Humans are really really bad at undestanding randomisation. Weighting dice is the best way to stop people thinking you're weighting the dice.
100% agree.
I won't believe that there is an actual problem with the current (pre patch 4) RNG until I see a detailed statistical analysis of a huge number of rolls. Humans just can't do this "by eye".
The thing is. We are playing an adventurer. We are aspiring Conans, Gandalfs or Galadriëls. Aragorn 's Andúril misses the orc 3 times in a row while being shot at with poisoned arrows and fire, but fortunately Iluvatar can reload the saved game and retry.
You can have a fine distribution after 100 rolls. Hooray, the RNG is good. Only, no combat lasts that long. So, we don't want to play mindles sword swingers who randomly hit opponents and reload when the dice go wrong. We want to be a hero, who can fail, of course. But not because of a streak of bad dice rolls. A DM will prevent this from happening.
And there are quite a lot of these streaks , too many, from my experience with the game. So, a RNG that performs OK over large samples is not necessarily OK for a game like this. (my opinion)