Matchmaker IS rigged..but not how you think

Competitive Discussion
41

SR and mmr be pretty much a whimsical fantasy by a broken calculation:

Season 16 end: 2747

Season 17 placement: 2851

Games played: 8
Games lost: 2
Games won: 6

Left every game within character select or if toxic in first minute. This totalled 6 games. Only 2 counted to games played, 2 seem to be invisible counts, and other two just gone in the wind I suppose.

-75xp (level 1660 so I guess yay?) and ban timer at 2 hours at end of final placement.

Moral - Don’t like the team comp in placement matches? Leave.

Such a good, clever, well thought out and infallible mmr system. Turns out leaving gets you more SR in the end than playing. My alt at the same level went 5/5 and drop about 100SR.

42

This creates lots of frustration for me. In almost every game i have players who dont even understand the job/basics of the hero they play. Some i can carry. But winning to many of those (carried to much) I just got even worse team mates who just are suiciding all the time, running 1 by 1 into full team. So completely unwinnable even if i get a card like having 60+% of team damage.

Remove that enforced 50% win chance… Let people climb or drop to a rank they belong to without having to farm lots of games first. The matchmaker should even group up people who are playing above the rank they are in(high MMR) and let them play against a random team. So they almost cant loose. Making them climb to the rank they belong to faster.

So create group by MMR(each player has about the same MMR) .
And use SR, or group SR to find opponents. (both teams have about the same group SR)

But this is soloQ with not a team ranking but a personal ladder. I rather have them use he same kind of system they use to rank individual players in sports. So the ones that rank people for each position to a certain rank, that scouts use etc…

Of course this would mean your rating will almost not change, so you will not climb that easily. (except when you actually get better). But the matchmaker could much better do his job by placing people of the same skill level together and against each other.

And i hope most players rather have good games then a rating that grows by playing a lot.

1 Like
43

I prefer the MMO matchmaker. True skill would be too variant and result in unbalanced matches.

Balancing high performing players is necessary as that reduces the chance of a stomp.

I wouldn’t say they’re expected to vastly outperform. They do have to play better though. From what I’ve seen, the matchmaker will estimate your “true” rank and check your progress on getting there. If you take too long it will recalibrate and estimate you to be lower. It adjusts SR gains so that even with a 50% winrate you will climb.

Players used to climb much faster back in the day but the loss streaks ruined people’s experiences of the game. As a result, they removed the SR bonus/debuff. It only takes months to climb if you need to take that long to improve.

The NFL example is a bit different. It’s more like comparing OWL. The coaches and managers (I’m not really sure because I don’t watch it) choose the players I think. This is based on their perceived skill level. There’s no other way of doing this than through the matchmaker. In season 3 a hard MMR reset was done and I was playing with everyone from bronze to GM and everything in between. It was an absolute mess. This is the reason I’m in favour of MMR balancing.

As for your questions:

  1. The fairest possible (probably) way to balance
  2. MMO
  3. Less variance and more balanced matches.
1 Like
44

Yes, it’s perfect. Please leave 3 more games.

45

Done. Removed. It doesn’t exist.

PBSR Does this already.

SR is likely already used to find a match. They try to get 12 people with the same SR.

Check out the referenced posts from the section “But I just read this post from a developer, and it said matchmaking is based on SR, not MMR”.

46

I give you a glaring issue with placements. Your response is to tell me to get banned.

2 problems: ban timer didn’t kick in until placed, and it resets so fast that if I leave now it’ll be back at 30 minutes at best.

If you did the above on a new account, you could possibly glitch to 4000SR. Or 500SR if you throw the games you play.

47

Matchmaking isn’t rigged but i do wish win and loss streaks came back as they where before

48

You weren’t very clear on your initial post. I couldn’t keep track of how many wins / losses / leaves you had.

Your SR gain looks completely normal for 6 wins and 2 losses. That’s a net 4 wins, or an average of 25.75 SR per win/loss. That’s completely normal.

The only thing odd is that you’re a leaver. I have no patience for leavers, they’re toxic and a blight upon the entire community. Yes, I want you banned because you said you intentionally left games.

