I have read repeatedly that sorting by headstamp is a needed step. When I do that, I get jackets with a nominal +/- 0.003” variation. The problem is the odd flier that is 0.008-0.010” out. If I don’t catch them before point forming I get a mess. Sorting by length of jacket after core seating has yielded much better results when point forming and removes the need to sort by headstamp to begin with. What am I missing?