45 vs 9mm interesting test

[Shadow9mm]

removed

[Sharkbite]

That assumes they both expanded the instant they contacted the gel. They did not. There is always some depth of penetration before expansion starts.

Which one starts to expand first, gaining some additional “tissue” damage?

Guys we are splitting hairs here. If there was a difference in the real world, we would have seen it in the OVER 100 YEARS people have been getting shot with each caliber.

We have not.

[JohnKSa]

This assumes that the .45 ACP numbers had a roughly equal or lower proportion of shootings with FMJ as compared to 9mm...

I find it interesting that you spent a good deal of your post continuing to make a strong case for the 9mm data set being heavily over-populated with FMJ but at the same time still object to the comparison with .45ACP because it might have more or equal FMJ in it although the preponderance of your arguments and Ellifritz's comments seem to point the other way in no uncertain terms.

...the larger data set for 9mm should mitigate that variable to a greater degree than the smaller data set for .45 ACP.

That makes absolutely no sense at all. How could the fact that FMJ is over-represented in a data subset be mitigated by the fact that there are more shootings in that particular data subset?

As such, details about the gun, caliber, and specific bullet are often omitted from the report as they're not deemed relevant to the investigation.

I'm curious what evidence you used to arrive at this conclusion? You're saying that when a police department investigates a shooting they don't record the type and caliber of the firearm or the bullets recovered from the body/victim? That seems very strange. I haven't read tons of police reports, but the ones I have that involve shootings certainly include information about the firearm type and caliber and if ballistics are a factor, they may describe the bullets as well.

Unfortunately this method of study inherently contains the highest number of uncontrolled variables...

YES. That is absolutely correct. Just like the real world, where, if we get involved in a shooting, it will take place. With lots of uncontrolled variables.

Ok, let me make an observation here:

"bullet type"
"the loading used"
"duty ammunition issued"
"specific loading"
"specific bullet "
"specific .45 ACP loading"
"FMJ ammunition"
"FMJ ammunition"
"ammunition is almost exclusively FMJ"
"FMJ ammunition "
"shootings with FMJ"
"expanding vs. non-expanding bullets"
"bullet types"
"expanding and non-expanding bullets"
""expanding" bullets"
"non-expanding ones"
"expanding bullet"
"FMJ and other non-expanding bullet types"
"expanding vs. non-expanding bullet"
"type and barrel length of handgun"
"modern JHP bullets"
"barrel length,"
"barrel length"
"premium 9mm JHP"
"fails to expand when fired from a snubnose"

You've spent a good deal of time explaining how failing to control for bullet differences and even barrel length is going to make it impossible to prove there are differences due to caliber in the real world.

Don't you see that can't be true unless bullet differences and barrel length matter much more than caliber differences in the real world?

Can you see that this is actually a variant of my argument that differences DUE TO CALIBER in the service pistol terminal performance class don't show up in the real world?

Something to think about.

So the .45 crushes 2.3 square inches more tissue which, depending on which way you calculate your percentage, is 26-34% more.

So, like maybe "a 27% increase"?

Anyone bother to consider the 5 questions I posed in my last post? I'm not asking for anyone to post the answers, but it would be useful for people interested in this topic to think about them.

[Pumpkin]

John,

Would a comparison of the 9mm vs the 40 have been a better one because of the 40 being used in more police shootings?

I realize all the uncontrollable elements would still be in place (except military) but at least the amount of data for the larger caliber would be increased.

This is assuming my thinking that the 40 would actually have been involved in more shootings than the 45, maybe I’m wrong

One last thing, would narrowing down the data to say one group, ex. “The Texas Border Patrol” help in removing SOME variables?

Yes, the the amount of data would be seriously reduced but could it give us a little bit more reliable data in the end?

[micromontenegro]

Cubic inches, gents, cubic.

[Webleymkv]

Originally posted by JohnKSa
Quote:
This assumes that the .45 ACP numbers had a roughly equal or lower proportion of shootings with FMJ as compared to 9mm...
I find it interesting that you spent a good deal of your post continuing to make a strong case for the 9mm data set being heavily over-populated with FMJ but at the same time still object to the comparison with .45ACP because it might have more or equal FMJ in it although the preponderance of your arguments and Ellifritz's comments seem to point the other way in no uncertain terms.

Ellifritz specifically stated that over half the 9mm data was with FMJ ammunition and that likely trended the overall 9mm statistics downward, which is certainly a reasonable assumption. However, he did not state what proportion of data from any other caliber was FMJ vs JHP bullets or even if he even had that information. Yes, I agree that the 9mm probably would have fared better in his study if so much of the data weren't FMJ bullets, but if we're going to compare the data to that of .45 ACP then, at the very least, we need to know what proportion of the .45 ACP data was FMJ vs JHP and that is information that Ellifritz hasn't provided and may not have himself.

As I explained before, there are perfectly reasonable explanations for why Ellifritz may have had more information about the bullet type used in the 9mm shootings as opposed to the .45. If, for example, a large number of the 9mm shootings that Ellifritz included in his data came from officer-involved shootings with the NYPD between 1993 and 1998, then we know to a high probability that those shootings were with FMJ ammunition even if the bullet type wasn't specifically mentioned in the report.

Originally posted by JohnKSa
Quote:
...the larger data set for 9mm should mitigate that variable to a greater degree than the smaller data set for .45 ACP.
That makes absolutely no sense at all. How could the fact that FMJ is over-represented in a data subset be mitigated by the fact that there are more shootings in that particular data subset?

It makes more sense if you include the entire quote:

Originally posted by Webleymkv
Even if we assume that both calibers had roughly the same proportion of shootings with FMJ ammo, the larger data set for 9mm should mitigate that variable to a greater degree than the smaller data set for .45 ACP.

When compiling statistics, it is well-known that the larger the data set, the more uncontrolled variables or outliers, like differences in bullet type, will be mitigated. If we are to assume, as Ellifritz did, that the inclusion of FMJ shootings in the data causes a downward trend on the overall calculation, then so long as both calibers had roughly equal proportions of shootings with FMJ the 9mm should have less of a downward trend to the final calculation due to the over twice as large data set. Unfortunately, Ellifritz doesn't tell us what proportion of the .45 ACP shootings, or the shootings in any caliber other than 9mm, were FMJ so it is impossible to say whether the 9mm data was more handicapped by the inclusion of so many FMJ shootings or not.

Originally posted by JohnKSa
Quote:
As such, details about the gun, caliber, and specific bullet are often omitted from the report as they're not deemed relevant to the investigation.
I'm curious what evidence you used to arrive at this conclusion? You're saying that when a police department investigates a shooting they don't record the type and caliber of the firearm or the bullets recovered from the body/victim? That seems very strange. I haven't read tons of police reports, but the ones I have that involve shootings certainly include information about the firearm type and caliber and if ballistics are a factor, they may describe the bullets as well.

OK, let's break this down a little. First we must assume that all police reports are accurate, which IMHO is a very big and erroneous assumption. Errors in police reports, as in every other aspect of the world we live in, are not unheard of. Secondly, you are assuming that the type of firearm, caliber, bullet type, etc. are always known by the police writing the report. If the police show up to a shooting and find a deceased victim with a through and through gunshot wound and are not able to recover the bullet, firearm, shell casings, etc. then it would be very difficult for them to identify any information about the gun, caliber, or bullet used. Third, you state that "if ballistics are a factor, they may describe the bullet." but what about cases in which ballistics are not a factor? If you have an eyewitness account of someone being murdered by an "execution style" point-blank gunshot to the back of their head, ballistics aren't really a factor as the distance of the shot is already known and the end result is unlikely to have been different regardless of what type of bullet is used, so the bullet type is irrelevant and more likely to be omitted from the report. As for my source on the information, or lack thereof, from police shootings, I draw that opinion from the comments of people like Darryl Bolke and Gary Roberts both of whom have, at various times on podcasts and such, described the difficulty in obtaining accurate information from police shootings.

Originally posted by JohnKSa
Quote:
Unfortunately this method of study inherently contains the highest number of uncontrolled variables...
YES. That is absolutely correct. Just like the real world, where, if we get involved in a shooting, it will take place. With lots of uncontrolled variables...

...You've spent a good deal of time explaining how failing to control for bullet differences and even barrel length is going to make it impossible to prove there are differences due to caliber in the real world.

