First of all you’ve got it a slightly wrong. If there were only SR, there wouldn’t be anything there for players to “game.” All we could to improve our rank was win the game, instead of worrying about hidden metrics that we could try to game. The metrics ARE there to be gamed.
You’re right on that.
That is true that Blizzard wants us to win and so they hide the MMR system. However, how can people game the metrics if they are hidden?
I don’t know what has you so convinced that what Top 500 players do has the slightest bearing on MMR at lower ranks. There is no motivation for Blizzard to do that, and they’ve said before that performance calculation compares you against people in similar situations/rank; why would that sort of logic not apply to MMR as well?
1 Like
Your SR and MMR are separate so even if you get put in a game with throwers etc you will lose but then you will end up in a win streak to put you back to the SR range that MMR dictates you belong.
The MMR curve that is used to determine who belongs where has the highest players getting heavy damage boosts. So its basically comparing everyones stats against players with high damage boosts, and determining they belong lower because you cant have as high damage as a damage boosted player.
I disagree here. LFG should cut way down on trolling. People can only troll for one game before being kicked from the group, and there is also some basic social pressure when a person is invited into a group.
There will be some new social issues, however, as people get mad about not being let into groups for silly (or not silly) reasons or kicked for silly (or not silly) reasons.
Most gamers don’t have the talent or experience to lead a group, but someone will have to do it.
People will still blame others for their losses.
etc.
We’ll have to see how it works out, but I am quite certain that it will not be the sunshine and rainbows that some are expecting.
This thread doesn’t state answers. It asks questions. But yes, I am leaning towards declaring all posts that talk about MMR obsolete, unless they are explicitly referring to decay. Because that is what Scott has implicitly done.
Your theory violates quite a number of posts as well. If you’d like to state (or link to) your complete theory again, I can link to all the ways that it has been contradicted by the developers.
Sure, it’s your theory “How Competitive Skill Rating Works (Season 10)”. You may recall that about a year ago I did say I believed you got pretty much everything right but were making assumptions with your interpretations when it came to SR vs MMR.
My only difference is that I hold that SR is part of MMR.
If you take every blue quote you reference with that line of thought in mind, nothing in any of the blue posts regarding SR/MMR needs to be ignored or reworked due to contradiction.
If true, it does give rise to how there is a deeply flawed concept within the MMR that would explain the weirder streaks people experience, and that there is indeed a handicap system.
I would like to know how you interpreted that post to contradict BouncyKnight.
Nothing in that link showed how MMR is calculated.
If you are referring to the “closely” again, then let’s try a thought experiment:
Recently there is a post about a 10 game loss streak, where the player complained of having high stats per game on soldier. I have experienced similar loss streaks.
If you assume they are telling the truth, which personal experience supports, then on that 10 loss streak it is likely they lost very little to no MMR per game (as per said in your quote).
Using ESR as the Effective SR based on MMR, and SR as SR:
Start:
Game 1: SR: 2500 EFR: 2500
Game 2: SR: 2475 EFR: 2490
Game 3: SR: 2450 EFR: 2480
Game 4: SR: 2425 EFR: 2470
Game 5: SR: 2400 EFR: 2460
Game 6: SR: 2375 EFR: 2450
Game 7: SR: 2350 EFR: 2440
Game 8: SR: 2325 EFR: 2430
Game 9: SR: 2300 EFR: 2420
Game 10: SR: 2275 EFR: 2410
So after 10 games, this player now has a difference of over 100 between his SR and his ESR. So Blizzard must mean a greater difference is still “close”. For a player that has gone out of whack, I believe it can be even an entire tier worth, of around 500ESR difference. After that PBSR gains help them to “chase” the difference.
the post you linked to actually backs up what I posted, it says you can win a match, gain sr but not gain MMR.
