If you sort by your case's headspace, each batch should have a pretty consistent AOL =<.001", if you sort them in groups with .001" difference. But, of course, there'll be more difference between the batches. And since you don't anneal, there can be variance in springback that can add an...