CV19 April 3

CV19 Pandemic Update for 2020-04-03

I did not make screenshots of my favorite graphs, but check out and play with log vs normal, and etc on the interactive graphs.  Deaths and confirmations log chart is very visibly bending in the right direction. We still have a ways to go.

Executive Summary:

  • Worldwide, 1,095,917 confirmed cases and 58,787 deaths, which is within 1% of the 1-day estimate.
  • US is 275k infected, and 7382 dead, which is also within 1% of 1-day estimates.
  • China probably has ~400k unconfirmed cases, and ~16,500 unconfirmed deaths.
  • We need our confirmed and death exponents to fall to as close to 100% as possible. (vs 113% & 119% respectively).
  • US & TX confirmed exponents are stalled in the 113% range. The death date exponent has not yet fallen, averaging over 120%.
  • US deaths per confirmed case continues to increase, and will do so until as far as 47 days after the last new infection.
  • Many places in the US extended their stay-at-home plans to May 20.

1-Day Estimate for April 3
Non-China 1,018,824/56,704; Italy 120,107;14,719; USA 277,775/7,382; TX 5,900/90; ME 411/8
SD Adjusted USA 287,625/7,557; Texas 6,139/100

Actual April 3
Non-China 1,013,406/55,461; Italy 119,827/14,681; USA 275,586/7,087; TX 5,734/100; ME 432/9

1-Day Estimate for April 4
Non-China 1,103,432/61,938; Italy 124,594/15,489; USA 311,960/8,475; TX 6,486/130; ME 496/9
SD Adjusted USA 318,885/9,052; Texas 6,853/129

7-day Estimate for April 8
non-China 1,917,505/109,549; Italy 169,991/23,851; USA 745,176/25,239; Texas 15,874/306; Maine 656/26
SD Adjusted USA 707,819/21,874; Texas 14,357/317
This is falling due to our curve getting flatter. US 500k, TX 11.2k


Day to Day deaths are still climbing nationally. No consistent major slowing there. CA and NY/NJ/MA areas have a lot going on. It may take some time for them to dig out of this.

After lockdown, confirmed cases take 9-12 days to be affected, and deaths another 8-11 days after that. Anywhere that is overloaded will have a higher death rate, which may offset this prediction

As expected with the slow-down in spread, our mortality rate keeps climbing. 1.44%, 1.55%, 1.67%, 1.75%, 1.84%, 2.06%, 2.23%, 2.43%, 2.57%.

Inflection point for outbreaks with an R0 of 2.2 is 60% of the population becoming immune. Divide that by whatever you think our detection rate is. I think we get about 38%, so maybe 75m people in the US showing as confirmed. In other words, we are clamping down on the spread. It is not slowing due to going through the bulk of the population. If we all started mingling again tomorrow, we would see a big surge in cases 2 weeks later.

Average is 6 days to onset and 14 days from onset to death for those who do not survive. Social distancing data lags by 3-5 days. Statistics lag by 1 day. Detection lags by 9-12 days. Impact delay of social distancing on confirmations MAY BE 9 days. The correlation is fuzzy, but may be around 9 days lag.

Research for improving mortality rates could put things back to normal, but we are not there yt. Even if it is solved now, logistics of getting it deployed are not instantaneous.

Spreadsheet is updated, and downloadable here:

My favorite charts:

Additional Data Sources:

China Underreporting Claims:
I do not think this is willful deception, so much as, they are only reporting what they can confirm. It just happens that they can confirm only a small subset of the infections prior to March.

China cremation rates, I lost the URL for the older data, but current data is here:
2019    Number of cremated remains    JZH014221000    56,007
It varies by about 2500 per quarter, so roughly 5k per month in Wuhan province.

The number of urns and cremations are very high, and “the funeral homes are unable to keep up.”,-endless-queues-for-ashes-of-coronavirus-dead-cast-doubts-on-numbers-49673.html

There are multiple sources indicating people who were not checked into a hospital were not counted.
Journalistic sources include

I did not save teh ones from February where the WHO felt they were underestimating because they had no way to count those otuside of the hospitals. They said 4-8 times higher rates were likely.

This exclusion has been common in other countries, including Italy, Iran, etc:

Comments are closed.