I’ve seen people who basically ruined their team getting cards at the end, for example dps-only Moira, so I really REALLY hope the Blizzard internal rating algorithm that decides your matchups works VERY differently from that one?
Also, how does the MMR even check things like that Mei wall that decided the team fight by separating 2 tanks off the rest of the team, or that perfect sleep dart, and stuff like that? I read that MMR basically just checks how often you died, lol. Which means you could be forced to play really bad just to keep a high MMR?
The problem I have here is that I know it’s Blizzard who coded this, so I have super high doubts that they were able to get it right -_-
This is one thing that seriously scares me as well because MMR is supposed to work in your favor if you’re too good for a certain rank, but I have performed many games very well and not got on top stats and I have won many games where the enemy gets all the top stats and we still win. I know that play of the game is determines value by basic stats, basically whoever gets high picks or the highest damage output will most likely get the play of the game unless you kill somebody while they’re ulting then that has some kind of high point value to it, however if MMR goes by the same stats the play of the game does and the stat cards I think I’m largely screwed on that one, I like almost never see adjusted SR no matter how good I do it seems.
Yeah but that was the whole arguments that if you were to just get bad teammates MMR would be there for you So that your teammates would not drag you down. You’re going to have a large loss number if your teammates are dragging you down and if they’re tracking you down that much MMR will never come into play to supposedly help you in the one thing it was meant to help you in. This is another reason why I don’t understand why Blizzard would only calculate MMR on the wins, as I admitted this in an interview somewhere.
Blizzard stated in an interview that your MMR is calculated using only the statistics from games that you have won. Of course your MMR does apply every game in terms of SR points, but I was talking more in terms of how that MMR is calculated to begin with. Help with undeserved losses and get you to climb faster and out of underserved ranks is what it’s supposed to do, but if it was ment to help with the losses caused by your team in the first place then I think it was a dumb move not to be able to calculate the stats on those losses to begin with. I didn’t save this interview page I’d have to look it up again if you wanted the source but The memory was very clear to me when they said this. It was during an interview when they were talking about how the matchmaker balances for stacks as well
I’m fairly certain that MMR is adjusted after EVERY game. If you have a source to the contrary then I would really like to see it. My assumption was made using Jeff Kaplan’s statement that it “only took 10 games for the players MMR to match their main accounts” when he watched a streamer do a Bronze to GM challenge and was watching their internal data adjust in realtime.
I’m fairly certain that MMR is adjusted only on WINNING games. If you have a source to the contrary then I would really like to see it.
I’m slightly kidding just wanted to point out that we both don’t have proof on hand for what we say, though I guess I’ll look it up if I have the time I don’t know lol
If you wall of someone off and freeze him, you should have kill which count toward stuff. We dont know details, maybe even number of damage your wall is considered in stats. Or if rein charged in it by killing someone. Certain POTGs shows that calculation for your performance must be pretty complex.
I hate to say it but that specific paragraph isnt talking about how your mmr rating calculated, but rather how it is applied after the fact as a “multiplier” to the SR rating in which case would apply to the wins and losses
That said you’ve done more to prove your point then I have so I’ma hit up google see if I can find anything.
Interestingly enough what you just quoted it incidentally proves that the matchmaker is in fact balancing matches against you using your mmr and not just your SR, which poses a problem since trying your hardest to win a match would also be counter productive as it will put you with worse players.
While the statement doesn’t come out directly and state how the MMR is calculated after a game the fact that “The faster we can update your matchmaking rating (MMR) to match your ‘real’ skill, the fewer unfair matches get created with a player with inaccurate skill.” and “upon these consecutive wins or losses.” should be proof enough the MMR is being manipulated on factors outside of just winning a game. For MMR to be adjusted on consecutive wins or losses means that it has to be adjusted on a loss as well as a win.
The matchmaker has always used MMR to match people and has never used SR. SR is just a “visual representation” of MMR and can be fairly inaccurate as a judge of skill.
