If one shoots 3, 5, 7 or 10 shots per test group, there's a 95% probability that all groups will be somewhere in the following percentage spread of what one group's extreme spread is:
3 shots, 41% to 244%
5 shots, 66% to 153%
7 shots, 74% to 134%
And you don't know if that single, few-shot group fired is at the large end of several or the small one. Ten or twenty shot groups are better:
10 shots, 81% to 116%
20 shots, 89% to 112%
If you've ever seen the results of a 100 yard benchrest match on a weekend, you'll note each person shoots 5 to 10 five-shot groups per day. There's a 5X or more spread from the smallest to the largest group shot by each of the top ten on the score board. Bigger spreads for the ne'r-do-wells. The first group shot each day per competitor is seldom his smallest; largest, either, for that matter. Single group records are under .03" but record aggregates (average) for several groups have biggest ones in the .200" to .300" range at 100 yards. These rifles are shot free recoil virtually untouched by humans.
Group shooting equates to rolling a pair of dice. Simple statistics. Odds of getting a small number are the same for a big one. How often do you get snake eyes? Or a double Duce?
If you're holding the rifle against a shoulder to test from a bench, 25% to 50% of the group is the rifle and ammo; you're the other 75% to 50%.