Looks like I should have brought my thoughts here instead of making a thread in the help section, but some of you may be interested. For one thing I describe a method of obtaining 'karmic dice' that is simple and works to achieve a 'less streaky' distribution that is otherwise equivalent to rolling dice.
https://forums.larian.com/ubbthreads.php?ubb=showflat&Number=860125#Post860125Interesting idea that the karmic dice distribution depends on what it's actually rolling for (i.e. on the AC needed to overcome for an attack roll). That's completely off the rails compared to what I thought the intention of the karmic dice system was.
Did you do the tests in some sort of order? (i.e. all the AC 15 non karmic then all the AC 17 non karmic, etc). I'm interested in obtaining a large data set that's in chronological order to test some time series features.