Don't you see that can't be true unless bullet differences and barrel length matter much more than caliber differences in the real world?

And when, exactly, have I ever stated that bullet differences don't make a bigger difference than caliber? That's actually kind of the whole point I've been trying to make: you can't make an accurate comparison between two calibers unless you compare similar bullet types. We know, from laboratory testing, that the performance of "service pistol" class handgun calibers is drastically different with JHP ammunition as opposed to FMJ. I have never argued that a .45, even if loaded with FMJ, is superior to a 9mm JHP because of a difference in caliber as that argument is not supported by either laboratory testing or the extremely limited amount of real world data that we have. As a matter of fact, I would argue that one of the defining characteristics of a "service pistol" class handgun cartridge is one which has its performance significantly enhanced by using expanding bullets. Comparing a FMJ bullet in one caliber to a JHP in another is pointless because it's and apples-to-oranges comparison.

Originally posted by JohnKSa
Can you see that this is actually a variant of my argument that differences DUE TO CALIBER in the service pistol terminal performance class don't show up in the real world?

Yes, but why don't they show up? Is it because there is no significant difference at all or is it because they're not compared using similar bullet types? My entire argument is that this is a question we cannot answer because we don't have sufficient data to draw any meaningful conclusion.

Originally posted by JohnKSa
Quote:
So the .45 crushes 2.3 square inches more tissue which, depending on which way you calculate your percentage, is 26-34% more.
So, like maybe "a 27% increase"?

Yes, if you run the percentage calculation more conservatively (6.7/9=.74 vs 9/6.7=1.34). What I was trying to illustrate was the difference if volume between the two in cubic inches (I misspoke as labeled it as square inches) as opposed to a percentage because, for me at least, the volume in cubic inches is easier to visualize. Whether or not one thinks that 2.3 cubic inches is significant, I did not comment on as I will leave that up to the individual.

Originally posted by JohnKSa
Anyone bother to consider the 5 questions I posed in my last post? I'm not asking for anyone to post the answers, but it would be useful for people interested in this topic to think about them.

OK, let's consider them:

1. How much weight of tissue is damaged, on average, by a single hit from a service pistol bullet. If you increase that by 27%, then what how much weight does it damage? What about if you double it, then how much weight is damaged? What are the percentages of those weights to a 180 lb body?

This is really impossible to answer because human bodies are not homogenous so the weight of the tissue damaged by a gunshot would depend greatly on the density of the specific tissue/tissues that the bullet strikes and the angle of the shot.

2. If a human body weighs 180lbs, how many pounds of that body can be damaged without resulting in physical incapacitation? What percentage is that?

Again, this depends on the specific tissue that is damaged the the body composition of the person who is shot as even individuals of roughly equal weight can have very different body types.

3. What is the percentage ratio of the weight answers from 1 to the weight answer from 2?

Again, impossible to answer because we can't answer 1 and 2.

4. What percentage of the time is the physical incapacitation of the attacker the outcome of a successful defensive shooting?

We don't know because we don't have a satisfactory way to differentiate between physical and psychological incapacitations in real-world shootings.

5. What percentage of the time does the person scoring the first "good hit" win a gunfight?

We'd need accurate gunfight statistics to answer this and those are notoriously difficult to get.

I see the point you're trying to make here: there are too many uncontrolled variables other than caliber to definitively state that caliber is a significant factor in the outcome of real-world shootings and that is something I've never disagreed with. The point I'm trying to make is that this cuts both ways and that, due to the lack of concrete data, we cannot rule out caliber as a significant factor either.

[Webleymkv]

Originally posted by Pumpkin
Would a comparison of the 9mm vs the 40 have been a better one because of the 40 being used in more police shootings?...

...This is assuming my thinking that the 40 would actually have been involved in more shootings than the 45, maybe I’m wrong

At least in the case of Ellifritz, the .40 wouldn't give us any better comparison because it was actually a lower number of shootings (188) than .45 was (209). I suspect that a higher percentage of the .40 S&W shootings were likely with JHP ammunition as I'm not aware of any large police or military organization that issues or issued .40 S&W pistols with FMJ ammunition, but we really don't know for sure because Ellifritz doesn't specify.

Originally posted by Pumpkin
One last thing, would narrowing down the data to say one group, ex. “The Texas Border Patrol” help in removing SOME variables?

It would reduce some variables as you could infer the the weapon and ammunition type used by the specific agency in instances where the officer did the shooting, but it would also severely limit your data as you would have far less, and in some cases probably no, data on guns and calibers not used by that specific agency.

[JohnKSa]

Would a comparison of the 9mm vs the 40 have been a better one because of the 40 being used in more police shootings?

9mm Luger
# of people shot – 456
# of hits – 1121
% of hits that were fatal – 24%
Average number of rounds until incapacitation – 2.45
% of people who were not incapacitated – 13%
One-shot-stop % – 34%
Accuracy (head and torso hits) – 74%
% actually incapacitated by one shot (torso or head hit) – 47%

.40 S&W
# of people shot – 188
# of hits – 443
% of hits that were fatal – 25%
Average number of rounds until incapacitation – 2.36
% of people who were not incapacitated – 13%
One-shot-stop % – 45%
Accuracy (head and torso hits) – 76%
% actually incapacitated by one shot (torso or head hit) – 52%

Here they are in Ellifritz's data. Ellifritz and Webleymkv are worried that FMJ is over-represented in the 9mm data set which Ellifritz says might make the 9mm data look worse than it really is. But even before trying to account for that, the two are more similar than they are different.

I don't think it matters, that's what I've been trying to say. What I've been saying is that all the service pistol calibers are going to perform the same in terms of real-world shooting outcomes.

Or, said another way, the many variables that affect shootings in the real world more strongly than caliber will dominate to the point that caliber differences won't even be detectable in the outcomes.

Or, said another way, the differences due to caliber that can be isolated by testing in sterile environments will have such a small effect on the outcome of real-world shootings that it won't be possible to correlate those differences to real-world shooting outcomes.

It looks like some folks want to assume, in the absence of evidence, that there is a difference but it's just that no one has been able to show it so far.

There's been plenty of time to come up with some way to prove that there's a difference in the real and no one has succeeded. The FBI's expert told us that it was a waste of time to even try 35 years ago.

As for me, I'm willing to take a position based on the following:

1. If someone was going to be able prove a difference in real-world shooting outcomes based on caliber differences in the service pistol class, they would have by now.
2. The fact that it's so hard to prove that such a difference exists is very good evidence that whatever difference may exist is so small that it's not worth bothering about and that other factors affect the outcomes of real-world shootings so strongly that they hide any effects due to caliber.

[JohnKSa]

It makes more sense if you include the entire quote:

Actually, it doesn't.

When compiling statistics, it is well-known that the larger the data set, the more uncontrolled variables or outliers, like differences in bullet type, will be mitigated.

That's not how it works. If you KNOW you have a bias in your data due to some over- or under-representation, you can't fix that by adding more data points that have the same problem. That just makes the problem worse.

The principle is that if you add more data points that don't have the same over- or under-representation, they will gradually dilute the effect of the biased data.

...then so long as both calibers had roughly equal proportions of shootings with FMJ ...

If we assume that, then we can compare the two data sets on an even basis--at least as far as the expanding vs FMJ basis is concerned.

OK, let's break this down a little. First we must assume that all police reports are accurate, which IMHO is a very big and erroneous assumption.

Your argument is that there will be errors in the data. That is EXACTLY the kind of thing that large data sets tend to minimize. As far as the police not knowing the caliber in some cases, that's grasping at straws. Those cases would, obviously, not be included in a shooting study designed to look at caliber differences. I mean, how would you even put the shooting into a caliber group if you didn't know what caliber was used?

I'm seeing a lot of this in your arguments. I can tell you think you're being objective, but what you're doing is making a pretty serious, and in some cases (as with the "sometimes they don't know the caliber" claim) laughable efforts to dismiss any data that leads to an outcome you don't like.

Yes, but why don't they show up? Is it because there is no significant difference at all or is it because they're not compared using similar bullet types?

You're circling back. They don't show up because they're too small to show up in the presence of REAL-WORLD variability.

This is what I was getting at when I said it doesn't matter why.

You are trying to argue that other effects are hiding the caliber difference you assume must be present.

