How often we see Internet snipers posting 5 shot groups to "prove" precision of their rifles, and other Internet snipers drawing thoughtful conclusions on that basis? I have seen too many; anyone of us who has shot more than one 5 (or 3 or 10) shot group in his life knows how different the results can get due to sheer random statistical variance. Furthermore, besides the warm feeling of posting a nice group to a forum, the knowledge of the size of a 5-shot group does have zero practical value. The main purpose of the exercise is to know the probability to hit a given size target at a given distance; to this goal, a 5-shot group size is totally useless. Sigma, R50 or R95 are good, practical indicators of the actual spread.
How do we properly measure rifle precision -- what method, how many shots? This question actually does have precise scientific answers. Many people have thought these thoughts before; for example:
This guy: https://github.com/lstange/mcgs
And these guys: http://ballistipedia.com/
And anyone listed here: http://ballistipedia.com/index.php?title=Prior_Art
(And pretty much any army in the world with enough money to hire a graduate engineering student.)
The science knows how to properly measure the precision of your rifle. The science also knows how to properly measure the precision of the measurement -- its confidence interval. Then why doesn't everyone do it properly? -- you may wonder. The answer is -- because it is complicated. We need to make it simple.
It all started as an attempt to scratch my personal itch -- to evaluate precision of shooter-rifle-cartridge combination with statistically valid and scientifically sound conclusions. Although it is possible to do it "by hand" with a ruler and an Excel sheet, it does require quite a bit of time, that I'd rather spend shooting.