Like I said, it calculates MMR based on how well you did vs everyone else, so if you did mediocre, you may win but still not increase MMR, that means you will eventually go on a lose streak to come back down to where MMR thinks you belong. the post you linked even says that SR chases MMR lol
This is where MMR is flawed. By having everyhone in the same leaderboards it is comparing your MMR to (for example) the players that have damage boost bonus so there is no way anyone single queing to pump out as much damage or ults as someone with a dedicated mercy damage boost. (as the post you linked to said your MMR will not change at times and its mainly due to that reason that it thinks you didn’t perform on par with everyone else who it thinks played better)
This skews up the stats for everyone as it now compares everyones stats to the top performers who are mostly all using mercy damage boost pocket. It prevents everyone from rising or climbing to where they belong as well.
Maybe changing Mercys damage boost or reworking it completely would help.
My other suggestion was, If on the winning team MMR should increase for the entire team no matter how well or bad everyone played. So if you have a mercy boosting a player, everyones MMR will increase for that game.
And if you are on the losing team, the MMR system stays in place how it currently is
I wonder if SR tracks MMR so closely that the difference doesn’t even register to Scott Mercer’s mind when communicating. In fact, it’s not clear why he would say MMR in this instance even if it was technically true. His audience isn’t us who have an intellectual curiosity regarding the difference. I think a good communicator would interchange the two in that post.
MMR obviously isn’t gone. You have decay, inactivity, and non-competitive modes, all of which would rely on a robust TrueSkill type MMR system.
However, three has never been any reason to think that outside of some rare instances that MMR and SR have any functional difference. SR is MMR, but prettier.
Note how he talks about gaining 2 SR on a 60% predicted win chance. That seems in line with what you would expect MMR to rise. Not 0 but really close. The “feel good” value may have been 15 SR with the 13 “extra” being made up on subsequent matches. This doesn’t seem to be the case. So MMR is SR, but prettier, but not by much.
This seems in line with what we’ve thought all along.
I can’t speak for Kaawumba, but I would think that the first line, “MMR works very similarly to SR”, would rule out Bouncy’s ideas that MMR is 100% stats based.
MMR moves on wins and losses, mostly. If you lose a bunch of games, whether you play well or not, your games get easier. That would not be the case if your MMR stayed the same.
1 Like
my personal theory is that MMR is not based on wins or losses, solely based on stats, particularly any of the ones that have “per 10 min”, as well as how many succesfull ults used in a game/round/match, and successful kit abilities used (for example the amount of sombra enemy hacks per round/match)
Steamroll losses where you have no chance to hold a point may lower MMR due to you are not able to do anything good to your stats , however the losses where the game goes on and is a very close call may increase your MMR if you did very good even if you lost. This all falls in line with “Sr chasing your mmr” which basically translate to your SR going up and down/around it as you win win win lose lose lose
So, when they directly state that performance stats are only a small factor where a win or loss is the primary factor, do you think they are lying or deliberately trying too confuse the matter?
I haven’t personally seen anything that says that performance is a small factor and win/loss is primary factor for MMR gains/losses. For SR gains or losses however yes it is, however SR doesn’t matter because if your MMR doesn’t change, your SR will always go up and down/around your MMR.
Either way they are too vague and purposely ignore direct questions about it. In this game and all the other blizzard games that use the hidden MMR formula.
As a minor factor, we also do evaluate how well you played the heroes you used in a match. The comparison is largely based on historical data of people playing a specific hero (not medals, not pure damage done), and we’ve done a lot of work to this system based on the community’s feedback. In fact, I’ve seen some people indicate that they don’t think we’re doing this anymore. We still are. While it’s a minor factor compared to wins/losses (The best way to increase your SR is still to play together and win as a team!), doing so does help us determine your skill more accurately and faster.
1 Like
I personally hate this from him… I wish it were more towards 45%. 15% is a much more malleable variable than 20% difference in odds against you
it is my understanding they are talking about SR gains. (if you notice parenthesis right after specifically mentioning SR gains)
again with the wording on all their posts they are very careful and vague and usually ignore direct questions related to how MMR works.
In fact that whole post you linked was supposed to answer a question about MMR, yet it does not say anything about all about how MMR works! It pretty much talks about SR only lol.
They confirmed what all of us said force 50%…but you never believed it
So the second option? You think they’re deliberately trying to confuse the matter?