OK. That's great. I'm not saying there's no difference. I'm saying there's no detectable/significant difference in real-world outcomes. That's what you're proposing as well, you're just trying to come up with reasons WHY it's not detectable to obscure the basic fact that it is not.

If you have to tightly control the shooting data to the point that it's artificially constrained, to see a difference, guess what--that's not a significant difference because shootings in the real world aren't tightly controlled.

And, by the way, you want to focus on bullet types and barrel lengths without being able to show that they have a significant effect on real-world shooting outcomes or that they are greater than the effect due to caliber. Again, this seems characteristic of grasping at straws.

Again, this depends on the specific tissue that is damaged the the body composition of the person who is shot as even individuals of roughly equal weight can have very different body types.

Same thing again. This appears to be grasping at straws so you don't have to consider/can dismiss things that don't work with your desired outcome.

Take an average value and assume there's variation. The idea isn't that we're going to nail down these kinds of numbers to the nth degree, the idea is to get a rough feel for what's going on. Is it 10%? Maybe 30%? We know it's not 90%. Get as close as you can using the vast amounts of information on the internet.

Same applies to the answers to your other responses. Besides, remember, I said the point wasn't to post an answer, it was to give you something to think about. You have successfully avoided thinking about them by raising objections that aren't material to the point.

Get a feel for the problem. Make some estimates. Use expert estimates. Work through the numbers with ranges of values. You will gain insight even though you won't get an exact answer.

I see the point you're trying to make here: there are too many uncontrolled variables other than caliber to definitively state that caliber is a significant factor in the outcome of real-world shootings and that is something I've never disagreed with.

Right. There is no evidence from real-world shootings to indicate that caliber choice in the service pistol class makes a difference.

The point I'm trying to make is that this cuts both ways and that, due to the lack of concrete data, we cannot rule out caliber as a significant factor either.

It cuts both ways, but not to the same extent.

I claim (accurately) that no one has proven that caliber choice in the service pistol class makes a significant difference in real world shooting outcomes.

You claim (based on assumption) that it could make a difference but argue that there are other things affecting the outcome so much that the difference is obscured and therefore the real-world data is not useful and therefore my claim is problematic. It's not, because if it were significant (in any sense of the word) it would show up in real-world data unless the real-world data is actually flawed.

You are arguing that a difference might exist. I'm not disputing that--I will admit it could exist. I'm saying that no one can prove it is significant using real-world data--a claim that is not in dispute.

You are arguing that other things obscure it and that's why it isn't detectable. That is consistent with my claim. I'm saying that no one can prove it's significant if they can't even detect it. Those two things are not mutually exclusive and yet you keep trying to make it sound like somehow my argument is flawed.

You want to keep arguing the details of WHY the proof that it could make a significant difference isn't there. I don't care because that doesn't call my premise in to question because if it's that hard to detect in the real world--it's not significant. Because, you know, words mean things and 'significant' is a word that has a meaning.

You are working hard to dismiss the real-world data you have at your disposal. That is creative, but it won't help you make your point because to refute my point, you must have real-world data to work with.

Here's my position in two sentences:

No one can show with real-world data that caliber choice makes a significant difference. (This can be refuted if someone can produce real-world shooting outcome data that are significantly affected by caliber.)

If caliber choice makes a significant difference in the real-world then real-world shooting data must show a detectable effect due to caliber. (This is a tautology--not refutable. If there's no detectable effect, it's not a significant effect.)

If you want to throw away the real-world data and explain WHY it doesn't show a significant differene due to caliber--feel free. That doesn't speak to either one of the sentences that compose my position.

And when, exactly, have I ever stated that bullet differences don't make a bigger difference than caliber? That's actually kind of the whole point I've been trying to make: you can't make an accurate comparison between two calibers unless you compare similar bullet types.

1. It's interesting that you are willing to assume that bullet differences and barrel length differences make a difference in the outcome of real-world shootings while arguing that we shouldn't assume things in the absence of evidence.
2. If things like a few FPS of velocity due to barrel length or the choice of bullet make more difference then caliber, how is that an argument for caliber differences having a significant effect on real-world shooting outcomes? Your argument is that you assume these effects are greater than the effect due to caliber so they are hiding the differences due to caliber that you assume exist and therefore we can't use real-world shooting outcomes as evidence.

I'm ok with that. Because although it takes the long way around and relies on some assumptions, it does nothing to invalidate my claim that there's no evidence that real-world shooting outcomes are affected by caliber differences in the service pistol class.

[44 AMP]

If there was a difference in the real world, we would have seen it in the OVER 100 YEARS people have been getting shot with each caliber.

I generally agree with this, but I need to point out that people have only been getting shot with expanding bullets in either caliber for about the past 50 years or so. AND that one can ID several generations of improvement in bullet performance during that time.

Many of the JHP bullets from the early decades of their use did not expand reliably, or consistently in all circumstances.

Also note, that the 9mm JHP that "failed" in the 1986 Miami shoot out met every requirement the FBI had, at the time.

The FBI has since changed their standards...

I appreciate the discussion of how the statistics work, but what I am not seeing is how one can "boil down" (ignore??) all the real world factors that can cause differing results, often significant ones.

How can you not take into account hits in non vital areas, attackers ranging from 100lbs up to possibly 300lbs? People shot multiple times before falling down? Things like that, and more, all rolled up into one data point, how can that possibly a valid, accurate representation of the real world performance of an individual bullet or load???

Generalities, sure, but detailed percentage points, or fractions of a percent? I don't get it. I keep coming back to the experience of a guy who had to defend himself with GI FMJ ammo, once with a .45 and once with a 9mm.

He felt they were the same, as he puts it, "I shot him twice, he fell down. Both times."

Right now, I'm thinking we get over involved with the complexities of statistics, that are actually oversimplified data from real world shootings.

[Webleymkv]

Originally posted by JohnKSa
Here they are in Ellifritz's data. Ellifritz and Webleymkv are worried that FMJ is over-represented in the 9mm data set which Ellifritz says might make the 9mm data look worse than it really is. But even before trying to account for that, the two are more similar than they are different.

I don't think it matters, that's what I've been trying to say. What I've been saying is that all the service pistol calibers are going to perform the same in terms of real-world shooting outcomes.

Or, said another way, the many variables that affect shootings in the real world more strongly than caliber will dominate to the point that caliber differences won't even be detectable in the outcomes.

So if it doesn't matter that Ellifritz data is known to be skewed by the inclusion of a lot of FMJ data in 9mm, and unknown amounts in other calibers, and all service pistol calibers are going to perform the same in real-world shootings, can we conclude that it will make little to no difference in real world shootings whether FMJ or JHP ammunition is used? You say that other variables will more strongly influence the outcome of a real-world shooting than caliber, but how do you know this? Do you have some way of quantifying the degree to which these other variables will effect the outcome of a shooting?

It looks like some folks want to assume, in the absence of evidence, that there is a difference but it's just that no one has been able to show it so far.

Perhaps some folks want to assume that, but I'm not one of them. The only point I'm trying to make is that we can neither assume that there is or is not a difference. The inability to prove a theory does not necessarily disprove it either and that's the only point I'm trying to make. For example, the Big Bang Theory (not the TV show) was first posited by Alexander Friedmann in 1922 and in over a century of research and study, it remains unproven yet it is not assumed to be false or incorrect because it can neither be disproven.

There's been plenty of time to come up with some way to prove that there's a difference in the real and no one has succeeded. The FBI's expert told us that it was a waste of time to even try 35 years ago.

And how many studies of real-world shootings have we had in the last 35 years? By my count we have exactly two: Marshall/Sanow and Ellifritz. Of those two, Marshall/Sanow has credibility problems and Ellifritz has incomplete data, yet we're supposed to assume that something does or does not exist based to the result of two demonstrably flawed studies?

1. If someone was going to be able prove a difference in real-world shooting outcomes based on caliber differences in the service pistol class, they would have by now.

And how exactly would they do that? We have ample laboratory testing which does prove that there are measurable differences in penetration, expansion, and kinetic energy, but you and I both agree that the laboratory testing cannot predict whether those measurable differences will be significant in the real world, so how do you propose that we prove or disprove this other than study of real world shootings of which, I argue, none have been done with satisfactorily reliable data.

