Accuracy Testing article in Shooting Sports magazine

Interesting article on the shortcomings of 5-shot group testing for accuracy....
https://www.nramedia.org/t/11155664/7491181/96881/20/
This should get a good discussion going!

Always struck me that to do a proper analysis what should be done is something similar to the heat maps produced by TargetScan.

https://targetshootingapp.com/
And basic statistical analysis is well aware of the mockery that a tiny sample size of 3 or 5 is. Not to mention the habit of throwing away the flyer.

As for the article, I particularly liked the summary of the randomness of 5 shot groups. Shoot 50 shots, choose the best 5 closest together shots, and declare that your precision. While you're at it, move the target and declare accuracy perfect as well.

So what is the "right" sample size? Six sigma and others simplistically go with 30. Seems to be derived from a rule of thumb that says 10 per variable.

Now, considering the multitude of variables in a firearm and cartridge...

Then it seems to me we're back at the idea of the heat map and taking data until saturation is achieved.
In my own amateurish way I suppose that is what I did when I was starting out pistol shooting. Would use the same backer for a while and locate the target identically each time, then when is resembled swiss cheese, adjust sights to center the cloud, replace backer, and repeat. Got me to the point of using just repair centers with the occasional shot outside.

Here's another question concerning a another sacred cow: Why does the ransom rest rotate with recoil? Small bore and air pistol/rifle testing (and ammo matching) is done with the action clamped. Granted they're testing bolt action weapons, but allowing for recoil rotation just adds to the variables it seems...
Reading the Ransom Rest booklet, the idea that high calibre weapons are allowed to rotate to vertical from recoil seems absolutely ridiculous.

Granted if the the recoil control is identical each shot then the resulting groups will simply be higher.

As I was asked by multiple masters from the 60's and 70's, "How many shots do you shoot at 50 yards?"

My minimum is 10. My average is 30.

The ransom rest is allowed to rotate in only one axis. You reset to the stop with every single shot. The recoil has to go somewhere and absorbing it into the spring seems like a really good idea to me.

Slartybartfast wrote:

Here's another question concerning a another sacred cow: Why does the ransom rest rotate with recoil? Small bore and air pistol/rifle testing (and ammo matching) is done with the action clamped. Granted they're testing bolt action weapons, but allowing for recoil rotation just adds to the variables it seems...
Reading the Ransom Rest booklet, the idea that high calibre weapons are allowed to rotate to vertical from recoil seems absolutely ridiculous.

Granted if the the recoil control is identical each shot then the resulting groups will simply be higher.

Way back when the Ransom rest was created, the rotary brake was the simplest method and producible. However rotary recoil isn’t the biggest issue with the Ransom rest, that belongs to the inserts and the other methods of mounting the gun itself. The polymer/elastomer used for these inserts is part of the extra randomness in the results. There are way better methods of mounting something rigidly. But those will mar the gun if not used properly.

Wobbley wrote:

The polymer/elastomer used for these inserts is part of the extra randomness in the results.  There are way better methods of mounting something rigidly.  

I've heard of it referred to as the Random Rest more than once.

As the world is all about statistics (and sample size issues aside), I've always viewed claims of accuracy/precision dubiously when they're expressed as simple measurements.

Is a 2" pistol claiming that 2" represents a standard deviation of 1" or is it claiming that 2" represents six standard deviations (with one standard deviation being 1/3"). Even in the latter case, 0.2% of shots will happen outside the 2" (6 sigma) circle.

Considering scores of 1-10 regardless of shot placement horizontally and vertically, the graph for multivariate normal distribution (https://en.wikipedia.org/wiki/Multivariate_normal_distribution) does illustrate well what we’re trying to do as markmen. Consider one normal distribution as the performance of firearm and ammunition, the other the performance of the marksman’s shot process.

The visualisation of how this translates onto a 2D surface of a target is difficult, as that becomes a 4D graph with normal distributions in X and Y for each of the generalised variables of shooter and equipment.

Interesting discussion, and it almost makes me want to start diligently documenting scores. Then I remember that I want to have fun and that I really hate statistics...

