Test criteria?
1 brick of what is available.
ID the cartridges 1-500 (sharpies work well for that, yes, been there done that)
Sort by weight, record by ID number
Sort by overall cartridge length, record by ID number
Sort by rim thickness, record by ID number
Sort by visual defects(dents, dings, chips, scratches, crimp problems, seating), record by ID number
If y'er gonna go OCD, go all the way.
Shoot indoors from a one piece rest at 50 yards, one dot-one shot
recording both point of impact relative to point of aim,
as well as the muzzle velocity of each cartridge, record by ID number.
Load into a spreadsheet for data mining.
Results will show sorting by culling visible cartridge defects, will provide the most consistent trajectories
Me? I'd rather just buy better ammo.