Accuracy Measurement Question

[EMTDAD]

I see a lot of posts about the "accuracy" of a specific firearm. Typically, these include all kinds of measurements of "tight" groups, and MOA (whatever that is), and various other "statistics".

As a scientist, I'm taught that in order to do a fair analysis, as many variables as possible must be controlled.. I'd assume, that in testing firearms for accuracy, that means taking out any wind effects, shooting rifles from a bench rest to remove human "error", and getting a big enough sample size to come to a definitive statistical conclusion.

How can someone say handgun X is more accurate than handgun Y, just by shooting a box or two of ammo... I can see saying something like "I'm more accurate with X than Y", but can't quite figure out if that's what folks mean.

Please educate the noob in me. Thanx.

 

[RImike]

Nah, I go out when it's blow dogs off chains and it's 10 degrees and compare my results taken from last August.


So my sarcasm meter was quite pegged there, sorry about it.

I can't speak for most people, but for zeroing in yes, you want ideal conditions and are looking for groups. In real world conditions, part of the enjoyment is factoring in wind, range and other environmental conditions.

There is two types of accuracy, accurate guns and accurate people. If you want an accurate gun, there is this:

 

[jasons]

I see a lot of posts about the "accuracy" of a specific firearm. Typically, these include all kinds of measurements of "tight" groups, and MOA (whatever that is), and various other "statistics".

As a scientist, I'm taught that in order to do a fair analysis, as many variables as possible must be controlled.. I'd assume, that in testing firearms for accuracy, that means taking out any wind effects, shooting rifles from a bench rest to remove human "error", and getting a big enough sample size to come to a definitive statistical conclusion.

How can someone say handgun X is more accurate than handgun Y, just by shooting a box or two of ammo... I can see saying something like "I'm more accurate with X than Y", but can't quite figure out if that's what folks mean.

Please educate the noob in me. Thanx.

If you really care you put it in a ransom rest.

 

[PaulR]

MOA=Minute of Angle. Must admit I don't really understand it either, I just like to shoot at targets and adjust my scope accordingly with those two knob thingies.

 

[milktree]

I see a lot of posts about the "accuracy" of a specific firearm. Typically, these include all kinds of measurements of "tight" groups, and MOA (whatever that is), and various other "statistics".

As a scientist, I'm taught that in order to do a fair analysis, as many variables as possible must be controlled.. I'd assume, that in testing firearms for accuracy, that means taking out any wind effects, shooting rifles from a bench rest to remove human "error", and getting a big enough sample size to come to a definitive statistical conclusion.

How can someone say handgun X is more accurate than handgun Y, just by shooting a box or two of ammo... I can see saying something like "I'm more accurate with X than Y", but can't quite figure out if that's what folks mean.

Please educate the noob in me. Thanx.

"MOA" is "Minute Of Angle". 1MOA is 0°1'0" of arc. It's about 1.08" at 100 yards. It's the standard way to describe the accuracy of a gun.
There's lots of ways people do that measurement. Obviously one shot is useless, and two is a poor data set. Most people shoot five shots in a row without changing the point of aim and measure either the largest distance between two hole centers, or the diameter of the smallest circle that encompasses all the shots. There's tomes written on the "right" way to measure groups.

When I "test" a rifle (I'm not that scientific about it) I put the rifle on the most sturdy bench with the best props (sandbags/bipod, whatever) and use a quality scope, and shoot at 200Yards. I take my time so I can eliminate things like jerking the trigger or having my eye not quite in the center of the scope or accidentally being off my aim point, or shooting during a wind gust. A better way is with a vice, or some other repeatable method that takes the human out of the equation.

To further complicate things, a handgun might be more accurate with brand "A" than brand "B", while another might be better with "B" than "A"

 

[jasons]

MOA=Minute of Angle. Must admit I don't really understand it either, I just like to shoot at targets and adjust my scope accordingly with those two knob thingies.

There are 60 minutes in one degree. One minute is roughly one inch at 100 yards, 2 inches at 200 yards, etc. At 5 inch group at 50 yards is pretty suck, that same 5 inch group at 1000 yards is pretty awesome. Really just a way of stating the accuracy of a gun relative to the distance it's fired from.

 

[BTSDOG]

This explains MOA.......
http://dexadine.com/WhatMOA.htm

 

[jasons]

"MOA" is "Minute Of Angle". 1MOA is 0°1'0" of arc. It's about 1.08" at 100 yards.

I think it's actually more like 1.047" but who's counting.

 

[RImike]

Also the term sub-MOA is quite common on long guns these days. I have a sub MOA AR that shoots average .75" groups at 100 yards.

 

[mac1911]

Also the term sub-MOA is quite common on long guns these days. I have a sub MOA AR that shoots average .75" groups at 100 yards.

OP is scientist.......so what did you use to come to your average. how many groups, how many shops per group and so on.
OP as far as some manufactures of ammo and firearms they will do a lot of median averages of x amount of test dropping the worst X % then factoring in the average.

