when I was starting to learn this method, and code my auto-tapes, one of my errors was assuming that a point 3 channel had to have an up-down-up tape sequence, or the inverse. I then started noticing that it is in fact more subtle than this. What I found is that not the *sign* of the slope, but...