/sci/ - Science & Math » Thread #11490218

23KiB, 544x327, itsjustaflubro.png

View Same Google iqdb SauceNAO

Please tell me I did this wrong /sci/

Anonymous Sat 21 Mar 22:00:09 2020 No.11490218 View Reply Original Report

Quoted By: >>11490228 >>11490253 >>11490267 >>11491180 >>11491232 >>11491601 >>11492231

I fitted an S curve to the global death rates from Covid 19 and got pic related. Please tell me I'm a low IQ brainlet who should give up math forever. No way this is true
>tfw 7 million deaths by September 2020

Data source: https://www.worldometers.info/coronavirus/coronavirus-death-toll/
Equation used to curve fit and forecast: a / (1. + exp(-c * (x - d))) + b

Anonymous

Anonymous Sat 21 Mar 2020 22:01:01 No.11490220 Report

Quoted By: >>11490224 >>11491594 >>11492470

Still too conservative. More people will die of coronavirus in 2020 than there are atoms in the universe

Anonymous

Anonymous Sat 21 Mar 2020 22:02:40 No.11490224 Report

Quoted By:

>>11490220
I know you're being sarcastic, but does this confirm that I am a retard with my S curve modelling? Please say yes.

Anonymous

Anonymous Sat 21 Mar 2020 22:03:12 No.11490228 Report

Quoted By: >>11490232 >>11491179

>>11490218
>Please tell me I'm a low IQ brainlet who should give up math forever.
You are. Extrapolating such a small and flawed dateset is meaningless.

Anonymous

Anonymous Sat 21 Mar 2020 22:05:07 No.11490232 Report

Quoted By: >>11490355 >>11491599 >>11492190

>>11490228
>Extrapolating such a small and flawed dateset is meaningless.
Isn't this exactly what governments worldwide are doing right now?

Anonymous

Anonymous Sat 21 Mar 2020 22:14:42 No.11490253 Report

Quoted By: >>11490282

>>11490218
Your fit is not unique. How did you fit it, what errors does your fit blackbox give you for the fit parameters? You need to plot the spread of the curves in order to know how meaningful your fit is. Spoiler: It's not going ot be meaningful, as you're fitting the beginning of a curve with large gradients

Anonymous

Anonymous Sat 21 Mar 2020 22:14:54 No.11490254 Report

Quoted By: >>11490282 >>11490312

your graph is using the data from known cases to extrapolate for actual cases. Known cases might be as low 30% of actual cases.

the real black pill is the cfr has grown 100% in under two months despite unprecedented lockdowns.

we get an average of 500k flu deaths globally per year from a cfr of .1%

So even if the corona cfr stopped increasing now and we only got as many symptomatic cases as we do with the seasonal flu you'd see 21,000,000 dead.
When the reality the cfr is only get much worse as hospitals get overwhelmed which the uk study showed was inevitable without locking down people 2/3rds of the time until nov 2021

It's just not unreasonable to expect 100million dead from this thing

Anonymous

View Same Google iqdb SauceNAO 1584809022333.jpg, 83KiB, 550x543

Anonymous Sat 21 Mar 2020 22:20:44 No.11490267 Report

Quoted By: >>11490273

>>11490218
You're extrapolating to 7,000,000 from <5,000? You're a moran.

Anonymous

Anonymous Sat 21 Mar 2020 22:22:01 No.11490273 Report

Quoted By:

>>11490267
3 orders of magnitude is nothing with exponential growth

Anonymous

View Same Google iqdb SauceNAO corned doggo is scared.jpg, 93KiB, 960x951

Anonymous Sat 21 Mar 2020 22:24:16 No.11490278 Report

Quoted By: >>11490286

Better question for this thread. Will the American healthcare system be able to handle this? Or will there be too many people and too few beds in the hospitals?