So did you just suddenly decide after a life of no toxicity whatsoever, and never leaving a game to experiment with leaving as a thing during placements, or is this normal behavior for you?

If you have a habit of leaving if the game doesn’t look right to you, then I have zero qualms about asking for you to get banned for the rest of the season.

Now, since I’ve already admitted that I don’t know exactly how many times you left, nor how those correlate to your record, if you were to clear it up with a chart like:

#1 - Played - Won - 1-0
#2 - Left - Disappeared - 1-0
#3 - Left - Lost - 1-1

etc, then maybe it would be clear what you’re communicating.

49

Clearly. Let’s just go with pure mathematics then.

2 + 6 = 10?

50

Is it 8 + 6 = 8?
Is it 6 + 6 = 8?
How about
6+6 = 10?
8+6 = 10?

If you’re admitting that you left 6 times, that’s the amount for a ban.

So now that you’ve demonstrated you’re a leaver AND a troll, I think I’ll just ignore whatever else you inflict upon this world.

51

Troll? I’m not the one being obtuse here.

I had left my account open for 24 hours after posting, so people could confirm I was simply posting my “stats” from my overwatch account.

I “played” 12 placement games:
6 won
2 lost (1 left in 3rd round of 2CP as 5 dps and best outcome was a draw anyway)
4 left (before game began so no one else suffered anything)

6 + 2 + 4 = 12

So either I did 8 placements and the system placed me before I did my full 10, or it placed me 100SR higher with a 6 wins to 4 losses, or it placed me 100SR higher with 6 wins 6 losses.

Only because everything is hidden within mmr, is this possible.

Edit: cool, looks like he just decided to ignore something that destroys the hallowed Kaawumba Bible, simply because he couldn’t understand it. I was expecting the “pics or it didn’t happen” line next anyway.

52

Well they themselves get all score data from scrims and the games obviously, as you can see some times blizzard pulls up the current stats in real time, which means they can access “profile” data in real time. The underlying system is there, and it is accessable by all OWL teams and their coaches ofc.

It is funny coz blizzard understans this, that any serious professionals needs data on their performance and provides it, they want their game to be taken “seriously” yet they do not wanna for fill the basic demands that comes with it unless you are an OWL-pro. While yes you can look at your profile but using and doing the math to get an game to game basis performance sheet is a drag, especially after a couple of games. It shows an average which is decent to judge your overall performance but you’ll never see how good you really did that game you felt awesome and hard carried.

1 Like
53

It does so within 10 matches a GM skilled player on a bronze account can be in the rank he belongs? So he only destroys max 10 games?

So all the rumors with prove are false? People with high MMR arent placed together with low MMR, to keep a more 50% win chance?
If this is false, then the MMR/SR system is even more faultier then i already thought. Since how else can you explain the big difference in skill levels in low tiers between players. I have seen hitscans that almost never miss. But also players who not even understand the basics of the hero they are playing in the same game.

3 Likes
54

Your MMR is translated to your SR and the only time there’s deviation is inactivity decay. And no, high MMR players aren’t placed with low MMR players UNLESS they volunteer to play with each other (within enforced limits of the highest and lowest members of the group). If you have a game where all 12 players are solo queue and you see that everyone (including yourself) is SR 2100 plus/minus 50ish, then all 12 players have basically the same MMR.

As for perceived “skill” levels between players of the same MMR, everyone gets to where they’re at in different ways:

  • an SR 2100 player can be mechanically decent but have terrible game sense
  • another SR 2100 player can also be mechanically decent but tilt like 65% of the time and thus perform poorly when that happens
  • another SR 2100 player may not be so mechanically gifted but makes great callouts and possess better game-sense than others of the same rank
  • another SR 2100 player may have been carried by a duo partner from SR 1700
  • another SR 2100 player may actually be a legit SR 2600 except the player tanked 500 SR after trying to play while very drunk

The examples go on and on and on. And that’s not even including illegitimate ways of getting to a different rank, such as intentionally trying to lose games to get to a lower SR or getting the account boosted by a different player.