During load development I test 5 to 10 round sets. Once I settle on the load, and mass-load ammo (several hundred or even couple of thousand rounds), I then do testing of samples from the entire batch to make sure that ammo is consistent from start to finish of the loading.

I usually combine RR test with LabRadar. As such I test 10 rounds minimum, but normally 20 rounds set on single targer. If I feel that test is inconclusive, I run a several additional tests, and that includes more than one gun. Targets then overlay-ed for comparison.

AP

My contribution to this is that when I visited the Eley Customer Range a few years back, they recommended forty (40) shot groups.  The Chief Range Officer, Bert Brookes, told me that Eley hired a statistician to figure out how many shots to fire.

Richard Ashmore wrote:My contribution to this is that when I visited the Eley Customer Range a few years back, they recommended forty (40) shot groups.  The Chief Range Officer, Bert Brookes, told me that Eley hired a statistician to figure out how many shots to fire.

I met Bert Brookes at Camp Perry once or twice. Great guy. I've always wanted to go to their test range.

My thoughts on the subject are similar to some of the others. When people (including me) say it is a 2" gun, I assume some, maybe most 10 shot groups will be that size, but at best it's a 1-2 sigma measure.

Similar to PhotoEscape, When I am testing loads, I shoot 1 or 2 10 shot groups with each potential load then a 30 shot group with the best load. Assuming it's good, it becomes a "reference" load that gets shot each time I test loads to (A) make sure the setup is reliable and (B) make sure nothing has changed in my loading process and (C) get a larger sample population for the load over time.

That said, I haven't tested much in the last decade. I found good loads back then, and have no reason to doubt their accuracy. If shots are significantly off-call, then it's time to investigate, but so far, it's never been the ammo. A few times I've set up a sandbag for a quick check, but in each case the resulting group was good enough to eliminate my concerns.

absolutely no less than 10 rounds per ransom test then 20 - 30 with the most promising load to verify..  Wobbley is right about the inserts.. even worse today.. the old inserts were much better.. much

Here's my testing from this past weekend of a new barrel I installed last year using a load suggested by my shooting buddy. I shot 200 rounds and watched thru my spotting scope as I shot.
I discovered mags can make a difference as shown in the pictures. Almost all of the wide shots came from the same magazine over and over. More testing to be done.

If you can, locate a copy of Walter F.Roper's book- Experiments of a Handgunner.

He was WAY ahead of his time with scientific testing of handguns.

Wes. Was that irons? Slide mounted dot?  If so what dot.

 by Wes Lorenz Today at 1:04 am
I discovered mags can make a difference as shown in the pictures. Almost all of the wide shots came from the same magazine over and over.


Interesting find Wes.
Not that I can find a way to blame the magazine for my occasional 7 or 6 even. But, I have noticed that when I change mags, when Ransom testing, that the first shot is out of the group. Even tho I load 5, shoot 4 and then change mags.

I noticed years back that one of the local guys has always used 1 mag for the entire match and just reloads that one. Never thought to ask him why. Perhaps he knew something that never occurred to me.

UPDATE 25 Sep19:

I gathered 15 (most, not all) of my 1911 mags & set out to do some comparison testing with the RR. The load used (4.0 clays, RP brass, Zero 200 swc, .469 crimp) wasn't the best one for the springfield, averages 2 1/2-3"@ 50 yds. Just happened to be what I had plenty of in my locker for the test. Waited until the last shooter had left for the evening & then set up @ 50 yds. 4 of them were really consistent. 8 of them were mediocre. and 3 were 6-7" patterns. No best brand, the good mags were a mixture of mfg. Barrel was scrubbed every 50 rds. Loaded 5-shot 4-reloaded 5 more with the same mag.

At the end I shot the Zero SWC load the Springfield likes (same except for TZZ brass instead of RP) and it was back to it's normal 2"+ @ 50. Then, just to satisfy my own curiosity of mixing lead & jacketed, I put 10 rds of Zero 185 Jacketed, 4.6gr BE without scrubbing the bore and it went right into it's normal 1 1/2" group. Not bad for a wad gun that's seen in excess of 83,000 rds.