For example 30-06 M2 ball military ammo is tested from a accuracy device....much like the bench gun in the previous post. from that testing device they will shoot X groups Y amount of shots per group and the median average can not exceed Z inches at A yards? which for M2 ball comes to about 4 moa-ish

same forb some rifles.... M1 garands are considered good if they hold a 1.77" group at 1000" for 6 consectutive shots with in 2 attempts

My rock river national match comes with a .75" moa accuracy claim. on their testing equipment im sure and possibly federal gold medal match ammo?

 

[RKG]

Accuracy can mean a number of things. Most commonly, when used with respect to rifles (and, to a lesser extent, handguns) is means “machine accuracy” or the capacity of the firearm to put repetitively fired rounds as close as possible to the same point of the target.

In theory, one would like to bound as many variables as possible, both those entirely extrinsic to the device under test (e.g., wind) and those not entirely so (e.g., barrel temperature). If we limited our shooting to perfectly calm days, there wouldn’t be an ammo shortage. And if we waited 30 minutes between shots in a string, we’d fall asleep and be accused of violating the firearms storage rules. So, since we are engineers and not scientists, we try to minimize the big stuff and understand that we can’t be perfect. (What’s an engineer? A guy who understands the value of the precept, “I’ll get close enough.”)

In theory, the “right” way to assess machine accuracy starts with identifying the spot on the target that exactly corresponds to the point of aim. Note that this is not necessarily, or even likely, either the geometric center of the group or, more commonly measured, the mid-point of a line drawn between the centers of the two apparently widest holes. We would then measure the distance between this real point of airm and the center of each hole on paper. If we then take the average of these radii, we can calculate a mean rifle-induced error for the sample of shots fired, and if we take the standard deviation of these radii, we can begin to get some handle on the tendency of the rifle (under perfect conditions, of course!) not ever to stray more than X distance from the point of aim. If our SD is reliable, and if we further assume that the factors that contribute to the slugs hitting anywhere but “in the middle” are normally distributed, then we can say that the accuracy of the rifle is equal to the mean plus about two times the SD.

We don’t do that. We take the mean of the radii and ignore its variability. Moreover, we don’t fire enough rounds for the statistics regarding the sample to become reliable measurements of the population (i.e., all the shots that will ever be fired through the rifle). If you apply MilSpec 105D to the task (a bit of a mismatch), group size should be 32 rounds (or the earlier point where SD becomes asymptotic to some value). Now we’re going to cause an ammo shortage (as well as go broke).

What does all this mean in the real world? Actually, not much. If we fire 5 reasonably well aimed shots with each of two rifles, under reasonably benign conditions and taking a minute or two between shots, and if after doing that we find that one rifle makes something like a single ragged hole and the other looks like a blast of #8 from a .410 shell, we can comfortably say that one rifle is probably more accurate than the other. Which in the real world is probably close enough.

 

[RoterTeufel]

Let's make sure we're all on the same page & using the right words...

Easier to correct inaccuracy than poor precision, IMHO.
If you have a good group, but are off target, it may be as simple as adjusting you POA, or your sights.

 

[economist]

Something I wrote up years ago when I must have been really bored:

http://www.northeastshooters.com/vb...g-Accuracy-why-I-have-stopped-shooting-groups

 

[mac1911]

I shoot a lot or old surplus rifles and often surplus ammo.
I judge gun ,ammo and myself as doing well when I can keep all my shots in the 6"ish black bulls eye
of a SR 1 target at 100 yards shooting prone,sitting and standing.
Some of my rifles are more than accurate enough to keep them in the 10 ring.
My AR and match loads should pound the X ring but I hold it back from doing so.

Now there are other things also as some have said better than I can. For example the M14 national match rifle standard below
M14 NM Rifle
The M14 National Match rifle is the same basic design and operation as the standard issue M14 rifle, except for the elimination of automatic fire capability. Each M14 NM rifle is required to fire 50 match rounds without a malfunction during targeting and accuracy tests. The extreme spread of these groups cannot exceed 3.5 inches at 100 yards

Now some will say that's not that great but that's what they (the folks building them) decided was a good bench mark.

 

[Varmint]

I see a lot of posts about the "accuracy" of a specific firearm. Typically, these include all kinds of measurements of "tight" groups, and MOA (whatever that is), and various other "statistics".

As a scientist, I'm taught that in order to do a fair analysis, as many variables as possible must be controlled.. I'd assume, that in testing firearms for accuracy, that means taking out any wind effects, shooting rifles from a bench rest to remove human "error", and getting a big enough sample size to come to a definitive statistical conclusion.

How can someone say handgun X is more accurate than handgun Y, just by shooting a box or two of ammo... I can see saying something like "I'm more accurate with X than Y", but can't quite figure out if that's what folks mean.

Please educate the noob in me. Thanx.

A scientific test should reflect reality or it doesn't have much use. Part of what makes a pistol accurate are how it feels, how it handles recoil, the grip, the trigger, how easy the sights are to use. If you set up a test with the gun anchored to a bench with a robot pulling the trigger, you're eliminating most of those factors, and the test is meaningless.

