Seems to be very loose or not a thing. I play with some people who have never been above gold ever and they got placed plat and now are getting farmed in these lobbies. We lose and lose and lose and they have a terrible time. I rarely have felt an evenly matched game. Why is it placing people so high? Seems like it’s done on purpose to make the experience feel bad.
Update: My friend who hasn’t been diamond since season 3 placed masters.
It’s exactly the same as 5v5.
It is using your open queue MMR. So there’s a lot of people trying it out that have never played open queue. So it’s calibrating them too.
But the match maker is the same for every mode.
6v6 is a mess because of the players. And theres nothing they can do.
The placements is the issue. People are placing much higher in OQ than they are in RQ.
And people are also playing OQ for the first time who are way better than the somewhat normal causals of OQ.
So the mix of people being higher ranked than they are supposed to be, plus the new influx of people who actually are high rank, is making for very lopsided matches
OW ranks were always just a farce. They never told us how their rating is calculated, so, it’s all random.
Every competitive game has some sort of Elo rating, which uses points for each win/loss. These points given are generally equal in terms of amount. So, there shouldn’t be lots of fluctuations until your rank stabilizes. And the amount given is usually just based on match outcome.
However, OW devs always said how their rating system looks at EVERYTHING in the game to decide how many points to give/take. So, god knows how it calculates it. Also because they have such vast ranges of ranks in a single match can further skew it even more.
So, basically rating system in OW has always been flawed and was never revealed how it is calculated. Many team games before used MMR point system like DOTA, LOL, CS and it was clear how they work. So, even after introducing ranked stars or whatever, we can still know that they look fair because points are always involved - and every loss/win point amount is usually same, which shows it works as expected.
OW, on the other hand is not like that, and points taken/given are wild. So, they don’t have a proper and fair Elo rating there. So, ignore the rank.
They literally said the opposite.
yes, in overwatch it just looks at the mmr of all the players in the match and whether you win or lose, plus the modifiers they show you on the rank up screen.
That’s the problem now isn’t it? How are modifiers justified? How are modifiers calculated? What are thresholds, etc.?
Any competitive Elo rating should only be based on the outcome and difference in rating. It doesn’t matter who you have on your team and how they perform.
If you lose the game you lose points, if you win the game you win points. How much you gain or lose should only be determined by the outcome and the difference in the team vs. team rating.
What’s the justification to lose more points because you’re on a loss-streak and vice-versa?
Take chess for instance. The simplest competitive 1vs1 game. You are judged by your rating and title/rank is given by what your rating is.
Default rating is given to you at first, then it adjusts based on your loss or win. Your rating is not adjusted by how many pieces you took or lost; it’s not adjusted according to how many good moves you made. It’s simple win or lose. The amount is generally the same either way with very small fluctuations based on the opponents rating.
A team game should be no different. The system should evaluate teams not individuals in teams, and there most certainly must not be so many modifiers. That’s just extra crap designed to justify a bad rating system.
The key part in Overwatch is building the teams. Because they designed their rating system in such a weird way (which I think is extremely unfair), they get away with building teams weirdly also. Thus, allowing them to create faster matches and, in essence, gamifying the competition. Essentially you’re playing an arcade game and not a competitive team vs team game instead. So, your chances of winning are not based on whether you have better rating or are better player, rather it’s based on whether your team was build better than the opponent by the algorithm.
This is what leads people to say “matchmaker is bad”, “smurfs everywhere”, “competitive is a joke”, etc…
Simple matter of fact is that if we have just a simple Elo rating system that evaluates only the outcome of team vs team fights, competitive would be much fairer and better, but queues will be longer.
For queues to be faster you’d need an extreme amount of players playing at the same time, because player ratings will be averaged to create teams and threshold you choose for building the teams will define how fair it will be.
Just look at the queue time between 5vs5 comp and 6vs6 comp. Even though former requires less players, it is still longer. Why do you think that is? Are people not playing the game? Is the algorithm having hard time building teams?
Overwatch just went downhill every since they started adding more game-modes, messing up with matchmaker/rating system and overall not knowing how to build a good competitive shooter game.
They kept dividing the playerbase to increase the marketshare, but messed up everything else in the process. Now it’s just an arcade game and it doesn’t matter that “the card says ‘competitive’”.
It never did. Some people just took it seriously.
the modifiers are:
calibration modifier
if you just did your placement matches so your rank is more uncertain you gain or lose more the next few matches after that since it’s still placing you. that makes sense.
uphill battle/reversal
just like in chess elo, if your team was at a slight disadvantage (based on mmr of the teams), you gain more for winning and less for losing and vice versa.
a modifier for a win streak or loss streak
if you’re repeatedly winning or losing could be a sign your rank is misplaced so it’s trying to get you there more quickly. also serves the purpose of getting smurfs out of lower ranks more quickly and boosted accounts back down etc
those are the only modifiers there are. these are all public, you can see them on your rank up screen. you don’t get more or less points for how you “perform” individually as you claim. the matchmaker takes into account only whether your team wins or loses, and accounts for the modifiers listed above. that’s it.
oh yes, and the reason queue times are longer for 5v5 is because you need 1 tank, 2 dps and 2 support on each team to make a match. you’ll see, if you queue tank in 5v5 you will usually get matches very quickly. there aren’t enough people queuing tank to meet the demand of the other roles as quickly. in role queue, it is actually the ratio of the demand for different roles that has a much bigger impact on queue times, rather than the total quantity of players queuing.
6v6 is open queue, so it doesn’t have this limitation. you just need 12 players, without any need to account for this balancing act between roles.
This makes me wonder. I wonder if they should experiment with 1v1 ranked on each of the role. If they can achieve creating a “fair” matchmaker based on that, they then can work their way up to 2v2, perhaps from 2 different roles. After that, 3v3 and eventually 5v5 (or 6v6).
If any bottleneck arises during the process, the devs could then better understand at what point, the matchmaker starts becoming unreliable and find a solution to it.
Aside from lack of players in certain regions, I think one of the main reason why the matchmaker is unreliable is because of different roles/heroes and how different the playstyles are between them. There’s so much factors that needs to be considered, and maybe this Elo Rating algorithm that’s been used wasn’t really designed with hero shooters in mind.
I’m concerned about the upcoming ban system potentially making the matchmaking quality worse as it could prevent one tricks from not being able to play their hero. If you get a one trick in your lobby and their hero gets banned, leading to the one trick player playing poorly, the matchmaker wouldn’t be able to predict that.
hell yeah, if they released how they rank us i am sure the game would be messy at first but become good. Open source it it Blizz 
I know you cant
Ranks were never a farce. Some people just think they are better than they actually are.
Opening sourcing it, would be interesting to see how it works. But the end result would be a dead game.