MM is beyond stupid

[Thelxi]

Games frequently being very one-sided is sucking the fun out of this game. People had the exact same complaints about HOTS and Blizzard did next to nothing about it, to the point where people are actually demotivated to play that game. I really hope they do the right thing this time and actually address the issue, because overwatch is truly a great game that I would hate see end up like HOTS.

I don't even care about transparency or community management about this issue. Just fix this shit.

[Very Tired]

You mean the expected outcome of a system is happening to people, and therefore is evidence of something else at play?

So you expected the system in place that matches people up against purposely higher/lower skill levels to force a 50/50? Not sure why anyone would be expecting or wanting a system that has never been used or wanted in any other game before.

[Jessicka]

So you expected the system in place that matches people up against purposely higher/lower skill levels to force a 50/50? Not sure why anyone would be expecting or wanting a system that has never been used or wanted in any other game before.

The game is barely 2 weeks old, the MMR probably needs a much larger dataset than the 20-30 hours people have put in at this point to get a stable average that isn't kicked out significantly by a relatively short winning streak. 6 wins is going to have a much bigger impact on your first 50 games win% than on your 1000th.

[Rudol Von Stroheim]

Well, to be fair, the game in general is stupid in my modest opinion.

You're going to see a ripple effect as they begin to nerf X, Y will be too strong, and then once they nerf Y, Z will be too strong, etc. It'll be almost a non-stop ride in trying to correct various blunders that all stem from an original blunder. It's a really, really poorly balanced game once you start tinkering underneath the hood. Then MM will not work very reliably because skill across the various heroes differs and if people cannot play their stronger points, they'll likely drag the team down trying to round it off seeming much more inferior than they really are.

It's like WM needs a nerf, for sure, but what about Pharah? I already noticed her becoming more and more the "go to" for countering tough defensive positions. So you nerf her only real reliable counter and what then? She's that much stronger. Mei has an iceblock that can heal her. When they nerf McCree's stun/hammer, will he be so reliable against her? In which case, she's that much stronger. This chain reaction can go on and on.

Don't even get me started on just how bad some of the map design is. Like the beginning/mid sections of the map feel somewhat fair but the end of many are simply ridiculous when you're doing escort mode. A lot of the time it feels like sheer blind luck breaking their line in the finale and scoring a win. Sometimes you end up having to cheese it hardcore with that five Winston stack. ;(

Back to MM, I'm not sure how you can really make it good in a game where people are constantly shifting heroes. What happens when someone is forced to play a hero they're not terribly good with? It's different for something like HotS(which took them a while to iron out anyway) where you can more reliably pick beforehand and then be rated by that hero and your overall stats. This isn't going to be that simple and I think you're going to end up with a lot of broken, one-sided matches anyway.

I feel like a jackass for not considering all this beforehand but I was just so enthralled by the presentation of the game that I completely glossed over the negatives early on.

[Artelia]

if i win too much i got bad team and usually end in lose and get 3x gold and 2x silver medal . Sorry but this is not balance with better opponents. You get bad team-mates. And this is horrible.

[Very Tired]

The game is barely 2 weeks old, the MMR probably needs a much larger dataset than the 20-30 hours people have put in at this point to get a stable average that isn't kicked out significantly by a relatively short winning streak. 6 wins is going to have a much bigger impact on your first 50 games win% than on your 1000th.

I can only imagine how difficult it would be to rank players seeing as everyone wins as much as they lose.

[Deleted]

Well, you've got damage done, healing done, weapon accuracy, kills, deaths, objective time, objective kills, and several other statistics that you can measure a player with to build a much more reliable MMR, an MMR that's based on the character being used, not the entire account.

I'm guessing their bright minds just went with "Wins give X and losses take Y and yay I figured MMR out!"

Anyone who thinks a proper MMR should simply be based on wins and losses isn't really thinking shit through on an intelligent level.

[Maxilian]

I mean, pairing players with the purpose of achieving a ~50% winrate for every person involved is retarded. In every game. The moment you queue to play and thinking that the following match itself is already rigged against you or in favor isn't joyful or an interesting concept at all. Another brilliant MM system developed by Blizzard Entertainment.

