Suppose you have 12 players all with an SR of exactly 2500. There are countless ways to arrange these players into a game on random teams. This is what should happen but instead the game ranks these players based on their MMR in an attempt to make the match as even as possible.
So for example it will never put the 6 highest MMR players against the 6 lowest MMR players, even though their SRs are the same as that would likely result in a stomp. But this should be possible to happen sometimes or else it is unfair to the high MMR players who are always forced to do more work than the potatoes on their team in order to gain the same reward and rank on the ladder.
Creating matches using only SR would result in more unbalanced games early on as the lower MMR players would fall and the higher MMR players would be able to climb easier. Over time the result would be a more accurate ladder where players are more deserving of their SR.
The difference is a system designed to maximize competitive integrity instead of a system designed to maximize player activity and profit. 3 things that would be different:
There would be no more handicapping or rigging of games
SR would become a better reflection of actual skill over time
MMR manipulation would no longer be possible since it no longer exists
This part, right here.
This part is eating at my mental health & I don’t think I’m the only one. It affects the health of the game & community. Ripple effects.
The matchmaker NEVER has 12 players with exactly 2500 SR available at the same time.
There’s no reason to believe that SR and MMR in the majority of cases are significantly different than each other. Only in the case of new players, thrower penalties, or decay.
The current system NEVER puts the 6 highest MMR players on the same team out of a group of 12. The developers have said explicitly that they attempt to balance the teams out so that their modeled prediction of win chance is as close to 50/50 as possible.
They could do the same thing with SR, and for people with a sufficient number of games, they ARE effectively doing this.
Please define “handicapping” and “rigging” and why they would no longer happen if the Matchmaker used SR instead of MMR.
The exact same thing would happen without MMR as it does now. They would try to balance the SR between the teams to make it as even as possible.
If there was no MMR, then one of two things would need to happen:
They would have to remove the restrictions on SR movement
It would take much much longer for people to get into the games where they belong when they are on a new account.
The developers have said that they do things to MMR to make it move in a way to make it more statistically accurate faster than they can make SR accurate. When a player has played a sufficient number of games, say 50 in a season, there’s no reason to believe MMR and SR are significantly different.
SR is already accurate after a sufficient number of games.
MMR manipulation (edit: my interpretation of manipulation is the concept of being able to change MMR values dramatically based on results of matches, which is a benefit of it being hidden) benefits the player base by changing skill faster than SR changes so that people get closer to the right matches as fast as possible.
You say it’s unfair for a high MMR player to do all the hard work… but due to how probability works you’ll only ever have to do that as much as any other player.
The bigger issue is your solution if you have random lobbies you won’t as much…
Yep because a bigger % of matches will end up so one-sided skill won’t matter. You want to play the game for 20mins plus where you have no impact cause you just got lucky or unlucky? Or in other words, you’ll have to carry less because matches will be less competitive and about skill and more about luck.
As it is right now if MMR does work correctly your skill is tested more often and your skill means more.
Now let’s get to your points…
Yea there will still be handicapping any game when you get lucky or unlucky with teammates you will be handicapped without MMR.
MMR doesn’t rig game simple as,(not proven at least) but players do.
Say I make a 4 stack with 3 smurfs that threw 2000SR to 1k and their mate 500 SR andy on his main.
In the current system, it’ll correct the SR quicker and mean they’ll rig less game because of MMR. Without MMR, they now ruin many more games.
Also boosting will be more reliable and profitable because of it.
SR is a good reflection of skill anyway.
Your argument is MMR alters your SR or slows it down but without MMR random lobbies mean you can lose SR just by being unlucky or lucky. It’s a hard sell that the removal of MMR wouldn’t just make SR more volitile.
People do manipulate MMR yes. That said MMR is there to stop people manipulating SR which is easier without MMR than manipulating MMR right now.
Lastly,
There is 0 evidence MMR is used for that outside of QP but that’s because a competive game is more fun than a random game.
The intent of MMR as explain by the devs is to maximize competitive integrity by fixing the problems in a pure SR system. Maybe you want to start by disproving that.
Probably not but they shouldn’t be different at all if SR literally stands for skill rating and is the only representation of skill that the player can see.
Handicapping and rigging is when the matchmaker uses hidden MMR in order to balance a match. If it used average SR instead, then SR would be a more accurate measurement of skill than it is currently.