Anonymous

View Same Google iqdb SauceNAO errorstuff.png, 20KiB, 440x530

Anonymous Sat 21 Mar 2020 22:25:03 No.11490282 Report

Quoted By: >>11490311 >>11490312

>>11490253
>>11490254
Does pic related make any sense? I just plotted the distribution of my predicted deaths and real deaths. The plot below is just predicted vs real

Anonymous

View Same Google iqdb SauceNAO usaforecastcovid19deaths.png, 22KiB, 559x333

Anonymous Sat 21 Mar 2020 22:27:04 No.11490286 Report

Quoted By: >>11491147

>>11490278
I also did a forecast for USA in pic related. Kinda looks retarded that we'll have 200k deaths in a month desu.

Anonymous

Anonymous Sat 21 Mar 2020 22:36:51 No.11490311 Report

Quoted By: >>11490396

>>11490282
I repeat, you have to assess the quality and uniqueness of your fit first. Do a Chi2 and a maximum likelihood estimation to get the parameter errors.

If you don't know what this means, then you're not fit to throw around "forecasts". But hey, it's 4chan, what did I expect

Anonymous

Anonymous Sat 21 Mar 2020 22:36:58 No.11490312 Report

Quoted By: >>11490340

>>11490282
you already know those two match up to today because you fit the curve by that metric.

To make a prediction we need to know the empirical relationship between actual cases and symptomatic cases which we will never know because we can't measure non symptomatic cases.
If we could get that then you would plot the graph of cfr until the total cases (symptomatic + not symptomatic) was = to the entire population (saturation).
then you would extrapolate the rate of deaths per known cases to the same date as saturation

but like I said >>11490254 very conservative assumptions lead to 21,000,000, reasonable assumptions to 100millino dead.

There isn't much utility in predicting total deaths past are the hospitals gonna get overloaded or not and the answer to that is absolutely.

Anonymous

Anonymous Sat 21 Mar 2020 22:49:17 No.11490340 Report

Quoted By: >>11490348

>>11490312
A 0.5% kill rate disease can't kill 100 million out of 7 billion people

Anonymous

Anonymous Sat 21 Mar 2020 22:55:48 No.11490348 Report

Quoted By: >>11490370

>>11490340
no one knows what the kill rate is.
the case fatality rate (cfr) has grown 100% over the past 1.5 months to 4.2%

there only needs to be 5 times the confirmed cases of the seasonal flu (32% of the population) to get 100million dead at that cfr

Anonymous

Anonymous Sat 21 Mar 2020 22:58:45 No.11490355 Report

Quoted By:

>>11490232
Did the government give you that dumb ass chart?

Anonymous

Anonymous Sat 21 Mar 2020 23:03:24 No.11490370 Report

Quoted By: >>11490376

>>11490348
>no one knows what the kill rate is.
Don't mistake your lack of knowledge for general lack of knowledge

Anonymous

Anonymous Sat 21 Mar 2020 23:05:49 No.11490376 Report

Quoted By: >>11490381 >>11490388

>>11490370
why is /sci full of this kind of hubris

you have no idea what you are talking about and you know it.
we can only measure symptomatic cases the mortality rate requires non symptomatic cases

Anonymous

Anonymous Sat 21 Mar 2020 23:06:53 No.11490381 Report

Quoted By:

>>11490376
It also requires we verify every symptomatic case which we are woefully behind on.

Anonymous

Anonymous Sat 21 Mar 2020 23:08:38 No.11490388 Report

Quoted By: >>11490393

>>11490376
Knowing more than you isn't hubris, if anything it's hubris to assume everyone who knows more than you is wrong

Anonymous

Anonymous Sat 21 Mar 2020 23:10:28 No.11490393 Report

Quoted By: >>11490399

>>11490388
source the 0.5% mortality rate then you brainlet

Anonymous

Anonymous Sat 21 Mar 2020 23:11:33 No.11490396 Report

Quoted By: >>11490685

>>11490311
Ok I did what you asked for. Took the square root of the diagonal elements of the covariance matrix (reduced chi squared) to get an estimate of the standard deviation of the fit parameters

a = 6758438.832 +- 961947400.960
b = 333.288 +- 196.683
c = 0.071 +- 0.010
d = 149.852 +- 2031.820

