Forced 50/50 not real. The reality is just as bad though

Forced 50% win rate does not exist. It’s confusing. There is nothing forced. It’s more like a prediction. Matchmaker looks at SR and MMR of 12 players and predicts how best to separate them into 2 evenly matched teams. This often FEELS like a forced 50/50, as many claim, but it isn’t.

The system does not look at a high MMR player and determine that that player needs a loss. The system looks at a pool of 12 separate MMRs and, unfortunately, high MMR players get placed with low MMR players to even out the overall MMR of the team.

The result is usually much the same as a forced 50% win rate would feel, since the higher MMR players have to work significantly harder and still frequently end up with a loss. The better you play, the higher your MMR, the tougher your games become.

To me, this isn’t fair at all and makes SR meaningless. My games should get tougher as I climb to my appropriate ELO and play with players the same level of skill as myself. Instead, the hidden MMR mechanism creates a literal roller coaster of matches all withing a small range of SR. At any given SR there is a huge range of skill.

Why is it this way? The stated purpose is to blunt the impact of smurfs. By placing smurfs with 5 other much lower MMR players, the game is sill enjoyable and winnable for the other team. The sad truth is that the hidden MMR system has resulted in such huge ranges of skill at any given SR that any decent player will eventually find themselves suffering loss streaks on teams with players of a lower MMR.

It’s not a forced 50/50. It’s just as bad though.

7 Likes

The way you word this assumes you are always the high mmr player being grouped with lower mmr players and made to carry.

You cant actually think that. You know its disingenuous and bias to make it sound like that. Just as often as you are one of the high mmr players you are also one of the low ones or middle ones w/e randomly.

Its fun for me as well to just play better and stay against lower skilled players. Too much fun.

Think of what you want here and think of the consequence of getting what you want. If you are playing better you are going to move up maybe only slightly and you are going to have harder games as a result.

3 Likes

You make it sound as if the closer you get to GM, the system will start putting you on a bronze team to even you out.

It’s nowhere that severe.

They get a group of 11 other players that are as close to your ranking as possible, then split it up from there. So, realistically, everyone is roughly the same skill level, so close that you shouldn’t notice the difference.

The only time there’s any noteable difference between MMR/SR is when you’re playing in a lowly populated rank/time/group queue so the system has to force some wiggle room just so you can find a game.

2 Likes

Its not how you describe it, the system doesnt just take 12 random players of the same sr and sort them based on their mmr. The system tries to find people of the same mmr and around the same sr first, and then broadens the mmr range as necessary while searching. You sometimes find people with up to 300 sr difference who get put together cause they’re similar in skill level and thats just the highest i noticed and i dont check very often.

The fluctuating in difficulty doesnt come from good and bad people getting put on the same team to create a “balanced” team. It comes from the fact that skill and performance are very situational and not everything is or can be considered in mmr.

A really common example: a match with 3 dps mains and one of them plays tank cause he wants 2-2-2. He doesnt have as much expierience on tanks as he has on dps, so his skill level as tank may be lower than as a dps, but because he got put with people on the same skill level as he has on his main he may appear to be a bad player compared to the rest of his team.

Just like that there are many more factors that arent considered, such as maps, team comps (including the enemy team) or simply having a good or bad day. All these can affect performance but cant be tracked by your mmr, so even if you put 12 people with exactly the same mmr together it can still result in one team being steamrolled by the other.

Bang on, OP! The purpose of Match Making Rating is to improve the optics of Overwatch games (boosting sales revenue), at the expense of players’ mobility in the ranking system. Matchmaking doesn’t seek to make Competitive Play fair; it seeks to give competitive players the illusion of fairness.

3 Likes

Yes, this is my experience.

Where is this stated?

This is how it should be, yes?

This is basically how it is.

There is a large variation in skill at a given SR, but that is because SR/MMR is inaccurate. It is not caused by any hidden MMR mechanism. Among other things, for active players, MMR and SR are in sync (Overwatch Forums). So inaccuracies in SR are matched by inaccuracies in MMR.

To summarize the system:

(4) Overwatch Forums “In Overwatch, whether your MMR goes up or down is contingent on winning or losing. But there are a number of factors that determine how much that rating goes up or down.”
(7) Overwatch Forums
(23) Overwatch Forums
(32) Overwatch Forums
(47) Twitch

I sincerely hope you understand the difference between a flat SR system where SR is the only metric, vs Overwatche’s hidden MMR system. Based upon the text you chose to cherry pick from my post, it appears that you do not.

I think matches should be determined by sr and sr only, they often are, but when people perform well above their rank the matchmaker tries way too hard to find fair matches when they should just let that person stomp until they reach their natural sr. An 50 percentage chance of winning is determined by two teams who are close to their elos, with no tempering with the use of stats.

Without performance based sr in diamond smurfs are chronic stuck and matchmaking does a an actual terrible job of accelerating alt accounts in this game. If somebody is good let them stomp.

1 Like

I understand your view of the system. It is incorrect. I only highlight the sentences of your post that highlight your incorrect view.

Specifically, your view is not compatible with Jeff’s statement that MMR and SR are closely linked.

And your statement of: “The stated purpose is to blunt the impact of smurfs?” By whom? Where?

1 Like

That’s actually one of the functions of a dual rating system.

If there was just one raw rating system then players could camp the top of the ladder each season after finishing placements. Decay is used to prevent this and force activity. However, if decay is applied to the same rating as used for matchmaking then it can be abused to forcefully keep an account especially smurfs queuing much lower than the players actual skill.

By using two ratings decay keeps players from camping, but no amount of decay queues players against lower tiers.

Blizzard adopted this system back in WoW because Glad rated players who provided boosting for others could just reset their rating to the default 1500. This would queue them against mostly noob/undergeared players, which they could easily crush. By adding a hidden MMR resetting the visible rating didn’t do anything other than abandoning their position on the ladder. They’d still get queued as if they were Glad level.

There are no rating resets in OW, and although camping the ladder is prevented with the current system, smurfing is not.

although most of your arguments are fine, here’s another thought.

player skill IS NOT CONSTANT

say I play 5 different heroes in competitive. My skill will vary depending on

  1. what hero I play
    I can be a 2200SR Tracer and a 2400 Dva
  2. how I am on that day
    my Dva can be 2100 on a bad day and a 2600 on a good day

my performance can also vary throughout the map. this is much more noticeable with heroes like Widowmaker. You might not get a lot of hits, then you get 2 headshots and you win the fight.

so, NO the matchmaker does not try to force smurfs out of the game at the expense of 5 other players.

1 Like