What your analysis does is it provides us with a baseline for discussion. It establishes the inherent mathematical outcomes within this edition, assuming all other things being equal. What it doesn't do, however, is put a period on the discussion of survivability. In order to examine this idea of survivability we need to build upon the math you provided and examine how human behavior magnifies or mitigates the mathematical disparity. And, unfortunately, this is a very difficult discussion to have as all experiences are anecdotal.
This line assumes human behavior will be identical (or at least similar enough to not affect the outcome). However, many of us have noticed within our games that the mere act of changing the initial ability scores will affect a myriad of decisions made throughout character creation and campaign play. I think this is where some of us are getting hung up on with your algorithm. Your algorithm doesn't match our experiences. So the question ought to be: What is different in our games?
Within my own game, I noticed two things: First, players tend to place higher ability scores on lower tiered characters. Second, players tend to play higher ability score characters more recklessly than lower ability score characters. The combination of those two decisions I believe is mitigating a portion of the inherent mathematical differences between any two sets of stats.
It is also important to consider that my players have a character stable and adventure within a sandbox environment. This may influence why my players choose to build characters the way they do. A campaign that runs a more traditional 4/5 person party through a linear adventure may have wildly different experiences.
Although I do have a road trip coming up and in theory my wife will be driving half the time ... 
)
