RegionRat
Gold $$ Contributor
Sometimes the reason a standard deviation isn’t as good as it should be for estimating the ES, is because the distribution isn’t a normal Gaussian distribution. If the distribution on the ammo batch is good, then the SD times six will be a very good estimate of the ES. However... Mother Nature can be cruel. When the ammo batch has abnormalities mixed in, the sample size has to be larger to capture a better estimate.
A clue that something is going wrong, is when your SD times 6 and your ES don’t match up well. If your sample is too small, you haven’t seen your full potential for ES, but when you start taking samples on the order of 15 shots and those two values are not converging, then something in that ammo batch is likely pointing at two or more strong issues that would not look like a normal distribution if you plotted the whole population.
A clue that something is going wrong, is when your SD times 6 and your ES don’t match up well. If your sample is too small, you haven’t seen your full potential for ES, but when you start taking samples on the order of 15 shots and those two values are not converging, then something in that ammo batch is likely pointing at two or more strong issues that would not look like a normal distribution if you plotted the whole population.