In comp it does already. They match by role across the 2 teams. In QP they don’t.
I’m going to be brutally honest with you.
If you can’t get out of diamond solo you are not overperforming and are just in denial. Until you lose that idea you’ll never improve and climb.
Anyone who deserves to be in high elo hard stomps everything down there and climbs with 80%+ winrates. Over lots of games it just doesnt matter what your team does because you’ll be killing so many of their team every single fight. Sure some matches may be lost regardless but over a season your team really is irrelevant.
2 Likes
This “loss streak prevention” was removed shortly after because it had no effect other then increasing queue times.
1 Like
You are right and wrong.
Yes, the matchmaker attempts to create matches that would be as fair as possible.
But a perfectly fair match would in fact mean one where both teams have exactly 50% chance of winning.
However, it does this based on simply on the mmr of players, attempting to create games where everyone is also on similar mmr on both sides.
So, basically, if your own mmr reflects your own skill, then the matchmaker is placing you accurately in an attempt to make matches that would be fair by its assesment, and this then goes for everyone else in your lobbies.
If you actually are better than what your mmr would suggest, you would “throw the matchmaker off” in a very noticeable way and swing your team winning chances by high percentages every match.
That’s why top 500 players who do U2GM’s have 80+% winrates in plat etc.
This is what it does right now. It does not care if someone has been losing or winning or whatever, it only looks at their current mmr at that time they hit queue.
And this is also not happening. ~50% winrate is the natural end result when you play with players of your own skill.
e.g. What else do you think your winrate should be when you are playing with and against people who are equally skilled as you? Higher than 50? Lower?
Do you have any proof that such a sophisticated system exists?
I don’t know if I’m bad or good at the game when I go on major lose streaks either in qp or comp. As soon I can do something in a match it feels the other team just hard focus me… So now I just play for fun and don’t care….
Are you guys getting 50% winrate?! Feels more like 30% winrate these days since the end of the last season. I miss season 1-3 of OW2. I was like 60% winrate playing a trash hero like Soldier.
2 Likes
It does do skill. your chance of winning should be around 50%, and thats what it aims for.
Wait really? Do you know when or have the patch notes for it? No biggie if not (or you don’t care), just curious.
I know that this system basically doesn’t matter and its effects got exaggerated by the “it’s EOMM”-people, but still.
It was never in the patch notes, but they confirmed it in one of the matchmaking interviews: https://www.youtube.com/watch?v=d_Ut8pCH9QM
1 Like
jesus these devs are so horrible at PR
announce this thing so poorly worded that it’s easily misunderstood, then delete it and don’t announce anything at all, just mentioning it casually in a streamer interview 
/end short rant
I think they less they talk about these kind of changes the better.
Some people will take everything out of context, no matter how you word it.
3 Likes
Why do people still not understand that this is the hallmark of a matchmaking system that (at least at a rudimentary level) is working.
You’re winning 50% of games because on average you have an equal chance of winning or losing each game. The alternative is low rank players lose more than 50% of games (because probability wise, they’re more likely to face opponents of higher skill) and high rank players win more than 50% of games (because probability wise, they’re more likely to face opponents of lower skill)
The only time you won’t have 50% win rate is if the mm is wrong - typically because you’re either improving or getting worse at the game.
I don’t know if it is S9 changes or what but now it just seems I’m either worse at the game or miss matched… Like Fiat panda Vs a F1 car…
1 Like
I really think it’s more of a feel thing at this point. Sometimes the games just feel terrible or mismatched in either direction. So individual game quality may be off, but I don’t think they’re moving mountains to try to knock players off their perch.
I started checking in on my W/L in QP every so often to keep me (and it) honest. Nothing crazy, just marking the total down and calculating the difference from last time whenever I happen to think about it (maybe some streak or random experience happened that specific day or whatever), and it’s been roughly 100-ish match chunks. Each time so far, I’ve been positive.
On individual heroes if I bother to check, sometimes I’ll trend down, but overall, the winrate keeps climbing, which surprised the heck outta me. If it’s trying to knock me back down to 50%, it’s not really pulling it off.
Me too, kindred spirits. I’m averaging a 66% win rate with Junker Queen and all the new heroes with over 30 to 40 hours played are averaging above 55%. You can see my profile and see the proof yourself, if you want.
Fact is, this past week I was under 30% win rate with Emre, and over 80% with Anran 3 weeks ago and above 70% win rate with Wuyang 2 weeks ago. I only play Quick Play, but I did run an experiment back during Season 1 of New Overwatch (this year) to see how far I could get on during Drives Week. Got into the mid-diamonds for Support and Tank, didn’t get enough games played on Damage-role to move past Plat. Though, much of my data gathering there suggested some players were ranking higher than their general skill level… or the general skill level of the playerbase has declined.
Basically looking for an average, but there’s also a shortage of players in two roles, so it’ll find what it can.
Also because there’s no tutorials, nor in-game instructions, players aren’t really given any sort of guidance towards learning how to play any of the roles. They have to “intuit” what each role does and how each character plays, and most of the time they “intuit” wrong or whatever works for their skill level - and never really move past it.
Correct. It will trend towards 50% or close to it.
Again you can see in my profile of some of my most players characters in OW1 are mostly trending towards 50%. I don’t really play them much anymore (save Winston and Rein).
Sorta. The problem is the Role Queue system is trying to find players of similar skill levels, but there’s a tremendous lack of players in the Tank and Support roles, so it can’t really find players of similar skill levels within the time frame Blizzard gives out.
If given enough time, the MMR will probably find players of exacting averages and do what it is supposed to do. The problem is, it doesn’t have that time, and Blizzard isn’t willing to give it, nor educate its playerbase about each role.
Both Competitive and Quick Play use the same system. Quick Play just has a shorter time frame to do it in, so it can be less discerning than Competitive.
Unfortunately, Blizzard’s always been bad at communicating. Overwatch isn’t any better or worse than their other games, and too many players often misunderstand their intentions.
This creates a feedback loop where Blizzard doesn’t want to communicate because they’re being misunderstood, so things like the loss-streak prevention removal go unmentioned because the players might misunderstand why it was removed.
Well you can see these forums when players see their favorite characters take a nerf or adjustment or misinterpret some changes that aren’t either.
As I mentioned above, the game has a lack of players in other roles outside of the Damage-role, so trying to fit you into a perfectly matched game of equal skill won’t happen often. More often than not, you’ll be fitted into a game of just about anyone because the MMR system doesn’t have the time to find the right players.
It explains why I keep trending towards positive routes when I play Tank or Support or queue into All Roles. The game just can’t find players of my skill level often, and I have more pull/value or leverage playing a Tank or Support over the Damage-role.
30%, that’s not bad at all. I’ve won like 10 out of last 50 matches. Granted, I’m playing a chracter that isn’t particularly great and I don’t know, but it’s tough for my teammates to carry my weight. 
1 Like
I have 38-44% winrate in QP for last 2 seasons. Does it mean MM is working…or not?
(post deleted by author)
Uh, that’s literally what will happen when you put players of similar skill against each other.