There is only one fly in the ointment here, which is that a test on SD transponders (for DTV that means SD signals on Ku transponders on a Ku satellite) can still show high or acceptable numbers even when the readings for HD (and those signals are on Ka transponders) may not.
IOW, if your dish is tuned perfectly for SD, HD will fall in line. If your dish is mistuned for SD, HD won't be good either and probably will be worse. But here's the big one: if you have acceptable numbers on SD you still may not be tuned in well enough to have acceptable numbers on HD, because the tuning for Ka (HD) is much more precise, and there is less tolerance for error. That somewhat invalidates a test on 101, because it is Ku and does not tell you how mistuned HD (Ka) can be. You could be mistuned and still get SD and still have good numbers for SD, but HD can still be bad because there is less tolerance in the tuning; the focal points for Ku are much larger than those for Ka, so you can be off a bit and still be in the ballpark enough for it not to matter for SD, but it still may indeed matter for HD.
Post the numbers for a HD Ka sat. (What are they these days, 103 and 99?). That will tell you what it really going on with how well tuned in your antenna is. And if you live in a market with HD locals on spotbeam (Little Rock probably qualifies) You should have at least one or two transponders that read 96-100 (100 if you are at population center, and 96 if you are a few dozen miles or more off center). You can ignore all of the other numbers even though some of them may represent CONUS HD, but at least one (your local spot beam for HD) should be right up there.
BTW, when you tune to a sat to test signals, the transponders don't all pop in at once. The tuner in your STB can only tune one channel at a time, so it has to check these signals in serial order, which takes a while. You will see the old numbers get replaced one-by-one with the true new readings after a minute or two, but the original numbers may at the start default to what it was in a previous reading of a previous sat, so won't be accurate until enough time has passed for the STB to poll all of the different channels. This is probably why you are seeing some high (false) readings jump to lower (true) readings.