2. The fact that it's so hard to prove that such a difference exists is very good evidence that whatever difference may exist is so small that it's not worth bothering about and that other factors affect the outcomes of real-world shootings so strongly that they hide any effects due to caliber.

I would agree if we had good reliable data, but we don't. As we can agree that Marshall/Sanow's credibility issues render their data moot, we're left with Ellifritz. Ellifritz include so many uncontrolled variables into his data, due to the limited amount of data available to him, that we can't even come remotely close to isolating the significance, if any, that caliber has on real world shootings. Not only do we not know, but with two flawed studies being our only source of data, we're not even in a position that we could know.

Quote:
When compiling statistics, it is well-known that the larger the data set, the more uncontrolled variables or outliers, like differences in bullet type, will be mitigated.
That's not how it works. If you KNOW you have a bias in your data due to some over- or under-representation, you can't fix that by adding more data points that have the same problem. That just makes the problem worse.

The principle is that if you add more data points that don't have the same over- or under-representation, they will gradually dilute the effect of the biased data.

This is correct if you can isolate a single bias in your data. However, Ellifritz's data includes multiple uncontrolled variables which could cause unknown degrees of bias. For example, if you had two shootings and one was with FMJ which performed poorly but the other was with a JHP bullet which performed well, but the the FMJ shooting was from a subcompact pistol while the JHP shooting was from a full-sized gun, you have a possibility that the poor performance of the FMJ was due to the reduced velocity from the shorter barrel or the shorter sight radius contributing to poorer marksmanship. However, if you have hundreds of shootings, as Ellifritz did with his 9mm data, with a wide variety of different guns and shooters for both FMJ and JHP shootings, then the reduction in performance of the FMJ loading would be easier to isolate as the other uncontrolled variables are more mitigated by the larger data set.

Quote:
...then so long as both calibers had roughly equal proportions of shootings with FMJ ...
If we assume that, then we can compare the two data sets on an even basis--at least as far as the expanding vs FMJ basis is concerned.

But we would also have to assume that all the other uncontrolled variables are equally represented, which we can't because we don't have any data on that. Perhaps a better way to phrase it would be this: So long as both calibers had roughly equal proportions of shootings with FMJ ammunition, the larger data set of the 9mm would mitigate other uncontrolled variables to a greater degree than that of the .45 data set.

Quote:
OK, let's break this down a little. First we must assume that all police reports are accurate, which IMHO is a very big and erroneous assumption.
Your argument is that there will be errors in the data. That is EXACTLY the kind of thing that large data sets tend to minimize. As far as the police not knowing the caliber in some cases, that's grasping at straws. Those cases would, obviously, not be included in a shooting study designed to look at caliber differences. I mean, how would you even put the shooting into a caliber group if you didn't know what caliber was used?

Obviously a case in which the caliber is not known wouldn't be included in Ellifritz's study or one like it. My example about not knowing the caliber was simply meant to illustrate how a police report might not include pertinent information. How about this example, the police respond to a murder in which the victim was shot with a through-and-through shot so the bullet was not recovered, but they find Federal .45 ACP shell casings at the scene. In this case, it can reasonably be inferred that the victim was most likely shot with a .45 ACP, but they don't know what kind of gun was used or, because Federal makes a wide variety of .45 ACP ammunition, what type of bullet was used. Would such information be included in a study like Ellifritz's? I tend to think it would due to Ellifritz's own statements:

Originally posted by Greg Ellifritz
I documented all of the data I could; tracking caliber, type of bullet (IF KNOWN), where the bullet hit and whether or not the person was incapacitated. I also tracked fatalities, noting which bullets were more likely to kill and which were not. It was an exhaustive project, but I’m glad I did it and I’m happy to report the results of my study here.

Obviously Ellifritz included data in his study in which the bullet types was not known, so why is it unreasonable to think that other pertinent information might be missing in some of his data?

I'm seeing a lot of this in your arguments. I can tell you think you're being objective, but what you're doing is making a pretty serious, and in some cases (as with the "sometimes they don't know the caliber" claim) laughable efforts to dismiss any data that leads to an outcome you don't like.

And what outcome, pray tell, to you think I would like? I have not once said that a .45 ACP is significantly more effective than 9mm in real world shootings, only that whether it is or not is unknown based on a lack of data. As a matter of fact, I actually own more 9mm pistols than I do .45's and neither is my preferred defensive caliber, but my own personal preferences or theories on what makes an effective handgun cartridge are immaterial to the argument I'm trying to make which is that we simply don't have sufficient data to say whether there is a significant difference between the two calibers or not. You are claiming that, because no one has been able to prove that .45 is significantly better, we should assume that it is not while I am claiming that due to the extremely limited amount of reliable data, we shouldn't make any assumptions at all.

Quote:
Yes, but why don't they show up? Is it because there is no significant difference at all or is it because they're not compared using similar bullet types?
You're circling back. They don't show up because they're too small to show up in the presence of REAL-WORLD variability.

This is what I was getting at when I said it doesn't matter why.

You are trying to argue that other effects are hiding the caliber difference you assume must be present.

I don't assume that any caliber difference is present, but I don't assume that it's not either. You claim that the other variables have a greater effect of the outcome of real world shootings that caliber, but on what data is that claim based? Among the rationale of the FBI and many other law enforcement agencies for transitioning back to 9mm firearms is that the 9mm's lower recoil impulse makes it easier to shoot quickly and accurately than larger, more powerful calibers like .40 S&W or .45 ACP. If then caliber makes no significant difference then shouldn't the 9mm be demonstrably more effective than these other calibers? Ellifritz's data certainly doesn't show that because, as you've pointed out repeatedly, they all came out roughly equivalent in his tabulations. Is it not a possibility that the larger caliber's greater effectiveness cancels out the fact that they're more difficult to shoot? We don't know because we don't have sufficient data.

Let's assume, for a moment, that 9mm had a much greater proportion of shootings with FMJ ammunition than .45 ACP did. That would seem to suggest that the difference between FMJ and JHP ammunition has just as little, if not less, effect on the outcome of real-world shootings as caliber does. Do you believe that JHP ammunition is not significantly more effective than FMJ ammunition is? Alternatively if, as Ellifritz seems to believe, the 9mm data had included a significantly higher proportion of JHP shootings it would have scored better in his tabulations. This would seem to suggest that 9mm is significantly more effective than .45 ACP is which would still contradict your claim that caliber makes no significant difference in real world shootings.

OK. That's great. I'm not saying there's no difference. I'm saying there's no detectable/significant difference in real-world outcomes. That's what you're proposing as well, you're just trying to come up with reasons WHY it's not detectable to obscure the basic fact that it is not.

I'm pointing out that there are many very real reasons why the difference isn't detectable in an effort to show that just because something isn't detectable, that does not prove that it does not exist. You claimed that the reasons why there are no detectable differences do not matter and that is my primary point of disagreement: I think that the why does matter.

If you have to tightly control the shooting data to the point that it's artificially constrained, to see a difference, guess what--that's not a significant difference because shootings in the real world aren't tightly controlled.

How then can you claim that other variables have a greater effect on the outcome of real world shootings than caliber? Can we agree that 9mm generally has less recoil than .45 ACP does and, if so, that a 9mm pistol is generally easier for most people to shoot well than a comparable .45 pistol? Since Ellifritz's numbers showed that 9mm and .45 were roughly equal in effectiveness, would that not suggest that the enhanced marksmanship that a 9mm offers to most shooters is of no greater significance than the difference between calibers? Does the fact that despite the majority of the 9mm data was with FMJ bullets it still scored roughly the same as the other service calibers not suggest that bullet types is of no greater significance than caliber? You say that I'm trying to throw out data that I don't like to "prove" that my opinion is correct, but I would argue that you are ignoring a lack of data that does not support your own opinion.

And, by the way, you want to focus on bullet types and barrel lengths without being able to show that they have a significant effect on real-world shooting outcomes or that they are greater than the effect due to caliber. Again, this seems characteristic of grasping at straws.

So if barrel length and bullet type don't have a significant effect on real world shootings, what are these "other variables" that you claim do have a significant effect? Are guns with shorter barrels not often more difficult to shoot than guns with longer barrels? It would seem to me that if one set of data included a significantly higher proportion of short-barreled guns, then the effect on marksmanship would skew the data. The barrel length argument is also relevant because not all calibers are equally affected by it. As I pointed out earlier, calibers like .38 Special and .380 Auto are quite drastically affected by the loss in velocity from short barrels while 9mm is affected less.

