Pretty confident there are two checks. First is to see if it triggers. And second to see whether the effect actually takes. (As far as I know, seems that negative effects can trigger even on a dodge, but not certain. If not, that would of course be a third check.)
As far as the tenacity v potency check, it is opaque and I don't know the math. If someone is willing to collect the data and reverse engineer it, that's doable.
As far as my working hypothesis, I'd assumed that the baseline is some arbitrary number, say 1 for simplicity. So 0 tenacity v 0 potency would be 1:1, or a 50% chance of applying. Then, 100% v 0 could be 2:1, or a 66% chance of application. Etc.
Of course, I just pulled this right out of my arse. Both the initial baseline (50%) and scaling function assumptions could be totally wrong. But if someone's willing to test hypothesizes against observations, it's crackable. Just curious whether someone has already bothered (virtually certain) and willing to share.
