Just be aware that if you roll 3 dice and then pick the best (aka
lowest) the chance for one of them being low is much higher. For example
to get a failure you have to roll: First die 14 or higher, second one
above 7+, third die 4+. That results in a failure chance of
(0,35*0.70*0.85) = 21% instead of 35% as it should. Unless I misunderstand what you're trying to do there. Even
if you say "Regular has to be a success, otherwise it's a failure. Then
if Good is also a success it's a good success, and if Regular, Good and
Amazing are successes are then it's an amazing", it doesn't work too
well. You'll now have the failure probability right, but good and
amazing are now too low. For Amazing Success you'd need 13 or lower on
the first, 6 or lower on the second and 3 or lower on the last die.
(0.65*0.30*0.15) = 3% instead of 15%. Hm, that gives me an idea. [[d20<13]] (Regular), [[d13<6]] (Good), [[d6<3]] (Amazing) and
it would have to be like I said above, only if all 3 show 1 then it's a
amazing success. If Regular and Amazing are but Good isn't, then it's
just regular. If Regular is 0, then it fails, no matter what the other 3
show. I don't think you can automate that step unfortunately. It's
not super easy to understand why, but the probabilities are still the
same. 35% for failure, 35% for regular success, 15% for Good success and
15% for amazing success, just as it should be. (if you really wanna know I can try to do the math here for you) (Basically
you already checked if it's 13 or lower on the first die, and only if it is you look at the second die, so you can
now ignore 14+ on that one as those are impossible results, instead on a
full d20. Then you see if it's also below 6, and so on)