# A second coronavirus death surge is coming

1. [3]
Overzeetop
18±4 day lag from infection tracking to death tracking - https://imgur.com/gallery/6uTvXCh I only did three states a week ago but and Arizona has the same cycle. I didn't have the second death...

18±4 day lag from infection tracking to death tracking - https://imgur.com/gallery/6uTvXCh

I only did three states a week ago but and Arizona has the same cycle. I didn't have the second death jump in Florida to chart, so I left it as ?? and called a spike starting 7/12. It might have started a day or two early.

1. [2]
krg
How do your projections line up with Healthdata.org's projections?

How do your projections line up with Healthdata.org's projections?

1 vote
1. Overzeetop
I wouldn't even consider estimating out past 1-2 weeks. Their curves look very nice, but there are too many inputs/variables to effectively account for. And I don't mean that we shouldn't be doing...

I wouldn't even consider estimating out past 1-2 weeks. Their curves look very nice, but there are too many inputs/variables to effectively account for. And I don't mean that we shouldn't be doing projections, or that we can't run simulations for an n-dimensional problem (Though the visualization of such a condition would be challenging. if not impossible).

The biggest threat is that people are impatient. They see days or weeks as proof of a trend, whereas the ebb and flow of this virus propagation (from patient zero to death-rate trends) is on the order of months. It's a control system with a feedback loop that has a reaction time of weeks and data which is incomplete and full of noise.

2. skybrian
From the article:

From the article:

Cases began to rise on June 16; a week later, hospitalizations began to rise. Two weeks after that—21 days after cases rose—states began to report more deaths. That’s the exact number of days that the Centers for Disease Control and Prevention has estimated from the onset of symptoms to the reporting of a death.

1 vote
3. [2]
Eabryt