So maybe sanding 0.001 or 0.002" off the bullet before seating would help you there.
I'm going to assume that was a joke.
Anyway... I have no problems myself with oversized lead and plunk testing. Have always been able to find a usable OAL. Whether or not early contact with the walls of the freebore has ever led me to load shorter than necessary is a different matter. I can't say with certainty.
What I have had happen, though, is that I once accidentally shot coated lead loads for my VP9 through my ShadowLine, which plunks more or less all bullets about .03 shorter than my VP9, so those VP9 loads were definitely into the ShadowLine's rifling, and they were super accurate. At the time, I recalled noylj having said he loads lead right into the lands, and since then I have often done the same for the CZ with good results. Maybe I'm not into the lands at all, but into a freebore squeeze. I'll spend some time figuring out how to test that. My assumption has been that loading lead into the lands has had the effect of "trueing" the bullet to the bore.
And to that end, I once read an article where a guy tested 9mm for accuracy, grouped by different case lengths. The shorter the cases got, the worse precision got. His assumption was that it was about headspacing on the chamber step VS the extractor, where headspacing on the chamber step squares the case to the bore and aligns the bullet axis with the bore axis. But if the case was short enough to head space on the extractor, that you lost the squaring effect of the chamber step, and the cartridge would rest on the case mouth in the chamber, not perfectly aligned. In his tests, case lengths of .750 - .754 were about the same in terms of precision, but below .750, things got progressively worse as cases got shorter, suggesting that cases were starting to hang/headspace on the extractor shorter than .750.
So while I personally think .004 shorter than SAAMI spec before extractor headspacing is a lot to ask in terms of tolerances, especially given that most 9mm cases don't even make it to .750, so I don't know that shorter than .749 is where the case starts headspacing on the extractor, but I'd say it's inarguable that there is some case length below which that happens. And I know that precision rifle shooters will take stock Remington 700 actions and have the chambers and barrels "trued" to achieve that perfect pre-ignition alignment, so I suspect there is some truth to that writer's hypothesis in principle, at least, even if his specifics are a bit off.
Food for thought.