Interesting thoughts, while I may throw out a group I know was bad, because of my error, I use all the data, averaging so that I can see what is happening through the test. Cherry picking is the best way I know to really screw the works up. My point was one group, no matter how many shots are in it, cannot tell a pattern.
I may be different from most folks, but posting pretty pictures of one hole groups, is not for me. I want to hear steel go clang, after I cause the bang, or see the game animal go down after a short run with a properly placed shot creating as humane a kill as possible.