Most optimization is a curve. Arguing for moving closer to the top of the curve is not the same as arguing for moving all the way to the minima on the other side. But why do I have to say that?