.5 difference in average DPR doesn't account for wasted damage during overkill, doubly wasted damage during crit overkill, and the exact round # where a monster dies at, which is the only metric that matters in D&D.
Until you show me the stats of a sim between a large squad of greataxe wielding dwarfs vs maul wielding dwarfs, vs wave after wave of orcs, I think you reducing the difference between 1d12 and 2d6 to 0.5 is simplistic and wrong (considering the new crit rule with is probably just double damage, thus exacerbating this difference even further).
The benefit of hitting average DPR more often vs less often (which 2d6 achieves over 1d12), is something you're not considering in your facile analysis. Show actual sim data vs a variety of monsters and we'll chat further then. I think the results will be far greater than .5 DPR. I want the results in # of rounds it took to kill X. Don't forget, for hard to hit opponents, when you roll a 1 on your d12, you might not get a chance to deal damage again for another round or 2. Hitting the average damage on a hit values hits more, because rolling a 12 on a d12 vs 2 6s on 2d6 doesn't matter as much as dealing 6 + mods (the mean damage), which is quite possibly sufficient to finish the job. At least finish it 1 round sooner. Most mobs don't last more than 2-3 rounds, so the quicker you converge to average DPR the more benefit it gets you on average, in terms of # of rounds it took you to kill that monster. All that matters is how much HP the monster has right now, and whether the average (considering the standard deviation) enough to kill it or not. For every time the d12 allows you to kill a foe faster than a 2d6, it also slows you down the next time. Hitting average DPR more often allows you to kill stuff more reliably, and thus better / faster / better / stronger. Unreliable stuff is crud.
2d6 is potent vs 1d12, and not simply because it gives .5 dpr benefit or +1 benefit on crits. There is more to the story here.