This post by Riou in another forum expresses what I feel:


At least in TF2 I click Quickmatch, I choose a server, I see how many players there are, I choose a map and off I go. It's a pub after all, so no MM to dictate rules; you will have awesome players, good players and gibus players. And if you're being stomped, you can always leave the game with no penalty (crazy right?). Not sure why a system like this isn't in OW.

Well that means that you will improve little by little and you will start winning those fights that have highter lvl players

[Deleted]

Oh no, not this "performance based MMR" pipe dream again. I would think that this would have died down a long time ago.

Are you doing things to help your team win? If so, you'll win and your MMR will go up. Are you not? Then the opposite happens. Just doing the most damage or being more accurate doesn't mean you're actually helping to win the game.

Yeah... There's also % of damage done when a kill actually happens, objective kills, objective time, headshots, and a million different things that the game pays attention to.

Just because you live inside a box where you can't see how an incredibly complex MMR system could and would work, doesn't mean such a thing doesn't exist.

The fact is, win% doesn't work. It's pointless. It's nothing. So basically, there is no MMR.

When looking at a complex system that you can't even imagine in your head because it would have to be worked out properly, it's much easier to just give up and get all jaded and say "that shit will never work." That's why people fail tests and drop out of school and don't excel in complex shit. They're not built to understand complex things and how those things can be achieved and accomplished.

But there are those kinds of people out there, people who do understand those things and people who can achieve and accomplish those things.

I'm not sure if Blizzard has them working there. I doubt it. I've seen nothing to suggest that they do.

[Skyfall]

Probably a little too simplistic, but I would hope/imagine medals have something to do with your MMR. I.e. if you lose with top objective kills and top objective time on your team, maybe your MMR doesn't take a big hit and maybe if you win with no medals at all it doesn't go up much at all. Obv it would need to be more complicated than that, but there are somewhat measures already in the game and visible to you that indicate if you actually helped your team win or not rather than just going on pure W/L ratio, so i would imagine those are included in it.

[Tziva]

Yeah, it's kind of weird. It may not be what it's trying to do but outside of a full premade group, it seems to constantly oscillate between matching against teams we steamroll, and teams that steamroll us. Yeah, that ends up being a 50% winrate also, but I'd rather be matched against similar teams and have a bunch of close games rather than cycling between stomp or be stomped. (I wonder how the inclusion of ranked will change this. Perhaps it will take a lot of the better players out of the pool and lessen the dramatic differences between teams? who knows.)

As Rivelle mentioned, on partial groups it seems to also move the ungrouped player between the partial groups based on the win/loss history. It could all be coincidental, and it is just my own anecdotal experience, but it does seem to operate that way.

(If we're a full team, it usually just keeps against the same other full-team until one side of us requeues or someone leaves and forces it to re-match, regardless of how we do. Although usually if one team is getting massacred, they requeue intentionally to get a new match)

[Konteil]

how does it stack up as far as players using a computer vs players that are playing from a console? is there a huge difference in how it plays? curious how the WOW pvp type contols vs the HALO paddle types stack

[Deleted]

Winning is the ONLY thing. If you're not doing things to ensure victory, then you're doing the wrong things. It doesn't matter if you're doing the most damage, or have the most headshots.

Yeah, except, you're not thinking outside the box, like I said. If there's this one guy, who manages to get a bunch of wins thanks to the five other guys, who is so absolutely crap that he just can't do shit. But he manages to get wins somehow, because he gets put in games with other winners, who are really good players.

Then, one day, five other completely shit players get put on his team, and there's now six shit players against six great players. All with the same "MMR", because the MMR system isn't an MMR system, but something so absolutely asinine and simplistically stupid that it's unbelievable, and now they get steamrolled.

Next game, he finds himself with three of those shit players and three good ones, but the other team has six good ones. Again, a blow out.

And this cycle keeps happening. People keep getting wins that they don't deserve. They keep rising in MMR, and then there keep being those fucking annoying blow out games thanks to this completely retarded and moronic so-called "MMR" system that isn't an MMR system.

Now, see, if it was 1v1. Always 1v1. Never anything else than 1v1. All games were 1v1. With the same character for both players, in a completely neutral map. Then your win% -based "MMR" would work.

But, no. People keep getting wins they don't deserve, then these 6 bad, 5 bad 1 good, 4 bad 2 good, 3 bad 3 good games keep happening. And, the way I've experienced it, getting actually even teams is much, much rarer than completely and absolutely lopsided teams.

