You won’t see any difference at steady state from different shapes, sizes, emissivity etc. at the same input power level.
You won't see any difference at all, in any state, with any of the shapes, sizes, emissivity etc. tested by Mizuno. You cannot distinguish between them, period. Except, of course, things like the surface temperature of the test reactor.
It is possible a person might accidentally come up with control reactors that do produce different results. Those would be invalid control reactors, and Mizuno would not use them. He makes sure there are no detectable differences before employing a control reactor. That is the whole point of testing and calibrating.
I am not saying what you describe is impossible. I am saying:
1. It did not happen. Not yet, anyway.
2. If it does happen, it will be readily apparent. He won't just ignore it and keep going.
You are saying that something hypothetical might happen, so therefore it did happen, so therefore there must be a mistake. You are letting your imagination run away with you.