What you are seeing is that any node that tends to improve consistency will tend to improve accuracy if all else is equal. There are just more than one kinds of node, and the trick is to get them all to play together for maximum effect. So, for example, you could have a pretty jagged Audette or OCW ladder that doesn't have the tightest groups but that is clearly identifying where muzzle deflection is optimal and then go on to tighten groups up by finding a bullet and powder and seating depth combination that also has a velocity flat spot coincident with that phase of the muzzle deflection. Things that improve consistency, be it in velocity or muzzle position, can stack up.
TimRB,
You are correct. Typical Audette and OCW ladders run over 20 shots, and if you want to see where your velocity flat spots lie, it takes the same sort of effort. I know there are guys on YouTube running just 10 shots, but that's because you can usually get at least one velocity flat spot identified in that number if you shoot carefully. But they then settle on that flat spot's load without knowing that if they'd have expanded the range they might have found one in a muzzle whip flat spot that further shrank the group considerably.
Then there are rounds that need to be fired to determine optimal seating depth. I've seen a YouTube video of a fellow starting 15 or 20 thousandths off the lands and changing seating depth in very small increments of a few thousandths, getting closer and closer to the lands. He shot, IIRC, five-shot groups at each seating depth, all of them pretty good, and then identified one a bit smaller than the others as best. The problem is, with five-shot groups he could have fired them all at the same seating depth and random probability says he would still have got one group as small as the one for which he concluded he'd found the best seating depth. It's easy to delude oneself in these matters. The real problem is he didn't try nearly a wide enough range of seating depths to know he was at the best point. Berger has found seating depths as much as 0.150" off the lands will shoot best with some bullets in some rifles. So you really want to move bullets in increments of at least 0.020" to fine the most promising region, and then refine down to smaller steps around that spot.
In the mid '90s, the late Dan Hackett described loading 0.020" off the lands as a matter of course because there was a general concensus at the time that this was optimal. He had a 220 Swift that would not average tighter than 0.5" 100-yard 5-shot groups no matter how he tuned the powder charge. Then one day, when switching between bullets, he inadvertently turned the micrometer head on his seating die the wrong way to compensate for a longer ogive, and wound up loading 20 rounds 0.050" off the lands before he discovered the error. He considered pulling and reseating them, but decided just to shoot them up in practice. To his amazement the 5-shot groups from the bullets seated that far back gave him two 0.25" groups and two true bugholes in the 1's (below 0..2").
So, there's a good deal to do in tuning. I may have 100 rounds down range before I'm really sure I've ferreted out all I can by testing.