Penetration Question

[rpenmanparker]

I won’t continue to argue with you about this. Most likely you all are right about how the FBI tested. But I will continue to maintain that the most sensible test would be to correlate gel penetration characteristics (depth, expansion, mass retention) of specific bullets with the data record regarding real life outcomes of shooting humans with them.

This is somewhat similar to cholesterol testing. You can say such and such a drug lowers cholesterol 50%, and that’s great. But saying that a drug which lowers cholesterol by 50% also cuts heart attack episodes by 50% is much, much better. The real outcome is what is important.

Besides you will never manage to make sense of the gazillion variables in the human body. Dead or alive is a much simpler proposition.

 

[481]

One small correction, the study excluded any shots involving bone.

Before I hitch my wagon to the Wolberg paper I want to know:

1) What tissues were traversed in each of the 25 valid shots. "Torso" means chest and abdomen and there is a big difference between lung tissue and the abdominal organs, even if bone is excluded. They opened the door to this line of questioning in the paper by posting differing expansion diameters of bullets fired in pig "abdomen" vs pig "muscle" vs water.
2) More variables. What was the range of fire? Were all the service weapons the same in terms of barrel length? What influence did clothing play? What was the angle of incidence for each shot?
3) How did the pathologist measure the penetration length? Specifically I am interested in trajectories that aren't straight, such as if a round is deflected by the diaphragm. It can go the other way too: you can get a false increased penetration distance if the bullet slips between two tissue planes.

Smells like a "massaged" selection to me.

If that is the case (that Wolberg's data is indeed 'massaged' as you suggest), then is it your position that Wolberg's conclusion (that bullet penetration in ordnance gelatin correlates strongly with that seen in human bodies) is faulty?

 

[Odd Job]

If that is the case (that Wolberg's data is indeed 'massaged' as you suggest), then is it your position that Wolberg's conclusion (that bullet penetration in ordnance gelatin correlates strongly with that seen in human bodies) is faulty?

Yes.

 

[Odd Job]

By the way that doesn't mean I don't think ballistic gel is not a valid test medium. It's the best medium we have at present, to evaluate ammunition in a reproducible manner.

 

[481]

If that is the case (that Wolberg's data is indeed 'massaged' as you suggest), then is it your position that Wolberg's conclusion (that bullet penetration in ordnance gelatin correlates strongly with that seen in human bodies) is faulty?

Yes.

By the way that doesn't mean I don't think ballistic gel is not a valid test medium. It's the best medium we have at present, to evaluate ammunition in a reproducible manner.

You cannot have it both ways.

Wolberg's conclusion, found on page 13 of that publication and quoted below exactly as it was written by Wolberg the article states quite clearly-

Based on comparisons of data from living tissue penetrations by the 147 grain Subsonic Winchester, 9mm Parabellum bullet with test shots of the same bullet into 10% Knox Ordnance Gelatin, type 250A, shot at 4 degrees centigrade, it is concluded that this gelatin can be a useful predictor of this bullet' s penetration and expansion characteristics in shots in the human torso.

-that he finds 10% ordnance gelatin to be a valid test medium.

Your statement, "By the way that doesn't mean I don't think ballistic gel is not a valid test medium..." within 4 minutes of your assertion that you find Wolberg's conclusion (that 10% ordnance gelatin is a valid test medium) to be faulty is thoroughly self-contradictory. It also flies in the face of three decades of research conducted by numerous researchers, all of whom I'd bet you are familiar with at least by name, if not by content, that supports Wolberg's conclusion that 10% ordnance gelatin is a valid test medium.

 

[JohnKSa]

One small correction, the study excluded any shots involving bone.

Interesting. I obviously missed that one very important bit of information.

 

[Odd Job]

You cannot have it both ways.

Yes I can.

Ballistic gel is a valid medium in which to compare ammunition performance in a reproducible manner. The homogeneous nature of the medium gives it that quality.

That doesn't mean I have to believe Wolberg's conclusion that "this gelatin can be a useful predictor of this bullet' s penetration and expansion characteristics in shots in the human torso."

 

[rpenmanparker]

Yes I can.

Ballistic gel is a valid medium in which to compare ammunition performance in a reproducible manner. The homogeneous nature of the medium gives it that quality.

That doesn't mean I have to believe Wolberg's conclusion that "this gelatin can be a useful predictor of this bullet' s penetration and expansion characteristics in shots in the human torso."

