A Question on the Meaning of Win-rates

So recently I've heard a lot of different opinions on what win-rates mean. In my opinion, they are a good way to analyze relative champion strengths, and how well you can expect to see said champion perform against you on average in your games.

But I've heard a LOT of different interpretations as well. Here are some.

1. Champion win-rates are "artificially deflated" by high pick/ban-rates (I don't know why high ban-rates means anything). If a champion is picked more, than it's more likely that there will be MORE BAD players playing that champion. Since there will be so many bad players playing the champion, the win-rate of the champion will drop.

This interpretation assumes not only that the majority of players are bad (moreover, even assumes relatively bad compared to their own elo? lmao) but also stupid. It assumes that people who have little success with a champion will continue to play that champion even though they consistently lose on that champion. In ranked. I will play the champion I don't know how to play IN RANKED because "that's what everyone else is doing" or because "I did it before and I mostly lose, so yolo, I'll do it again." This seems pretty fallacious to me.

2. Champions that are near perma-ban status have their win-rates "artificially inflated" because in cases where they are picked, people are more likely to dodge than play against them. Only when the team has very high hopes in their own composition will they actually go forward with the game. Thus, perma-ban champions would win more games, but those loses are "pushed-out" because of dodges.

Again, everybody knows that compositions mean next nothing in more than half your soloQ games. Lane-phase can go whatever way, even if its caitlyn vs. vayne. Shit happens in soloQ than nobody would suspect. If anything, the "artificial deflation" argument should hold here INSTEAD because champions that are near perma-ban have this aura of "OPness" around them that attract people to pick that champion, even if they have no idea how to play it. Again, seems pretty fallacious either way.

3. Champions with low pick-rates have their win-rates "artificially inflated" because only people who main those champions play them. Thus, they are more likely to win because they know the champion very well.

Okay, now I have A LOT of problems with all of these interpretations. Before we get back to circle-jerking which one is actually the correct interpretation, let's take a step back and realize that they are ALL ASSUMPTIONS. It seems to me like somebody arbitrarily got up one-day and said "Hey, I have a great way to explain win-rates that SOUNDS REALLY GOOD. It's probably right." And everybody agreed with them.

Has anybody actually gone out and done the statistical analysis to determine which (if any) of these interpretations are correct? Has Riot ever responded to this? Because if not, then all 3 interpretations seem like nothing more than assumptions with no basis in fact. It's seems pretty ridiculous that we can generalize things like "more bad players pick X champion in ranked than good players because their pick-rate is high."

Can we just take win-rates as they are instead of putting these unnecessary and arbitrary interpretations on them?