"The 100K-250K deaths number was a very plausible estimate": a thread.

There are 18 bazillion ways to look at the COVID data coming in and we're all getting that rare chance to look behind the scientific veil and glimpse the messy nature of data gathering (1 of 13)

There have been a lot of weird data moments in this thing and there will be more.

There was the week where CA had 50K+ pending tests, where WA stopped reporting entirely for 4 days while they re-jiggered their reporting system. /2

We're used to seeing data after it's been gathered, estimated, cleaned, and referenced against other proxy data sources.

We have none of that here and it's confusing and frustrating to people who have never been this close to the ground in data gathering before. /3

There's a lot to say about the "right" way to gather and report this data, but this thread isn't for that.

It's to talk about what we knew and what we know now. /4

Two weeks ago, there were the barest hints that Seattle might be doing ok, while NYC was getting very bad very fast.

It was still too soon to be certain about *anything*, but we were getting hints.
https://twitter.com/politicalmath/status/1240784235717210112
/5

Since that day (3 weeks ago today), Washington has seen deaths increase from 66 to 421 (6X)

New York has seen deaths increase from 12 to 6,268 (522X). And that doesn't count New Jersey, which is part of the NYC metro.

Something different happened between these two places. /6

What exactly was different is not the topic of this thread. It is to point out that, for a while early in this thing, Seattle was worse than NYC.

But also we had to assume (at the time) that what we are currently seeing in NYC was *possible* in other areas. /7

When the 100K-250K projection was announced (late March, early April) I started a spreadsheet that roughly estimated how many deaths per state we should expect to see the 100K and the 250K scenarios.

Yesturday, NY passed the 100K ratio. It could still hit the 250K ratio. /8

Put another way: If every place in the US ended up like NYC, then we would hit 100K deaths and very possibly 250K.

And, at the time, Cuomo was telling us that NYC wasn't the worst, it was just the first. He said the NYC scenario was coming to every city in the country /9

But we're three weeks later and we have a lot more information now.

Washington looks better than we expected. So do California and Oregon, both of which got this thing early.

Boston got this early, but isn't in a NYC situation. Louisiana might be closer to NYC. /10

In short, we are seeing BOTH pessimistic and optimistic numbers (and everything in between) all over the US.

Each region is getting this thing slightly differently and there are so many variables, there isn't a good simplistic narrative to which we can retreat. /11

The models aren't everything, but they are something. They're attempts to set expectations and convey danger and they suffer from the fact that we have a know-it-all culture of hind-sight second-guessers who pose as experts. /12

I'm not saying "trust the experts blindly" but more "think very hard, digest information slowly, wait until you're sure, check back on what you said before"

Oh... and show grace. Most everyone is trying to do their honest best. /13