Awhile back, I posted a thread on this forum asking how closely case weight correlates to internal volume. No one seemed to have a definitive answer.
Recently, I tested around 100 milsurp 30-06 cases, from the same headstamp and production lot, all fired from the same rifle. I weighed each case, then measured the actual internal case capacity in grains of water. All but a handful of cases fell within + or - 1 grain of the overall average (70.2 grains). Of the cases that fell outside that range, only one of them would have been detected by weight sorting the cases. I found no correlation between case weight and internal volume.
Based on my own experiment, I believe the only effective way to sort cases is to actually measure the internal volume of each case and sort accordingly. I believe the best way to do this would be to use once fired brass that has been re-sized in your preferred manner (full length or neck -sized), with the de-capping pin removed to leave the spent primer in the case. The cases should then be trimmed to a uniform length, measured for volume, and sorted accordingly.