Parameter plateau, not parameter peak

I used to call this the wiggle test, which undersold it. The thing you are looking for has a better name: a parameter plateau. The idea is the same but the image is more useful. You are not trying to find the highest point on the surface. You are trying to find a flat region, somewhere the results barely change as you move the setting around, because a flat region means the parameter is describing something with a shape, and a peak means it's describing a coincidence in this particular sample. Practically it changes what I do with the result. I no longer take the best-performing value. I take the middle of the widest flat region, even when a sharper setting nearby tested better. The sharper one will not survive contact with a different period, and I've stopped pretending otherwise. If your surface has no plateau anywhere, that's a result too: the parameter isn't real and you should remove it rather than tune it.