First, do you regard Soundofmind's Audyssey mic technique of clustering the positions closely around the MLP centern point valid? If not, then that would explain why you would not want to verify the results of Audyssey with the same mic technique with REW. That technique uses points "within a few square feet" and that is a small area...like your head and how you might move shifting your weight around in your seat, leaning to one side for a while, etc.. And you don't have just one ear.
I think that Soundofmind's techique will provide good results for MLP listening. Verifying it by putting the REW mic in the same positions for an averaged look at those postions seems to be quite logical; like I said earlier...apples to apples.
What say you, professor(s)?
The fly in the ointment is that Audyssey inserts their proprietary weighting calculations after the mic feeds and before the calculation/imposition of the corrections. In other words, even with the same mic positions, Audyssey is "listening" for something that REW/XTZ/Omnimic/etc. can not. The only way to avoid this is to use a single, constant mic position for all.
That said, as long as the independent measurement system is in the same positions, with and without Audyssey, it will be a valid assessment of the effects. So, trying to emulate the Audyssey positioning or choosing another arrangement is equally acceptable.