Exactly. But if you knew how many inches a variety of bullets penetrated gelatin, how much they expanded, how much mass they retained and how effective they were historically as defensive rounds, you could create a rule that would allow you to predict the effectiveness of other brands and models just based on their behaviors in gelatin. No humans would have to be harmed in the testing. Nobody really should care how deep a bullet penetrates a human body. They should only care what is the bottom line effect on the human.

 

[481]

Yes I can.

Ballistic gel is a valid medium in which to compare ammunition performance in a reproducible manner. The homogeneous nature of the medium gives it that quality.

That doesn't mean I have to believe Wolberg's conclusion that "this gelatin can be a useful predictor of this bullet' s penetration and expansion characteristics in shots in the human torso."

Then your position contradicts that of 30 some-odd years worth of research (which is not solely limited to Wolberg's work, by the way) that successfully correlates the terminal behavior of bullets in tissue simulants (10% and 20% gelatin, water) with that found in human bodies.

In fact, one of those numerous researchers, Duncan MacPherson, dedicates one full chapter (Chapter 5) of his book, Bullet Penetration, to that topic prefaced with his introduction to that chapter:

One of the issues making demonstration of of bullet effectiveness so intractable by either test or analysis is the inhomogeneity of animal (and human) bodies. Even minute differences in impact can alter the tissues in the bullet wound track and produce WTI changes that are larger than those due to the bullet physical and dynamic parameters. Even if this difficulty did not exist, widespread extensive testing on animals (much less humans) is not practical on humanitarian and other grounds. Both of these reasons have created a longstanding desire to find a satisfactory inanimate substitute for soft body tissue in bullet penetration and performance testing. Any belief that tissue must be used in testing bullet terminal ballistics in tissue penetration is fallacious because there are no mystical effects associated with the results of bullet penetration in tissue; in fact, tissue is inferior to a good tissue simulant for this purpose. Tissue simulants are discussed in detail in this chapter.

So then, you disagree with MacPherson that 10% ordnance gelatin is a useful—and even a preferable—predictor of bullet performance in human tissue as well?

 

[rpenmanparker]

T
In fact, one of those numerous researchers, Duncan MacPherson, dedicates one full chapter (Chapter 5) of his book, Bullet Penetration, to that topic prefaced with his introduction to that chapter:



So then, you disagree with MacPherson that 10% ordnance gelatin is a useful—and even a preferable—predictor of bullet performance in human tissue as well?

You misunderstand. Saying that ordinance gelatin is a "useful---and even a preferably---predictor of bullet performance in human tissue" does not mean that the behavior in ordnance gelatin and live bodies must be identical. It only means that there must be a correlation, a formula, function, a mathematical relationship which reliably relates the two sets of data. It may be possible to find such a relationship for things like penetration in both media, but I prefer to think that it is easier to relate the measurements from ordnance to the end result of shooting the live body.

I won't draw you any graphs, but to keep it simple, let's just look at a few imaginary data points. Let's assume that 9mm bullets that we can show penetrate on average 5 in of gelatin have historically killed 2% of the people who had been shot with them one time. And that loads which penetrated 15 in gelatin killed 33% of the victims of one hit. And that a complete graph of all the data showed a peak mortality of a single hit with 9 mm bullet occurred with loads that penetrated to 13.5 in in gelatin. Those would all be average penetrations of say 20 shots. Then you could say that 9mm bullets that penetrated 13.5 in plus or minus one standard deviation would have the best chance of killing someone shot with them, and you would look for bullets that had that penetration profile. So the gelatin would be very useful. It just wouldn't be expected to give the same penetrations in the two types of media. And frankly that is no great loss, because if you don't know what a certain penetration in a human body means with regard to bullet effectiveness, what good does it do to be able to predict body penetration from gelatin penetration. What you should want to do is to predict mortality from gelatin penetration.

 

[Odd Job]

Then your position contradicts that of 30 some-odd years worth of research (which is not solely limited to Wolberg's work, by the way) that successfully correlates the terminal behavior of bullets in tissue simulants (10% and 20% gelatin, water) with that found in human bodies.

Conversely, I can claim your position is because I have problems with the Wolberg paper, I am against all the other research involving ballistic gelatin.
I certainly do have a problem with some of it, but not all of it.