No. Win% doesn't work. There's nothing you can possibly say that will make any sense that will convince anyone that win% works.

[Skyfall]

Yeah, this does exist in some systems I believe and I don't mind the idea TOO much. The idea that you perhaps take a slightly smaller MMR hit if you lose but performed well.

You still lose, and you still take a hit, but perhaps a slightly smaller one.

I think there is definitely value to analyzing individual performance in a loss/win relative to how much you contirbuted to your team to some extent, especially in hidden MMR used for matchmaking. because in reality, if someone is really good, but has a streak of bad teammates, it hurts the experience for everyone if he keeps getting put in worse and worse games (and likewise with a bad player who gets a lucky streak, and suddenly is in way over his head). In a broader context, people have been trying to quanitfy individual contributions to team activities for a while, think of PER in basketball or WAR in baseball. There is value in those metrics as long as you trust blizzard to calculate the equivalent correctly. Im not saying if you run around without caring about the objective and rack up a ton of kills in a loss you should go up, but if they have a way to see your team gave you no help at all you got 5 gold medals and lost, you probably shouldn't lose as much MMR as someone who, say , played reinhart and got 0 objective time or something.

[Deleted]

Except none of that happens in the long term.

Of course it does. The games are 6v6. Of course the bad players will regularly be carried by 5 better ones. So of course it will keep happening, ad infinitum.

[Riistov]

Also, if the MMR is working, you'll be at around a 50% win rate. Most people complaining seem to think they're pros and should have higher, but that's not how a properly-working MMR works.

This is the perfect response to the majority of the MMR complaints I see about the game.

People seem to think they should be matched against inferior competition most of the time.

[classic username]

Except that they kinda do. They represent experience, and while there are always the outliers like the low level person who plays a lot of FPS games or is coming from TF or something and does awesome, (and the inverse, the high level person who has played a ton... but just -sucks-) if you put an average Joe gamer who hasn't played at all against average Joe gamer who has played 100 matches its pretty obvious who will win.

thats why you just play and get better. what does it matter when someone has more experience. if getting stomped in the beginning makes someone quit.. the Game isnt for them anyway

[Deleted]

People seem to think they should be matched against inferior competition most of the time.

Wait, what? No... The exact opposite of that. The MMR doesn't work, because that 50% win rate that you're "supposed to have" comes from getting blow out wins and blow out losses most of the time, whereas actually even games, which are the actually fun ones, are rare as hell.

Even games is what we want. Not inferior or superior teams.

[Deleted]

Getting to a stable 50% win rate with blow outs is like being thrown in the ring against the world champion heavyweight MMA fighter one day, and a toddler the next, and then continuing this forever.

Who the hell wants that?

So all the ranking systems in all the other team games that exist now don't work?

Ok.

If they work exactly the same way it seems to be "working" in Overwatch, then yeah, none of them do.

[Xar226]

I don't mind the 50% at all. It's the lack of closeness in the fights, whether being rolled or rolling the opponent. You don't learn anything from those fights, either because there was no challenge or you spent too long dead to really see what's going on.

I'm not sure that's necessarily a matchmaking issue though. I'll give you an example from Sunday night with a couple of friends.

We went up against this team on offense, and we got absolutely, embarrassingly crushed. It was one of those soul-consuming asswhoopings that make you just want to close the game and never play it again. We (barely) took A early on and then never sniffed B. One of their players ended up with a 35 kill streak and 43 eliminations.

Then the map changed. We were on defense, they were on offense. To the best of my recollection the teams didn't change; I definitely know the top two guys on their team were still there and there were four of us grouped on my team, and I know I was worried about another whooping. Anyway, they didn't take the first objective. They really only threatened it once.

Our skill levels didn't change. We didn't suddenly become gods and they didn't suddenly become scrubs. It just happens that, for whatever reason, we both whomped each other on defense. Maybe they're good defensive maps. Maybe both sides are better at their defensive heroes.

Now, you're right that I didn't learn a lot from that whomping. I don't know if it was a map issue, a hero issue, a skill issue, or what combination of all of those things. But it's not the matchmaking system's fault. It certainly wasn't matchmaking that made me go from like a 0.3 KD to one near three.