You right. I guess my hatred of New England clouded my view of them retroactively back when I was continually telling myself that Brady was more lucky than good.
Regardless, I still don't understand how needing an offense factors into a secondary being effective or not. Sure, if you're behind big and the other team stops having to throw your secondary is pretty pointless, but how is that any different than saying being able stop the run is useless at the beginning of a game because the other team can still pass a lot and can just ignore the run or a pass rush is worthless when the opposing team is up big and doesn't have to pass anymore as well? If your secondary is really good, it makes it so that the other team does not get ahead big from passing in the first place which keeps the game closer to your "lesser" offense.
Analytics is analytics and that's not some interpretation of events. Teams with better secondaries on average win more than teams with better pass rush. That's objectively factual. How you (not you, but previous posters) decide to tie offense into that turns it into a subjective argument at that point.
I think you have a misconception about analytics.
It isn't nearly as straightforward as you state.
"Having a passrush" or not, is way too much of a simplification. It's not an either or proposition. Even if you rate it on a scale of 1 to 10, it still becomes subjective, and that doesn't even to take into account strength of the opponent, where those opponents strengths lie, score differential, opponents QB efficiency, etc.
While analytics have an extremely valuable place, and are great for generalities, specific situations, not so much.
I'm not argueing your premise either way, just pointing out that once you go deeper than directly comparable statistical "facts", analytics become, to one extent or another, nebulous and subjective.
Statisticians can look at the exact same data set, and arrive at far different conclusions.
If it were just a matter of mathematicians crunching numbers, and getting the answer that 2+2=4, that would not be the case. There are entirely too many variables to make blanket statements.
Now, if you told me the raw data showed a 4 game difference, thats a lot more conclusive. 1 game is only a 6% deviation. Not conclusive enough to make it fact, given the high number of variables across 32 teams, IMO.