I have MacPherson's book here.
I agree with his premise that tissue is not homogeneous and therefore it is not possible to conduct reproducible testing and meaningful comparisons of ammunition in a prospective manner, using live tissues (ethics considerations aside). That was never my position.

Further on in Chapter 5, on page 69 MacPherson writes:

It is not easy to prove that the forces produced on the bullet when it traverses tissue simulant are really similar to the forces produced by tissue under the same conditions. Dynamic equivalence can never be exact, but similar penetration depth in simulant and in "average" soft tissue over a range of velocities is a very good demonstration of equivalence.

Fair enough, but we must then state what that average tissue is. Do we exclude inelastic tissues such as liver? Do we exclude trajectories mainly involving low density tissues such as lung? Do we exclude shots involving tissue planes which give rise to deflection?

It might well be that for a given soft tissue trajectory, 10% ballistic gel will provide the same influence on the expansion of the bullet and give rise to the same penetration. It might be that this medium does not do it for a particular shot, but another medium does. That determination can only be made retrospectively.
Note that MacPherson excludes bones in his discussion of this "average" tissue. He also makes this point about lungs on page 71:

The density of soft tissue varies somewhat, but is generally a few percent less than the density of water; the lungs are an exception because they are gas-filled. Lung tissue density varies during the respiration cycle, but the density is always much lower than that of other tissue. Most popular tissue simulants have a density near that of water; the density of gelatin is slightly greater than water.

There is also gas in the abdomen. Have a look at a CT scan of the chest and abdomen and you will find out just how much there is!

Wolberg's paper doesn't tell us anything about the trajectories through the torso, and I have a problem with that.

Other publications support ballistic gel as a simulant for a tissue "average" and I have issues with aspects of that too, in terms of predictive modelling (prospective vs retrospective determination of equivalent behaviours of bullets in gel vs tissue).

To put it another way: I have seen more than 3000 acute gunshot wounds up close. Not just the wounds, but the clothing and the imaging and the clinical findings too.
I have seen too many strange trajectories, even when bone is not involved, to be able to hitch my wagon to a Wolberg-type conclusion where it is asserted that gel is a good predictor of a certain ammunition's performance in "shots to the torso."

My real world observations do not support this, unless there are major caveats! And with those caveats come reductions in the available human trajectories to be analysed with reference to whatever simulant is being tested, for comparison purposes.

 

[481]

You misunderstand.

I seriously doubt that.

Saying that ordinance gelatin is...

By the way, it is 'ordnance' (related to weapons, ammunition, etc.), not 'ordinance' (legislation enacted by a municipal authority)...

 

[rpenmanparker]

I seriously doubt that.



By the way, it is 'ordnance' (related to weapons, ammunition, etc.), not 'ordinance' (legislation enacted by a municipal authority)...

I’m glad we got those two issues out of the way. Cleaning up the spell check is a never ending chore.

 

[481]

Conversely, I can claim your position is because I have problems with the Wolberg paper, I am against all the other research involving ballistic gelatin.
I certainly do have a problem with some of it, but not all of it.

I have MacPherson's book here.
I agree with his premise that tissue is not homogeneous and therefore it is not possible to conduct reproducible testing and meaningful comparisons of ammunition in a prospective manner, using live tissues (ethics considerations aside). That was never my position.

Further on in Chapter 5, on page 69 MacPherson writes:

It is not easy to prove that the forces produced on the bullet when it traverses tissue simulant are really similar to the forces produced by tissue under the same conditions. Dynamic equivalence can never be exact, but similar penetration depth in simulant and in "average" soft tissue over a range of velocities is a very good demonstration of equivalence.

Fair enough, but we must then state what that average tissue is. Do we exclude inelastic tissues such as liver? Do we exclude trajectories mainly involving low density tissues such as lung? Do we exclude shots involving tissue planes which give rise to deflection?

It might well be that for a given soft tissue trajectory, 10% ballistic gel will provide the same influence on the expansion of the bullet and give rise to the same penetration. It might be that this medium does not do it for a particular shot, but another medium does. That determination can only be made retrospectively.
Note that MacPherson excludes bones in his discussion of this "average" tissue. He also makes this point about lungs on page 71:

The density of soft tissue varies somewhat, but is generally a few percent less than the density of water; the lungs are an exception because they are gas-filled. Lung tissue density varies during the respiration cycle, but the density is always much lower than that of other tissue. Most popular tissue simulants have a density near that of water; the density of gelatin is slightly greater than water.

