Indeed, the Hornady engineers have presented data which suggests most common load development methods don't survive contact with actual science. One particularly polarizing episode presented results which show that the Satterlee curve (derived from Creighton Audette's Ladder method ~50yrs before) doesn't hold up. They presented data which showed that powder charge nodes - flat spots in the velocity vs. charge weight curve - don't survive from barrel to barrel. Alex Wheeler & Erik Cortina put out a podcast a couple of years ago which also agreed - there's no such thing as a velocity node, but rather we experience relatively coincidental nodes correlated to what really amounts to experimental bias which occur because competitors use the same components (bullets, brass, powder, barrels, cartridges) for the same purposes/courses of fire, so it should be natural that doing the same thing yields the same results... Bryan Litz & Doc Beech's team at Applied Ballistics have a similar podcast with similar findings, also published within Volume 3 of Applied Ballistics Modern Advancements in Long Range Shooting (chapter starts on page 57 for those following along).
In my own testing, having conducted the same tests over multiple different barrels which both confirm repeatable results within a range of barrel life for a specific barrel, but ALSO confirm that powder charge nodes don't last forever and don't uniformly transfer from one barrel to the next. If I lay out 5 or 6 tests (15-18) on top of one another for a given barrel, I get nodes repeating in the same location, with ever so slight variability in the velocity result for each (hence why I confirm my nodes and correct my velocity before matches), BUT, when I overlay all of the data I have using the same components, which now spans 11 different barrels, I get one big, fuzzy catepillar shaped data set which perfectly matches a simple linear regression (Change in potential energy available = change in kinetic energy yielded by the system).
Alternatively, before a match last season, I swapped barrels at midnight on Wednesday night when I returned from a business trip and used absolutely arbitrary load data - I texted a buddy from Wisconsin who I knew was using the same bullet and brass - both new to me - and I dropped his load into my barrel, and at 9am Thursday morning, I shot the smallest 20rnd group I've ever fired in my life, and ONE of the smallest 10rnd groups I've ever fired (technically 2, or maybe technically 3, since the first 10 and second 10 of the 20rnd group would have also been on the short list). I shot 35 rounds at a 100yrd zero board 3 for a group, moved the scope, 1 more, moved to zero, 11 on one target, then 20 on another, estimated a velocity, then moved to an 800yrd KYL and took one shot each to walk down from ~2.5moa to 3/4MOA, and put 6 rounds on the 3/4moa target at 800... Complete SWAG on velocity, brand new barrel which had been on the rifle for less than half a day, ammo loaded STARTING in the middle of the night on virgin brass of a brand I'd never used before, and a bullet I'd never used before... I taught an Intro to PRS Competition course that day, then used the same load to shoot a 1 day match that weekend...
Don't overthink load development - but don't chase 3 shot groups at 100yrds and pretend the results are meaningful either...
Be surprised then.
The Hornady Podcast is hosted by a figure head, Seth Swerczek, but the meat and potatoes of the discussions are lead by Senior Ballistician Jayden Quinlan, Ballistician Jacob Morrow, and Engineer Miles Neville (productive competitive shooters themselves). The data-driven science they share, for FREE, through this podcast - much akin to the Applied Ballistics Podcast - is an exceptional resource for shooters, and has invaluable information being shared freely online for the betterment of our engaged community. Narcissism and cynicism aren't part of their program.