This is so true. While I don't believe there is an accurate mathematical model which accurately estimates handgun "power" nor predicts effectiveness, due to the wide variations in placement, physical characteristics and temperament or attitude (of the subject being shot) there are some generic stabs that work out very roughly:
They are not limited to (I know of more than 30 "scales" or "models" to calculate power) but include The Hatcher Scale, Cooper Short form or Taylor (Chuck) RSP. Another possibility, though not invented for pistols, would be the Taylor (John) Knock Out - or even the "power factor" (which is relative momentum) might do, but it disregards cross sectional area which is a mistake, I think.
Now where you draw the line is also problematic and is affected by "who you are shooting" - A Cole Younger (shot 11 times - some with 45-70 and .58 cal M/L), Emmet Dalton (shot 21+ times), Michael Platt (shot 12 times) or a Richard Blackburn (who took five .357s center mass) is harder to stop than a Mark Coates (who succumbed - slowly - to single .22 fired from a mini-revolver) and there-in lies the rub - it is more about the man than the gun. It also depends on where those were placed - Younger, Dalton and Blackburn survived their encounters.
Still, while experience looking into this tells me it is highly variable, I have toyed with the idea of where I draw the line. For me, it is somewhere around 60 on the Hatcher Scale (or 30 on the Cooper Short Form).
Some, notably George Bredsten, believe expansion is a factor and calculates Hatcher (which he considered the best of over 30 mathematical models) using the expanded diameter. I don't really disagree, but that begs the question just how to you establish how much a bullet expands - it gets back to "who your are shooting" (is he "hard" or "soft") and where you hit (do you hit bone going in - if you are a good shot, there is a 90% chance you hit bone with not much soft tissue over it). In my experience, modern expanding pistol bullet don't (expand) if they hit something hard right after they pass through clothing and skin - at least not until you get past 1800 fps at the impact range).
That is actually more questions than I can answer so I don't mention it a whole lot.
Apologies in advance if you've already seen this, umm, speech from me in another thread or subforum.
I am pretty much on the same page with you in all that you said.
I diverge a little bit when it comes to the subject of hollow point bullets, and the question of how much a bullet expands.
Without laying out a bibliography, I'll just say that I feel like even the best modern hollow points are not sufficiently reliable. They get plugged and don't expand. They overexpand and become basic slugs. They're going too slow and they don't expand or don't expand fully. Lots of things can factor in that mean they simply don't work as intended, in real-world scenarios.
But, nowadays, we have available the Lehigh Defense Xtreme Defender bullets (and maybe other similar bullets, I haven't researched them). They are solid copper. They don't expand. But, thanks to CNC machining, solid engineerign, and fluid dynamics, in animal tissue (and ballistic gel), they produce a permanent wound channel that is equal in diameter to modern hollow point rounds (of the same caliber) that "worked". And, because the XD rounds don't expand, they penetrate much further, creating a permanent wound cavity (i.e. total volume of wound) that is easily double what a "working" hollow point produces.
To ME, these bullets address many of the questions you brought up and make them moot.
Immediate incapacitation pretty much requires hitting a major portion of the central nervous system - i.e. the upper spine or brain. Near-immediate requires hitting a major organ, like the heart. I think even those worst-case scenarios would have been people for whom hitting the brain or upper spine would have shut them down right away. But, unless the target is facing away from you, that CAN be difficult. There can be a lot of mass between the initial impact and the upper spine. Hollow points simply may not penetrate far enough. That is why the FBI concluded in their 1989 white paper on handgun wounding characteristics that penetration was, by FAR, the most important characteristic (after placement, of course) in the ability of a handgun round to achieve immediate incapacitation. And that is why my EDC is loaded with XD rounds and not hollow points.
I think 90gr of 9mm, at 1550fps, is able to penetrate any real-world self-defense target sufficiently well, that I don't need a bigger caliber, or a faster muzzle velocity. But, I also don't want a smaller caliber - like 30 SC. I don't think cross-sectional area is as important as you seem to, but I do think it does have some importance. Or maybe we're really on the same page there.
38 Super is really my ideal carry round. I carry the same bullets, but in 38 Super +P they are 1640 fps. But, 9mm is close enough that I am willing to lose 90 fps in exchange for the extra capacity I get in a modern double stack, versus the 38 Super single stack that I have. If somebody made a micro compact double stack in 38 Super, I'd be all over it. But I'm not aware of such a beast.