I found a good article on pulley kinematics which gave me an alternative solution to fixing the tripod kinematics math to compensate.

On the effects of the design of cable-Driven robots on kinematics and dynamics models accuracy

Summing up: add another pulley wrapping in the same direction, and the compensation math goes away.

Rather than adding pulleys, I think I'll just duplicate the shape of the pulley on the outer surface of the end effector.  Unfortunately as I've designed the end effector currently, this entire surface would end up being an overhang when printed.

EDIT:  Um.. no it won't.  I just re-read that article (and what I said).  I was stuck in thinking the sphere shape needed to be on the bottom to compensate.  It needs to be inverted and on the top!

I also know my upper pulleys not rotating to align with the end effector will introduce some error.  Now do I duplicate this error on the bottom shape, or go for the ideal shape?  I'm not sure which will be better at this point, so I'll probably just end up modeling them all and finding out in practice.

Rapid iteration is one of the advantages of 3D printing after all.