Al

It is very common for different magazines to shift point of impact. The pressure from the follower on the bottom of the slide can vary as well as the feedlip release of the cartridge causing inconsistencies between the two magazines and feeding. So basically it is changing the cycling consistency of the pistol. This has been known for a very long time. When I was young shooter I was taught to use only one magazine for the 50-yard line. I have only used one magazine to shoot 2700's for this reason. Many many years ago I was doing some testing what two magazines and was consistently getting a 3in point of impact change for the group from the machine rest.
Jon

Jon Eulette wrote:It is very common for different magazines to shift point of impact. The pressure from the follower on the bottom of the slide can vary as well as the feedlip release of the cartridge causing inconsistencies between the two magazines and feeding. So basically it is changing the cycling consistency of the pistol. This has been known for a very long time. When I was young shooter I was taught to use only one magazine for the 50-yard line. I have only used one magazine to shoot 2700's for this reason. Many many years ago I was doing some testing what two magazines and was consistently getting a 3in point of impact change for the group from the machine rest.
Jon

So to clarify (got to love the english language) you literally mean 1 single magazine, not the same type or brand of magazine. And while it is a simple thing to reload a single magazine in slow fire, you also do the same thing between strings of sustained fire?

Yes. If you test your mags you might find less variation, but to play it safe and spend more time shooting and less time testing just use one. Majority of your HM use 1 magazine.
Jon

Do you reload 1 handed or put the gun down and regrip? and do you reload after firing 4 or 5?

Shoot 4. Reload with both hands and regrasp pistol.
Jon

I presume this is only applicable to Slow Fire strings, right? Asking for clarification because of the original question by Weber1b was "And while it is a simple thing to reload a single magazine in slow fire, you also do the same thing between strings of sustained fire?"

AP

Yes for SF. I still use 1 magazine for short line.
Jon

Jon Eulette wrote:Yes for SF. I still use 1 magazine for short line.
Jon

I seriously thought I was the only person in the world that did this.  Looks like I need to get out more.

I stood there and shot 50 rounds at this 25yd target. I just used one magazine. Reloading was the only break I gave myself. I was taking Gunny Zins challenge to fire 50 rds and keep them all in the 10 .A requirement before your ready to shoot timed and rapid fire. I was doing good until I hit 41 then faded fast. I'm not sure firing 50 rds from a Ransom Rest would prove much more then 10 rds. Except, there wouldn't be much of a center left. I'm rambling but to get to my observation. I think standing there and shooting tells you more about your gun and you then clamping the pistol down. I also noticed that I'm getting older. That drill used to be easy.

Greetings,

Personally, I feel testing gives you an air of confidence when shooting.

Example: I bought a full case (2000 rounds) of commercial 32 WadCutter ammo. Shooting it off hand I was getting quite a few shots in the white.

I then tested the ammo in a barrel tester and discovered it would hold the eight-ring at 50 yards.

My "tested" hand loads generally stay inside the ten-ring with high X counts at 50 yards.

So, if I held hard and shot what I thought was an X, the shot, in reality, could land anywhere out to the eight-ring.

Result? Was it a bad call; or, do I make a sight correction? With eight-ring ammo it is near impossible to sight in the gun at 50 yards.

I agree that the best way to bring up your aggregate is to shoot a lot in practice and matches.

Testing ammo helps keep your shots "on call".

Cheers,

Dave

With respect to Ransom Rest movement/rotation from recoil, it moves fairly consistently in relationship to how much recoil a given round produces. 

That was tested in this article:

https://www.shootingtimes.com/editorial/measure-relative-handgun-recoil/99442

Same author, by the way.

My favorite article on accuracy is:
http://accuracy.martinchick.com/handgun_rifle.html

Seems to me the people suggesting 40 or 50 rounds minimum will get the most realistic answers.

It also seems to me that the end result, group size, etc.,  will depend mostly on "the shooter" (person, or testing machine).