M14 Forum banner

1 - 16 of 16 Posts

·
Registered
Joined
·
612 Posts
Discussion Starter #1 (Edited)
Most every rifleman at one time or another measures group sizes as a means of seeing how how well the rifle and ammo perform. How to do so can be controversial.

I started at 3 then later 5 shot test groups. Anymore, my minimum test group size is 10 shots.

Like most folks I used to measure extreme spread. If there was a wide hit, I often tossed it out of the data as being a fluke. Lately I've switched to measuring "mean radius" instead of extreme spread. For mean radius I include the "flyers" because chances are, they really belong there.

Yup - 10 shot groups and including the flyers can make that magic 1 MOA group hard to come by. But bad data can be worse than no data at all.

Part of the reasons I've gone to 10 shot groups and mean radius are in the articles linked below. There are some statistics and equations but the core message is easy to digest. I used statistical quality control methods on the job and know that sort of analysis is valid.

See Figure 7 on the shot group analysis website for why larger test groups are important.

The AR15.com link has a good explanation on how to measure mean radius and why it is a better way.

http://the-long-family.com/group_size_analysis.htm

http://www.ar15.com/forums/t_3_16/512887_.html&page=1

Are big test groups and mean radius measurement splitting hairs? Maybe.

But I do not enjoy burning powder up and barrels out doing testing.

Getting reliable data out of the effort is worth it to me.

How best to measure shot dispersion depends entirely on what you expect out of the rifle and ammo.

EDIT: (added additional links ref: how to measure precision)

Digging around some more and found the website 'Ballistipedia'. Looks to be some good info.

http://ballistipedia.com/index.php?title=Describing_Precision

http://ballistipedia.com/images/9/9...st_Measure_of_Accuracy_by_J.E._Leslie_III.pdf
 

·
Administrator
Joined
·
17,858 Posts
Lately I've switched to measuring "mean radius" instead of extreme spread. For mean radius I include the "flyers" because chances are, they really belong there. Yes.

Getting reliable data out of the effort is worth it to me.

How best to measure shot dispersion depends entirely on what you expect out of the rifle and ammo.
NMC_EXP,

I started some 15 years ago with 5 round groups, then it was ten before I end up with twenty rounds to get the mean radius that took all the play out of the rifle and ammunition for the true accuracy of all three together. Rifle, ammunition and youself.

DI5
 
  • Like
Reactions: Louis1332

·
Registered
Joined
·
1,304 Posts
Ten shots minimum for my testing.
Circle in photo represents 1" dots from last Springs M1A challenge here on the forum. Faces represent 14 shots fired at 14 separate 1" dots at 100yds.

[URL=http://s1180.photobucket.com/user/nf1e/media/IMG_0120_zpsfjalepsi.jpg.html][/URL]
Well, I am not very good at this, but I will give it a try:


1st: The group appears to be less than 1.2 MOA at the widest point, but less than 1 MOA average; demonstrating pretty good shooting and equipment.

2nd: Since all shots were in the circle or broke or touched the outside plane of the circle they are all considered hits on the one inch dot----I would be happy with that.

3rd: 80% of the shots are in the top half of the circle with very minor horizontal stringing....Stringing could be caused by wind or shooting technique---but still very nice shooting.

4th: Since only 20% of the shots fell in the bottom half of the dot indicating little or no vertical stringing, ---I would keep the ammo/recipe as is, at least until you can tighten up the horizontal stringing.

5th---If I was to make any adjustments, I might drop the scope 1/4 to 1/2 MOA. and try and focus on breathing and trigger technique.

ALL that said, If this was the result of my rifle and my shooting---I would be quite pleased.


Feel free to critique or dispute my guesswork.


Bob
 

·
Registered
Joined
·
2,723 Posts
Anymore, my minimum test group size is 10 shots.

Like most folks I used to measure extreme spread. If there was a wide hit, I often tossed it out of the data as being a fluke. Lately I've switched to measuring "mean radius" instead of extreme spread. For mean radius I include the "flyers" because chances are, they really belong there.
Well said. I think that anyone serious about measuring group sizes eventually settles on at least 10 shots per group. Anything less doesn't yield adequate repeatability.

While extreme spread is easier and quicker to measure, mean radius data is a lot more reliable when making a decision. Those darn fliers totally mess up extreme spread measurements. Rick
 

·
Premium Member
Joined
·
17,308 Posts
Well, I am not very good at this, but I will give it a try:


1st: The group appears to be less than 1.2 MOA at the widest point, but less than 1 MOA average; demonstrating pretty good shooting and equipment.

2nd: Since all shots were in the circle or broke or touched the outside plane of the circle they are all considered hits on the one inch dot----I would be happy with that.

3rd: 80% of the shots are in the top half of the circle with very minor horizontal stringing....Stringing could be caused by wind or shooting technique---but still very nice shooting.

4th: Since only 20% of the shots fell in the bottom half of the dot indicating little or no vertical stringing, ---I would keep the ammo/recipe as is, at least until you can tighten up the horizontal stringing.

