BTW, that link with the 15 shot groups? The method they uses is NOWHERE NEAR as precise as what natoreloading used. Nowhere near. And you can tell this from their standard deviation and extreme spread numbers.
Incorrect and irrelevant. We can't control the standard deviation (SD) or extreme spread (ES). It is what it is. Some powders tend to have narrow SD and ES, some powders tend to have wide SD and ES.
And the SD and ES will not be exactly the same from one 15-shot group to another even when shooting the exact same ammo. That is clearly demonstrated at the American Handgunner link.
BTW, 15 shot groups are flawed too, because you could do 1 group, that is 100+ shots instead huh?
Groups can be any number of rounds, 5, 10, 20, 100, whatever. But you need more than 1 group for the test at the American Handgunner link.
A good way to be consistent with the experimental design, is to make all the groups have the same number of shots (e.g. 15), which is what the author at the American Handgunner link did.
That author was testing a specific hypothesis; is group size correlated with velocity standard deviation or extreme spread. To do this test, one needs to shoot multiple groups (samples), at least 2 for correlation statistical analysis (e.g. two 15-shot groups would be 2 samples), and the more samples you have, the more confident you are in the results from statistical analysis.
The author had two independent sets of data, one with 20 samples (twenty 15-shot groups; 300 rounds) and the other set of data had 12 samples (twelve 15-shot groups; 180 rounds).
To recap, that author found no correlation between group size and velocity SD or ES.