Digest this.....and come to your own conclusions: http://riflemansjournal.blogspot.com/2010/07/statistics-for-rifle-shooters.html

I believe I can say with a 99.5% confidence level based on a normal distribution and two tailed T test that hand selecting a photo of the tiniest group ever shot seems to be the preferred method. -- richard

It may be....but that does not DISCOUNT....the LAW OF PROBABILITY!! Shoot enough groups....and every now and then...a SCREAMER will pop up! But THAT is not probably indicative of the accuracy of the rifle/load!

Discussion of the central limit theorem would have been of some value here... He dances all around it, but doesn't actually come out and say it.