Log Interpolation of US Covid-19 Body Count for April

[presence]

https://infection2020.com/​

 

[kahless]

On that site I do not see a way to display past 30 days. Is that your > 30 day interpretation ?

 

[kcchiefs6465]

Your modeling is that one million will die over the next 25 days?

And that in a span of eight days it will go from 1 million infected to 100 million infected?

 

[presence]

yes just image editor linear interpolation of this:

 

[presence]

Your modeling is that one million will die over the next 25 days?

And that in a span of eight days it will go from 1 million infected to 100 million infected?

yes that's correct

 

[kcchiefs6465]

yes that's correct

The assumption is that this virus has been here for as long as it has elsewhere or how would you explain the data sets for countries that are a month or more advanced in the progression of this disease not having realized these figures?

And granted, the Chinese numbers are most certainly shit.

 

[kcchiefs6465]

Wouldn’t it also be the case that more sparsely populated regions aren’t going to see projected growth figures in line with New York?

Modeling based primarily on NYC or greater New York and then extrapolated upon the whole of society seems flawed from the get go.

 

[kcchiefs6465]

And then too for people who are asymptomatic or untested due to a lack of severity in their symptoms who are uncounted, wouldn’t the numbers you’ve used all be effectually skewed?

 

[presence]

The assumption is

the only assumption is that the line of best fit on logarithmic scale points as it points.

if you look at death toll is actually concave UP from linear on log scale.

that means the dead are piling up faster; its a "hyperbolic" curve,

so the log linear projection is conservative on the 30 day derivative terms

 

[presence]

Wouldn’t it also be the case that more sparsely populated regions aren’t going to see projected growth figures in line with New York?

Modeling based primarily on NYC or greater New York and then extrapolated upon the whole of society seems flawed from the get go.

well there's always this, so we shall see:

 

[kcchiefs6465]

the only assumption is that the line of best fit on logarithmic scale points as it points.

if you look at death toll is actually concave UP from linear on log scale.

that means the dead are piling up faster than linear on log

its a "hyperbolic" curve, so the log linear projection is conservative on the 30 day derivative terms

I appreciate the response but to be fair, countries have not seen the growth as it is being or has been projected.

Obviously changes of behavior, etc. will alter trajectory.

One million cases to 100 million in a matter of eight days seems rather unimaginable.

I would be curious to have the graphs of countries in this point in time of the pandemic matched to where we are at (and then extrapolate that while keeping the graph you have created).

ETA: It would also be nice to see countries’ population densities, etc. accounted for.

I just think the data profile is incomplete and can lead to some rather unimaginable conclusions.

Not to conflate apples and oranges, but it is reminiscent of NY will be underwater by 2016, etc.

 

[pcosmar]

Computer algorithms all seem to be hockey sticks these days,, since that shit was accepted as science.

Bullshit will be illuminated by reality soon.

current dead 195 in Washington State,, and that includes at least one Fatal Head injury,,that tested positive on autopsy..

Most are from Old folks homes with questionable care.

 

[JJ2]

I appreciate the response but to be fair, countries have not seen the growth as it is being or has been projected.

Obviously changes of behavior, etc. will alter trajectory.

One million cases to 100 million in a matter of eight days seems rather unimaginable.

I would be curious to have the graphs of countries in this point in time of the pandemic matched to where we are at (and then extrapolate that while keeping the graph you have created).

ETA: It would also be nice to see countries’ population densities, etc. accounted for.

I just think the data profile is incomplete and can lead to some rather unimaginable conclusions.

Not to conflate apples and oranges, but it is reminiscent of NY will be underwater by 2016, etc.

100 million cases that soon is not even possible because we won't have that many tests available.

 

[presence]

And then too for people who are asymptomatic or untested due to a lack of severity in their symptoms who are uncounted, wouldn’t the numbers you’ve used all be effectually skewed?

at some point something will slow the trend we're on. It could be "herd immunity at 60-80%" with lots of asymptomatic people.

The diamond princess numbers are pretty telling of % that die though. I think they all got tested to get off the boat.

so 712 test positive
99 still sweating it out
10 dead

assuming the rest live if you look at the ten dead relative to the 712 that's 1.4% death rate, but lets say that's extreme of death toll; be more conservative at 1%

the average time of death is 18 days after you catch it. so you could throw in another trend line on 3/12 "how many are actually infested"

4000 "dead now" / 0.01

= 400000 "actually infected" 18 days ago!

so this will mitigate as we approach the 330,000,000 US population

but, although future "infected" and "official infected" will curve off over time,

future deaths are 18 days delayed from "curving"

so not much will change in death projection terms until late April

so on account of the fact that we actually have 40M infected right now; we an be fairly assured close to well over 100k and close to 400k will die in 18 days

 

[pcosmar]

yes that's correct

Sounds like the Hockey Stick algorithm

are you sure that is correct?

why is algorithm so much like ALGOREithm?

 

[pcosmar]

[MENTION=36577]presence[/MENTION]

Are you seriously presenting Known Liars and Liars Projections as factual information?

or are you just presenting the obvious lies as documentation?

It is hard to tell.

 

[ClaytonB]

 

[swissaustrian]

There are so many unknowns that any model inevitably has a huge margin of error, no offense OP:

1. some variables that the goverment or any human for that matter can NOT influence
What is the quota of assymptomatic patients, ie what is the actual number of infected, not the number of tested?
How are environmental effects (temperature and humidity) affecting the spread?
How is pollution affecting the death rate?
How are genes, including blood type, affecting the death rate?
Why are men representing 60-70% of deaths?
What is the age pyramid looking like in a given area?
Spread of pre-exisiting conditions, including obesity and smoking, in a given population?

2. Variables that the goverment or any other human can/could influence:
Social distancing?
Proliferation of masks?
Hygiene/disinfection of surfaces and hands?
Quality of health care, including prevention?

And on a more fundamental point: just because a person tests positive does not mean that their actual cause of death is the virus. In Italy, the first 2500 deaths had 99.2% pre-existing conditions.

 

[Okie RP fan]

Does anyone else think the U.S. gov't has confirmed and/or have their own estimates of how many really died in China and that's why they've jumped up their estimates?

 

[swissaustrian]

Does anyone else think the U.S. gov't has confirmed and/or have their own estimates of how many really died in China and that's why they've jumped up their estimates?

Dr Birx (Trumps advisor) based her projections solely on Italy in yesterday's press conference. The Italian numbers are probably not falsified (contrary to the Chines), but still, why pick the worst possible country and not include countries like Germany that have much lower numbers on a per capita basis.