Some doctor dropped a video online that says 1 egg per week (whole egg) can prevent dementia by like 47%. I guess I need to eat more eggs. Doh!

What I don't get it what the 47% actually means. Is this a chance of a person getting something in a population? So lets say the chance of dementia is 10%. Does that mean eating an egg will reduce this chance to 6.3?

As far as I understand, yes

eggs got sulfur

sulfur is a healing agent, like how methyl blue (sulfur) cures autism

eggs were demonized to increase cholesterol but because they increase natural anti oxidants and your cholesterol goes up cuz you're vitamin d deficient with all that cancer causing sunscreen and everybody not being allowed outside when it's sunny your body keeps making cholesterol because cholesterol deposited under your skin is what gets converted into vitamin d... and instead you make tons of it, never use it, and eat a bunch of toxic shit that over time oxidizes the cholesterol and now suddenly cholesterol causes X and Y issues when it doesn't

I been eating 3 organic eggs a day since new yeaarrrss

colostrum (which is a higher dose of enzymes from milk they cooked out of that raw milk) seems to be doing more for my immune system though but sulfur is pretty good also

I stopped eating colostrum and got dementia again, mania in the night. took it again last night woke up being able to think again

Statistics are tricky.

Without any further info, what it means is that, given a sample population, of which an unknown percent developed dementia, we establish a prior hypothesis, whether everyone, dementia or not, eats an egg a week or more.

When calculating the odds of developing dementia taking into account this prior, you apply Bayes' theorem and get the compound probability, which seems like it comes to half of the value without the prior.

But that's the tricky part, yes, you may be half as likely to develop dementia, but you already had a slim chance to begin with.

I ignore such stats 100%. He can't know that. It's that easy.

Well too late for anyone who follows that dogshit bc anyone who does already has a brain going to shit

There's actually a really good example of this very phenomenon
(If anyone else knows statistics, yes, it's not the same but I'm trying to illustrate)

When dealing with a binary classification problem, just saying if something x belongs to a class a or b, seems simple enough, right?

Well, it's not about whether x belongs to a or b, it's really four scenarios.

You thought x was A, and it was, or wasn't A.

Same for B.

Naive reasoning would say they are complementary. They are not.

The two metrics for statistical accuracy are, stupidly, accuracy and recall.

You can't get perfect in both. One measures how many guesses turned out right, and the other, how many guesses actually matched the right result, which is surprisingly tricky.

Say. You are a developer at Tesla's autonomous cars division.

Your car comes to a point where it must decide whereas there's a person in front or not, but it's not sure.

What do you do?

Exactly, you stop the car at the bare minimum suspicion that it might be a person.

You can not allow a false negative. That means your company going bankrupt.

False positives? Sure. Might be an inconvenience, but no harm done.

Tesla wants recall, not accuracy.

And in any case, I shall simplify.

Tesla doesn't care if every time the car stops it's because there was a person in front. That's accuracy

Tesla cares about the car not stopping coz it didn't think it was a person.

That's recall. It matters. More than most of you think.