Lenses I'll look to do multi-row panos are the 24/1.4L, 50/1.4, 85/1.8
I don't have the 24mm, and I have not used the 85mm with the bracket (my own making).
The 50mm f/1.4 does change the entrance pupil location with focusing.
Measured from the mounting flange (which sits on the body) forward:
10mm @ infinity,
11mm @ 4m
Add to this the distance from the mounting flange on the camera to the center of the screw hole, then you get the distance the camera has to be set back from the center of swiweling (i.e. the center of the tripod column, if there is one).
My guess is I can nail it within a mm one way or the other
That's my best one too. At longer focal lengths this is a non-issue., for example with the 200mm f/2.8L, three mm more or less does not make any difference. In fact, I am unable to find the location with a higher accuracy, for the difference in effect is practically none.
However, with the 24mm you have to be careful, particularly at close focusing if the location changes with the focusing distance. When measuring the EF-S 17-55mm f/2.8 IS, I looked at the test images in non-demosaiced format in order to verify the result looking single pixels.
Btw, one advise I often need to tell: one can find the approximate location already through the viewfinder, particularly if there is enough light to stop down the aperture and still see objects close and far good enough. However, when you think you found it, then you have to make a test shot and verify it, because *the position of your eye over the viewfinder influences, what you see*.