I’m genuinely confused as to what you mean when you say this.
But, yes, you have some evidence! It indicates, very obviously, that something is suspicious. And that’s a good thing.
You looked at the numbers a bit, saw Terran MMR averages low and went to saying Terran is the hardest race! … and just kind of stopped? Instead of continuing to look at it and break down the data, and the only reason I can see to do that is because it’s what you wanted them to say?
I agree the conclusion is a logical one from that data. I said pretty similarly in my first long post of a previous rendition of this thread –
And I’ll polish up an incomplete reply I composed there -
The thing is - and the reason I care so much - is that that actually is a good basis of an argument.
We have a game with three races. A naive prediction says we should see vague thirds as the representation for each race. This ends up not being the case, by a significant margin.
Given that, from my understanding of player profiles, we should then see one of two things:
First, that despite the odd global representation, the number of experts of each race is closer to a third, so the most populated race thins out as you climb the ladder percentage-wise;
or second, that the distribution of player skill is even, that the number of players of a race at a rank is roughly equal to the population percentages.
And neither of these are true at all. The number of Terran players cramped into the leagues below Diamond is tremendously out of line with either of these naive expectations; nor are they a ‘fair’ percentage of masters+ given how much of the population they are.
And that’s wacky! That’s indicative of something strange. But you then didn’t do any further leg work, because you had a conclusion in mind and not a hypothesis.
You wanted the data you acquired to affirm your pre-existing belief and the moment that you saw that it did, you stopped.
Now, I will not pretend for a moment, that once I found the data said what I wanted I didn’t lose significant motivation to continue - but I didn’t stop. I kept looking at those numbers and doing math on them to see if anything didn’t line up further. Nothing I came up in the next hour or two of looking at it was maligned with that conclusion.
And you didn’t pay attention to it, because this is not the thing that I pointed out that it says. Arguably, the most important part of what it says in relation to this.
The whole of the stats can indicate one thing and then the subsets can indicate something directly contradictory.
Data can indicate multiple things and those things can even be contradictory. What matters is how you use those indications and control for variance, making sure that the way you control it isn’t just a way to reconfirm your bias.