D&D and other RPGs grew out of wargames that used to solve fighting with a single die roll on a CRT. If you had 3:2 odds, then you had 66% chance of winning a single die roll. Lose it, and you lost your unit. You only got one shot at it. Fortunately, a wargame was strategic -- you had other units to use if this one got killed.
Before games exploded with multiple mechanics for multiple die rolls, there were gamers who meticulously described every action they performed -- how they looked behind them, how they opened doors, how they held their swords, how they opened the chest with a ten-foot-pole -- because by describing what they were doing, they didn't have to roll any dice. When gamers fail a roll, they don't think that disorder can be forestalled -- they think about how, next time, they will prevent ANY dice from being rolled.
Even the first edition of D&D had the core conceit that the game-master should be the last word on everything, up to and over-riding randomness when it was inappropriate. That's because, in a role-playing game, arbitrary one-shot randomness is the enemy of good story-telling, good narrative, and good genre.
Given that Next seems to be hewing pretty close to that early edition playstyle referred to above, it's interesting to think that players might have adopted that approach not because it was fun or because it was right, but because it overcame an inherent problem with the d20 resolution mechanic.
It makes me wonder whether Next would be better served by looking at different resolution mechanics (perhaps roll 3d20, take the middle) to take away that inherent swingyness and emplower players by giving them more certainty as to what their characters can achieve.
WotC has moved in that direction with the rogue's skill mastery, but maybe they should be looking at the issue more widely?