Yeah, more exactly the surge arrestor only has to deal with the voltage difference between the center and shield that remains when the strike comes to the entry panel. I suspect there's not much point in over-designing Polyphaser EMP arrestors to handle higher voltages than they already do since at one point the differential will spark through the coax dielectric, or jump acros the feed point anyway. Or turn it into a fluid.
A long run of (preferably buried) coax will work to limit strike energy all on its own, and is hopefully augmented by ground kits (connection from shield to the grounding system) along the way, so the arrestor only has to deal with any differential that remains.
On the flip side, if the lowest impedance path from the antenna to ground goes through the arrestor and into the house, it doesn't matter if you had the world's best arrestor there since the most of the strike will enter the house anyway - through the shield.
Now theoretically, even if you have the strike going through your station, if you have a ground system that ties everything together with low enough impedance, everything should rise in potential at approximately the same time, but I sure don't want to be there to test it out if it happens.

PS: I'm an engineer, but not an electrical engineer, so take the above with some salt.