How I believe the Matchmaker works

Matchmaking rank but the SR tries to be as close as possible to the MMR

SR is just the way blizzard tries to explain your MMR in a way that you can understand it

1 Like

No way. My lack of aim and choice of aim intense heroes can’t be why I’m lower.

1 Like

Then riddle me this, Kristal. Why have a hidden MMR with calculations we are not privy to? Just for decay? Then why not have it only affect those of diamond or higher.

Complete transparency of the system would settle the debate, and might actually support the White Knights arguments. Makes me wonder why the defenders of comp aren’t clamoring for this with the rest of us.

1 Like

Literally, Blizzard has stated that they try to make matches 50/50, meaning that you have a 50% chance of winning any match you are in according to your MMR, which is different from your SR. They also will not elaborate on how MMR is calculated, so nobody knows how matchmaker works except for Blizzard. If anyone tells you otherwise, it’s just an opinion. My guess is that they find 12 players in an SR range then try to balance out the teams by MMR, which is why you will see team SR’s being pretty close.

If you are at 2000 SR, you will be put in a match where you have 50% chance to win. If you get are 2500 SR a few days later, you will be put in a match where you have 50% chance to win. If you are 3000 SR a few days later, you will be put in a match where you have a 50% chance to win. If you are 1500 next week, you will be put in a match where you have a 50% chance to win.

If 2 people are the same SR, but MMR believes 1 is higher than 2, it will give 2 the better teammate, but why does 1 have to carry because he’s better than 2? By this way of thinking, 1 is under ranked, but if he wants to be at the correct SR, he has to fight for it, while 2 is overranked, and might get lucky and get an even bigger boost. If SR were accurate, they would be of equal skill and nobody would ever have to carry. Personally, I think they should scrap the MMR system and just group people randomly by SR.

1 Like

Highly simplified, yes. More detail here: How Competitive Skill Rating Works (Season 13).

2 Likes

What blizzard means by 50/50 is not you win then you lose the next match. They mean 50/50 chance of winning the match. Do you want it to be 100/0? That wouldn’t be fair would it?

Decay is something that I think shouldn’t be at 3000 but instead 3500. Why should diamond players decay? It should be at masters and above. And I’m sure that there is hidden MMR in every competitive game. And if Blizzard is completely open with how the MMR and SR system works, then people will find a way to abuse the system to win more matches. What the MMR system looks like to me is an extremely bare bones system. You win, you gain SR. You lose, you lose SR.

1 Like

I can see three main reasons.
The first and most important is that it acts like a mask to hide the direct impact of certain in game actions. If players break the code of A = stats = better SR reward then that’s all they would do. Example:Mercy rez issue and players just farming rez over anything else.

Second is it allows for behind the scene tinkering with the core math of the system without needing to tell or show anyone.
Third is, as the dev have said, the MMR and SR being different acts like a buffer when it comes to win and lose streaks. Being the SR might be at a peak or valley but the MMR knows the sort of player you really are and looks to even up the numbers a bit more. SR moves quickly, MMR moves slowly. Forcing players to get better over all vs just get cheap carry me sort of wins.

the reason higher SR player don’t whine about SR/MMR and mostly focus on match making issue is players past 3k understand that golds play like golds, masters like masters. “Why am I stuck in gold” is because that player makes gold SR plays and errors.

2 Likes

So both not what you claimed in OP. If SR chases mmr then that means they are not similar btw.

If true then:

Can’t be true.

Everything is based on the assumption that:

Which is unknown. What if the opposite is true? If you make that assumption then the yoyo streaks are explained as the system constantly over estimates and underestimates players, leading to completely uneven (not 50/50) matches while thinking they are 50/50.

2 Likes

My statement is not based on the idea that MMR moves slowly, but more so that MMR and SR move at the same pace. SR is simplified MMR, and MMR is not just one number but many, explained by Kaawumba. If you wanna know the most accurate idea of what MMR is, look into Kaawumba’s study.

Which would mean mmr = sr which is proven false. Any jump of rank between seasons while going 5/5 means the two are not always the same.

Anyway, I used your quote really to point out that every person claiming “there is no handicapping” falls over each other with muddled arguments that contradict each other. But still rely on those arguments they contradict.

Care to elaborate? Placements are not just matches with hidden SR

Not in my case. I’m a <200SR player and I get matched vs. silver & gold all the time.

1 Like

In which case mmr and sr do not match, as if they move at the same pace then placements would be just hidden SR matches. Good to see you’ve come around.

How do you get this anywhere from my statement? Your MMR is more volatile during Placements, there are chances for bigger change based upon your performance.

Citation needed for volatility? You just agreed mmr can move more than SR, and that placements put SR to your mmr after completed. This isn’t special to placements, it’s just the SR movement limitation removed.

I said your MMR is more volatile, as in during placements. Your MMR will reflect this. Your SR moves along with your MMR, and the way you lost SR without MMR is by either leaving matches, or decaying.

Citation? If you’re going purely by observation then my model still fits better.

I could say the same, it’s a matter of opinion there not fact. You don’t have any facts to backup your own, so don’t try to call me out on them.

Except I never claimed my model as fact, only that it explains a lot. As it fits all the time, not some of the time, it is more likely.