5th---If I was to make any adjustments, I might drop the scope 1/4 to 1/2 MOA. and try and focus on breathing and trigger technique.

ALL that said, If this was the result of my rifle and my shooting---I would be quite pleased.


Feel free to critique or dispute my guesswork.


Bob
Thanks, my feelings exactly.
It was the best I could do with what I have to work with.
I shoot prone most of the time, and moving from dot to dot was , to say the least, a challenge. Were all 14 shots fired at the same dot without changing position, things would have been much tighter. Oh yea, I single loaded each shot so position had to be completely re-acquired for each dot

Semper Fi
Art.
I don't think it was too bad for a 67 year old used to was.
 

·
Registered
Joined
·
612 Posts
Discussion Starter #7
NMC_EXP,

I started some 15 years ago with 5 round groups, then it was ten before I end up with twenty rounds to get the mean radius that took all the play out of the rifle and ammunition for the true accuracy of all three together. Rifle, ammunition and youself.

DI5
Could not agree more that precision is a three part equation. That is the reason for the Crossman quote sig line I use.

With that in mind, I should have posted this in "accuracy" as opposed to "ammunition".
 
  • Like
Reactions: XXIV Corps

·
Premium Member
Joined
·
17,308 Posts
Art,

Nice shooting! I'd give anything for groups like that!

Semper Fi,

Wes
Thanks Wes. 50+ years of practicing and practicing. Just gets harder every year to get up off the shooting mat. A few years ago I used to draw a smiley face on a 1" paster with my Savage model 12. Those days are past, but it's still the most fun and old dinosaur can have with his clothes on.

[URL=http://s1180.photobucket.com/user/nf1e/media/March272012accuracyweapons001.jpg.html][/URL]

Semper Fi
Art
 

·
Rest in Peace
Joined
·
908 Posts
IMO, there is no single best way to judge groups. Likewise, there is no single best number of shots. The "best way' may be a combination of all.

Each method of measuring has advantages, and disadvantages. Group shooters (such as Benchrest) rely almost exclusively on extreme spread measurements. But extreme spread can give a distorted value to a single shot. If that shot is vertical it can tell you one thing, if it's horizontal it may mean something else altogether. To a score shooter it may mean one dropped point. To a group shooter it may mean the entire match is lost.

Figure Of Merit (F/M) may be a better way to judge a group based on spread.

Mean radius (MR) is more difficult to measure and usually requires a higher number of shots to give useful information. But it's not perfect. Since it's the product of averaging, it's possible that two groups having the same mean radius can have very different extreme spreads.

Big game hunters care only about the location of the first one, two, or three shots. Everything else is a waste of ammunition to them.

And then there is group shape. ES, MR, or F/M tell you little about shape. For a competition shooter whose shots have to be in a certain sized scoring ring, that may be the most important factor of all. But it's impossible to measure shape. Like fine art, you know it when you see it.

Almost makes you want to take up another hobby, doesn't it? Almost.
 

·
Registered
Joined
·
612 Posts
Discussion Starter #11
IMO, there is no single best way to judge groups. Likewise, there is no single best number of shots. The "best way' may be a combination of all.

Almost makes you want to take up another hobby, doesn't it? Almost.
Ray

All I'm interested in anymore is NRA highpower mid range prone and smallbore prone.

Question: For these types of shooting which method of measuring precision has the most value in your opinion?

As to taking up another hobby - me giving up on 5 shot groups and extreme spread is the result of my frustration with misleading data.
 

·
Registered
Joined
·
2,723 Posts
IMO, there is no single best way to judge groups. Likewise, there is no single best number of shots. The "best way' may be a combination of all.
I tend to agree with Ray. Some methods of measurement are better than others for specific applications but all methods bcan ring some additional information to the discussion. If in doubt, measure more than one way. Relative to group size, what we are trying to do is get a large enough sample size that the sample accurately reflects the population. Anything beyond about 70 shots adds little to the quality of the data. However, shooting 70 shots per group is time consuming and can be expensive. Experience has indicated to me that less than 10 shots per group is generally insufficient. That still forms a pretty broad sample size range. The question you must ask yourself is "How precise do I really need to be?" A high level of precison is necessary when one is comparing two loads or rifles or modifications that are expected to be quite similar. A much lower level of precision is required when one doesn't expect the two loads or rifles or modifications to be very similar. When high precision is potentially required, increase your sample size to some number well in excess of 10 shots. Rick
 

·
Registered
Joined
·
11,954 Posts
In layman's terms, using the mean radius instead of the group size allows you to include "flyers" without making your rifle look like a POS, since you are taking an average measure from the center of the group.
 

·
Registered
Joined
·
44 Posts
I believe from basic statistics class that a population of 30 is needed to use statistical probabilities. Less than 30 is called a SWAG.
 

·
Administrator
Joined
·
17,858 Posts
I believe from basic statistics class that a population of 30 is needed to use statistical probabilities. Less than 30 is called a SWAG.
SWAG, Scientific Wide Ass Guess


DI5
 
  • Like
Reactions: nf1e
1 - 16 of 16 Posts
Top