The damage variance between a Critical Hit and a Non-Critical hit.

[Equitable_Remedy]

Gonna have to disagree here. While yes, variance can be fun. Too much is simply annoying. It's ridiculous when a Ninja can lose 80-100k on Hyo because the game said "screw you." I have runs on Paladin in P8S where I didn't Crit Goring or most of my Blade combo once the entire fight. Auto DHCs are necessary but kind of defeat the point. Direct Hit itself should be removed and Crit split. You still have some variance but not these insane swings pull to pull.

For your NIN example, what is that 80-100k over the course of a 10-minute savage pull? What is it over the course of a night of 10-minute savage pulls? Did the NIN's group hit enrage at 80k? What exactly is the "screw you" here other than you didn't do the theoretical maximum damage on a given pull? We should ask all of the same questions for your slightly below average PLD pulls.

Are any of these things making the difference on clears? (The answer is almost certainly no except for the edgiest of cases, though it's the internet, so I'm sure people will say that they consistently wipe at 73 HP or something.) Since there's no fine measure of damage or DPS in the game beyond threat order, are you constantly finding that order varying wildly pull to pull because of crits (as opposed to variance in gameplay by the player)? If yes (which would truly surprise me), do you think that it would be better for the game to be more static from pull to pull?