You yourself stated that .380 Auto isn't included in "service pistol" class cartridges because it cannot reliably meet the FBI penetration and expansion standards with expanding bullets. How does this matter when, in Ellifritz's study of real world shootings .380 did not score demonstrably worse than "service pistol" class cartridges. Should we take from Ellifritz's data that the FBI's penetration and expansion standards are irrelevant?

The more uncontrolled variables you have, the more difficult it is to determine the significance, if any, that they have on the outcome. This is a big part of the reason why proponents of laboratory testing are often critical of studies like Marshall/Sanow and Ellifritz. Now, it's certainly true, as you pointed out, that the larger your pool of data, the less effect uncontrolled variables will have but you can't compare two pools of data which are drastically different in size and expect to have the same mitigation of uncontrolled variables. Because the 9mm data had over twice as many shootings as any other caliber for which Ellifritz compiled his statistics, we cannot assume that the other calibers had their uncontrolled variables mitigated to the same degree. As such, we can only conclude that these uncontrolled variables had greater effect on the .45 ACP, and all the other calibers, final tabulations than they did on 9mm.

Quote:
Again, this depends on the specific tissue that is damaged the the body composition of the person who is shot as even individuals of roughly equal weight can have very different body types.
Same thing again. This appears to be grasping at straws so you don't have to consider/can dismiss things that don't work with your desired outcome.

Take an average value and assume there's variation. The idea isn't that we're going to nail down these kinds of numbers to the nth degree, the idea is to get a rough feel for what's going on. Is it 10%? Maybe 30%? We know it's not 90%. Get as close as you can using the vast amounts of information on the internet.

Let's look at the specific question you asked:

1. How much weight of tissue is damaged, on average, by a single hit from a service pistol bullet. If you increase that by 27%, then what how much weight does it damage? What about if you double it, then how much weight is damaged? What are the percentages of those weights to a 180 lb body?

It takes only a basic knowledge of anatomy and physiology, which a college degree and over decade long career in healthcare has given me, to understand that different body tissues have different densities and thus a wound of a given volume will damage a varying amount of weight of tissue depending on which specific tissue is struck. Also, because body types and compositions can vary greatly from one individual to another, even a bullet which strikes the same general area of the body will likely damage a different weight of tissue depending on the body composition of the person who is struck. As a simple example suppose two people are shot in the abdomen, if one is 5'2" tall and weighs 180 lbs, that person is likely to have a significantly higher percentage of body fat than someone who is 6'4" tall but also weighs 180 lbs. Fat has a different density than muscle, bone, and various tissues from internal organs like the stomach, intestines, liver, spleen, etc. As such, the weight of the tissue damaged in the 5' 2" person will likely be quite different than the weight of that damaged by the 6' 4" person.

Even if we tried to calculate an "average" as you suggest, we'd have to have a large set of data for people of many different body types and know what internal organs, bones, muscles, etc. were damaged. If you have such specific information about a large number of shootings please share it, but if such information were available I don't think we'd be having this debate.

Same applies to the answers to your other responses. Besides, remember, I said the point wasn't to post an answer, it was to give you something to think about. You have successfully avoided thinking about them by raising objections that aren't material to the point.

Well, since your subsequent questions were based on the response to the first, if the first question can't be answered then what sort of response did you expect to the subsequent ones? What exactly are we supposed to be thinking about if not the amount of tissue damaged by a gunshot wound?

Quote:
I see the point you're trying to make here: there are too many uncontrolled variables other than caliber to definitively state that caliber is a significant factor in the outcome of real-world shootings and that is something I've never disagreed with.
Right. There is no evidence from real-world shootings to indicate that caliber choice in the service pistol class makes a difference.

So if I was able to see the point you were attempting to make, then how exactly did I avoid thinking about the question to begin with? Or do you think I didn't think about it sufficiently because, even though I understood your point, I still don't agree with it?

Quote:
The point I'm trying to make is that this cuts both ways and that, due to the lack of concrete data, we cannot rule out caliber as a significant factor either.
It cuts both ways, but not to the same extent.

I claim (accurately) that no one has proven that caliber choice in the service pistol class makes a significant difference in real world shooting outcomes.

And that is not a claim I've disputed. I agree that the difference, if any, between service calibers remains unproven, I am simply unwilling to assume that it is non-existent.

You claim (based on assumption) that it could make a difference but argue that there are other things affecting the outcome so much that the difference is obscured and therefore the real-world data is not useful and therefore my claim is problematic. It's not, because if it were significant (in any sense of the word) it would show up in real-world data unless the real-world data is actually flawed.

I claim that, due to the the various factors that we've discussed extensively, that the real-world data is in fact flawed. I fail to see how we can claim that a significant difference in the outcome of real world shootings does or does not exist when the only data we have to draw conclusions from is demonstrably flawed.

You are arguing that a difference might exist. I'm not disputing that--I will admit it could exist. I'm saying that no one can prove it is significant using real-world data--a claim that is not in dispute.

Thank you, that's the entire point I've been trying to make. There might be a significant difference and there might not, but we have no way to know one way or the other because all the available data is flawed.

You are arguing that other things obscure it and that's why it isn't detectable. That is consistent with my claim. I'm saying that no one can prove it's significant if they can't even detect it. Those two things are not mutually exclusive and yet you keep trying to make it sound like somehow my argument is flawed.

If this was your argument in its entirety, we would have no disagreement, however, such is not the case.

You want to keep arguing the details of WHY the proof that it could make a significant difference isn't there. I don't care because that doesn't call my premise in to question because if it's that hard to detect in the real world--it's not significant. Because, you know, words mean things and 'significant' is a word that has a meaning.

And here is the crux of our disagreement, I DO care WHY the proof isn't there. You claim that because the difference, if any, isn't detectable, it MUST be insignificant. The problem is that we don't know the significance of the other uncontrolled variables in the data, so we don't know to what degree they may or may not mask a difference due to caliber. How can we say that the difference in caliber is insignificant because other variables obscure it when we cannot define how significant or insignificant those other variables are.

You are working hard to dismiss the real-world data you have at your disposal. That is creative, but it won't help you make your point because to refute my point, you must have real-world data to work with.

Your point is based on an assumption: if the difference due to caliber can be obscured by other variables, then it must be insignificant. I will concede that the other variables may have a greater effect on the overall outcome than caliber does. However, just because one factor is smaller than another, that doesn't mean that it cannot be significant, because, well, you know words mean things and lesser does not necessarily mean insignificant. Falling off the roof a two story building probably won't injure you as badly as falling off the roof of the Empire State Building, but that doesn't mean that the injuries from the lesser fall won't be significant .

No one can show with real-world data that caliber choice makes a significant difference. (This can be refuted if someone can produce real-world shooting outcome data that are significantly affected by caliber.)

Based on the, IMHO, flawed data that is available, I do not disagree with this.

If caliber choice makes a significant difference in the real-world then real-world shooting data must show a detectable effect due to caliber. (This is a tautology--not refutable. If there's no detectable effect, it's not a significant effect.)

This I do disagree with because I fail to see how one can make such a claim when the only data available is deeply flawed. Your argument basically boils to to the claim that the difference due to caliber is undetectable therefore it must be insignificant. My argument is that we cannot now the significance or insignificance of a difference due to caliber because we don't have the sufficient means to detect it. History is replete with examples of erroneous beliefs that were held because the means of detecting the true explanation was unavailable at the time. It seems rather short sighted to me to believe that, because we cannot detect something with the means available to us at this time, then that thing must be insignificant or non-existent.

If you want to throw away the real-world data and explain WHY it doesn't show a significant differene due to caliber--feel free. That doesn't speak to either one of the sentences that compose my position.

If you want to ignore the numerous demonstrable flaws in the real world data that I've described extensively, you are free to do so. That doesn't change the fact that the data is deeply flawed regardless of whether it's the only data we have or not.