Anonymous

Anonymous Sat 21 Mar 2020 23:13:29 No.11490399 Report

Quoted By: >>11490403 >>11491182

>>11490393
It's the maximum estimate assumption, a reasonable assumption would be 0.2%

Anonymous

Anonymous Sat 21 Mar 2020 23:14:26 No.11490403 Report

Quoted By: >>11490409

>>11490399
that is not a source, you're just saying it

Anonymous

Anonymous Sat 21 Mar 2020 23:16:14 No.11490409 Report

Quoted By: >>11490412 >>11491182

>>11490403
And yet that's how it is

Anonymous

Anonymous Sat 21 Mar 2020 23:17:06 No.11490412 Report

Quoted By: >>11490414

>>11490409
confirmed retard

Anonymous

Anonymous Sat 21 Mar 2020 23:17:32 No.11490414 Report

Quoted By:

>>11490412
based retard

Anonymous

Anonymous Sun 22 Mar 2020 01:00:22 No.11490685 Report

Quoted By:

>>11490396
love it

Anonymous

Anonymous Sun 22 Mar 2020 01:24:21 No.11490736 Report

Quoted By:

>naively applying a generic function to a scarce amount of initial data
please don't do this

Anonymous

Anonymous Sun 22 Mar 2020 04:31:12 No.11491147 Report

Quoted By:

>>11490286
Why do you say that? All it would take is ~10% of seniors to get it and for 5% of those to die.

Anonymous

Anonymous Sun 22 Mar 2020 04:48:49 No.11491179 Report

Quoted By:

>>11490228
/thread

Anonymous

Anonymous Sun 22 Mar 2020 04:49:33 No.11491180 Report

Quoted By:

>>11490218
You did it wrong. You arrogantly assumed you know all the variables.

Anonymous

Anonymous Sun 22 Mar 2020 04:50:49 No.11491182 Report

Quoted By:

>>11490399
>>11490409
>can't provide a single source
>thats just how it is
Seriously grow the fuck up

Anonymous

Anonymous Sun 22 Mar 2020 05:10:14 No.11491232 Report

Quoted By:

>>11490218
7.8 bn * 2/3 * 0.02 = 104 million

Anonymous

Anonymous Sun 22 Mar 2020 09:12:29 No.11491594 Report

Quoted By: >>11492470

>>11490220
There's reason to believe this virus even infects atoms.

Anonymous

Anonymous Sun 22 Mar 2020 09:16:20 No.11491599 Report

Quoted By:

>>11490232
yes, but they are low IQ brainlets who should give up life forever

Anonymous

Anonymous Sun 22 Mar 2020 09:18:58 No.11491601 Report

Quoted By:

>>11490218
you are a low iq brainlet. but math is just perfect for those kind, just learn to suck academic dick well and agree with what you're presented regardless if it makes sense or not. try to develop photographic memory too then you're perfect.

Anonymous

Anonymous Sun 22 Mar 2020 15:39:40 No.11492190 Report

Quoted By:

>>11490232
Yes, but they actually know how to model this type of data

Anonymous

Anonymous Sun 22 Mar 2020 15:59:25 No.11492231 Report

Quoted By:

>>11490218
You're model only takes into account unchecked growth assuming theres no mitigation. Don't get me wrong it's still going to be absolutely fucking dreadful, but more on the order of say 1-200 million dreadful than billions.

Anonymous

View Same Google iqdb SauceNAO Kazuma.jpg, 51KiB, 500x500

Anonymous Sun 22 Mar 2020 18:09:07 No.11492470 Report

Quoted By:

>>11490220
>>11491594
nice

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

Your latest searches