There is also gas in the abdomen. Have a look at a CT scan of the chest and abdomen and you will find out just how much there is!

Wolberg's paper doesn't tell us anything about the trajectories through the torso, and I have a problem with that.

Other publications support ballistic gel as a simulant for a tissue "average" and I have issues with aspects of that too, in terms of predictive modelling (prospective vs retrospective determination of equivalent behaviours of bullets in gel vs tissue).

To put it another way: I have seen more than 3000 acute gunshot wounds up close. Not just the wounds, but the clothing and the imaging and the clinical findings too.
I have seen too many strange trajectories, even when bone is not involved, to be able to hitch my wagon to a Wolberg-type conclusion where it is asserted that gel is a good predictor of a certain ammunition's performance in "shots to the torso."

My real world observations do not support this, unless there are major caveats! And with those caveats come reductions in the available human trajectories to be analysed with reference to whatever simulant is being tested, for comparison purposes.

I believe that the answer to describing what constitutes ''average'' tissue is posited by MacPherson as quoted above, "Dynamic equivalence can never be exact, but similar penetration depth in simulant and in "average" soft tissue over a range of velocities is a very good demonstration of equivalence."

In another chapter (I won't bother reprinting the salient portions of Chapter 6 here since you obviously have the book) MacPherson, explains that so long as a tissue simulant has the density and an internal sonic velocity that is very close to that of soft tissue (which would also suggest that the bulk modulus is also the same or very close to that of soft tissue), the simulant, in this case, ordnance gelatin, possesses dynamic equivalence to soft tissue. This is what Wolberg's paper confirms, albeit not with the intense degree of scrutiny that you suggest is needed to validate Wolberg's conclusion.

 

[481]

I’m glad we got those two issues out of the way.

So am I. If we are imprecise with our spelling and our words, how can we ever hope make a relevant point?

Cleaning up the spell check is a never ending chore.

So it is.

 

[RecoilRob]

Many years ago a bunch of live goats were shot with various handgun calibers and bullets then timed to see how long they took to fall down. Was this a good predictor of caliber and bullet performance? Some would think so but there were some caveats that were illuminated by this testing. If you went by this test results....we'd all be using Glaser Safety-Slugs because they were, hands down the winner. And some people were influenced enough by these tests to adopt the Glaser's long enough to get some 'real world' usage with them...where they were found wanting in some situations. As great as they performed in some scenarios was about as pitiful as they performed in others.

What is needed is a true 'Jack of All Trades' because you can't pick and choose your shot or shooting situation. Wouldn't it be great if we had a pistol that could carry Glaser, FMJ, fast opening HP, and slow opening HP.....then have sensors to analyze the target a millisecond before launching the projectile. The density would be probed and the appropriate load fired to assure optimum performance! Simple! Or not.... We don't have such technology available to us...and likely never will, so a gelatin medium that is repeatable and able to replicate a bullet impact for testing is a valuable tool for all to use. It is never going to be able to duplicate any and all possible impacts, and I can't think of anything that could. The current FBI protocol tests seem to be a pretty respectable effort to find bullets that will work decently in most situations you might find yourself encountering. Nothing is perfect...but if what you carry can pass the FBI tests...it'll work fine if you deliver it where it needs to go.

 

[Zendude]

Do the 4 layers of denim do a good job of simulating something like a tough leather jacket?

 

[rpenmanparker]

Many years ago a bunch of live goats were shot with various handgun calibers and bullets then timed to see how long they took to fall down. Was this a good predictor of caliber and bullet performance? Some would think so but there were some caveats that were illuminated by this testing. If you went by this test results....we'd all be using Glaser Safety-Slugs because they were, hands down the winner. And some people were influenced enough by these tests to adopt the Glaser's long enough to get some 'real world' usage with them...where they were found wanting in some situations. As great as they performed in some scenarios was about as pitiful as they performed in others.

