Firstly, you have been doing this a lot longer than I have.
In terms of gun nerd standards, Bill Reihl made an argument a few years ago here that I think makes sense for vetting a new unproven design. His argument was that each cartridge position in the magazine is unique, so to have a statistically valid test, you would load the entire mag a statistically valid number of times and fire it, tracking any anomalies. My statistics professor told me that, in general n=30 is the smallest statistically valid sample size for a normal distribution.
For me, this makes magguts and the shield mags too expensive for serious use.
I got what I do from JodyH. Use something that other people have validated (G19 for now), and when I get a new mag, it gets a box of practice ammo and a mag of carry ammo through it, then it is loaded with carry ammo and lives in the gun until the next time I rotate ammo.
Hope this helps.