Ok, I just ran through a series of attacks in a scenario, video recorded it, watched the recording and wrote down the exact series of attacks, etc. with who attacked who, the hex numbers, die rolls, etc., so I could do the exact same thing in another run through of the scenario. I did the second run through (and video recorded it also) and checked the die rolls of the exact same series of attacks. The results, 53 die rolls, were exactly the same (which I verified again when I review the second recording).
After reviewing some statistical blogs (I took other math courses instead of statistics in college), the chance of rolling a certain number on a six-sided fair dice is 1/6 or 0.16667. That's the same probability each time the dice is rolled. I'm going to take a flier here and say that the probability of the same results happening (i.e., the same number on consecutive rolls) when two sets of rolls are compared with a six-sided dice is 1/(0.16667)^n where n is the number of rolls in the series.
In other words, if I rolled the dice six times and got, consecutively, 6,6,4,3,4,5 (that was one AI attack in the scenario) and then rolled the dice six more times and got 6,6,4,3,4,5 that would be a 1/(0.16667)^6 or 0.00002143 or 1 in 46,656 times occurrence. The chance of the same numbers coming up in to series of 53 is, as we used to say in engineering class when we didn't have time in class to calculate it, a BIG number.
Even if my guess at the math is incorrect, I'm still saying that the dice rolls are not random. Is this a bad thing? Well, not really in one sense as if you ended up doing the mistake I did and want to get back the exact same state prior to your screw up then this is great. However, if you're thinking that the game has a truly randomness to it, you are mistaken; at least in my opinion. I could be wrong but doubt it from my work.
Anyway, take care and enjoy the game.