It is plain to me that, after making the straightforward modifications to make sure that the characterization algorithm deals with adjusted ceiling squares of the correct “polarity” (according to m being congruent to 1 or -1 modulo 4), I need to look again at how I am choosing nice, smooth, linearly increasing ranges. These values I’m seeing are all over the place, and the noise factor, which I was trying to eliminate, is not edifying or informative at all.
Fear not, folks. It’s not back to the drawing board. I haven’t actually left the drawing board. What it is, equally metaphorically, is sleeves-rolling-up time. Oh, heck, it’s football season, so those of you who fancy that sport can come up with a better metter for.