people an average person knows well, is acquainted with, can recognize, etc. Using Dunbar numbers, that means in the worst-case scenario, on average ten people you know would die, including at least one who is in your close circle. Now of course deaths will not be distributed evenly, but it's a useful way to get our heads around what it might mean for each of us personally.

But not everyone agrees with that. Ten days ago UMass Amherst's Flu Forecasting Center of Excellence surveyed 18 infectious disease modelers and came up with a prediction of 195,000, with only a 0.06 chance deaths would exceed 1.5 million. If 195,000 die in the U.S., then that means that on average one person you kinda-sorta know might die, and in a typical U.S. town there would be about a half dozen deaths. Nobel Prize winner Michael Levitt of Stanford also expects the toll to be much less, as does a team from Oxford University.

Why are the numbers so all over the place? Well, part of it is the nature of modeling off a small data set. That's risky business. The Hall of Fame pitcher Hoyt Wilhelm hit a home run in his first at bat. A model would have predicted he'd be the next Babe Ruth, but he never hit another one.

The biggest unknowns with coronavirus are when it really started, how many people are at risk, and what the mortality rate is. If the contagion started a month ago, then it's rising at an incredible rate. But the team from Oxford thinks it might have started two months ago, and if that's the case, it suggests that it's much less of a risk than we think, perhaps because a significant portion of the population is immune. That's what Levitt thinks. He notes that even on the Diamond Princess, only 712 of the 3,711 got sick and 8 died. His inference is that some portion of the population must have a natural immunity. Of course if eight out of every 3711 people in the U.S. died, that'd still be 700,000.

But the bigger reason for the wildly divergent projections is because the outcome depends to a large extent on what we do next. It's hard for most of us to get our heads around scenarios with names like "Suppression" vs. "Mitigation." Last week a team of modelers from the University of Texas led by Remy Pasco and Xutong Wang presented some plain language estimates to the city of Austin. Their model concluded that if the city does nothing, 10,908 people could die. But if it closes schools and people strictly adhere to social distancing, that number drops to 267. Wow.

Now, the best-case UT model assumed 90 percent of the population practiced social distancing, which is probably optimistic, so a more realistic expectation is probably somewhere in between. And of course, there's a feedback loop involved. If we don't see people we know getting sick, we're more likely to let down our guard. Scaling up those numbers to the U.S. would equate to either 40,000 U.S. deaths or 1.6 million, depending on what we do.

No matter what happens, people are going to spin this for their own purposes. If the death toll is high, Democrats will spin it as an indictment of the incompetence of the current federal administration. (And all those Doomsday Preppers eating canned chili in bunkers in their backyards right now will say, "We told you so.") If it's low, then Republicans will say that the crisis was exaggerated as a way to attack the administration. Since it will likely come out somewhere in the middle, expect to hear both arguments.

Still, here's the bottom line: If we treat the Corona virus like a big deal it won't be—but if we don't treat it like a big deal, it will be.

Sam Hill is, among other things, a Newsweek contributor and best-selling author. His most recent Newsweek story is Black China: Africa's First Superpower is Coming Sooner Than You Think