Christopher F. Rufo  ⚔️

In fact, its conclusions aren’t very different from what Imperial researchers conclude will occur if we maintain radical social distancing measures The fact that the methods of the two models are radically different and yet converge on similar conclusions adds credibility

The U of Wash model is real-time. Murray says he updates his model every day. “We were talking with one of a governor’s groups in a state last night where they had a big bump in deaths,” said Murray, “more than expected. And so we want our forecasts to reflect the new trends.”

Imperial’s model is “like measuring the impact of a fire on a town if nobody changed their behavior—everybody would be burned in their homes—as opposed to how many would die if everyone runs away when the fire is coming. Some will still die but most people change their behavior.”

Our model is meant to be real-time,” Murray said. “We fit a statistical curve to the available data and to the variables that explain the curve, and the key variable is social distancing.”

I asked Murray if one model was better than the other. “Both are legitimate,” Murray said. “The first model is a great call to action and super useful. But if you’re trying to help the hospital and fire-fighting service, you need our model.”

But hadn’t the Imperial model been too apocalyptic and alarmist? “There’s a value to apocalyptic alarmism because as we saw some places were slow to do anything,” Murray said.

“And there are still some countries that are slow to do much,” he explained. “I’ve been on the phone with political leaders in Europe who still don’t have stay at home orders because they are worried about the economic effects.”

Murray warned against ending social distancing too early. “If we get over the peak of deaths and people say we’re over it, and we end social distancing and get a second peak in May or June, then we’re in a long period of deaths and far more damaging.”