The Problem with Overwatch's MMR

Competitive Discussion
1 2 3 7 Next
Highly Rated
Jeff and the Overwatch team have been transparent that they "balance" a match by giving both teams, not individuals, a 50% chance of victory.

What this means to the individual:

I'm going to form a short hypothetical.

Let's say Player 1 is doing amazing, 8 game win streak. The MMR system gives them a 70% chance of victory for the next match.

Player 2 is doing awful, lost 10 games in a row, terrible stats etc. The MMR system gives this player a 30% chance for victory (remember, hypothetical).

Just between those two players, we have reached a mean of 50% chance of victory.

The basic concept is that the better you do (especially consistently), the more harshly you're punished.

Any person who has played competitive for an extended period of time knows that this phenomenon is palpable.

You see it in the forums all the time : " Carried hard and the system gives me potato team" usually these posts get down voted into oblivion by the White Knights but there is sound logic as to why this is actually happening.

Overall:
This is a terrible idea for any game that wants to have a competitive ladder that is a true representation of skill and creates a very frustrating experience for all, except those fortunate ones that get carried.

**Bonus**
Not only do I notice that if I play exceptionally well i'm more likely to get bad teammates, the games I win that I get carried in i'm more likely to continue getting decent/good teammates.... weird.

Unless you can solidly carry well above your rank, it actually punishes you.

Here is a fellow community member's experience trying out this theory in action (spoiler alert it works just as hypothesized) : https://us.battle.net/forums/en/overwatch/topic/20759379241?page=1#post-7

High elo player experiencing the same phenomenon, refuting the fallacy that it's only a mid to low rank skewed perception:
https://us.battle.net/forums/en/overwatch/topic/20753337220
Highly Rated
The game punishes players for being good? And the team wonders why the competitive side of OW is so toxic.
Totally agree with the OP. I feel like this is issue overshadows everything else.

I dunno if you saw my thread on this already but let me add a link to yours in it: https://us.battle.net/forums/en/overwatch/topic/20759237722

I spent a while trying to describe this problem with a concrete example but I was missing the language, thanks for doing it here!
10/17/2017 02:50 PMPosted by Zendayo
The game punishes players for being good? And the team wonders why the competitive side of OW is so toxic.


I know it's a hotly disputed/entirely different subject, but I think having a role selection pre que would fix a lot of the toxicity.

I understand Overwatch's philosophy of change heroes etc. etc. but at some point you have to be practical.

Most people are tilting before that match even starts!!!, due to !@#$ed up team comps. I would be much happier waiting longer to guarantee a "balanced team" ( might as well have 2/2/2 be the meta because it already is) and my role or 2nd role of choice.

Maybe have two ques, one for a 2/2/2 role selection que and one free for all que, but I think we all know which one most people would choose.

League of Legends use to not have a role selection and it was a toxic nightmare. You would enter a match and people would spam - "their role" or feed - after league implemented a role selection I almost never get tilted before a match... Food for thought.
10/17/2017 02:58 PMPosted by Cuthbert
Totally agree with the OP. I feel like this is issue overshadows everything else.

I dunno if you saw my thread on this already but let me add a link to yours in it: https://us.battle.net/forums/en/overwatch/topic/20759237722

I spent a while trying to describe this problem with a concrete example but I was missing the language, thanks for doing it here!


I did see your post and it inspired me to finally put into words how I've feeling. Thanks for the great feedback to the community and sharing my link.
This is so insanely true. Winning more than 3 games in a row is tough and blizz makes it even tougher, its honestly cancerous how the odds are so stacked against climbing unless you get really lucky or really good at the game.
every idea is good.

