Talk:Prior Art

A sample of 2 follows Chi-squared distribution, but it is very statistically inefficient (assuming Rayleigh shot distribution and no fliers.)

Shooting groups of two might be a decent "shooting-range" method in presence of a high percentage of fliers. Say something like 20%.

Assuming say a manufacturing problem with ammo, then the "best" method to detect fliers as outliers would depend on the nature of the shot distribution of the flyers and the percentage of fliers.