I addressed what @AltisticRight said. In fact, I dragged up the rest of the graphs that supported the argument that the data was fabricated to fit a curve.
I was going to let his last post slide unchallenged, but since you sassed me up, now I have to go back.
You dragged up nothing actually.
What the twitter handle is doing, is in fact completely circular. It's a non-argument. It proves jackshit, or rather it shows the possibility to fit a curve to any given data.
Let's go through this.
- Epidemic starts, the numbers are on the rise
- Twitter user fits a curve for this data
- R-sq=1 because no shit, he fit a curve, a function to a plot
- This is exactly why his function started to fail as things get jumpy
- He adjusted his function, and it still failed
- He admits that the function needs to be adjusted, while you're passing it off to suit your autistic conspiracies
Why is this method flawed?
- I can do the same with numbers for the USA and etc. I will be able to fit some kind of a polynomial to the increasing data
- I am using cumulative cases, because fucking reasons. Or rather, if I did use daily increase, fitting a curve is harder, results in lower R-sq, the numbers won't consistently increase, and renders my preconceived conclusion invalid
- Most people don't understand what I'm doing.
So I question these economist's capability to actually understand regression analysis, something that's taught in a year 2 statistics course, easy as fuck.
- Is the data normal?
- Residual plots?
- Then, linear or nonlinear regression?
None of these are substantiated, but I'm a phd and I'll pull a curve to fit some dots.
Excel is shit, fuck Excel.
Here's what I've done:
1. Tossed in a bunch of numbers
2. Fit a curve to these numbers
Oh look, 99.64% R-sq! The numbers I've plucked out aren't random, I must have generated it with a formula!
If I use an order 2 polynomial, the R-sq is still high:
What about increasing the order of the polynomial?
Let's generate some random autism:
I pasted the results as numerical values, then ordered them.
I removed the duplicates.
Oh look, Excel's randomly generated sequence must be based on this formula, because it has an R-sq of 99.95%.
I read this from a statistics for nigggers book, trust me. 1352 is a lie. Most bronies don't clop and don't molest children.
Now let's apply my Statistics for Nigggers book to the American cases of Coronachan:
(Source: WHO fuckwits, 2020)
Oy vey, 98.07% R-sq, the American numbers must be generated from the formula on my screen! Let's add one more degree of precision to counter that annoying trough at the beginning.
Oh look, this is my order 3 polynomial to generate the fake numbers, 99.94% R-sq, the numbers must be generated via a spreadsheet!
“I have never in my years seen an r-squared of 0.99,” Goodman says. “As a statistician, it makes me question the data.”
I question your qualifications.
For context, Goodman says a “really good” r-squared, in terms of public health data, would be a 0.7. “Anything like 0.99,” she said, “would make me think that someone is simulating data. It would mean you already know what is going to happen.”
This claim is fallacious at best. Public health data isn't CUMULATIVE infection data. Cumulative means it will consistently increase, which makes fitting a nice polynomial infinitely easier.
"She" said, maybe leave it to the men.
"The numbers are increasing, people can fit a polynomial to a set of consistently increasing numbers, therefore it must be a simulation".
Maybe leave the math to mathematicians?
Way to compare fucking apples to dildos, m'lady.
Let's apply the same method to newly reported cases:
(WHO is gay, 2020)
I deleted the entries with zero reported cases which is definitely an error.
Chinese new cases, an R-sq of 70.19%.
America new cases, an R-sq of 94.495%!!!!
So, what's your conclusion, David Clop.
You seem to think high R-sq means fake data. The Chinese R-sq is low, America ones are high. So will you accept the premise based on your own logic and accept the conclusion that the American numbers are generated from a spreadsheet?
Or is R-sq an inherently bad measure, and should not be used to claim the adequacy of a model?
I have yet to gather the numbers for China for the month of February and January. If I did, the fitted polynomial will have a good looking R-sq as well, because the new cases will be consistently increasing.
I'm also comparing apples to oranges. Chinese numbers are droplets now since the epidemic is over.
I like how you pretend that these data sets are in any way equivalent.
South Korea and Singapore have comprehensive screening programs in place, while the authorities in China have basically stopped counting cases.
What this means is that South Korea and Singapore will be counting a much, much larger number of asymptomatic and mild cases than anywhere else.
I like how you pretend to know anything while I've demonstrated the opposite.
China still has screening programs. Get off the train, enter a major station? Body temperature still tested, cases are still counted. Sorry if facts collide with your delusions.
South Korea will be counting larger numbers, yet the statistics for newly diagnosed is in single digits. Nice.
Not only does this mean that they'll be better off at preventing the spread of the virus, it also means that their ratio of total cases to total deaths will look better than in other places. I find it amusing that the 1 in 55 death rate in Korea matches the presumed IFR of the disease almost exactly, about 1.8%, squarely in the middle of the 1 to 3% figure I've been hearing. This means that South Korea is detecting a very, very large number of the infections.
So is China, Taiwan, Singapore, Japan...
And no, the rate being in the interval of 1-3% doesn't mean they are detecting large numbers of infections.
Stick to fucking pony plushies, David Clop.
Did you guys know that
@Drain Todger has been
researching a cure for COVID-19 without telling us? I thought he cared about us.
We shall call the cure "David-19", or "Star of David"... maybe "Star of Covid".
Upload your twisted sexual deviancies guys, if you're not a brony, you get no cure.
I doxxxxed David Clop's Coronachan vaccine, it's patented, don't steal it.
No user manual required.
Oh no, this is much better.
1. The drama keeps the thread bumped and the info on the virus clearly visible.
2. I don't actually give a fuck about being doxxed. Imagine being afraid of a pair of calipers.
A pair of fucking calipers.
Oh calipers? Really?
How about this?
A very painful torture applied almost exclusively on witches, liars, blasphemers and homosexuals.
www.medievality.com
The Pear of Anguish was used during the Middle Ages as a way to torture women who conducted a miscarriage, liars, blasphemers and homosexuals.
A pear-shaped instrument was inserted into one of the victim's orifices: the vagina for women, the anus for homosexuals and the mouth for liars and blasphemers.
The instrument consisted of four leaves that slowly separated from each other as the torturer turned the screw at the top. It was the torturer's decision to simply tear the skin or expand the "pear" to its maximum and mutilate the victim.
The Pear of Anguish was usually very adorned to differentiate between the anal, vaginal and oral pears. They also varied in size accordingly.
This torture very rarely provoked death, but was often followed by other
torture methods.
Because this is what you've been experiencing, getting assblastered for the amount of autistic drivel and doomer sperging, wrapped up in your lack of understanding of basically everything and severe autism.
In the case of the instruments description, you need 2 of these. One for lying, the other for being a massive doomer faggot.
