Even when you're shown an example where popularity does not make as much of an impact, you come up with the excuse that it was only because "popularity was very close".
... No, that's not an excuse, that's an objective statement of what occurred. I don't know any other way to state it. It was very close because.... it was very close. It didn't make an impact because the numbers were too close to make an impact. I don't know what else you expect to happen
Rollercoasters won, but they won fairly. A narrow margin, but still fairly
the is a higher chance that the most popular team would get the most wins anyway.
No, there isn't. Not at all. The game mechanics and mathematics do not work that way:
1: Only matches between two different teams are counted. Matches between same team are assigned for fast match making, but they're not counted towards the team's win rate. Thus, each team has the same number of rounds counted in their score. You would be right if this mechanic wasn't in place, though. I thought this at first as well until I learned about this mechanic
2. Since there is no correlation between skill and preference over X or Y similar thing, each team has the same average skill level, and thus, their win rates will not deviate further than a few percentage points of 50/50. Again, see law of large numbers.
There's also the issue of people deliberately stacking the teams by all joining the most popular team, making the problem worse. I've already seen people trying to do it, just not at a large enough scale to make a difference
