Fix MMR range for Toss and Zerg

[Eliwan-1354]

@llllllllllll-12402

No body does this, lol. Find one study that stratifies data like this and then uses them to apply individual causes to an entire population.

I did this, I linked you to me doing this, embedded the quote of me explaining it, and explained it another unique time.

This is literally what Simpson’s Paradox was about in the first place: When you look at the whole of the data it says one thing but if you look at the subset it says different things; and the entire point here and in the wikipedia article is that it is that different thing was the actual reality that was reflected by the suspicious data observed initially.

en.wikipedia.org/wiki/Simpson's_Paradox

Simpson’s paradox is a phenomenon in probability and statistics in which a trend appears in several groups of data but disappears or reverses when the groups are combined.

Read and understand the article. This is the first sentence in the damn thing. Do you see this language it uses?

Did you see the language I’ve been using explaining how when a trend line disappears or emerges; and how that reflects truth value of that trend?

@Eliwan-1354

@llllllllllll-12402

Now, please. Find one study where any reputable scientist was studying a disparity in a variable, then decided to split the population by said variable in order to draw on conclusions.

this hurts to read, you know

but fortunately i did this already, and in one of these threads; way back in July:

@Eliwan-1354

For a practical example, Simpson’s Paradox. To summarize it, in 1973, the whole pool of UC Berkeley applicants - if divided only by gender - said that men were more likely to be accepted than women.

But if you divided by the applied department, most of the departments had a higher application acceptance rate for women.

The gap here came about because a huge number of women applied for locations with limited seats - for example, no matter how many applicants the college gets, it can only accept, say, 500 applicants in each department.

If you can take 500 applicants whether you get 1000 or 10,000; then the huge number of applicants that are rejected washes away a lot of the other data when you look at the data at a different angle.

The example you used about male and female heights is a good example as well, since in that example we would not observe the dramatic difference.

Get off the horse.