Saturday, March 21, 2020

New US Cases Not On the Exponential Curve Today

There is a slight hitch in new US coronavirus cases today. Hopefully that's the start of something:


The US now has more cases per capita than China has/had:
Italy  777 cases per million
US    60
China    58
Canada    29
But the US is still 15 days behind Italy, same as yesterday.

10 comments:

Layzej said...

This is great news.

Ned said...

Sorry, DA, but actually the new cases are right on the exponential curve; it's just that yesterday's new case count was slightly above the curve.

David in Cal said...

In a period of dramatically increased testing, one would expect new cases to be steeper than the curve. Current number of today is 4759, but that's an incomplete figure. It goes up throughout the day.

BTW according to today's White House press conference, the number of completed tests with results provided is 190,000. That does not include tests done by 15,000 small, independent labs.

Cheers

David in Cal said...

Long, very interesting article and analysis at

https://medium.com/six-four-six-nine/evidence-over-hysteria-covid-19-1b767def5894

This author concludes that we are over-reacting to the risk and that some of our steps are not needed. In particular, he believes we're closing down too much. He points out that "None of the countries the global health authorities admire for their approach issued “shelter-in-place” orders, rather they used data, measurement, and promoted common sense self-hygiene."

I hope he's right. If so, we should see a rapid improvement.

cheers

Ned said...

In a period of dramatically increased testing, one would expect new cases to be steeper than the curve.

Why? Testing is increasing on a cleanly exponential trend, and number of known cases are also increasing exponentially. Neither # of tests nor # of cases has significantly departed from the exponential trend in recent days.

For the last three days, the # of new cases has been slightly above the exponential trend each day. The # of new tests has gone from slightly above the trend to slightly below it.

For today, the exponential models and actual values are as follows:

New tested: Predicted 49,129 .... Actual 44,176
New cases: Predicted 5580 .... Actual 6165

Note: data from https://covidtracking.com/us-daily/. Numbers may differ slightly from other sources due to timing of reporting.

Ned said...

Here is a chart showing the data & models for the previous comment:

chart 1

and here are the residuals:

chart 2

Entropic man said...

Now this is worrying for you. The USA is now in third place in the world total cases ranking behind China and Italy.



https://www.worldometers.info/coronavirus/

Ned said...

Well, yes, but the reliability of countries' case counts varies widely due to differences in testing.

The number of deaths is probably a better indicator, and by that metric we're in 6th place, behind Italy, China, Iran, Spain, and France. Based on the numbers of deaths, I'd assume that at least Iran and Spain are also ahead of us in cases, maybe France too.

Ned said...

David in Cal writes:

Long, very interesting article and analysis at

https://medium.com/six-four-six-nine/evidence-over-hysteria-covid-19-1b767def5894

This author concludes that we are over-reacting to the risk and that some of our steps are not needed.


The author of that post has no expertise in epidemiology. It's been taken down by Medium and moved over to Zero Hedge, which is probably an appropriate venue for something of its quality (i.e., utter crap).

Actual epidemiologists are not pleased:

thread
twitter

Ned said...

This is one figure from Carl Bergstrom's takedown of the "Medium" (er, now "Zero Hedge") crap that David in Cal was promoting:

https://i.imgur.com/K7nvZCc.jpg

Oops!