What is needed is a true 'Jack of All Trades' because you can't pick and choose your shot or shooting situation. Wouldn't it be great if we had a pistol that could carry Glaser, FMJ, fast opening HP, and slow opening HP.....then have sensors to analyze the target a millisecond before launching the projectile. The density would be probed and the appropriate load fired to assure optimum performance! Simple! Or not.... We don't have such technology available to us...and likely never will, so a gelatin medium that is repeatable and able to replicate a bullet impact for testing is a valuable tool for all to use. It is never going to be able to duplicate any and all possible impacts, and I can't think of anything that could. The current FBI protocol tests seem to be a pretty respectable effort to find bullets that will work decently in most situations you might find yourself encountering. Nothing is perfect...but if what you carry can pass the FBI tests...it'll work fine if you deliver it where it needs to go.

That is why I advocate using thousands of data points representing a random selection of conditions recorded for actual human shootings. The correlation between gelatin penetration and % effectiveness of bullets should give the highest probability of success. The sheer mass of the data should smooth out any peculiarities. What the person selecting ammo will have to face, however, is that the best they can hope for is a moderate probability of success. Picking the rounds the gel penetration of which corresponds to the highest human mortality probability will be the best bet, but it won’t be a sure thing. That will be more obvious than using other methods of selection.

 

[JTQ]

I think you're largely splitting hairs when choosing ammo, especially when trying to decide between, lets say the top five or 10 in any particular category.

I've posted this elsewhere in the past in a .45 ACP ammo debate ...

"Imagine a bar graph showing performance (if you could quantify performance) of just .45 Auto rounds. You'd possibly see some significant performance gaps between the best and worst round.

Now let's add .380, .38 Special, .357 Magnum and .44 Magnum to the chart. I suspect the differences between the best and worst .45 Auto rounds would look less significant as they would bunch closer together compared to the outliers at the top and bottom of the spectrum.

Now lets add .22 LR and .308 and .30-06 to the chart. You'd most likely be unable to tell the difference between any of the .45 Auto rounds as the chart would have to accommodate the very low relative performance of the .22 LR and the relatively high performance of the .30 caliber rifle rounds.

It makes fun debates, but you're really just splitting hairs when you're comparing .45 Auto rounds. Are some better than others? Sure. Am I losing sleep over what choice I'm making in .45 Auto ammo? No."

 

[481]

Do the 4 layers of denim do a good job of simulating something like a tough leather jacket?

Z-

The 4 layers of 16-ounce (per square yard) denim as a barrier was an IWBA mechanical test meant to simulate a ''worst case scenario'' for projectiles passing through clothing/textiles, etc. Since a thick, tough leather jacket might rival the resistive strength of 16-ounce denim, it would likely be a reasonably close match for that obstruction.

If math is your thing, you could also use this model (Hudgins, US ARMY, RDECOM, ARDEC, AETC, Aeroballistics Div., 2005)--

https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2005/garm/tuesday/hudgins.pdf

--if you wanted to examine the two materials for equivalence.

 

[RetiredUSNChief]

I won’t continue to argue with you about this. Most likely you all are right about how the FBI tested. But I will continue to maintain that the most sensible test would be to correlate gel penetration characteristics (depth, expansion, mass retention) of specific bullets with the data record regarding real life outcomes of shooting humans with them.

This is somewhat similar to cholesterol testing. You can say such and such a drug lowers cholesterol 50%, and that’s great. But saying that a drug which lowers cholesterol by 50% also cuts heart attack episodes by 50% is much, much better. The real outcome is what is important.

Besides you will never manage to make sense of the gazillion variables in the human body. Dead or alive is a much simpler proposition.

But the problem is that you CAN'T account for all the variables. And that's the point.

You even allude to this at the end of your own posting.

And even if you do manage to correlate accurate, meaningful information on ONE bullet design at ONE particular loading from ONE particular gun, how many such shootings would this take in which ALL the other various factors are known to the degree of accuracy required?

Then ask yourself how anybody who developes a new design/loading can gather the SAME relevant data? How many shooting would this require? Because remember...those shooting have to match those from every OTHER actual shooting in order to be meaningful in any way.

Controlled and proper use of ballistics gel does this. That's it's beauty.

 

[Jenrick]

We're doing this one again? If I feel motivated I'll see if I can find the reply to another thread of the same content previously that is more detailed.

The short version:

I was in charge of doing ballistic testing for a large metro policy agency when we selected a new rifle round. This involved a LOT of background reading, talking with people, bugging the FBI, etc in addition to shooting a lot of rounds into and through a lot of things. The FBI does a presentation at LE conferences where they show a round recovered from an FBI shooting, and a round recovered from ballistic gel, and the audience is asked to identify which is which. In general it's pure guess work as they look pretty dang identical these days (modern duty/defensive rounds). I don't remember the exact percentage off the top of my head, but it was well over 90% of rounds which penetrate 12"-18" of in testing, performed to spec in the real world (by this I mean expanded to approximately the same diameter, same weight retention, and penetrated far enough to reach vital organs).

