No recent epiphany, but a few years back with pour over, it was mating grid size to average pour rate (not brew time, nor flow rate).
As for repeatability, you might need to be more specific to what you mean by that? Mechanically speaking percolation brewing is very repeatable in the quantitative sense. If you mean does this make every bean/roast taste great? Then, no…it just means they brew repeatably (say, 95% or more in the target range)…but if the bean/roast isn’t great to start with, that will limit what you can realise from a brew flavour wise.
Thinking about it, that does lead me to a more recent and subjective epiphany, to recognise which origins I personally like most and to spend less time trying make origin I have a less than stellar history with, to my liking. I still have the odd pleasant surprise, of course (all origins seem to produce coffee I really like at least occasionally), but I spend much less time trying to bang square pegs into round holes.