At this point I think every contribution to a futher change of the MM is good.}

in competitive a minimum of 2 healers would be obligatory.
Couldn't explain it better myself
I noticed this when playing in bronze because it's a lot more obvious. Dominate for 3 or 4 games carry hard, game 5 is complete potatoes that do nothing and I have to do it all from team kills to standing on the payload. It's a win win for the matchmaking either I'm going to lose to get closer to 50% or I'm going to carry and win and bring them closer to 50%.
Gaben plz fix this game
This is a little too true.
08/20/2017 08:36 AMPosted by Kaawumba
Popular Myths
08/20/2017 08:36 AMPosted by Kaawumba
Matchmaking pushes a 50% win percentage using broken criteria

It varies what this broken criteria is. Recent win percentage is popular. For example, a 75% win percentage player would be matched with a 25% win percentage player (on the same team). Another popular one is that a high stats player would be matched with a low stats player. The theories typically have a few things in common:

1) Blizzard has never confirmed (and sometimes has denied) them.
2) They would lead to obviously broken matchmaking.
3) Their proponents never have solid data backing them up.
4) They usually are not compatible with Jeff’s statement, “All the system does when it comes to matching on skill is attempt to match you with people of a similar number” (32).

The supposed support for these theories are that Blizzard has indicated that they desire and are happy with a 50% win percentage (32). However, there are many ways of accomplishing a 50% win percentage, and much better ways than the theories these forum writers ascribe to. Specifically, a 50% win percentage can be accomplished using a procedure similar to that which Blizzard describes (7): As a player wins matches, he is placed with and against stronger players. As he loses matches, he is placed with and against weaker players. With time, his win rate will converge to 50%, with some random oscillation around 50%. If he gets better (compared to the rest of the community), his win percentage will go up a bit until he finds his new level.


(7) https://us.battle.net/forums/en/overwatch/topic/20753625906#post-13
(32) https://us.battle.net/forums/en/overwatch/topic/20745504371#post-3
10/17/2017 04:37 PMPosted by Kaawumba
08/20/2017 08:36 AMPosted by Kaawumba
Popular Myths
08/20/2017 08:36 AMPosted by Kaawumba
Matchmaking pushes a 50% win percentage using broken criteria

It varies what this broken criteria is. Recent win percentage is popular. For example, a 75% win percentage player would be matched with a 25% win percentage player (on the same team). Another popular one is that a high stats player would be matched with a low stats player. The theories typically have a few things in common:

1) Blizzard has never confirmed (and sometimes has denied) them.
2) They would lead to obviously broken matchmaking.
3) Their proponents never have solid data backing them up.
4) They usually are not compatible with Jeff’s statement, “All the system does when it comes to matching on skill is attempt to match you with people of a similar number” (32).

The supposed support for these theories are that Blizzard has indicated that they desire and are happy with a 50% win percentage (32). However, there are many ways of accomplishing a 50% win percentage, and much better ways than the theories these forum writers ascribe to. Specifically, a 50% win percentage can be accomplished using a procedure similar to that which Blizzard describes (7): As a player wins matches, he is placed with and against stronger players. As he loses matches, he is placed with and against weaker players. With time, his win rate will converge to 50%, with some random oscillation around 50%. If he gets better (compared to the rest of the community), his win percentage will go up a bit until he finds his new level.


(7) https://us.battle.net/forums/en/overwatch/topic/20753625906#post-13
(32) https://us.battle.net/forums/en/overwatch/topic/20745504371#post-3


By that logic, why have MMR at all? Why not just use SR? Oh ya, because MMR is affected by a large number of things, including specific things you are trying to falsify with zero evidence.
10/17/2017 06:35 PMPosted by JesseMcCree
10/17/2017 04:37 PMPosted by Kaawumba
......

(7) https://us.battle.net/forums/en/overwatch/topic/20753625906#post-13
(32) https://us.battle.net/forums/en/overwatch/topic/20745504371#post-3


By that logic, why have MMR at all? Why not just use SR? Oh ya, because MMR is affected by a large number of things, including specific things you are trying to falsify with zero evidence.


Read the rest of the post: https://us.battle.net/forums/en/overwatch/topic/20758885930 and all of the (blue post) references, and then talk to me about evidence.

