He goes deep on small sample testing, but focuses on tuners with his testing. He shows both that small sample testing is bogus, and that tuner adjustments don't make any observable impact on precision. However, I consider it nearly the same as saying that powder and seating depth tweaks don't alter precision since the theory says that tuners, seating depth, and powder charge are all ways of adjusting bullet exit time to coincide with quiescent periods of barrel harmonic motion. Others might disagree but that's my take.
For those who default to saying that any truly statistically robust test would be so time and money intensive that it's just not worth doing, I implore you to think more like an engineer and less like a physicist. You can do a DAMN good test with 100rds, and that is absolutely in the realm for all of us here to afford. And think about all the subsequent ammo it will save you once you determine you can stop "tuning"! Oh whoops, I spoiled the ending.