Scott wrote a long and informative post on Groups and Matchmaking in Overwatch. However, rather glaring in it’s absence was any mention of MMR.
So is MMR gone? Are matches now made entirely by SR? That’s certainly what this post says.
But the post doesn’t say that things actually changed, so…?
If MMR is indeed gone, how is matchmaking handled for decayed players? Are decayed GM’s being matched against diamonds now? [Edit: No. Seagull ground up from decay this week, and he was placed in GM/Top 500 matches while still in “diamond” Twitch] Or does “MMR” still exist, but only for decayed players?
Scott does say that the SR buff on returning from decay is still in place.
An aside: This post definitely changed things, without asserting that it was a change. Here, it is asserted that group size has no effect in determining win probability and hence matchmaking.
This did not use to be the case:
The other issue is how do we model the synergistic effects of players being together in a group. As you noted, they have access to voice chat. Now here’s where things get interesting. This “massive” advantage actually differs based upon the skill rating of the group members. Based upon the data we’ve seen groups of low to mid SR players don’t see that much improvement to their win %. Higher SR players do see more notable improvements, but it’s not as huge as you might think. Still, we do take this into account when we predict the win% for each team. Regardless of how the data looks, we do know there’s a perception of a large advantage for groups. That’s one of the reasons why we explicitly try to match similar sized groups together.
I just assumed he used “SR” interchangeably with “MMR” (even though they’re different).
The change to how groups are handled is interesting! I think it’s likely more accurate with the update. I never thought being grouped was as big an advantage as others thought.
One hypothesis, which does have some support, is that Scott is just a bad communicator.
Neither did I. I was actually convinced that being in a group (unless you really stayed together for many gaming sessions) was a net loss of SR. It was really nice that Scott gave numbers, and that the numbers were so small: 6 stack vs solo is only worth a few percent in win percentage points.
Pretty sure MMR is still there. They never said they were doing away with it, they just cleverly avoided saying yes or no. Reason being is because its a flawed system.
While MMR is, in my opinion, a flawed concept, I can see Blizzard’s reasoning behind it. They don’t want people to game the SR system for stats; they want people to focus on winning instead. So, I can see that reasoning.
But, it’s clear that Blizzard doesn’t want to reveal their MMR system. You can tell by the language that they are using. In short, anytime they try to explain the MMR system, it all boils down to git gud —> rise in MMR, even though we specifically don’t know ourselves how to increase MMR.
This one is really difficult to analyze, actually.
Why? There are two distinct types of stacks.
Serious stacks, ones we call pre-mades. They have voice chat, coordination, they’ve been playing together for quite some time.
And then there’s the meme stacks. The kind of “lul we won a match once let’s group up ha ha”, which typically does not have voice chat, does not have coordination, and plays the same way it does in solo queue.
The former stack will always beat the latter stack unless the latter stack has significant individual skill advantage, but the game considers both kinds of stacks equal, so if you have a latter stack fighting against a former stack - gg.
At some point you will realise I’ve been right all along, and your interpretations were faulty exactly where I called it.
As for closely - it’s disproven by out of season matches where the SR limit as described is inactive and people play matches 400+ SR outside their last ranking.
And, as this very post you made admits, something said years ago may not currently be valid or as accurate as it once was.
This is definitely an issue. It will be interesting to see how LFG works out. Serious six stacks out there will farm LFG groups. Which will cause LFG groups to fall apart. If there are enough LFG groups out there that they are mostly facing each other, it should be fine. But if there aren’t, people could start to decide that the LFG tool is a trap, just like the “stay as team” button is.
Do you have actual data here? The season ending SR of every player in many matches during the off season?
Regardless, though, the off season is pretty memey with the low population and the lack of consequences. I’d really hesitate to draw much in the way of conclusions from it.
Sure. If the developers say things have changed, I change my understanding. If they just contradict themselves, I’ll eventually come around, but it will take longer (assuming they are consistent with the new explanations for a while).
I don’t expect to end up saying that SR is part of MMR. Because they never said anything like that, and it would be a silly thing to do.
However, if they suddenly did say that SR is part of MMR, then I’d dutifully update my guide, and add a comment about how silly/broken/dubious the design is.
People are gaming the mmr. Top 500s create 2 or 3 person stacks and have mercy damage boost one of them consistently. This causes them to get ults every 30 secs or so, which raises the mmr cap for everyone else.
No way anybody single queing will ever get the same amount of" damage per 10 min" with that same hero/map without having a dedicated damage boost.
They end up getting vwry high damage and maybe 2 or 3 times are more ults than anyone single queing. This causes matchmaker to believe anybody else getting less damage belongs further down in the sr ladder.
It is a perpetual cycle about the mmr issue, and I will admit that those types of problems can happen. Until Blizzard reveals how mmr works or scrap the system, there is really nothing that we can do about that besides report and hope for the best.
Luckily, the mmr system quickly promotes smurfs and anyone queuing with that smurf out of the tier so they can stop messing around with other people. Then again, that gets into the problem of boosted players and smurfs intentionally throwing, but more on that another day. The thing is, all of it, the mmr system, throwing, smurfing, all happens because Blizzard doesn’t know how to specifically find the type of people that you are talking about within a large playerbase.
the performance based calculations you are talking about are for SR only, (not for MMR).
The removal of PBSR (performance based SR) was literally pointless because MMR is still in the background and dictating where it is you belong on the SR scale.
Your MMR is based on a number of stats including elims per game, damage per 10 min, how many successful ults you got in each game/round, etc. It is calculated based on the hero you played, the map, and how well you did VS everyone else who played the same hero on the same map.
My comment refers to the top 500 stacks more like top 10/20 etc who are getting heavy damage boosts, causing them to ult every 30-45 seconds consistently, and this raises the cap so high that even a gm or top 500 player who solo queues will never get the same amount of damage or amount of ults.
They need to remove MMR and do flat SR gains/losses ONLY. Or split the queue into single players and team stacks. (2 leaderboards).
Having it the way they do is unfair for anyone that single stacks as there is no way their stats can ever match the stats of someone playing the same hero on a stack (and with a damage boost bonus from mercy)