But yes, having MMR separate from SR is an unnecessary complication.
10/17/2017 02:47 PMPosted by JesseMcCree

*Bonus*
I speculate that the reason Mercy mains typically tend to rise above their skill level and climb the ladder (knowing she is an easy hero and the period where she had huge SR gains) is due to the fact that the MMR system knows that the Mercy player is less skilled statistically and gives them teammates that are more likely to carry.


No, we (Mercy players) still get terrible teammates the same as anyone else. Only we have to rely on them not being dumb to make good progress.
Basically this MMR is the definition of COMMUNISM...
10/17/2017 02:47 PMPosted by JesseMcCree

I'm going to form a short hypothetical.

Let's say Player 1 is doing amazing, 8 game win streak. The MMR system gives them a 70% chance of victory for the next match.

Player 2 is doing awful, lost 10 games in a row, terrible stats etc. The MMR system gives this player a 30% chance for victory (remember, hypothetical).

Just between those two players, we have reached a mean of 50% chance of victory.

[/quote]
Theres 12 players per overwatch game.

Even if this is the case(which it is not, they DON'T track your match history, period. Read the blue posts.). People on both teams have the SAME possibility of experiencing this.

No "one" player is singled out and forced to carry. Everyone is using the same MMR and same match making system.

Yes the match quality is crap, the SR variance is huge, the matchmaker does not identify player roles or toxicity history. But one thing we are sure of is, EVERYONE experiences this, no one is singled out and forced to have badluck or have to carry.
10/17/2017 08:49 PMPosted by Stebung
10/17/2017 02:47 PMPosted by JesseMcCree

I'm going to form a short hypothetical.

Let's say Player 1 is doing amazing, 8 game win streak. The MMR system gives them a 70% chance of victory for the next match.

Player 2 is doing awful, lost 10 games in a row, terrible stats etc. The MMR system gives this player a 30% chance for victory (remember, hypothetical).

Just between those two players, we have reached a mean of 50% chance of victory.

Theres 12 players per overwatch game.

Even if this is the case(which it is not, they DON'T track your match history, period. Read the blue posts.). People on both teams have the SAME possibility of experiencing this.

No "one" player is singled out and forced to carry. Everyone is using the same MMR and same match making system.

Yes the match quality is crap, the SR variance is huge, the matchmaker does not identify player roles or toxicity history. But one thing we are sure of is, EVERYONE experiences this, no one is singled out and forced to have badluck or have to carry.


We all have the same chance to get the same items from loot boxes too. Do we all have the same items at the same levels?
At this point all hope is gone.Since matchmaking is balanced around mmr just show us a rating of mmr in stars for each player in the game.So for example 6 stars is the player with the higher mmr 5 stars the next one etc down to 1 star.Dont show the number if you dont want to but do show who the better player is so people can pick roles accordin to that at least.

I will happily fill if the rest of the team is better than me.But the amount of games i get where i feel my teammates are bots/trolling/not trying but most importantly the team has no idea how to counter enemy composition is just mind blowing.

I cant even put it in words.I ve had a game hard carrying as pharah and the team would actually ask me to change to tank.Are they trolling or are they THAT bad i really dont know.

I ve read the forums i already know what i am gonna be told.If i win half my games then this is where i belong.Maybe it is true maybe this is where i belong.But what drives me crazy is the fact that i have to fill just to get that 50%.It feels like noone else wants to fill.How am i lower sr from the guy that plays only one dps hero?I can understand being lower ranked from tank one tricks and supports one tricks cause they are needed in every match but the dps one tricks??How the hell a dps one trick is higher ranked than me and he cant even get kills...I prefer playing dps myself but that is only cause that way you feel you hold your own fate.If you lose its cause you didnt kill the enemy and not cause someone else didnt kill them.

I am the crazy guy talking alone in team chat.I try to ask nicely.I dont talk down to others.

I do have friends online and the games are better with them but they also feel harder if we are 4stacked or above.So i guess duo queue is the best chances for me to climb.

https://masteroverwatch.com/profile/pc/eu/BlackOut-1570

I just want to find flex players and get carried to platinum.I cant climb the ladder and am sad cause bad.Add me.

Join the Conversation

Return to Forum