ed, i spent ages thinking over all this crap when i got lower than expected at surrey. Then i read loads on the web about dyno dynamics in particular and comparisons with dynojets and dynapackks.
Long and short of it was dyno dynamics have a reputation for either reading low or being inconsistent, due to these correction factors and other settings and parameters that go into setting them up at each place. Unfortunately the only real way to know for sure would be to do a run in the morning on whichever dyno you choose, then fit the new set-up, then dyno it again in the afternoon on the same dyno! lol, not exactly practical.
Failing that going back to the placeyou first went is next best. I know where you're coming from with the base runs seeming consistent etc, but there's too much to it.
Personally if i was looking at mods and gains now i think i'd use a dynoet or dynapack, forgetting the high/low figures thing, purely on them apparently being more consistent.