Quote:
And when, exactly, have I ever stated that bullet differences don't make a bigger difference than caliber? That's actually kind of the whole point I've been trying to make: you can't make an accurate comparison between two calibers unless you compare similar bullet types.
1. It's interesting that you are willing to assume that bullet differences and barrel length differences make a difference in the outcome of real-world shootings while arguing that we shouldn't assume things in the absence of evidence.
2. If things like a few FPS of velocity due to barrel length or the choice of bullet make more difference then caliber, how is that an argument for caliber differences having a significant effect on real-world shooting outcomes? Your argument is that you assume these effects are greater than the effect due to caliber so they are hiding the differences due to caliber that you assume exist and therefore we can't use real-world shooting outcomes as evidence.

I'm not making any assumptions at all. My argument is that we do not know what the significance of any of these uncontrolled variables is so we cannot assume that one being masked by another demonstrates that the first is insignificant. In order to determine the significance of a variable on the outcome of a study, that variable needs to be isolated at least somewhat, but the data we have available doesn't allow us to do that.

I'm ok with that. Because although it takes the long way around and relies on some assumptions, it does nothing to invalidate my claim that there's no evidence that real-world shooting outcomes are affected by caliber differences in the service pistol class.

I'm afraid it's you that is making assumptions here. You assume that I'm trying to intentionally ignore data that doesn't fit my opinions or preferences when that is not the case. As a matter of fact, my preferred defensive handgun caliber is .357 Magnum which actually did fare noticeably better in Ellifritz's study so, had I a bias, Ellifritz's work would support rather than contradict it. You also assume that, because no one can detect something, that means that it must be insignificant, an intellectual leap I'm not willing to make. Finally, you say that I'm ignoring and dismissing data that does not support my imagined agenda yet you continue to ignore and dismiss the numerous flaws in the data upon which your position is based. You are assuming that because no one has yet been able to detect a difference due to caliber, then any such difference must be insignificant while my position is that, because the available data is deeply and demonstrably flawed, we have no way to know how significant the effect of differences in caliber might be.

[Koda94]

Improvement in bullet technology has not favored the 45. If there was a measurable difference in stopping power it would be detected regardless of the bullet used as long as the same type was used.

[Sharkbite]

Where you hit is WAY more important than what you hit with. An extreme example of this would be a 22lr through the eye socket face-on would be more incapacitating then a 44mag to the big toe.

A 9mm to the sternum bisecting the heart is going to cause a physiological stop sooner then a 45 through the outside of the thigh (and vice-versa).

So if shot placement trumps “power factor” (whatever that is), why not carry what you shoot best (accuracy, recoil control, etc) and what gives you the most ammo on board. Assuming adequate penetration.

[sigarms228]

And the decades long debate yet again rages on!

[Webleymkv]

Originally posted by 44 AMP
I generally agree with this, but I need to point out that people have only been getting shot with expanding bullets in either caliber for about the past 50 years or so. AND that one can ID several generations of improvement in bullet performance during that time.

Many of the JHP bullets from the early decades of their use did not expand reliably, or consistently in all circumstances.

Also note, that the 9mm JHP that "failed" in the 1986 Miami shoot out met every requirement the FBI had, at the time.

The FBI has since changed their standards...

The advances in bullet technology would likely result in less difference between calibers than more and the FBI is a big part of the reason behind this. After the FBI changed their standards in the wake the the 1986 Miami shooting, the majority of work in bullet design has been to make all calibers more consistently meet these standards. As such, today's JHP bullets perform more alike across calibers than those of yesteryear because they're all designed to do the same thing: meet the FBI standards. Of course, there are still differences between what a 9mm and a .45 will do with the same bullet in terms of expansion and penetration, but those differences are smaller than they used to be.

That being said, not everyone is using the latest bullet technology and not every caliber has received the same "upgrades", so once you move outside of 9mm, .40 S&W, and .45 ACP, you might be more likely to see more pronounced difference because fewer of the bullets in other calibers are designed with the same "goal" of meeting the FBI standards.

I appreciate the discussion of how the statistics work, but what I am not seeing is how one can "boil down" (ignore??) all the real world factors that can cause differing results, often significant ones.

How can you not take into account hits in non vital areas, attackers ranging from 100lbs up to possibly 300lbs? People shot multiple times before falling down? Things like that, and more, all rolled up into one data point, how can that possibly a valid, accurate representation of the real world performance of an individual bullet or load???

Honestly, I agree with all of this. While I think that real world data is useful, I don't really think that trying to distill it down into a "one shot stop" percentage or some other statistic is the most useful way to use it.

Generalities, sure, but detailed percentage points, or fractions of a percent? I don't get it. I keep coming back to the experience of a guy who had to defend himself with GI FMJ ammo, once with a .45 and once with a 9mm.

He felt they were the same, as he puts it, "I shot him twice, he fell down. Both times."

Right now, I'm thinking we get over involved with the complexities of statistics, that are actually oversimplified data from real world shootings.

Again, I generally agree. I personally think that a more useful approach is something of the inverse of how it's been done for the last 35 years or so. Now, let me preface this by saying that what I'm about to state are simply theories that I have based on my own observation and reading over the years and I make to claim to have scientific data to back them up, again, these are just my own theories and anyone reading this can consider them and/or dismiss them as they see fit.

An approach that I think might be more useful is to look at what specific loadings have widespread reputations for being particularly effective, comparing these loadings through laboratory testing, and looking for trends or similarities among them. If you're able to establish a trend or similarity, then investigate into whether you can make a loading which has similar attributes in other calibers.

For example, the various .357 Magnum 125 gr semi-jacketed hollowpoint loadings, Federal's and Remington's in particular, have reputations for being particularly effective in gunfights. Now, while this reputation is, by definition, anecdotal, the fact that it was so widespread among so many different police agencies which used these loadings makes it a bit more reliable than anecdotes like "my grandpa shot a Jap in the hand with his .45 and he fell over dead." Also, because this reputation comes primarily from officer-involved shootings, we know that the majority were from medium or large frame service revolvers with 4" or longer barrels. We also know that, because the majority of the shooters were cops, the majority had received some degree of firearms training and thus it would be reasonable to expect that variances in marksmanship would be less impactful than amongst the public at large.

OK, so we have data on some similar loadings in one caliber (.357 Magnum), can we find any loadings in other calibers with similar reputations which were used under similar circumstances (police shootings, full sized guns). As it happens, we can as both the Federal 9BPLE and Winchester Ranger 9mm +P+ 115 gr JHP loadings have similar reputations and were also used by large police agencies such as the Illinois State Police and DeKalb Count GA Sherriff's Department for many years. Like the .357 Magnum, we know that the great majority of these shootings were from cops and most were from service pistols with 4" or longer barrels.

Now that we have two loadings with similar reputations under similar circumstances, let's look at laboratory testing on both and see if there are any similarities. Low and behold there are: both loadings seem to expand very aggressively to the point of moderate, though not excessive, fragmentation and penetrate in the neighborhood of 10-12" in ballistic gelatin. It would seem that a bullet which behaves this way correlates to an effective loading in the real world.

The next step would be to determine why these particular bullets behave the way that they do. Well we know that lower velocity 9mm loadings with similar bullets such as the Federal 9BP standard pressure 9mm 115 gr JHP don't behave this way and we also know that .38 Special loadings with similar 125 gr SJHP bullets don't behave the same as the .357 Magnum loadings do. Given the very similar, if not identical, bullet construction between the low velocity and high velocity loadings we have to assume that the relatively high velocity is a large part of these loadings' expansion and penetration characteristics. So what other similarities can we find? Well, as it turns out, they both the 9mm +P+ (115 gr at 1300-1350 fps) and .357 Magnum (125 gr at 1400-1450 fps) have roughly the same ratio of velocity to cross-sectional density (approximately 10,000 to 1). So what can we do with this information?

Well, if you wanted to make a loading in another caliber which performs similarly, it would seem reasonable to drive a bullet of similar construction (cup-and-core JHP) to a velocity which achieves the same velocity to cross-sectional density ratio. However, this isn't necessarily possible in every caliber without going to extremely light-for-caliber bullets which don't seem to have particularly good reputations and are known to perform poorly due to extremely shallow penetration. So, it would seem that we are restricted to calibers that have pressure limits high enough to drive a sufficiently heavy-for-caliber bullet to high enough velocity to get the performance we're looking for. For example, you're probably not going to be able to get there with a .38 Special because you have to drop down to a bullet of 110 gr or less to be able to reasonably drive it fast enough and stay under the SAAMI pressure limit. However, a 155 gr 10mm at 1400 fps or a 180 gr .44 Magnum at 1400 fps would give you the 10,000 to 1 ratio you're looking for and would be easily achievable within the pressure limits of those cartridges so, if you used a similar cup-and-core bullet I think it's reasonable that you're likely to achieve similar performance to the .357 Magnum and 9mm +P+ loadings and possibly even greater effect due the the greater energy possible with 10mm and .44 Magnum.

