Yeah, but we're not shooting for accuracy when working the load variable of the combination.Statistically, a three shot group is worthless. Shooting them in a random pattern does not make them any more so. Reliable accuracy testing needs more than 3 shots per group.
5grn increments with BP is a standard not random and as demonstrated is quit telling and useful.
Once that area of "harmonic balance" aka; the sweet spot,, is found the shooter can go back and work/tweak the other variables of the load within that sweet spot, even varying the load +/-2grns. But patch thickness/ lube properties/human conditions of the load and sight adjustment are all what's done to get "accuracy".
The Load Development technique as @Flint62Smoothie shares with us here will get a feller much closer too accuracy, much sooner than;
"What's the max load for my gun?"
p.s. It's way different than loading CF, where 1/10 of a Grn can be huge!