And while we’re still on the topic of perception, there are definitely a lot of players out there that are perceiving that they have bad teammates in certain games, when it turns out not even to be the case. Take these two players for example:

Both of these players claimed to get bad teammates after winning too much and even tried to prove it on video. Problem is, it became very clear that both players were very quick tilters AND their teammates were actually good, but they were so blinded by their own tilt that they threw away very winnable games.

55

Yeah, its a trap. There is no way to leave safely as every leaver gets punished. I don’t know if its intentional but they have had three years to fix it and haven’t.

1 Like
56

I always leave as soon as i can. I never had a low SR after such game.
So it is a very random bug.
And i rather have that chance of loosing one game of + SR, then to waste hours in games i am sure i get - SR. For each game like that where i can have like a 10 minutes more play time, i can farm lots more SR then i would maybe loose by this bug.

57

This is true because if you create a new account it is way easier to climb because the Blizzard match making system dosen’t know how good you are yet. If you have a high level account it honestly pairs you with dummies and you often have to play beyond your best just to win.

Previously the match making system would give you more SR to compensate for playing better than average, but this was driving abnormal behaviour because everyone was more focused on individual stats than on winning…

2 Likes
58

I took a 2-3 week break from the game. Came back and the first day, it was a different game. The matchmaker didn’t seem sure where to put me. So, i got INSTANT queues, and just tossed in what seemed like “normal games” to me.

Won 6 in a row.

The next day, this “TRY SO HARD TO CREATE AN EVEN MATCH BASED ON YOUR LATEST GAMES” crap kicked in and i lost 4 in a row. Dumbest fricking games you could imagine. So sorry i did well the previous day. S**t. If i knew, i wouldn’t have even tried. Congrats on getting me to that 50% win rate though.

So, i’d much rather see a matchmaker that isn’t TRYING to do a single F’ing thing. Toss me in with some players who are in my rank. Done.

Instead, we have this… whatever the hell matchmaking that’s literally trying to do too much, you can just see it over a three year span and learn the subtle hints.

I play in the same room as my duo partner. Sometimes, we queue solo. I can tell you, the longer the queue, the harder the Matchmaker is trying to find idiots for you to play with. We queue, one person gets instant queue, the other gets like some ridiculously long 4 minute queue and that game is over before it begins. You’ll literally see a leaver in that game. You’ll literally see 5 dps. The instant queue game? Fine. Good game. Real game. Un-rigged game.

To me, this is the most damning evidence i’ve seen for the case of the rigged matchmaker. Instant queues seem to be just “normal games” and when you’re regularly getting longer queues… look out. Matchmaker is pissed at you for some reason. Something i’ve observed for a long long time. Testing it out with my duo partner in the same room, looking at the different games we got and how long the queues were for them.

If you ask me, they should ditch this entire rigged a*** matchmaker and simply stick Gold vs. Gold. Randomly. Don’t try to make even teams even. After a little time, the players will settle to where they belong. Right now, no one is able to settle into where they belong because the matchmaker is interfering and causing bad players to mix with good, experience players. Over and over and over in little streaks, you can see it. Horrendous.

1 Like
59

That’s even worse though, because matches will be completely unbalanced then. Sure, you lost games after winning a bunch, but you should be able to climb back up there if you’re a player of that SR.

That’s interesting. The devs mentioned that they look for a game as perfectly balanced as possible, until the queue times get too long. That would make sense as to why you’re placed in suboptimal games after long queue times. I don’t see how this leads to the conclusion that the matchmaker is rigged though.

This is completely untrue. Making it random would be far worse as there’s a skill variance even within gold. If it was truly random then there would be no checks to determine one’s true SR and then it would take incredibly long for one to reach their true SR (what would the definition of that even be in that case then?). It’s quite the opposite, over a large number of games people will be exactly where they belong. While games have more variance than I would prefer (imo a 3-1-2 flex-tank-support arrangement would help the game out greatly), it’s accurate in getting you to where you belong.

60

Well, add more ranks then. Have Goldish and Gooooooolllld. Randomly fighting players within 250 SR, to me, would feel better than having the matchmaker try so hard.