I'm the "statistician" in question. I'm actually a physicist with a math minor who now makes his living as a software developer. The guys at CG had me run probability checks against two different dice rolling systems, one of which was eventually used. I decided to brute force it because a) I was a bit rusty with my probabilities (it's been 16 years since I took that class, and I've mostly used it for statistical and quantum mechanics problems since then), b) I was comparing each die set for multiple interpretations of the rules they'd given me, and c) they weren't paying me to do it; they just wanted the results. The program I wrote gave very detailed output for each of the two systems so that I could check any result if I so desired to make sure it was sorted correctly for each rule interpretation. I was originally asked to provide probabilities for dice pools between 2 and 20 in size. The smaller dice pools were trivial, but as the pool size grew, so did the time needed for that run. It was originally running with some graphic output so I could see progress, but that was slowing it down too much. With that, running a pool of 9 wound up taking most of a day. When I dropped the bells and whistles I was able to get up to 14 dice before the times needed for brute force got to that point.