Here is the post from the old forum, archived through reddit, where Jeff Kaplan addresses matchmaking.
Well I was unable to find the article which essentially leaves me with no source to the “only calculated on wins” thing so I’ll just take the loss there.
Such a shame too really, due to the fickleness of words and use of “depending” on the last quote, It could have technically been otherwise
Naw your probably right, I could have equally read it wrong back when I had read it initially🤷
Matchmaker is balancing team MMR so both teams have same average MMR. You can be easily highest or lowest MMR players in two games in row with same performance.
Why should matchmaker give you worse team mates when you play well? Makes no sense. Even if matchaker will give you weaker MMR team mates, the enemy team will be equal to your team in most cases unless you play in very low populated elo or are grouping up with players from different ranks.
It’s a problem because let’s say you decided to slack off and barely do enough so that you can win. The MMR system will assume that that’s your level of competency and give you teammates accordingly, and you will be judged as still winning but having a low competency you’ll still gain SR and get more competent team members.
Whereas if you were to try your best every match and be the star player every game then the MMR will assume that that’s your level of competency and again you get the same SR just like the previous example but in this case you’re going to get lower competency team members because of your high mmr.
This would impose a problem if they just used SR alone or just use MMR alone but because the pairing is based both and it cannot Accurately know when your slacking off or trying really hard. You will unintentionally get punished for trying harder.
This is because two people can technically be for examples 2000SR in rank but one of these can be a lot better than that and currently climbing and the other can be a lot worse and currently declining So beat since being paired with a person cannot give accurate results even if they are the same rank This is supposed to be what MMR does but you can get away by doing the bare minimum and now you have a window for when you actually do have to try hard if it works came to worse whereas the person who’s always been trying hard cannot try harder than his maximum.
This whole thing is just a conspiracy these forums spun. Nothing supports it. MMR is not a handicapping device contrary to baseless assertions otherwise; it’s a matchmaker. It wants even matches is all. If you do better, you ascend until you’re properly challenged again. That’s all. Winning does not give you worse teammates but better ones because your skill went up. If the theory folks posited was true, they’d never whine about their massive loss streaks. Playing your best always is the best way to climb.
Matchmaker doesnt create games for you. Matchmaker is creating team vs team situation, he is using only your MMR to accomplished same average team MMRs. There is no such thing as giving your team mates based on how you perform. If your MMR is higher, its higher and thats it, there is no advantage in it, matchmaker will just generaky speaking use and place you in games preferably closest to your MMR.
People for some weird reason think that they are special so matchmaker does care about them and create games just for them. It does not. Matchmaker care only about team vs team numbers.
Thats why players still were in high ranked games when they decayer in past. Because they had high MMR but low SR. They still played with and against people from their elo.
You’ve misread what I’ve written I didn’t explicitly say winning gives you worse teamates, It’s because if MMR is measuring your performance and will use that rating to determine what an “equal” match looks like, if you pretended to be mediocre or generally do not try then matcthmaker will cast you being mediocre even if your potential is a lot higher. However another person of equal skill as you tried his best he will have a higher MMR rating and the the matchmaker will match him accordingly. So in this case the matcthmaker does it’s job, It’s not so much that it’s out to get you It’s that it matches you by MMR primarily and SR is just a way to track your progress/climb. MMR is The one rating we can’t see and probably more representative of your skill level.
Really though so long as your below masters (maybe below diamond if u dps) I think that if you are able to, solo q’ing 1-3 ranks below your current rank will demonstrate it better than I could, and the reality of it is more the point anyway.
MMR is a factor we know little about but to our knowledge a 2400 SR= a 2400 SR even if how that is calculated is different. Any speculation on there be like a 2400+ v a 2400- is entirely unsupported fabrication. IF it is the case, we have zero evidence of it. It’s one of the many, many myths spouted as truth on these forums.
Good thread. Players should be asking more questions like these. I don’t think that MMR should exist at all, because it is handicapping competitive play:
