I agree. 12 characters required.
I disagree with you and your “evidence” that MMR differentiates and handicaps people of the same SRs.
MMR is used to make matches “even” but it also checks the SR range you are in to make a match: "The first change is limiting the maximum Skill Rating (SR) difference between the highest SR player on a team and the lowest SR player on the same team. " disagrees with you.
Me, Personally, would rather have the system set me a MMR number, display it, and then match me with that then a “Psychologically Nice” number that bounces like a rabbit.
You should, statistically prove yourself. You should be able to see what stats work for you and what areas you need to improve on to rank up.
I, myself, happen to like the separated skill ranking that uses stats and heck I may not see an improvement in my rank after a match, but if I am not improving that’s the measure.
To me, wins and losses, especially if you are continuously given 50/50 matches, should not affect your skill ranking in a team game.
According to some of the way things are in chess, you may actually rank up in a losing set of matches, because you were deemed to lose 70% of the matches against a higher ELO player(s), but you managed to lose only 60% during an event.
So far, as my observation, ranking up means you need a net positive wins to losses. At least anecdotally.
Unfortunately, the developers have been pretty sloppy with their language here. See How Competitive Skill Rating Works (Season 11) → “But I just read this post from a developer, and it said matchmaking is based on SR, not MMR”, for my full discussion. If you’re keeping score, I wrote that after the post you are quoting. This thread is a necro.
Ah, ok then. twenty!
Except that they are not as closely linked as you allways think:
Evidence: One of my accounts is just under 2700 SR the other is 2200 SR Wich one represents my MMR?
The answer is none.
Placements is more evidence of MMR/SR gaps - I once placed diamond with one account while having another account in gold. The difference in SR on that Diamond one from previous season was over 400SR - that is, I’d been playing with diamond MMR in my plat account. Never got to play any diamond players during the previous season tho.
Except in the offseason and placements where over half my teammates and enemies were Diamond.
Now Kaawumba, i never want to see you say “MMR/SR is closely linked” again, because it isn’t true.
1 Like
Are you using MMR as a synonym for your true rank? Because it isn’t. Any inaccuracy in SR would be mirrored by an inaccuracy in MMR. I don’t see how comparing two accounts would tell you anything about MMR vs SR.
What was your ending and starting SR? What was your placements record? How many games did you play on that account in the previous season? Without more information it is difficult to interpret anything from this story.
Go wrestle Jeff.
“With the exception of top players who have decayed, MMR and SR are closely linked.” – Jeff Kaplan Overwatch Forums
Its garbage all the way around and people are sick of it.
2 Likes
The purpose of comp is not to place you accurately at your skill and create good games, it is to create engagement. They want the app to be running for the longest possible amount of time to create the highest possible odds of you spending money on a lootbox.
The comp system as it currently works creates engagement, and that is likely the metric they are following around like a false idol when they designed the highly misguided MMR system.
They are trying to make you sink the most possible time into the game, period. A matchmaking system that read the stats and quickly yanked people up and down based on performance and mechanical skill wouldn’t create that, but a system where you yank the lever in what amounts to a slot machine many times waiting for a good break to watch your SR number rise has people sitting like gamblers yanking the lever again and again waiting for the payoff.
The people that get to a very high rank do so because all they do is sit at the slot machine, and by the effect of that become extremely good at the game. Even then people that have gotten lucky enough to get a good break are always grasping at them ready to pull downward, and even then its still a slot machine game.
Does this technically make it so the high ranks have skilled players? Yes. Is it a good system for a competitive game? Not in the slightest.
you can quote your rear end off Kaawumba, it doesn’t change the fact that placements often give you different SR than your ending SR or your typical SR. It fixes the problem every season. Or atleast TRIES to fix it.
My ending or starting SR is irelelevant - you only need to know that you placed, in my case, 400SR over my SR in the previous season. The numbers themselves are not important - tho I have given them several times in the Handicap thread (wich you skip over every time).
The closed linkage of SR and MMR is a myth. Placements prove beyond any reasonable doubt that they are.
Note that I’m not saying they never are linked, but that the system isn’t as robust as YOU think to the extent that saying “SR and MMR are closely linked” is a lie.
A lie you should stop perpetuating.
