IMHO, nailed it. The downfall of most existing tuner related work is the lack of validating data. Anyone that uses competition results as data that certain equipment is better is making a fallacious argument.
I've tried the 2 shot/3 shot/5 shot/dissimilar loads/group size/group shape/relative impact tuning methods, and while you could sort of draw a correlation between best/worst group size with one or two groups at each setting, that disappears once you shoot a 5x5 and the settings become interchangeable. Even more so, when you look at mean radius or overlay all groups with the same point of aim. Individual groups may show a slight edge for one setting vice another, but when you have one group left of aim and one right of aim for the same setting, has that really increased precision?
I'm not saying tuners don't or can't work. I just haven't seen compelling data that they could in any but a "narrow operating range" and I haven't been able to apply that in my own testing. I suspect that any effect is marginal, such that it's easily washed out by other variables ("the rifle isn't good enough," "your shooting wasn't good enough" "your load wasn't good enough to start with" "you needed an accuracy fixture" "you didn't use the tuner correctly" "you can't pick settings like that"). I'm more than willing to be convinced otherwise - I'm usually decently teachable.