That is the point. We can see our SR but we can not see MMR. So the system should use SR to create matches instead.
Both of these are incorrect.
It is fairly accurate, but the system could still be improved greatly. The biggest flaw with handicapping isn’t that is stops people from climbing. It is that it causes insane win/loss streaks and therefore expands your possible range of SR. I would assume this is deliberate to make players feel as if they are not stuck in a rank but it is artificial manipulation.
No one knows exactly how MMR works since Blizzard refuses to give details so this is all speculation but a large portion of the player base have experienced the same thing. I personally believe you can manipulate MMR by playing poorly in order to get put on a better team in future games. I can’t say whether or not I have actually tried this since throwing is a bannable offense. Also SR can change drastically just like MMR. On a new account you can either gain or lose 100 SR on your first game after placements.
Yes it works both ways. But regardless there will always be someone expected to carry the team more than others. And if that person is not the highest SR player on the team it is unfair to them.
Most games are already stomps anyway. At least my solution will create less stomps over time due to SR becoming more accurate than it is currently.
Your skill should only be tested more as your SR goes up.
Both do in my opinion
The first few games would be even more rigged because of the smurfs who threw 2000 SR also tanked their MMR along the way. So it probably treats them as though they dropped 2500 SR instead and gives better teammates or a easier opponents. I see what you are saying but SR could still go up a lot for these wins without MMR. The game can still use your stats to adjust SR. I am not saying they should take that out of the game.
I think removal of MMR would make SR less volatile and that is the point. Instead of having a SR range of 1000 it should be 250 or 500. You wouldn’t get as many win/loss streaks and your SR wouldn’t move as much. You wouldn’t be able to peak as high as you do with MMR but you also wouldn’t be able to drop as low.
Anyone can manipulate SR by deranking or getting a boost. There is unfortunately no solution to this.
I think myself and others have provided a lot of evidence. I agree we want the game to be more competitive and in order to do that you need to have as many players as possible ranked correctly according to their skill.
I fully agree with this statement. Assuming of course that by the use of the word “fixing” you mean rigging. And by the word maximize you mean minimize.
I mean, the SR of all 11 other people in the game is allowed to be hidden from you based on their own settings. The only thing you are likely to see is SR Tier (Gold, Plat, etc).
My statements are based upon the statement of developers for WHY they hide MMR. Please read the guide for all the provided evidence of those statements. Summarized here:
I’ve seen Samito on game 6 go from 3140 to 3249, so this is true. His average SR gained in 5 placements was 158. This is information I have reviewed since my previous post.
This does suggest that my argument about brand new accounts is incorrect, as there are already mechanisms that allow SR to move rapidly.
This means that we’re just left with Leaver Penalties, SR Decay, and concern about performance based adjustments being abused for reasons that MMR is separate and hidden from SR.
A system being able to be improved does not mean that replacing MMR with SR would be one of those improvements. If there was exactly the same Matchmaker, but that Matchmaker used SR instead of MMR, only in the edge cases would match quality be impacted at all.
The matchmaker can do exactly what the developers say it does and there can still be streaks. Random chance is a sufficient explanation of that oscillation, and changing from MMR to SR would not impact that.
You genuinely feel that random chance is an insufficient explanation for fluctuations in SR?
Do you have a spreadsheet to share where you’ve recorded information showing that most of your games are stomps? If you don’t even have that, then why would we argue about it?
What exactly is your solution? You seem to be banking quite a bit on the problem that MMR is hidden.
Basically, the “hardness” of the match is mostly determined by the randomness of the fact that 6 people are on a team that may or may not actually synergize with each other. It will also fluctuate an increasing amount based on how inaccurate your SR / MMR compared to your real skill. SR and MMR are just estimates of your skill, and when expressed accurately they would be expressed as a range to demonstrate a margin of error. If your SR / MMR is lower than your real skill, then matches should be on average a little easier. If your SR / MMR is higher than your real skill, then on average the matches will be harder.
It’s important to realize that the Matchmaker doesn’t need to succeed in every single match to be statistically accurate. Statistical accuracy comes from hundreds, even thousands of games of data averaged out over time.