Edit: Maybe the system is designed to try to achieve this linkage, but that doesn’t mean it succeeds.
1 Like
You should check out Cuthbert’s post about MMR and handicapping.
The dude takes a Scott Mercer quote completely out of context from a completely different topic and calls it evidence that there’s “handicapping”
The reason these numbers are important is that I have a model for what happens at season transitions: Season 9 Placements Analysis, High Variance Explained, and I’d like to see if your experience is compatible with this. If you don’t have them anymore, or don’t wish to give them to me, then I don’t really have anything else to say as there is no way to distinguish from the data between your theory and mine.
I no longer comment in Cuthbert’s thread (or other threads like it) unless I am addressed directly.
SR isn’t needed to rig games for profits, or predicted outcomes, all that is needed is MMR. People read the phrase “matchmaking is done by MMR alone” and think that this must mean it can’t be rigged, when it’s the opposite. It makes it easier to rig matches by using MMR that is invisible, otherwise you could look at the MMR of the other team and see how they are doing it.
The matchmaker using hidden MMR (which has individual performance baked into it) is at the heart of the maximizing profits/rigged matches/handicapping etc.
Think about the quote “MMR/SR is closely linked”, it means nothing. I could say “human males and females are closely linked” and it would sound true, I could also say “human males and females are very different” and it would sound true. The quotes that are spouted off on this forum board are about as meaningful as reading your horoscope - they are generalized quotes that have so much room for interpretation that they are useless.
The proof of how the matchmaker works is right in front of your eyes every time you play. The matchmaker is terrible and everyone knows it, only people who work for Blizz/have relations with Blizz using their personal forum account defend it on a daily basis.
It’s true. I would be lucky to be Silver if there were an actual skill-based score. I could run into walls and still win 50% of my games.
In my experience I found that MMR is very stat based.
Guys ALWAYS remember that Blizzard removed bonus streaks and sr loss/gain modifiers based on stats but they NEVER said they removed it from MMR.
Community was blaming Blizzard for streaks and stats modifiers and in my opionion they were always very smart pretending to give to our community what we ask for while nothing really changes.
Guys, atm Blizzard got NOTHING to judge the attitude of a player in cooperation, timings, or whatever else regarding ow as a team based game.
STATS are the individual final results and the only valuable thing they can actually judge as they are comparable in tiers and are a very good way to “understand” the carry potential of a player.
Every time i lose some match with bad stats i find more idiots and tilted players in my matches as if they try to put togheter ppl they think they deserve to go down and viceversa.
I realized I was really able to shorten loss streak durations by farming stats with my best plays instead of filling the classic 2-2-2
Im afraid to tell you that this system actually rewards personal skills but don’t hope they are gonna admit it.
Look at how oversumo works, it can accurately tells you if you are likely to go up or down BY ONLY WATCHING STATS, and in my case it never failed.
The mechanism is very simple, oversumo compare your sessions stats to other players in the same tiers and tells you:
stats under 25% of the rest of the tier = you are going down
stats above 75%, you are going up
stats between 25% and 75% you will fluctuate in the same tier depending on which side you are holding (under or above 50%)
Now it comes my exploit as I think that MMR can move in both direction regardless of the final esit of a match.
When you lose a match you are likely to have bad stats even if you are not that bad but your team simply get stomped.
So, focusing on farming in these cases is the key.
Pick the hero you think is better in that specific match for you to farm stats and if you succeed you will easily see how the MM will be tender in next matches.
Pratic example: you are filling by tanking or healing but enemies are wrecking you with pharmercy and tou know you are a very good hitscan? Screw this team and go for it, you are probably going to lose the game anyway and maybe the rest of your team will see it as throwing, ignore them, your job is raising your stats toward 50% or even more to make your MMR less punished.
It’s the same for other roles, if you feel the lack of healing is embarassing go even for the 3rd healer and farm good healer stats.
Sad truths, you can come here and tell me I’m stupid but Im not gonna stop with this attitude until I see it perfectly works.
Ofc this is not an exploit to raise your rank but it can definitely helps you in reducing loss streak duration and the volatility of your rank, and this means you will be more likely to stabilize at your cap and improve faster (instead of going through an insane and tilt-like loss streak).