Just a question to create a little discussion on a boring quarantine day. Now this is really for the average shooters but i imagine it works for everyone.
What i do to evaluate my rifle and load I shoot 5 five shot.groups on the same target and average them. That way I count all the groups including any that I don't really like. Try to pick a day with good conditions where wind is not the deciding factor. That 5 group average pretty well tells me what I can expect.in the future. Now they don't have to be shot back to back in fact they don't even have to be shot the same day. Give you barrel time to cool and even clean if you want.to.
How do you do it.
The downfall of this method is that you are moving the POA/POI within your sample. A 5x5 is not nearly as instructive as a single 25-shot group all shot to the same POA.
When we speak of precision, we really mean what is our statistical confidence that our points of impact will fall within some distance from the point of aim. The military uses the Circular Error Probable idea as the circle that will contain 50% of the points. It's a circle that represents the median error from the point of aim.
As a shooter there are two ways to use the interrelation of confidence and group size. You can either fix the confidence first and then assess group size to that confidence level. Or you can fix a size reference first, then assess the confidence in those terms. But here's the key thing: WITHOUT BOTH DEFINED YOU CANNOT COMPARE.
The reason I took up midrange shooting was because I wanted real data on the actual precision of my rifle and loads. On an NRA target of known scoring rings, I can easily know this data. If I shoot a 200-14x on a sling face, I know with a 20 shot sample that 14/20 are ~sub-moa, so this is 80% confidence. A higher score might take me up to 88% or 92% confidence as X count rises. And in this case, it was 100% confidence of sub-two MOA. But that's only a sample size of 20 shots. Unless you are agg 600 with regularity, you don't have that 100% confidence in reality.
The more difficult method would be to calculate the size of the rings that would contain 90% or 95% or 50% of your shots, or whatever it is you want to use for reference. Either way is valid because it contains both the essential elements of confidence value and error value (group size).
I'm fairly certain that at one point or another, almost everyone on this board as fired a particularly satisfying bughole in the 2s or less. But, given enough shots and attempts, ANY RIFLE can and will do this. It just might take several thousand groups for one to do it or a single attempt for a very excellent rifle.
This is why group size is not only part of the picture, but it's probably the less important part. Without sample size and demonstrated confidence values, the group sizes themselves are next to meaningless. This is why those who point to a single five shot (or worse, three shot) group as demonstrating a rifle's capability are kidding themselves first, and us second.