You can’t look at individual games and complain about the matchmaker, you have to take a step back and evaluate the matchmaker across all your games over the course of a season. You also have to evaluate your own progress over a longer period of time as well.
But why do you think that? I’m genuinely trying to understand. Why would the outcome be different if we used the other number in a pair of numbers that are intended to reflect the same knowledge except for the situations of Leaver Penalty and Decay?
I do not have an SR range of 1000. I do have an SR range of around 250. A range of 250 is basically a net difference of 10 games or so from high to low in the season. Are you trying to say you go for a streak of net 40 more losses than wins from your peak down to your season low, and then another streak where you have 40 more wins than losses to get back to that peak? (40 * 25 = 1000 SR where 25 is the average gain or loss per game rounded up from 24)
Please try to help us understand what needs fixing, the method you suggest using to fix it, and why that would help.
I mean from this it seems you feel MMR is rigging games. Which is not its intent or purpose and there really is no solid evidence to counter that. Yes you have presented your experience as evidence and all I can say is my experience doesn’t back that up and that’s as far as your evidence can take you. I get the benefit of the doubt since MMR has been a part of the game since release and changing that because you feel something isn’t really an argument it’s entitlement.
The rest of the stuff you write, I can put down to you having a closed mind. If MMR doesn’t rig. Removal of MMR would increase SR variance and make games more about luck than skill and therefore less competitive from a basic logic perspective. The fact you think MMR doesn’t help when people rig SR, because it would be wrong for 2-3 matches before it started to corrected over in a pure SR system that wouldn’t react for 100s of matches, shows how one-sided and therefore flawed your opinion is. You are saying because someone ruins only 50 matches with MMR on because it’s wrong for 3 matches, we should turn it off and let them ruin 100 matches. It really shows you don’t consider the positives and take the negatives out of context. One more example of that is how you think MMR can correct 500 SR it’s only been shown to do that via decay or start of season GMs.
I think it doesnt matter if you think 50/50 Matching is handicapping/rigging or not. I still think its a trash system and should be removed. Its absolutely crippling for mobility and match quality. At this point in the games life im less worried about moving up and more about fun comp games. Since skill levels at ranks cannot stabilize with this system, match quality is miserable.
As a side note, Other thank COD no other games use a matching system like this and for good reason. 50% chance for both teams winning is only fair from the outside looking in. If MMR dials you in, any increased skill level you have gets washed out in the matching when they throw someone on the other team to balance you out, same goes for that other person. Unless your skill level far exceeds what mmr can calculate for you ±1000sr.
Wrong…. If you genuinely out perform your peers, teammates or enemies, the performance based system moves you higher.
How much higher is determined by how much more you outperform your peers AND how consistently.
Yeah cool, you had one game where you top fragged and two games after that where you did “okay.” To achieve high ladder mobility, even on aged accounts, you really need to be exceptional.
Valorant uses an SR and MMR based system. League of legends…. There’s a few others.
Is it possible that skill variance among the average ranks is inconsistent enough that it’s overly complex to determine one gold player to another?
Stop spreading misinformation. It’s inappropriate and offensive.
You are not outperforming your peers, you are outperforming a cherry-picked lobby, where the odds/outcomes are tampered with independently of the rank. They are rigged to be 50-50 in expectation.
It’s literally the opposite of outperforming your peers or your rank.
The only thing genuine is outperforming the rigging. Over and over without proper payout. It’s boring, rigged, and makes the entire experience a fake sham.
Thank you giving an example how misuse of the term algorithm handicapping does manipulate people to the point of name calling.
An algorithm is a set of rules that is followed.
Handicapping is anything that makes a game harder than usual.
Any system that makes a game harder or easier is algorithmic handicapping.
Which means even a chess tournament in real life uses algorithmic handicapping.
Simply not his problem. It’s how it’s done and if it’s anything it’s mmr.
Thanks for the reply receipts.
Funny I rewrote the post like a week go because it was full of unnecessary stuff but didn’t wanna repost or bump it cause not really a fan of spam.
It all boils down to the hope that you are “secretly” better than your SR suggests. Chances are good that you are not.
And that weak players, in your model, are suddenly much more consistently bad than they have been.
But it’s that inconsistency that makes them worse players. sometimes they play like someone two ranks above them and sometimes two ranks below them. This means that the external circumstances play too big a role for them.
This part, right here.
