verify keyway position

I believe the "correct" way to measure this is to measure points on both planes - the pattern of points must be the exact same - and then create midpoints between each point corresponding to the same point on the other plane (ie. mirrored pattern). From these midpoints, construct a plane that you use for evaluation/s.

vpt.se would you create each plane as a BF Recomp or something else, and would you set it as "least_Sqr" or "min_Sep"? then for the mid plane if you were to use GeoTol to dim the TP, by constructing a width the only options are best Fit and BF recomp then least Sqr, Max inscribe and Min circsc, which would you select in this case with no datums?

vpt.se would you create each plane as a BF Recomp or something else, and would you set it as "least_Sqr" or "min_Sep"? then for the mid plane if you were to use GeoTol to dim the TP, by constructing a width the only options are best Fit and BF recomp then least Sqr, Max inscribe and Min circsc, which would you select in this case with no datums?

If using the points method (you are NOT creating planes with this, but ONE plane out of the midpoints):
If you are using GeoTol, it should do this according to the standard you have chosen (and selected "STANDARD" as evaluation), provided that the toleranced feature in question has hitdata (which it will in this case). So, it shouldn't matter what calculation you choose for the plane constructed by the midpoints as GeoTol will (should) recalculate the toleranced plane according to the standard chosen (the more hitdata, the better, if low hitdata, choose Gauss).

If inner keyway, max inscribed for sure, as that is the largest key that will fit in the keyway.