This is all, however, a theory as I've stated before and not one which I've seen tested nor do I have the means to test myself, so take it for what you will.

[Super Sneaky Steve]

The debate was settled back in 1846 with the introduction of the Walker pistol. 45 is just better.

[Pumpkin]

Enter, the .357/125 Sig. So much for anecdotal data (357/125 Magnum) not being respected enough by a large manufacturer to make a large investment in a new caliber.

If any police cartridge had a chance to break the 9mm +p+ or 40S&W hold on the police caliber market it would have been the Sig and I’m pretty sure lab testing did not disappoint.

Add to all these positives, a bottleneck case design to aid in eliminating feeding issues and the deal should have been closed.

So much for that.

[stagpanther]

I put together a 9 x 25 dillon based on a Glock 20--similar concept to the 357 sig but significantly more powerful than the sig or even 38 super--if you want to shoot 9mm really fast--that's the way to go. Pain making the brass though.

[Pumpkin]

Originally Posted by stagpanther

I put together a 9 x 25 dillon based on a Glock 20--similar concept to the 357 sig but significantly more powerful than the sig or even 38 super--if you want to shoot 9mm really fast--that's the way to go. Pain making the brass though.

stagpanther,

Was this the same round you were using in an AR for deer hunting?
I seem to remember reading about this?

[stagpanther]

stagpanther,

Was this the same round you were using in an AR for deer hunting?
I seem to remember reading about this?

It's the round I tried to do an AR build for, but I couldn't get a barrel maker to make a barrel for me. I have built blowback ARs in both 9mm and 10mm so I thought it would be possible, but I guess not. Even my 8.something" Wolf glock barrel gets pretty heady velocities and energy with the dillon.

[Pumpkin]

I’ve read up on the ballistics, it’s a bad dude!

[JohnKSa]

I appreciate the discussion of how the statistics work, but what I am not seeing is how one can "boil down" (ignore??) all the real world factors that can cause differing results, often significant ones.

That's exactly the point. Those real world factors affect the outcome far more strongly than differences due to caliber. That's precisely why no one has been able to prove that caliber makes a significant/practical difference in the outcome of real world shootings.

The only point I'm trying to make is that we can neither assume that there is or is not a difference.

Actually, that's not the only point you're trying to make.

For one thing, I'm not assuming there is no difference. I'm saying that no one has been able to demonstrate that there is a significant/practical difference in shooting outcomes due to terminal performance differences related to caliber selection in the service pistol class. Because, you know, that statement is 100% true.

You are coming up with all kinds of reasons why that is true, why it's so hard to detect the difference. What you don't understand is that every time you provide a reason why it's so hard to find the difference in the real world, you are actually emphasizing THAT the statement is true and explaining why it is extremely likely to remain true for all time.

Ellifritz include so many uncontrolled variables into his data, due to the limited amount of data available to him, that we can't even come remotely close to isolating the significance, if any, that caliber has on real world shootings.

I agree. All those uncontrolled real-world factors affect the outcome of shootings so strongly that "we can't come remotely close" to being able to detect the difference (if any) that caliber has on real-world shootings.

This is correct if you can isolate a single bias in your data.

And that's exactly what you were trying to do--isolate the single bias due to the over-representation of FMJ in the 9mm data set.

However, Ellifritz's data includes multiple uncontrolled variables which could cause unknown degrees of bias. For example, if you had two shootings and one was with FMJ which performed poorly but the other was with a JHP bullet which performed well, but the the FMJ shooting was from a subcompact pistol while the JHP shooting was from a full-sized gun, you have a possibility that the poor performance of the FMJ was due to the reduced velocity from the shorter barrel or the shorter sight radius contributing to poorer marksmanship.

Excellent points, all. There do seem to be many, many effects that are going to affect the outcome far more strongly than caliber.

However, if you have hundreds of shootings, as Ellifritz did with his 9mm data, with a wide variety of different guns and shooters for both FMJ and JHP shootings...

His concern was that his entire 9mm data set had FMJ over-represented in it. That makes its size irrelevant. A big data set with FMJ over-represented is still a data set with FMJ over-represented. To fix the problem would require adding valid data points where FMJ was NOT over-represented to dilute the over-representation, or somehow culling the data set to eliminate the over-representation. Of course, throwing away data is always problematic, so the former approach would be much better.

But we would also have to assume that all the other uncontrolled variables are equally represented, which we can't because we don't have any data on that.

Those uncontrolled variables are the REAL WORLD. That's the whole point. The real world is uncontrolled.

What you're saying is that to see what happens in the real world, we have to first control what happens in the real world. That's nonsense. If you control all those variables, it's no longer the real world. You've now generated another bunch of data like all the gel data we have. It makes for good comparisons but no one can tie it to what happens in the real world.

You claim that the other variables have a greater effect of the outcome of real world shootings that caliber...

You are too. That's practically all you've done on the thread. You keep explaining this and that effect that have to be eliminated before we can see caliber differences and then referring to the rest of them as uncontrolled variables.

Let's assume, for a moment, that 9mm had a much greater proportion of shootings with FMJ ammunition than .45 ACP did.

I already did that thought experiment when you first brought it up. Assuming that the over-representation of FMJ made the 9mm data set results worse than they should be, actually makes the 9mm numbers MORE similar to the .45ACP numbers, not more different. You didn't like that result. Remember?

Alternatively if, as Ellifritz seems to believe, the 9mm data had included a significantly higher proportion of JHP shootings it would have scored better in his tabulations. This would seem to suggest that 9mm is significantly more effective than .45 ACP is which would still contradict your claim that caliber makes no significant difference in real world shootings.

So you're assuming that the elimination of the over-representation in the 9mm data would make the 9mm data MUCH better? I figured, based on the generally small differences between calibers in the data set, even between quite disparate calibers, that if the FMJ issue made any difference at all it would be a small difference. I don't see any justification for assuming it would make a big difference.

I'm pointing out that there are many very real reasons why the difference isn't detectable in an effort to show that just because something isn't detectable, that does not prove that it does not exist.

I know all about the reasons. That's why I am confident my claim is correct. And, again, my claim isn't that the difference does not exist; I claim that no one has been able to prove that it is significant in the real world. You know, because they haven't.

How then can you claim that other variables have a greater effect on the outcome of real world shootings than caliber?

You're claiming that. You keep saying we have to eliminate them to be able to see the difference due to caliber. For what it's worth, I think that's a good point and I agree with you.

You yourself stated that .380 Auto isn't included in "service pistol" class cartridges because it cannot reliably meet the FBI penetration and expansion standards with expanding bullets. How does this matter when, in Ellifritz's study of real world shootings .380 did not score demonstrably worse than "service pistol" class cartridges.

What? Who said it mattered? Go back and read what I wrote. This is starting to seem a lot like a Gish Gallop.

It takes only a basic knowledge of anatomy and physiology, which a college degree and over decade long career in healthcare has given me, to ...<redacted for the sake of space> ... and various tissues from internal organs like the stomach, intestines, liver, spleen, etc. As such, the weight of the tissue damaged in the 5' 2" person will likely be quite different than the weight of that damaged by the 6' 4" person.

You can't ignore the word "average" in my question and expect your answer to make sense. What a total waste of your time.

If you have such specific information about a large number of shootings ...

I made no claim to have "specific information about a large number of shootings." Nor did I make any claim that could reasonably be interpreted to imply anything remotely similar to that. The question I asked is actually very simple.

I am simply unwilling to assume that it is non-existent.

Again, I'm not claiming it's non-existent. Please read what I've written.

You claim that because the difference, if any, isn't detectable, it MUST be insignificant.

If the difference isn't detectable in the real world, how is it helping anyone in the real world? If it isn't helping anyone in the real world, how could it be significant?

If you claim it is detectable in the real world, please provide data to support the claim.