I see scientists making this mistake all the time - they focus on how to set up the test, and record and analyze the data, never realizing their entire test is producing meaningless data because they made a mistake in some basic premise.

 

[EMTDAD]

A scientific test should reflect reality or it doesn't have much use. Part of what makes a pistol accurate are how it feels, how it handles recoil, the grip, the trigger, how easy the sights are to use. If you set up a test with the gun anchored to a bench with a robot pulling the trigger, you're eliminating most of those factors, and the test is meaningless.

I see scientists making this mistake all the time - they focus on how to set up the test, and record and analyze the data, never realizing their entire test is producing meaningless data because they made a mistake in some basic premise.

I disagree with the first part of your statement... how it feels, etc.. makes the SHOOTER more accurate, not the pistol. There's a difference.

Take away the variances in the shooter's ability, and environmental variables, and the pistol should theoretically put every bullet in the same hole each time. The laws of physics that govern how a bullet gets from point A to point B don't change. It's the variances in the initial conditions (and to some extent the environment) which alter the end result. If you set up a pistol to hit a specific point in a controlled environment, and automatically fire, and it doesn't repeat the result to within an acceptable error, then the pistol would be inaccurate, mostly due to the variability in the machining/assembling of the unit, assume that the ammo underwent acceptable quality assurance.

 

[pupchow]

It would lack precision. A pistol an be accurate without being precise. And, vice versa.

 

[EMTDAD]

by this, a pistol can only be precise, it cannot be accurate. Accuracy comes only from the shooter.

Let's make sure we're all on the same page & using the right words...

Easier to correct inaccuracy than poor precision, IMHO.
If you have a good group, but are off target, it may be as simple as adjusting you POA, or your sights.

 

[Varmint]

I disagree with the first part of your statement... how it feels, etc.. makes the SHOOTER more accurate, not the pistol. There's a difference.

Take away the variances in the shooter's ability, and environmental variables, and the pistol should theoretically put every bullet in the same hole each time. The laws of physics that govern how a bullet gets from point A to point B don't change. It's the variances in the initial conditions (and to some extent the environment) which alter the end result. If you set up a pistol to hit a specific point in a controlled environment, and automatically fire, and it doesn't repeat the result to within an acceptable error, then the pistol would be inaccurate, mostly due to the variability in the machining/assembling of the unit, assume that the ammo underwent acceptable quality assurance.

i disagree. If it were only the shooter that determines accuracy, every shooter would shoot every pistol with the same accuracy. They don't, it varies wildly. This is due to the factors I mentioned.

Again, you're thinking like a scientist, envisioning a test that has no purpose in the real world - things like the machining/assembly of a pistol are nearly insignificant compared to the ergonomics, recoil handling, sights etc.

 

[EMTDAD]

i disagree. If it were only the shooter that determines accuracy, every shooter would shoot every pistol with the same accuracy. They don't, it varies wildly. This is due to the factors I mentioned.

Again, you're thinking like a scientist, envisioning a test that has no purpose in the real world - things like the machining/assembly of a pistol are nearly insignificant compared to the ergonomics, recoil handling, sights etc.

but what I'm saying is it's not the pistol's fault.. it's the shooter's.

Therefore the pistol cannot be accurate or inaccurate.. it's the shooter. If I take away the shooter, and fire it with a robot, then the pistol should become accurate as all human error is removed. If it's manufactured properly, and undergoes proper quality control, it should be precise as well under the same conditions.

Of course when I say manufactured properly, it means the firearm is properly sighted, the barrel is unflawed, etc.

in a nutshell, the goal is to put the bullet in a specific spot. if it doesn't go there, whose fault is it, the shooter, or the pistol? I see a lot of posts about "this gun is accurate".. to echo what you said, it is more likely that the shooter is accurate with that particular gun and setup, rather than the gun being "accurate" in and of itself. Therefore, a review of a gun's accuracy is pretty useless to anyone else as everyone's ideal setup is likely different.

 

[Varmint]

but what I'm saying is it's not the pistol's fault.. it's the shooter's.

Therefore the pistol cannot be accurate or inaccurate.. it's the shooter. If I take away the shooter, and fire it with a robot, then the pistol should become accurate as all human error is removed. If it's manufactured properly, and undergoes proper quality control, it should be precise as well under the same conditions.

Of course when I say manufactured properly, it means the firearm is properly sighted, the barrel is unflawed, etc.

in a nutshell, the goal is to put the bullet in a specific spot. if it doesn't go there, whose fault is it, the shooter, or the pistol? I see a lot of posts about "this gun is accurate".. to echo what you said, it is more likely that the shooter is accurate with that particular gun and setup, rather than the gun being "accurate" in and of itself. Therefore, a review of a gun's accuracy is pretty useless to anyone else as everyone's ideal setup is likely different.

Yes, except that some guns are more accurate for a larger percentage of shooters than others. If it were entirely the shooters fault, that wouldn't be true.

The robot test is completely useless in the real world, it's of interest only to scientists. A much more useful test would be to get 10 shooters to test 10 handguns off hand, and report their findings, both subjective and objective.