There is. And you actually see that its a pretty useful thing: By your way to only look at 100% of the population you could perceive imbalance (or just skewed distribution). By further analysing the data (splitting by mmr) you can actually see that there is no difference between the races.
They do create labels. They do categorize their data. They dont look at 1 whole group and say thats it.
“A label is a category into which a record falls, usually in the context of predictive modeling. Label, class and category are different names for discrete values of a target (outcome) variable. “Label” typically has the added connotation that the label is something applied by a human to model-training data, so that a predictive modeling method can learn to assign labels to similar unlabeled data. For example, a paralegal might label a sample of documents as “relevant” and “not-relevant” so that a machine learning algorithm might learn from the sample, and apply the same labels to other, unlabeled data.”
But they DO split people by BMI. Age…gender…field of work, ethnicity, nation, etc. etc. etc. They do create labels or categories and split them. They dont say: a human has on average a bmi of xyz. That would be dumb and of limited use. They create labels and catagories to further analyse the data.
So yeah its a pretty important observation that for example office workers have on average a higher bmi than construction workers even tho that bmi is also influenced by higher muscle mass. Thats also a very notable thing. A high bmi doesnt mean that the person is fat.
You desperatly trying to make splitting into mmr sections (bottom 50% and top 50%) look like its a dumb thing to do. Which it isnt. The actual dumb thing to do is only to look at 1 side of the story especially if the distribution of the population is that lopsided.
Dont know if its usually, but if you say so…meaning its not a no no thing to do it nevertheless.
Yeah. Best we can do is saying that the avg mmr is not a valid argument because the statements we would draw out of them would contradict each other.
So yeah we dont get any definite anwer out of that.
I really like the height argument of yours: men are on avg taller than women. This can be seen if we split them into 2 parts (smaller humans and taller humans) and just the whole population. In both cases we get the same result: men are taller than women.
In this situation with race and mmr we dont get the same result.
Nope, you would still get the same result if there is an actual (factual) difference. And ofc given your sample size is big enough/the quality of data is good enough.
There’s an actual, factual difference. You said it yourself. The population sizes. That’s the exact difference is that the population sized are distributed along the MMR lines unevenly.
Please cite a single study where they split populations along the metric they were measuring in order to prove or disprove the impact of said metric.
Yeah And why would you conclude anything Out of that ? Please prove your assumption that you can conclude anything meaningful Out of that If the distribution is Not random.
Assuming Something Out of the blue without any Proof is Something borderline moronic. If we dont know anything about the Population of sc2 why would we Care about their distribution ? It Just doesnt make any Sense
Oh, they practice more? Well practice doesn’t have anything to do with skill (yes, you actually made me show that GM players played more than Silver)
Oh, well, the ratio of games played per league might be high but the total number is negligible
If Terran players had been found to play less games they’d be more casual, but the fact that they play more games means that they’re more casual
If I split the players by MMR, they have the same MMR
Tell me that if you were watching this as a debate about anything (let’s say, the wage gap) you wouldn’t be laughing at a guy trying to make these arguments
I have the average MMR data, and Terran is lower by a mile.
What he did was split them into two groups, based on MMR, found that the two groups were similar (because… they were split by MMR, lol) and decided that means something…
But to give you an obvious answer: the balance council/ the balance team said often times they balance according to top pro level. Doesnt matter if its right or wrong. They simply do that.
You however look at something that is not randomly distributed. In fact it is heavily biased and conclude something out of it.
It means something. Like your pretty illogical height example that would exactly show what i mean: If you look at height between men and women based on population you would see men are taller than women but you will also see that if you split them into 2 groups of smaller or taller people (because women are more likely to be smaller than men, they will more often be placed on lower spectrum of said group, meaning they would drag down the average, hence you will see men are taller than women)
And actually you mentioned the bmi too, which actually also shows what i mean: you cant conclude that much when you look at something in 1 dimension. 2 people with the same bmi dont need to have the same bodyfat meaning the bmi isnt even a very good indicator in itself. also, there is a height bias and the bmi for taller people is calculated differently. they rather go for a more cubic approach meaning its weight / height^2.5 and not weight / height^2. Meaning yeah its pretty good to split people by bmi and further look at them to point out possible flaws of used method.
All in all i can say: You are looking at things at a very very very naive and simplistic approach and you dont get the obvious mistakes that occur with it.
You only look 1 dimensional and only your view is correct. You view your opinion as the absolute state of truth.
Looking at a biased race selection and conclude anything out of it because bronze and silver players just dont advance and therefore conclude that its because of the difficulty is simply nuts. but you do you i guess.