"US covid deaths have been flat since august!" has become one of these odious refrains.

the problem is that is flat out not true.

US deaths are down by 60-70% since august when you use the real day of death data, not the day of report data favored by covid tracking project etc.

fortunately, the CDC provides this data on a weekly basis so we can use theirs and see that their data confirms & conforms to the fact that cases are WAY down from the april peak once you adjust for sample rate

(note that the recent weeks are incomplete) https://twitter.com/boriquagato/status/1320077624564174849?s=20

note that it also maps to the % positive on testing.

there is only one chart breaking out to new highs in the US instead of being basically on the lows: testing.

and this testing rate is driving the "casedemic."

the US is doing 1.2 million tests a day. this is an insane number. it's more flu tests than we do in a year.

in 2020 we will likely do more cov tests than the US has done flu tests in its entire history.

deaths are a bit more complex.

"but muh worldometer/CTP data looks like this!"

yes, i know it does. but it's bad data.

it's astounding to me that at this point, we're still having the day of death vs day of report issue, but it appears that we are.

worldometer/CTP do not report the real day of death

they report the death on the day that death is reported, not the day it happened. so it does not map to any real epidemiological curve

it works as shown below. they report the top series. but the bottom one is what happened

as can be readily seen, this can make a drop look like a spike and this is precisely what's going on in the US.

hospitals and states are "mining" for covid deaths and adding them from months ago. this looks like it happened today in the CTP data. but it didn't.

why would they do this? because it pays. bigly. the CARES act allocates state funding and massively increases payments to hospitals for finding covid/covid deaths.

so they are busily looking for it and going back months. they're still adding deaths from april and may.

this has the noxious effect of making it look like deaths are not going down when, in fact, they are.

they are at ~2900 a week now, down 83% from the highs and ~70% from august.

but, apparently, that does not make for good news copy or fodder for "experts" to sow fear with.

but, at the end of the day, the facts are the facts.

the data on covid is still a mess and the media outlets that report it have made zero effort to change this fact. real date of death data is available, yet they make no effort to use it or even make people aware of the issue.

anyone mistaking such selective use of data and the choice of a dataset known to not map to reality in preference to one that does is going to come up with the wrong answers.

and this is by design.

there's just no way this is an accident anymore

this is not "data" it's propaganda.

it's not like it's difficult to get real data.

go download it yourself.

it's not like this issue is not knowable and widely known.

it is. it's been discussed since march.

https://data.cdc.gov/NCHS/Provisional-COVID-19-Death-Counts-by-Week-Ending-D/r8kw-7aab

using this misleading data is a choice, and it's not one with your best interests at heart.

so make your own choice.

this is not a reality filter you want to fall prey to.