when the matchmaker puts together the teams does it (ideally) purely match players of similar SR so that both teams have close enough the same SR overall?
so in theory you could have 6 healer mains vs 6 dps mains?
if it does matchmake like that, wouldnt it be better to match both teams purely on the aggrigate stats? so that both teams overall stats are as evenly matched as possible?
that way the matchmaker would have a kind of have a role select built in by making sure both teams would have in theory the same aggrigated damage, shielding, healing output potential.
This might be a nice idea for solo Q but it depands heavily on stats. This could have bad side effects due to the diverse hero roster.
I’d give it a try but more as an additional (soft) condition, simply to avoid pure dps vs pure support situations but I don’t know how often that actually is the case (due to private profiles).
Matches are actually made by internal MMR, not SR. See How Competitive Skill Rating Works (Season 14) for details. However, for the purposes of your question, the distinction is not important.
Yes.
Matching on stats tends to make people chase stats rather than the win, making the game kinda toxic and broken.
As your system is imagined, if someone plays healer once during their first game of the season (to be nice, they don’t actually won’t to heal) the matchmaker will type cast them as a healer until they throw enough games (by playing dps or tank when the matchmaker expect them to heal) to get whatever role they prefer. This is broken.
1 is irelevant really. The incentive is still to win. The aim is to just make both teams equal in stat potential.
2 yes but you demonstrated a potential to heal at that level meaning you could possibly do it again so swapping roles wouldn’t really matter.
To me a matchmaker that can put 6dps mains against 6 healer mains and consider it fair seems strange.
I wouldn call it broken, rather improvable. I thought about the type cast situation but than I remembered how we currently match teams, randomly (all possibilities).
Depending on Blizzards trust in the validity of their data, you could possibly use this method in order to avoid a significant portion of the random match ups, you know are “bad”, basically the extremes.
A simplified example: 6 supports vs 6 dps players, I’d assume that the average healing and damage dealt on heroes would differ a lot when all of them are close in MMR.
As a player: I don’t trust data I’m not aware of hence the skepticism towards PBSR.
Out of a devs POV: I’d check the date. It might turn out that you can’t differentiate between any of those cases, even the extrem once but I’d still give it a try.
I’ve been thinking about this idea. If the stats were also associated with heroes it’s not such a big problem, although the other issues with such a system mentioned above still apply.
But if it only looks at stats and not which hero they’re on then it simply won’t work. A really good Widow gets one-shot kills often. They don’t do a huge amount of damage. A Zarya or D.Va would typically do a lot more damage. So the damage stat doesn’t really tell the system what the person plays. The same would occur for other stats too. If the system doesn’t match stats with heroes then a Roadhog main could look like a support player focusing too much on damage and not enough on healing.