If you can not show that it is detectable in the real world, please explain how an effect could help someone in the real world and remain undetected.

If you can not show that it has helped anyone in the real world, please explain what significance it has.

Falling off the roof a two story building probably won't injure you as badly as falling off the roof of the Empire State Building, but that doesn't mean that the injuries from the lesser fall won't be significant

Do you understand the meaning of detectable? If the injuries are detectably different, then how does this analogy relate in any way to anything I've said?

This I do disagree with because I fail to see how one can make such a claim when the only data available is deeply flawed.

I get it, you don't like the data. I wouldn't either in your position.

But the fact remains that you have no real world data with which to refute the claim. That's no reflection on you--the fact is that no one has such evidence. In fact, in the 35 years since Urey Patrick essentially said it was a waste to try to show caliber difference with real world data because any difference would be obscured by uncontrolled variables, no one has proved him wrong.

More than that, in all the years that people have been arguing about caliber, and it's more than just 100, by the way, no one has been able to come up with real world data to refute the claim.

It seems rather short sighted to me to believe that, because we cannot detect something with the means available to us at this time, then that thing must be insignificant or non-existent.

You keep wanting to throw in the strawman.

"The means available to us at this time" is whether or not anyone can show people being helped by terminal differences due to caliber choice in the service pistol class. Which is, actually, a VERY useful and relevant means of detection since that's what everyone who is interested in this topic cares about. Any way you slice or dice it, it comes down to one inescapable fact. If no one can show that it's helping people then it's not significant. Period. No matter how many novels you write on the topic.

An approach that I think might be more useful is to look at what specific loadings have widespread reputations for being particularly effective, comparing these loadings through laboratory testing, and looking for trends or similarities among them. If you're able to establish a trend or similarity, then investigate into whether you can make a loading which has similar attributes in other calibers.

This is more or less what M&S were doing. They were looking for loads that performed well and then trying to correlate that to something measurable in the laboratory. They claimed that their results correlated very well with kinetic energy, something that was intolerable to the Facklerites who made it their business to destroy the credibility of M&S and their data. To be fair, M&S were easy targets because they were inexpert in the gathering and manipulation of the data sets they were dealing with and, at the least, made mistakes that were easily exploited.

[44 AMP]

While I am personally not qualified to judge the accuracy or validity of statistics, from my point of view in the peanut gallery (or out in the weeds if you prefer) it seems that sometimes decisions get made based on certain numbers in statistics that don't pan out the way expected due to other factors not accounted for in "stopping power' statistics.

As an example, I point to the FBI and their thumbs up, thumbs down, thumbs up again, thumbs down again, over various rounds over the last 40 years.

The 9mm is great! Oh, wait, the 9mm isn't good enough! The 10mm is great! oh, wait, the 10mm is too much,, the 10lite is great! oh, wait the guns are still too big! The .40 is great....oh wait, you've improved the 9? The 9mm is great, again.....

Wonder what the next chapter will be???

[Pumpkin]

Not sure there’s enough ink

[Webleymkv]

OK, in an effort to keep this thread from turning into any more of a "wall of text" than it already has, I will no longer rebut each and every statement. That being said, here's what I have to say further on the matter:

We have no way of determining to what degree, if any, caliber effects the outcomes of Ellifritz's study because there are numerous other variables that Ellifritz either did not know or does not provide the data on in his write-up. We cannot say that the effect of caliber is insignificant because we do not know how significant the effects of other variables for which we do not have data are on the ultimate outcome. Even if caliber is a significant factor in the outcome, the other variables for which we have little to no data may simply be more significant to the point that they obscure the effect of caliber.

We cannot claim to know whether a variable like caliber or bullet type plays a significant role in the outcome of a shooting unless that variable can be somewhat isolated. Given the possibility that all the variables in a shooting have the potential to affect the outcome, we cannot even begin to know how significant the effect of one variable like caliber is unless we either already know how significant the effects of the other variables are (we don't) or we can segregate our data points sufficiently to make the one variable we're trying to determine the significance of the only uncontrolled variable (Ellifritz didn't and/or couldn't).

Now, while it is true that uncontrolled variables can be mitigated to greater degrees with increasing sample size, this is only useful in comparing two sets of data if both sets of data are of at least approximately equal size or both sets of data contain at least approximately equal proportions of uncontrolled variables. Due to the numerous uncontrolled and, in some cases, unknown variables in Ellifritz's data, comparing the 9mm results to the .45 ACP results isn't useful because they are not of approximately equal sample size (9mm had over twice as many shootings) and Ellifritz doesn't/can't tell us what proportion of the data from at least the .45 date if we're talking about bullet type and both sets if we're talking about the others is affected by uncontrolled and/or unknown variables.

Quite frankly, the more I think about it, the more I think that trying to boil data from real-world shootings down to something as simple as a "one shot stop" percentage is not only extremely difficult or impossible given the quality and quantity of data available, but is an over-simplification and not really the best way to utilize the data in the first place. While "one shot stop" percentages and other calculations are attractive because they're simple and easily understood, real world shootings are often not simple or easily understood and I'm really not sure that we can quantify them nor that we should try.

Now, while I said I'm not going to try to address or refute each and every statement tit-for-tat, there are a few things I do want to address more specifically:

The insinuation that I have some sort of agenda or that I'm trying to ignore/cherry-pick the data to fit my own preferences/opinions/hypotheses is becoming rather tiresome. I have not once claimed that any caliber is demonstrably "better" or "more effective" than another, rather I have claimed that we do not have sufficient data to determine what the significance, if any, of caliber is on the outcomes of real world shootings. Not only that, but I own and have relied upon for the defense of myself and my family in some capacity, a handgun (and in some instances multiple handguns) chambered for every single handgun cartridge included in the Ellifritz study as well as a few other calibers that were not included. I'm not sure how I could me any more "caliber agnostic" nor do I know how I'd ignore/cherry-pick the data to fit my "agenda" when I don't even know what that "agenda" might be.

The claim that I'm engaging in "straw man" arguments is ironic to the point of hilarity because the statement used to claim this was taken completely out of context. If taking someone's statement completely out of context in order to refute it isn't a strawman, well I don't know what is.

Finally, there is one particular statement that I will address very specifically:

Originally posted by JohnKSa
Quote:
It takes only a basic knowledge of anatomy and physiology, which a college degree and over decade long career in healthcare has given me, to ...<redacted for the sake of space> ... and various tissues from internal organs like the stomach, intestines, liver, spleen, etc. As such, the weight of the tissue damaged in the 5' 2" person will likely be quite different than the weight of that damaged by the 6' 4" person.
You can't ignore the word "average" in my question and expect your answer to make sense. What a total waste of your time.

No, John, your original question was a waste of time and so was trying to pick apart my answer and the explanation for why I answered it in the way I did. You specifically asked for the average weight of the tissue which would be damaged by a handgun bullet. Now, I don't know what you do for a living but as I mentioned I have spent many years working in healthcare and, while I will not post such information on a public forum, I can provide my credentials via PM if you feel the need to see them.

If you had done and seen the same things as I have, you would know that the sorts of things you ask for are not typically measured by weight. If, for example, you were to read a post-surgical note, of which I've read hundreds if not thousands, you would know that the measurements are usually units of volume rather than weight: milliliters of blood lost, excised a 4 centimeter by 6 centimeter tumor, etc. It isn't quite as simple as saying "on average, a 180 lb individual will have x grams (medicine primarily uses metric measurements) of tissue damaged by a gunshot wound" because such things simply aren't measured that way. In the real world, a gunshot wound would be described as x millimeters wide by y centimeters deep and what anatomical structures were impacted.

So, in order to answer your question, it would be necessary to convert the measurements that are actually taken of such things (units of volume) into the units of measurement that you asked for (units of weight) and such conversions are not possible without information such as specific tissue densities. Also, unless you have specific data about a shooting on one individual who is considered "average" (and extremely few people are "average" in every anatomical/physiological aspect), you would need to collect data on a large number of shootings, convert the units of volume into units of weight for each individual shooting, and then compute an average. Oh, and by the way, such information isn't all that readily available among the "vast amounts of information on the internet" due to privacy and confidentiality laws like HIPPA. If I wasted my time by explaining this, then it was because I was trying to answer a question asked by someone who didn't understand what would be involved in the answer to begin with.