I suppose if you really wanted to find a curve for resist probability as a function of magic resist, you could take some data - cast some magic-based spell a lot of times (50ish?) on your druid with 80 magic resist, divide the number of resists by the total number of casts to find P(80), the probability of a resist at 80 MR. Then do the same thing at 90 MR, 100 MR, 110 MR, ... , until you start getting pure resists. I'd imagine it's probably simple and linear, but who knows!
Of course, level difference is probably a factor too, so you should also repeat these tests after your druid dings 66, 67, 68, 69 and 70
, manic mechanic of <Temerity
Refuse, rubbish, filth, slime, muck and other alt-putrescence: