There are most definitely secondary and tertiary effects that would be nigh impossible to layer into the actual randomization process itself which is why I mostly ignored those.

To some extent, you’re right — if a player is strong on a certain champion, it does not mean everyone will be, but if we’re making a statistical argument about an entire population and not edge cases, then overall win rates for champions are a decent estimator of the true strength. While there are other variables that would further improve this system, I am not convinced that weighing on wins alone would yield a less fair environment. The statistics don’t check out, but would love to see a more detailed proposal of your thoughts!

Or maybe we can just hope that someone who works or worked for Riot will let us know!