[{quoted}](name=Nagoh Shan,realm=NA,application-id=3ErqAdtq,discussion-id=GVYk7tkA,comment-id=0001,timestamp=2016-09-17T19:39:49.122+0000)
Autofill would be ok in a different system, but in the current system it's the reason riot is hitting none of their matchmaking goals. Those being, speed, quality, and role selection.
Scarra and Gbay were too nice to that smug little bastard in the dynamic queue roundtable, and you can tell that guy's got no clue, so I imagine that the 2017 iteration they have been talking about will miss the mark as well.
But as it stands, the person with the higher MMR will be put off-role more often due to the way autofill systems need to work. That's unavoidable.
It seems to me that Riot's systems were more designed to balance MMR pretty much blind to roles, with LP rewards/loss being adjusted based on the difference. This worked with the old systems more because players were reasonably expected to know all roles. (Since Riot promoted pick order if people didn't get the role they wanted)
Dynamic Q skewed that system by allowing people (some of the times) get their main role. Thus their MMR isn't representative of the average of all roles. (if one's skill with top lane was two divisions lower than their mid lane skill for example) Which would mean that the average MMR for the team, could be skewed as well.
And that's completely compounded by the constraints of the Dynamic role Selection, where people looking for specific roles will end up increasing the Q times. And Ultimately, the issue with Riot's goals (speed, quality, role selection) is that it's ultimately a compromise between them. So my question is, With the obvious disparity between low and high ELO, along with the role miss match I just mentioned, Can Riot actually maintain Dynamic role Selection, and still maintain ranked integrity? Or would ranked integrity be better served with the old system where everyone was still expected to know all roles?