Walt,
You write, "you can be 95% confident that 95% of your shots will penetrate the gel at least 10.7 inches. In short, the IWBA is saying 10.7 inches in gel should be enough to get the job done on a human body." You may be right, but I'd argue that getting the job done on the human body may not be that simple, especially if you want THAT human body to stop shooting at you. Most of these analyses seem to be based on best case scenarios.
I apologize for the colloquialism "get the job done." By that I meant penetrate deep enough to be highly likely to reach vital tissue in a human target regardless of shot geometry. Whether or not the adequately penetrating bullet hits vital tissue, of course, depends primarily on the shot placement, which is almost entirely a function of the shooter's skill, but that's an aspect of handgun wounding effectiveness that has little to do with selecting effective ammunition (although a year and a half ago the FBI acknowledged that people shoot 9 Luger more accurately than .40 S&W).
The other factor that can affect bullet trajectory inside a target is deflection from hitting hard tissue. .22 LR rounds are notorious for bouncing around like a pinball after striking bone; thus, a well placed shot aimed at a bad guy's heart could be deflected by bone and miss vital tissue. Deflection is less likely with a bigger bullet. I can't recall any terminal ballistics guru focusing on this, but the potential for deflection (and poor shot placement) is a good reason to keep pulling the trigger until the threat is neutralized.
On the other hand, I've seen very little attention paid to how accurately some of these rounds can be used -- which addresses more than just the innate accuracy of the round, but also evaluates how well a shooter can fire the round). What role does the gun actually play in the results, or barrel length. Ballistics performance alone addresses part of that question, but most of that is measured from test barrels.
The IWBA's ammo spec boilerplate (I wish I could find the link to it) is intended to have each department conduct penetration testing to find an optimal round for the specific gun carried. Thus, a backup gun, even if the same caliber as the duty gun, might function from a terminal ballistics perspective with a different round.
I also think many shooters practice with one round but carry another. They'll shoot a box or two of the "good" stuff at the range to be sure it functions well in their weapon, but don't really stringently evaluate their accuracy with their carry ammo. (Guys who role their own probably do a much better job of evaluating their carry loads.) If you have to use more rounds because you shoot it less well, how does that affect the choice of ammo?
As TunnelRat says above, the key to cost-effective range shooting (for those of us who don't roll our own) is to find a FMJ round that has equivalent exterior ballistics of your chosen self-defense round. For example, I like to carry Federal Tactical 147-gr HST in my 75 Compact, because it's a good penetrator and expands reliably (the IWBA recommends choosing heavy for caliber for good penetration). But, the Federal American Eagle 147-gr Flat Nose FMJ has the same exterior ballistics as the HST for a fraction of the price.
The IWBA was concerned only with terminal ballistics. General exterior ballistics is something one typically considers before selecting the pistol caliber.
I'd prefer 'real-world results' if they were available, but real-world results are almost impossible to properly evaluate: 1) there aren't THAT many shootings that are properly assessed/measured, 2) the specific loads used are seldom reported, 3) the scenarios in which they occur aren't always known or reported [time of day, distance between parties, indoors or outdoors, weather, if out of doors, etc.], and 4) the proficiency and experience of the shooter is almost never recorded (or known).
That last factor may be the most critical one in any of these evaluations -- and it may be far more important than the gun or the loads used. We just don't know.
You obviously know that the validity of street data is a myth. The IWBA's terminal standardized ballistics testing protocol allows for meaningful comparisons between rounds.
Martin Fackler formed and headed the now-dissolved IWBA, and he was the father of scientific terminal ballistics science. He was a Colonel in the US Army Medical Corps, a battlefield surgeon in Vietnam, and the director of the Army's Wound Ballistics Lab. He standardized terminal ballistics testing, identified the overwhelming importance of penetration, and identified the importance of fragmentation in military rifle calibers. He was a key participant in the FBI's Quantico workshop on handgun wounding effectiveness. After the FBI adopted the IWBA criteria with minor modifications and elaborations in regard to barrier penetration, Fackler not long afterwards dissolved the IWBA, mission accomplished.
But Fackler et al. were focused on terminal ballistics, and for a handgun penetration über alles, with reliable expansion being important, too, and a big hole being better than a small hole.
Exterior ballistics is a separate issue, but nowhere near as important as terminal ballistics for self-defense distances.
Round accuracy, meaning finding a round that is rather optimally tuned to the pistol, is another factor, but probably not critical for self defense.
Round reliability is important -- it has to feed, fire, and eject reliably when the trigger is pulled. If I fire my carry round 59 times without failure in a given magazine, nonparametric statistics tells me I can be 95% confident that that round/magazine/pistol combination is at least 95% reliable. Then I can practice with the cheaper ammo that is exteriorly ballistically equivalent.
Shooter skill is another, separate, but vital issue. No way around this but to practice.
You and others here seem to feel our current SD ammo testing standards are good enough -- arguably the best that we can come up with -- and you may be right, but I'm not convinced.
The statistician who founded the Stats Department at my alma mater, George Box, is known for saying, "All models are wrong, but some models are useful." All models are simplifications of reality; thus, they are, by definition, wrong. I think Fackler's testing protocol, as elaborated by the FBI, is very useful. A diamond can always be polished, so I'll admit thete is room for improvement, but sooner or later a point is reached where further polishing doesn't result in noticeable improvements. For example, the FBI ballistic gel model is quite general. It can be refined by makng it more specific, for example, by embedding rib bones in the gel. But, specificity is typically gained only by sacrificing generality or increasing cost dramatically. My opinion is what exists now is fairly optimal.
).