Hmmm. What I was trying to say with the "T" thing was that if the vertical part was fixed so it could only slide vertically, and you then applied a force on the crossbar, it wouldn't matter where on the crossbar the force was applied- at either end or in the middle right under the vertical support- the vertical force acting on the vertical support would always be the same.
This is the same as with the lower balljoint /upright. The vertical support of the "T" is the spindle and the crossbar is the plane that the contact patch travels along when you change wheel offset. The upright can only move in a (roughly) vertical motion, and vertical force on the upright = vertical force on the lower balljoint. Therefore the lever length is to the balljoint, not the contact patch.
On the maths, the conversion from wheel rates to spring rates at the shocker mount I got wrong because it's not a simple ratio on lengths,
Wheel rate = Spring stiffness / (motion ratio)^2
This means that my calculations are even further out.
Scott