TL/DR: FBI protocol gel testing results, closely match real world results, based on large volumes of FBI data available to closed LEO sources

In case you're wondering, the official FBI data is restricted to LE agencies only, so no I'm not going to put any out here.

-Jenrick

 

[rpenmanparker]

But the problem is that you CAN'T account for all the variables. And that's the point.

You even allude to this at the end of your own posting.

And even if you do manage to correlate accurate, meaningful information on ONE bullet design at ONE particular loading from ONE particular gun, how many such shootings would this take in which ALL the other various factors are known to the degree of accuracy required?

Then ask yourself how anybody who developes a new design/loading can gather the SAME relevant data? How many shooting would this require? Because remember...those shooting have to match those from every OTHER actual shooting in order to be meaningful in any way.

Controlled and proper use of ballistics gel does this. That's it's beauty.

You aren't getting my point.

The method is called Statistical Experimental Design. You don't try to account for all the variables. Instead you just let the results of the shootings and the results of the gel testing tell you what is best to use without understanding all the underlying effects. Yes, it would be a lot of work, but statistics don't lie.

Here is how you would do it. Say you could identify 1,000 (or more) center mass shooting records for which you could know the details of the bullet used and the outcome of the shooting. Actually I am sure you could obtain many thousands. Say of that 1,000+ records, there were 225 different bullet identities with regard to brand, bullet design, caliber, etc. You don't need to worry about all those numbers, just the identities of the bullets so you can buy 20 rounds of each and test them with a standard gun (of your choosing) in standard gelatin. First you characterize each shooting outcome on a scale from 1-10 with 1 being no effect despite a center mass hit and 10 being instant death. You make a spreadsheet of the results with the shooting outcome results being in the column labeled "Output". Then you shoot the bullets into gelatin, and record four average "Input" variables for the 20 shots, muzzle velocity, penetration depth, terminal bullet mass, and terminal greatest bullet cross-sectional area. You enter that same set of data across four columns in each row of the spreadsheet where there is a shooting result (not all the same mortality results) for that same bullet identity. At that point, forget about the bullet identities. They don't matter any more.

Then you let your computer do some mathematical magic called multi-variable regression on the four measured Inputs and one Output. The computer will give you an equation expressing the mortality Output number as a function of the four Inputs. It will tell you statistically how each of the four characteristics of a fired bullet affects the mortality outcome. So if you grab ANY cartridge (new design, old design, any caliber it doesn't matter), shoot 20 of them into gelatin, and make the measurements, you can plug the results into the equation and get a good ESTIMATE of the AVERAGE mortality result of that cartridge.

Is it a lot of work? Yes. Does it tell you exactly how a bullet will act? No. It tells you what the most probable outcome would be. The good news is that you can keep making the regression equation better and better by adding more data to the spreadsheet as it becomes available. After a while you have so many data points that the regression equation becomes very, very effective at predicting the mortality results.

But wait, are the four variables measured from the gelatin testing the right things to measure? I don't know. Seems right to me right now. But the method will automatically kick out variables that don't matter. The statistical significance of an important variable will be high. For an unimportant variable the significance will be low and you can actually recalculate the regression equation leaving that unimportant variable out. That simplifies your work later on down the road since you don't have to measure that characteristic of the bullets fired into gelatin from then on. What about variables I haven't thought of? If you think of one, just start measuring it and see if it is significant to the regression equation. For example, maybe you want to include starting bullet mass and caliber. Add that data to the table, compute the new regression equation, and voila, you immediately know if those things are important to mortality. The statistics answer the question about what is important to a cartridge being deadly. And they don't lie. You don't have to guess.

At the end of the process you can create a table of cartridge brands and models with their mortality score listed. Easy-peasy to pick a viable self-defense round off the table. And interestingly you would see that caliber isn't as important as final mass and final cross-sectional area. Aren't those all related. Sure, but how exactly? Is FMJ as good as JHP? Isn't higher energy better? +P anyone? Forget about all that. The method just tells you the score for every bullet on the list and every bullet you could ever test.

