I think we are all saying the same thing. 'Mostly'.
I would be a hair more concerned with the rated Vmp since the Imp are nearly identical.
The concern comes from some experiments regarding a tracker concept for another member I did a few years ago, though I won't mention her by name.
The experiment involved various silicon structures, from various manufacturers, shipped in the same factory packaged skids/boxes (so presumably from the same production lots).
The results showed I remained relatively stable, but when I was the same (within 1~5mA?), E varied wildly under slightly differing conditions.
All that leads me back to ULR's "a nontrivial change in voltage will give you a small change in power",
and my "will operate at the same amount of current, just a little different Vmp voltage".
Back to 'mostly'.
No matter the configuration, they will each do about 4.84A Imp.
In like series pairs, 72.2V and 70.4V Vmp. That is far more than the Imp percentage.
I would be more concerned with the Vmp under these circumstances.
The numbers are so close, I doubt it will make any difference at sunset.
ULR said "If the controller has an "auto" setting it may hunt for the actual max power point by cycling around it and measuring the result. If this is the case it will find the max power point of the combined array. In that case the auto setting may do better than any fixed setpoint voltage."
I agree. A lot.
Vmp varies, and it can be the opposite of what is expected.
G-