Also, the RNG IS bad ^.^ This has been tested and demonstrated with a sufficiently large data set already, a long time ago... but it's beside the point here.
For what it's worth, I think the issue has been addressed since.
A few months ago, I published a couple hundred successive d20 rolls. (I think I recorded them in patch 8, but I can't find the post to make sure). I promised analysis which never materialized for 2 reasons : I'm no statistics expert; I couldn't find the sine pattern in the data which had been demonstrated a few years ago.
As far as I can tell (which, again, isn't all that far) there's no correlation between successive rolls. That's not the only metric of randomness, but it does suggest that streaks of bad rolls are just bad luck.
Interesting caveat: One way to generate random numbers in a video game is to base them on player actions. For example, you could take a poorly generated random number and multiply it by the fourth decimal in the main character's X-axis position. The extra noise would contribute to cancelling out the faults in the initial number generation.
In my testing, a few misclicks meant my PC moved around the target rather than just hitting it. Assuming positioning is relevant to the generated d20 value, that may have added the noise the game needed to hide the faults in its RNG.
Niara won't blame me for accusing Larian of a quick, dirty fix - though she could for doing so without any actual proof. The takeaway is this: If you're superstitious and you notice you're failing a lot of rolls in combat, try moving your team around just a little. The extra noise might just help you out.