That is how gelatin should be used to select effective self-defense ammo.​

 

[rpenmanparker]

And another very significant outcome of the statistical experimentation is that ammo makers could fine tune their products to maximize the characteristics that the regression equation says are important. So they could maximize expansion at the expense of muzzle energy if the equation said that was the right thing to do to assure effective target stopping. That’s just an example. Also you could decide on a lower caliber and be confident in it if it had a characteristic that the equation showed would make up for the smaller diameter. This is very powerful stuff. But then what would we argue about?

 

[481]

You aren't getting my point.

The method is called Statistical Experimental Design. You don't try to account for all the variables. Instead you just let the results of the shootings and the results of the gel testing tell you what is best to use without understanding all the underlying effects. Yes, it would be a lot of work, but statistics don't lie.

Here is how you would do it. Say you could identify 1,000 (or more) center mass shooting records for which you could know the details of the bullet used and the outcome of the shooting. Actually I am sure you could obtain many thousands. Say of that 1,000+ records, there were 225 different bullet identities with regard to brand, bullet design, caliber, etc. You don't need to worry about all those numbers, just the identities of the bullets so you can buy 20 rounds of each and test them with a standard gun (of your choosing) in standard gelatin. First you characterize each shooting outcome on a scale from 1-10 with 1 being no effect despite a center mass hit and 10 being instant death. You make a spreadsheet of the results with the shooting outcome results being in the column labeled "Output". Then you shoot the bullets into gelatin, and record four average "Input" variables for the 20 shots, muzzle velocity, penetration depth, terminal bullet mass, and terminal greatest bullet cross-sectional area. You enter that same set of data across four columns in each row of the spreadsheet where there is a shooting result (not all the same mortality results) for that same bullet identity. At that point, forget about the bullet identities. They don't matter any more.

Then you let your computer do some mathematical magic called multi-variable regression on the four measured Inputs and one Output. The computer will give you an equation expressing the mortality Output number as a function of the four Inputs. It will tell you statistically how each of the four characteristics of a fired bullet affects the mortality outcome. So if you grab ANY cartridge (new design, old design, any caliber it doesn't matter), shoot 20 of them into gelatin, and make the measurements, you can plug the results into the equation and get a good ESTIMATE of the AVERAGE mortality result of that cartridge.

Is it a lot of work? Yes. Does it tell you exactly how a bullet will act? No. It tells you what the most probable outcome would be. The good news is that you can keep making the regression equation better and better by adding more data to the spreadsheet as it becomes available. After a while you have so many data points that the regression equation becomes very, very effective at predicting the mortality results.

But wait, are the four variables measured from the gelatin testing the right things to measure? I don't know. Seems right to me right now. But the method will automatically kick out variables that don't matter. The statistical significance of an important variable will be high. For an unimportant variable the significance will be low and you can actually recalculate the regression equation leaving that unimportant variable out. That simplifies your work later on down the road since you don't have to measure that characteristic of the bullets fired into gelatin from then on. What about variables I haven't thought of? If you think of one, just start measuring it and see if it is significant to the regression equation. For example, maybe you want to include starting bullet mass and caliber. Add that data to the table, compute the new regression equation, and voila, you immediately know if those things are important to mortality. The statistics answer the question about what is important to a cartridge being deadly. And they don't lie. You don't have to guess.

At the end of the process you can create a table of cartridge brands and models with their mortality score listed. Easy-peasy to pick a viable self-defense round off the table. And interestingly you would see that caliber isn't as important as final mass and final cross-sectional area. Aren't those all related. Sure, but how exactly? Is FMJ as good as JHP? Isn't higher energy better? +P anyone? Forget about all that. The method just tells you the score for every bullet on the list and every bullet you could ever test.

That is how gelatin should be used to select effective self-defense ammo.​

That was an excellent description of what, I believe, current mathematical penetration models (both of which are modified forms of the Poncelet penetration equation) by both MacPherson and Schwartz acheive. At the present time, I believe that MacPherson has 400+ such gelatin test data and Schwartz has 800+ gel test data correlated to their respective models.

Once you have figured out a way to predict, or at least model, the four input variables (muzzle velocity, penetration depth, terminal bullet mass, and terminal greatest bullet cross-sectional area), it appears that you have gone to a dimensionless rating system for the ''Mortailty Output'' value. Do you think that a percentage or probability value might be a better way of doing that?