how to lie with correlations: harley davidson limited edition.

there has been a lot of aggressive talk about how sturgis was a massive super spreader event

charts like this make it tempting to agree.

but be very careful about doing so: there's a big issue with this graph Image
it's from 5000 miles away and across an ocean.

as one can rapidly see "cases" measured here are in eastern europe, not the american midwest.

hard to imagine the harley folk having caused this no matter how hard they partied. Image
interestingly enough, south dakota and czechia look incredibly similar in terms of disease curve (deaths per million population) despite having totally different responses.

(the CZE data is real day of death, so it lags. the last 10-14 days are likely incomplete) Image
their outcome is so similar you'd be tempted to say they must share a biome, but clearly, they don't.

what they do share is an approximate latitude and climate.

prague has avg temps for nov of 34-43
sioux falls: 23-42
covid seems to come in regional waves and those waves cluster by latitude. the april surge in the US was mostly northeast to IL.

the summer surge was southern.

we're now getting mid and mountain west, esp in the north.
and the EU is seeing the same. eastern europe did not get his in the spring. they have been getting their surge now.

masks, lockdowns, distancing all look irrelevant.

croatia has done none. they have the same curve as all the neighbors.

correlation is not causality. when your covid season hits, it hits.

ask peru and argentina who everyone thought were NPI success stories and now lead in deaths per capita.

we keep trying to claim that region first surges are some sort of second wave and it's misleading.
a lot of very shoddy science is being used to underwrite a lot of very costly and thoroughly baseless policy.

so, be very careful about what you accept as "evidence."

if a picture speaks 1000 words, imagine how many lies a graph can tell.

food for thought. Image

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with el gato malo

el gato malo Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @boriquagato

23 Nov
@PiranhaCapital @real_MikeBarnes @GCA_Worldwide there is lots or hard evidence for biome specific immunity.

SARS-1 never spread to the US. neither did MERS.

this study of pre-covid 19 blood samples shows 10-15X the prevalence of sars-2 antibodies in africa vs US

@PiranhaCapital @real_MikeBarnes @GCA_Worldwide regional virus exposure is not that uncommon and similarity to sars-2 does not necessarily imply contagion like sars-2.

you could spend years generation localized cross resistance from endemic or source based viruses with low R.

then you get a breakout variant like sars-2
@PiranhaCapital @real_MikeBarnes @GCA_Worldwide the idea that lockdowns work looks pretty fraught. every standing set of pandemic guidelines from 2020 said they do not and no contrary evidence has emerged since.

asia has varies responses, all got same results.

no way this was just NPI.

Read 4 tweets
23 Nov
there seems to be a lot of grandstanding by the lockdown governors (who are almost invariably democrats as this issue has become distressingly partisan) but when one looks at red state vs blue state outcomes, it raises some serious questions.

(graph from david steinmier, PhD) Image
the blue states are faring noticeably less well than the red.

this has been true all along.

interestingly, red states are noticeably fatter which is a significant cov risk enhancer (and links to others like diabetes)

so, ceteris paribus, you'd expect more red state deaths. Image
the rejoinder to this has always been "but density!" but this has been a surprisingly poor predictor of death rates globally.

most of the data is highly uncorrelated with a few outliers driving outcomes. county level is a little better, but not great. globally, it's worse. Image
Read 11 tweets
21 Nov
anyone still presuming that a + PCR test is showing a covid case needs to read this v carefully:

even 25 cycles of amplification, 70% of "positives" are not "cases." virus cannot be cultured. it's dead.

by 35: 97% non-clinical.

the US runs at 40, 32X the amplification of 35. Image
a lot of people still seem to not understand what this means, so let's lay that out for a minute.

PCR tests look for RNA. there is too little in your swab. so they amplify it using a primer based heating and annealing process.

each cycle of this process doubles the material. Image
the US (and much of the world) is using a 40 Ct (cycle threshold). so, 40 doublings, 1 trillion X amplification.

this is absurdly high.

the way that we know this is by running this test, seeing the Ct to find the RNA, and then using the same sample to try to culture virus.
Read 19 tweets
21 Nov
several months back, i posited that the low covid deaths in pac rim could not possibly be from lockdowns or masks.

the differential was simply too large, the policies across the region too varied, and the results to internally similar.

it has to be pre-existing resistance.
i'd like to now revisit this hypothesis as quite a bit of new evidence has emerged and i think it has been increasingly supportive of this idea.

the gaps are simply too large and too geographically consistent.

i am becoming convinced that this is the only plausible explanation.
i'm going to use deaths as a metric because testing in asia has been very low in many places (1-10% of western per capita levels) and trying to adjust for that makes gibberish out of the data.
Read 29 tweets
20 Nov
what's most disappointing about the pics of lockdown gavin standing around maskless at french laundry w/ his pals is not that he'll doubtless find some epidemiologist to say "it's safe because proles can't get reservations anyhow"

it's that no one can even muster outrage anymore
stop and think about that.

does this even surprise you one, teensy eensy little bit?

complete moral bankruptcy and frank, forthright hypocrisy have become the base case for american politicians.

it's a noise so deafening we cannot even hear it anymore.
you only notice the rare cases when it stops for a minute.

how many stories of "pulled own mom out of NH then filled it with covid patients" or "imposed lockdowns but kept work going on their lake house" or "banned salons but went anyway" or have you heard?
Read 6 tweets
19 Nov
i see we have reached the "aztek priest" stage of the mask superstition

demand sacrifice!

if the rains do not come, demand a bigger sacrifice!

if the rains do not come, bigger still!

eventually, it will rain.

claim you caused it and demand a promotion…
this is literally what epidemiology and public health has become.

just keep using bigger hammer theory and pretend that lockdowns and masks do anything at all to stop covid (they don't) and wait for the natural gompertz curve of community resistance/seasonality to bail you out
ignore the fact that lockdowns and masks are probably making it worse and that there is anything resembling a "cost" side to this cost-benefit equation and that it's massive to the point of "great leap forward" bad.

speak only in moral absolutes.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!