I really like the way you approach these issues Lee, very nice analysis of the system. I do hope this solves your problem with the cars once and for all, I do think they're a really neat way to go for animation of the streets.
"I tested two Faller wires sie by side ... was twice as strong."
Two pieces of the same roll of wire from Faller? Implying they have huge variance in their product?
"I tested doubling the height of magnetic stack doubled the attractive force. Not even close, more like the square root of 2."
To be expected. Magnetic force declines in proportion to the square of the distance - for an ideal infinitely long wire generating the force. I suspect the magnets are nearer a point source and the force would decline even faster. That suggests one larger magnet or two magnets laid front to back, not stacked.
I think it's inevitable for a wear pattern to develop on the roadway from the friction of the steering contact. Stronger paint types may resist it longer but all will wear eventually. Any insight how to deal with that?
In an earlier thread you spec.'d the oil based enamel at ~mil; an order of magnitude thinner than distances you measured that deteriorate the steering. Hopefully you won't have to strip the roadway after a few touch-up coats. It will depend on the rate of wear.