Forced Winrate is Semantics

Win 10 games or lose 10 games in a row and then see how your teammates get worse and better, respectively. You’ll see that you are wrong and why many pros swear that some games are simply unwinnable.

2 Likes

That is literally how it works. If it wasn’t it would be impossible to climb or fall rank. All you need is a single data point to prove a point wrong. Unfortunately for you there are millions of them. That’s how many people have climbed rank and fallen rank. You will not have a 50% winrate until you reach the ELO you actually belong in.

2 Likes

I’m trying to give you good news here, but you won’t take yes for an answer. We already have the benevolent matchmaker that you wanted. Everything you’ve witnessed can be explained in such a system.

Every time you win a game, your MMR goes up, and the next game gets a little bit more challenging. That’s because you are being matched both with and against players with slightly higher MMR than before. Your actual skill hasn’t changed in such a short time, so every time the quality of players around you goes up, your chances of losing the game go up too. If you get a lucky streak of 10 wins, the matchmaker is now overestimating your skills.

In reality, skill based matchmaking in a team game as complex as this one is a difficult problem, and nobody’s MMR is perfect. So a game that looks fair to the matchmaker might be unbalanced in either direction, purely by random chance. Maybe players are playing heroes they aren’t as good at this game. Maybe they’re one-tricking and encountering a bad matchup. Maybe the team’s play styles just don’t work well together. Who knows.

I agree that some games are unwinnable. But it’s absolutely not because the matchmaker designed them to be unwinnable. It’s just the result of an imperfect system trying its best.

1 Like

Spend a minute browsing through the forums and you’ll see many posts of people complaining that they won 6-7 games in a series and didn’t rank up. You’ll also see people talking about the opposite, where they lose 6-7 games and somehow rank up. This is a separate issue from what I have been discussing in this thread, and it appears that ranking up is more dependent on the stats of individual players (deaths, elims, damage, healing, and so on) than winrate.

Comp doesnt feel forced but QP is 100% forced win rate this is how it works, say you go on a win streak, you would expect to be playing with more and more skilled players as you go. The fact that you dont is proof. If you win too much in QP the matchmaker will place new players on your team that shoot the ground, stagger in, ult all alone, and die 20 times. Do you honestly expect me to believe this person was a skill set above me 5 games ago?

how the system SHOULD WORK.

  1. pick 10 players in the same range of latency and skill.
  2. assign them a match rank order based on their hidden MMR score. Rank 1 - 10.
  3. Place odd number players on team 1 and even number players on team 2.

that simple. Devs admitted to putting new players in matches with veteran players so they could show off their skins and skills to impress the new players.

1 Like

There’s not a lot of concrete information about MMR vs SR, and anything that wasn’t explicitly stated by a developer I consider to be nothing but speculation. They haven’t discussed it in a while but this thread is the best overview I could find:

Jeff specifically says here that your MMR is the sole determining factor in who you match against. SR is merely decoration, it’s a psychological trick. All you really need to know is that MMR is the game’s best effort to estimate what your skill currently is, and it is always trying to match you against other players who have a similar MMR. But as I explained above, this is a fundamentally imperfect process and stomps cannot be completely avoided.

“Jeff specifically says here that your MMR is the sole determining factor in who you match against.”

Now you’re getting warmer. MMR is the number that I was referring to that I simplified by saying it is between -10 to +10 to make the concept easily explainable in a previous post. The individual MMR of each team member is used to calculate a number for the team, let’s refer to is as “team MMR.” This is the mechanism used to “balance” games and create a 50% chance for each team to win each game. The logical extension of this concept is what I’ve stated a few times already: at any given rank, the better you are the worse your teammates are, and the worse you are the better your teammates are.

One thing though, I’m not sure why you have been claiming that I’m supporting this system when this entire thread has been oriented around how this type of “fair” matchmaking is undesirable. I don’t want to be punished for playing well relative to my rank or rewarded for playing poorly for my rank; what I want is a system that doesn’t factor anything, including MMR, into matchmaking other than my rank.

MMR is literally the Match Maker Rating. It is SR after the scaling maths have been applied. The matchmaker doesn’t see SR.

Noone said it did. Your post is nonsense.

What you’re saying makes absolutely no sense. You don’t get punished for playing well or rewarded for playing playing poorly. The matchmaker is literally factoring in your rank. Rank is nothing more than an arbitrary range, where your actual MMR is specific.

You are the Alex Jones of this board lol

No, this is completely wrong and not at all logical. Let’s focus on solo queuers for the moment. If you have a 2500 MMR, you will be matched with 9 other players who have 2500~ MMR. If you improve to the point of having 3000 MMR, you will now be matched with 9 other players who have 3000~ MMR. This is how the system is intended to work.

The only way you can get people with dramatically different MMR in one game is with premade groups. In QP especially, you could have a duo with a Bronze and a GM, or a five person group with three Bronzes and two GMs. What can the matchmaker possibly do with that? Not much, except to fill the opposing team with players who match the average skill of the premade.

Maybe you’re a gold player, and your teammates are a 4 man group with two bronzes and two diamonds. You look at your bronze teammates and are baffled by how bad they are, while the enemy team is full of golds. But the matchmaker did the best it could.

Your entire understanding of the goal of the matchmaker is wrong. In every single game, it is trying to create two teams of equivalent skill. That is the sole objective of the matchmaker.

4 Likes

For it to act like you think it does, it would need to.

