The reward dice and combat dice are absolutely, 100%, non-random. Without fail, I will consistently get the same numbers rolling multiple times in rewards during certain periods. This is simply not probable in a purely random system. And what’s more is that they’re always rolls that work against my gear goals. And let me be clear here, I’m not talking about anecdotally or by a small sample size, I’m talking that I’ve recorded play session reward results and found very distinct probability curves weighted against attainment of necessary gear. For instance, let’s say I have no need for the 20 result in Harvestshield Mountain but I really, really need that 1 x epic reward on a 10. I’ll run a hundred odd games during the challenge reward period. In that period there seems to be at least three distinct results ‘patterns’. In one of those, I’ll roll a lot of 20’s. In another, I’ll roll lots of 17’s, 11’s, and 4’s (this is just an example, not exact data). But never a single 10. In another pattern, it’ll be 14’s every so often (for that epic reward), and lots of 19’s, but otherwise no 10’s or other big numbers.
These patterns are incredibly improbable to the point where I have no doubt whatsoever that there is an algorithm or, more likely, an A.I. steering these results.
You can also see lots of patterns in the combats. Misses tend to come in lots of 2’s and 3’s. Crits are almost always 2 in a row. Misses will almost always come at crucial junctures where a hit is vital for success. And there also seems to be a “Marked for Death” on certain characters (usually the bard or the rogue, I find) in some dungeons where you’ll get persistent focus-fire on that character even if you exit the app and restart it to reset the dungeon a dozen times.