^ I think I have, but I'll expand on it if you want.
It's a bit unclear what you mean when you say "QB rating". The initial context made it seem like you were using it to mean passer rating, but your last comment makes me think it could have meant any type of QB rating (hence your use of plural). I'm going to keep assuming passer rating for the time being.
That said...
I don't consider passer rating to be a very good measure of quarterback ability, but a decent measure of the overall passing unit's performance. Even ignoring the tremendous gulf between ability and performance, passer rating is influenced by any number of issues that are not related to the quarterback. I've mentioned the obvious ones (receivers, blocking, and play-calling), but even defense plays a role as well, since their performance affects play-calling (if your defense can't stop the offense, then you need to take more chances on offense to keep up).
If passer rating is so closely tied to winning, with the causative direction you imply and the quarterback as the sole provenance, then it's remarkable that new coaches in the league improve, on average, more than 1 win a season over the first few years, even after removing all the coaches who don't last at least 5 years. This would, using your claimed causal direction and magnitude, imply that quarterbacks see a significant improvement by having the same coach for a number of years. This is something Tannehill, he of 4 different offensive coordinators in 5 years. hasn't had. And, on topic, something that isn't described by passer rating.
And so on and so forth for any of the other innumerable factors that affect passer ratings and wins.
The bolded portion is what I'll focus on here. When you take a group of the league's best QBs and compare them to a group of the league's mediocre QBs, over five-plus seasons, the between-group variation is significantly greater than the within-group variation in QB rating (yes, passer rating).
What does that tell you?
What it tells me is that the surrounding variables you mentioned (coaching, offensive lines, receivers, defenses, etc.) likely account for the season-to-season variation in QB ratings in both groups, but it isn't enough to make the within-group variation equal to or greater than the between-group variation.
What do we have left to explain the significantly greater between-group variation?
Author