This implies it knows your rank, which it doesn’t. It DOES know your MMR, how long you have been waiting for matches, and what your ping times to various servers are, and which game mode and roles you have selected.

In a timely way, but yes, Dark Hobbit indeed doesn’t understand the matchmaker. But they are not alone there.

It has to do with psychology and getting players addicted and well keeping them playing. Loser’s queue has been going on for ages and if I remember I think it was a dev at EA or some type of worker but basically they wrote a document about the pros of having a losers queue.

It didn’t state that games have them but it stated games “should” atm i can’t remember where that doc was at but if it ever pops up again I’ll link that.

2 Likes

The in game effect of this is to match players with lower MMR with players of higher MMR at the same ranks. What you are implying is that every game, from silver to diamond, will either be full of 10 super players or 10 terrible players. Though SR isn’t the determining factor here, it is clearly a variable if only because of the max 1000 SR limit difference allowed between players. Your ten good players in every game or ten bad players in every game scenario, which is what would happen if matches were created by putting 10 players with the same MMR in, is clearly not the case and makes me wonder how much you have even played this game because, much more commonly, you will see one or two players on each team “popping off” (the players of roughly the same rank with higher MMR) and one or two complete duds on a team (the players of roughly the same rank with lower MMR), regardless of rank. Tired of beating a dead horse here, but this is again because the matchmaker is balancing teams by putting players of roughly the same rank into teams that have players with varying MMR. Once again, MMR is the pivotal factor and that it is included at all in matchmaking is my issue here. Another poster on this referenced how this is also the case in quickplay, and I agree 100%. I’m not sure why it is so hard for you to comprehend this distinction.

@Poundtown By punishing I mean matching you with worse players on your team the better you play, by rewarding I mean placing you with better players the worse you perform. You are clearly intellectually deficient and have contributed nothing to this thread. You aren’t worth replying to anymore.

Bingo. This is exactly why such a system is in place. It ensures that the lower MMR players will keep spending money.

1 Like

No! This is completely, unequivocally wrong! Rank, aka SR, is irrelevant. The devs have said this explicitly, see the forum thread I linked above. MMR is the only thing worth discussing here. Just forget about SR altogether.

You are saying the game matches by rank and not MMR. This is 100% confirmed to be false by Jeff Kaplan. There’s no room for ambiguity here. Players are being matched by MMR, and whenever possible, the system creates games populated entirely by players of the same skill.

Your gut feeling is to say this is impossible, because if it were true, why are there stomps? The fact is that stomps happen all the time when you randomly throw together two teams of people at equal skill. There are a number of reasons for this but I’ll focus on two. First, premades are a thing, and they complicate the matchmaker’s job because they make it impossible to have everyone at the same MMR. Second, everyone has different levels of skill with every character, but they only have one MMR per role. So depending on which character they choose to play, and how good or bad a matchup they have against the enemy team, their results will be very different. This inserts a lot of randomness into the result of what should be a balanced game. The matchmaker is not capable of accounting for this.

Even in a system of only solo queuers, where everyone plays their mains, and the matchmaker perfectly matches by MMR, you’ll still get stomps. It’s just the nature of the game.

1 Like

I have been saying the opposite of this the entire thread. Read my posts.

I only mentioned that the maximum 1000 SR difference allowed between players in ranked is a factor in matchmaking. Are you saying this isn’t the case?

It is not only possible, it is commonplace. Let’s reverse this and you can explain why “stomps” are the norm and not the exception, as they would be if every player in the game had the same MMR.

I agree with you here. Forming a premade of five is the best way to diminish the influence of the matchmaker arranging for each team to have a 50% chance to win because it no longer has influence over the MMR composition of your own team. In this scenario, the only factor that influences who can or can’t be on your team is the max 1000 SR allowed between players.

Remember the point of this thread was to point out that a forced 50% winrate and a forced chance for each team to have a 50% chance to win each game are effectually the same thing and, in my eyes, this a hugely negative aspect of Overwatch. We’re getting a bit off track but I’ve enjoyed the discussion.

Goodnight for now.

1 Like

It doesn’t seem like you’re registering what I’m saying. Let me summarize. Every single game, the matchmaker is only trying to do one thing: fill the game with players of the same MMR. That’s it, that’s all there is to it. Aside from premades, there is no reason for it to do otherwise. SR does not factor into it one bit.

In a perfectly “balanced” game with equal MMR across the board, stomps will still happen. They’re not the norm, but they’re not unusual either. Overwatch is a complex game with many unpredictable factors - no matchmaker can ever prevent stomps from happening.

1 Like

And they conveniently made it impossible to tell at a glance whether people are grouped or not, so players have even less information to work with. How is that helpful at all? I know exactly why they did it, but again, all they do is try to fix a symptom of toxicity and pretty much ignore why it’s happening in the first place: because their game is designed to be toxic. That’s all.

No, it’s not sole objective of matchmaker. It has secondary objectives as well, such as:

  • retention of players, regardless from their skill;
  • sales of cosmetics;
4 Likes

Quite simple.

Whenever you get too many unfavourable outcomes; disengage; i.e. switch off the game and do something else. If you have been playing long enough, it knows your habits and what you will tolerate. If you tolerate being a punchbag (i.e. suffering huge losing streaks but still playing) then it will give you that. Keep in mind it is a zero sum game; people must lose for others to win, and if some gamers only accept a greater than 50% win rate, then some others must accept a less than 50% win rate.

Start to disengage when you suffer multiple losses; see what happens.