Again I wonder about hypering the paper. Is that nutty?
Just to add to my prior suggestion of pulling the darkslide in increments to do an exposure check: you could then take this graded exposure sheet and develop it in increments in direction perpendicular to your exposure steps, just by pulling it out of the developer at successive time intervals. So then you have constructed an exposure vs. development matrix and you can just pick which square works best.
The big wildcard, I think, is the colour temp of the light you use. Methinks that could affect the "speed" of the paper substantially.