I was thinking more in terms of behavioral characteristics, as your edit suggests. In my mind, a strictly 'P' controller should produce an oscillation, which t

I was thinking more in terms of behavioral characteristics, as your edit suggests. In my mind, a strictly 'P' controller should produce an oscillation, which the example did not. I then surmised it 'hides' the 'D' part of the equation by modulating the error down each step. Strictly speaking, it seems like your summary is more accurate, and I'm still not quite sure if my way of looking at it is skewed, or completely broken.