Regarding new shells -- Roster's test showed that a new, unfired shell had lower velocity (10-20 fps) than those fired subsequently. His hypothesis is that the interior of a previously-fired shell was scuffed up, and held the wad back a bit, versus a slick-walled new shell.