HOW MANY SHOTS AND HOW TO TEST FOR SIGNIFICANCE - A DIFFERENT PERSPECTIVE

[CharlieNC]

This chart can be used to determine how many shots are required, and to test for statistical significance. The writeup provides the details if interested.

 

[kelbro]

We used those formulae to evaluate the outputs of machines. Never considered using them when there was human action and environmental variables.

 

[davidjoe]

This is a very interesting and controversial subject.

The significant issue embedded is:

Do we attribute results to the gun or ammunition; the set up, shooter, or environment?

There is only one gun, and it “could” be mechanically held such that the only difference shot to shot is incremental additional fouling and slight additional erosion. Time the shots out to shoot at the same barrel temperature, and literally the only thing that has changed on the gun’s side is unavoidable wear.

Anything less than a comprehensive machine rest in the nature of a vice is for scientific purposes, not holding the gun the same way each shot, and it is difficult to blame the gun for resulting variances.

At the same time, it is not very satisfying to acknowledge the truism that rounds of ammunition cannot be identical, and might explain all the differences being observed.

A single gun does not have this problem; it is essentially just a pipe, and when held in such a way that it is forced to stay exactly in place, there is no leeway for a tight fitting bullet to take a different path than its predecessors did.

We don’t shoot a gun held in this ideal manner, but the point is that once a gun is secured in such a way, there is no physical reason why it needs to be shot dozens of times to determine group size capability.

I would suggest that if one does shoot it dozens or hundreds of times, (even if spaced out by days) all that is being discovered is how much inherent variance was built into the ammunition used, and what effect barrel wear has on group size.

 

[CharlieNC]

This is a very interesting and controversial subject.

The significant issue embedded is:

Do we attribute results to the gun or ammunition; the set up, shooter, or environment?

There is only one gun, and it “could” be mechanically held such that the only difference shot to shot is incremental additional fouling and slight additional erosion. Time the shots out to shoot at the same barrel temperature, and literally the only thing that has changed on the gun’s side is unavoidable wear.

Anything less than a comprehensive machine rest in the nature of a vice is for scientific purposes, not holding the gun the same way each shot, and it is difficult to blame the gun for resulting variances.

At the same time, it is not very satisfying to acknowledge the truism that rounds of ammunition cannot be identical, and might explain all the differences being observed.

A single gun does not have this problem; it is essentially just a pipe, and when held in such a way that it is forced to stay exactly in place, there is no leeway for a tight fitting bullet to take a different path than its predecessors did.

We don’t shoot a gun held in this ideal manner, but the point is that once a gun is secured in such a way, there is no physical reason why it needs to be shot dozens of times to determine group size capability.

I would suggest that if one does shoot it dozens or hundreds of times, (even if spaced out by days) all that is being discovered is how much inherent variance was built into the ammunition used, and what effect barrel wear has on group size.

A large part of the controversy is that we have been told that we must shoot very large sample sizes to achieve statistical significance; hopefully, I have put this error to rest and the impetus can be focused on effective shooting evaluations. To evaluate seating depth and charge weight optimization I test 5 levels of each via a trial involving 9 combinations x 3 shots each =27 total shots, and often replicate one combination for 30 total. Unless there is a major problem this has always provided a clear answer for me and a couple of associates as well; setup and shoot in 30minutes during consistent conditions.