There's a lot of stuff you're asking. I'll itemize.
(1) In rifles, it's considered fairly common to find "accurate" velocities. In other words, if Powder X is most accurate loaded to 2700 fps, then Powder Y will also be most accurate at 2700 fps, presuming there's no great mis-match in burn rates. HOWEVER, Powder Y may not be as accurate as Powder X. While I believe in accurate velocities within a given powder, I haven't had the chance to really delve into matching velocities between powders.
(2) Long range rifle loading is a different animal. Consistent velocity is more important than absolute accuracy. So it's quite common to test ascending charges to find flat spots in the velocity curve where a couple tenths of a grain of powder doesn't move the velocity much, and then test that small subset of charges for accuracy before moving on to seating depth.
(3) The reason the long range rifle shooter doesn't give a damn about maximizing accuracy is that
*the best-shooting load might be .5MOA, and the worst might be .7, in his very nice gun, with his excellent components
*he's shooting at 700-1200 yards, and .5MOA doesn't produce anywhere near as much dispersion as variations in velocity, both for wind drift and drop at extreme distances
(4) Ergo, you may or may not care about what the long range guys do.
(5) There's a whole different school of thought that revolves around optimum charge weight, and and another seeks to equalize bullet dwell time in the barrel and other voodoo I don't understand. I was going to study it, but honestly, the techniques I understand now already produce great results for me. My time would be better spent reloading and shooting more.
(6) Barrel harmonics is real. I can't locate it atm, but there's a series of 1-shot targets from an indoor range, showing how varying bullet dwell times caused POI shifts up and down, and slightly side-to-side. But frankly, so long as the bullet is leaving the barrel when it's at a consistent position, it doesn't matter. But it does explain why groups shrink and open in cycles.
(7) Sometimes, depending on the cartridge, the most accurate powder is simply well-known. In 6.5CM, for instance, it's Hodgdon 4350 (unless you want to be a damn hipster and use IMR 4350). Everyone pretty much starts there There's a reason so many .223 varminters lean on Varget.
Personally, I test velocity with ascending charges in .1-grain increments, hand-trickled. The .,1-grain increment (as opposed to more traditional 10-round, .2-grain increments) combats some of the variation while still letting me see accuracy nodes pretty easily. Especially because I can see obvious inconsistent velocities at the very bottom, and compare them to flat accuracy nodes at (usually) 90% max. It's not uncommon for me to find single-digit extreme spreads this way, both across spans, and then in five-round test groups later. Whether by accident or by design, this has resulted in excellent loads in less than a single box of "test" ammo--25 rounds to look for the node, and then another 25 to test 3-5 charges later.