Why can't we see our own MMR?
I don't think the argument that if we knew what our MMR is, it'll become some sort of 'bullying' tool or wtv.
I think it's important for us to see our personal MMR and have guidelines as to what's a 'low' MMR and what's a 'high' MMR.
The reason is, at the moment, there is zero incentive to do 'well' at the game. You can still reach rank 1 without escaping through gates as survivor and without getting 4K as killer.
Knowing the value and having some other form of compensation for reaching such value might make it worth it.
It'll also make certain games less frustrating when it's matchmaking not doing its job.
Example: you get 0K as a killer even though you're doing your best, but this team seems insanely good. If in post-game you saw that your MMR value is 1000 but theirs was 2000, you can blame matchmaking for this match and not doubt your skill level.
Same for survivor, if your MMR is 500 but the killer was 1500 post-game, you can understand why your 'usual' strategy wasn't working.
Making these numbers public will inevitably bring criticism towards matchmaking, which I believe is the root of a lot of the issues in dbd, but criticism is good and will lead to making it a better system.
Thoughts?
Comments
-
It's worth mentioning that it will absolutely become a bullying tool, and it will lead to toxicity. That's not up for debate, the only question there is if the benefits would outweigh it being used in that manner, which is a question that is very worth discussing!
So, in that vein, let's look at the benefits. Certainly, it would satisfy idle curiosity of players who'd like to know where their MMR is at, but I don't really see what else it'd bring? Other games don't show your MMR unless it's tied to a competitive mode, and that's mostly (I assume) because there's really not much reason to.
You also expand halfway through your post to not only your MMR, but also your opponent's MMR, and there's definitely no reason you should be seeing that outside of a competitive mode, surely? All it'd achieve is people getting mad if they think the gap between them and their opponent is too high- which it may be, but I don't know what the acceptable gap would be and I'm willing to bet the vast majority of people who'd take to the internet about it wouldn't know either.
Personally, I just don't see the benefits, and there'd be undeniable downsides to deal with if they made this change.
7 -
I think their logic is that there's no ranked ladder, so there's no reason to show it. They're not using it as a ranking tool.
I understand that, but my counterargument is that many players are going to treat it that way regardless. And it's easier to give feedback about things when there's a better reference point for match balance. Of course, BHVR may not care at all about balance relative to MMR brackets, so that might be moot.
I'd rather have ranks in addition to MMR. Not pip-based ranks, but more of a skill rating on top of MMR.
0 -
We already face toxicity from killers/survivors who'd call you in post-game a baby killer or baby survivor, or saying you aren't good at the game, etc etc. Do you really think having a number to throw out would make that worse?
I would agree knowing your MMR value (or your teammates) would not be relevant if matchmaking was doing its job how it's intended (without the expansion that happens).
We saw that when they did the matchmaking/MMR tests, the matches varied greatly from day to day depending on what values they chose. So clearly, it has such an impact on the type of games you encounter. Currently, you can go into a game, win it easily, feel great about your abilities only to get completely crushed in the match right after without any explanation as to how that's possible.
2 -
I would like to have ranks related to your MMR as well. With possibly good rewards. I don't play League of Legends, but I hear that's what they do more or less.
0 -
I could tell you how I would abuse it if I was able to see my own MMR. After I reached rank 1 every month, I would stop escaping until I found the "sweet spot" where I could juice killers but my teammates weren't too bad.
0 -
Yes. We would have the current toxicity, and the toxicity that comes from rank elitism. One doesn't replace or lessen the other, they both exist in harmony; the only blending would be calling someone "boosted" instead of "baby" some of the time.
As for your other points, I'm not really following- how would seeing your MMR fix that problem? It'd still be a problem with the matchmaking system afterwards, so they'd still have to fix it anyway- at which point you're back to seeing your MMR having absolutely no benefits. Could you elaborate?
3 -
Because they don't want people to have a number assigned to their skill, because most people would end up lower than they expect. They also definitely don't want to encourage all the mental gymnastics that come with it. either "why is it so low" "oh, thats just because x/y/z" "I must have hit the soft cap" etc etc.
This community is far too egocentric for any type of verification about standing. Which is a shame, because im sure there are plenty of people who are honest with themselves who might appreciate that info.
3 -
I think MMR is better as hidden as the intention is really just to aid the matchmaker in creating fun matches. This is also why the matchmaker got changed to include more variation so you'd get a mix of competitive matches and matches where you were just better than your opponent and could get a more chill win. It's really not supposed to be used as a kind of competitive ladder and the game would be worse if it was revealed. I can't think of any reason for it other than mild curiosity or obsession over a number. Mild curiosity is cool and I'd like to know mine, but I know others would obsess over the irrelevant number and go into every match with "number must go up or I have panic attack." It's just not healthy.
On the other hand, I'd also like MMR to be revealed just so that I can stop hearing how everyone on the forum is at the absolute top of MMR, they are above the MMR softcap, and how MMR doesn't work because most people are at high MMR or the MMR softcap despite this being statistically impossible. We not only have confirmation from the developers this isn't the case, we know the exact algorithm they used and it's mathematically impossible for that algorithm to produce the results people on this forum claim it does.
If you are playing against killers or survivors who are better than you, the game will always feel like you must be at the highest MMR. If you are matched against people worse than you, it will always feel like you're beating on babies. That's just how skill in things works. And, sometimes, the RNG of the game will determine more of the match than anything else. You might play against the same survivors on Midwich with Nurse or you might play against them on Garden of Joy as, well, anyone.
2 -
You could already do that now though. Without finding out your MMR level. We know that the only thing that MMR takes into account are escapes, so you can theoretically do very well in a match to get your pips but still let yourself die in the match. Your MMR will go lower.
0 -
Except I have a tinfoil hat theory that you lose less or no MMR at rank1. I have done a couple of experiments and the basic outcome is that you can die for 2 weeks straight and the games feel exactly the same.
0 -
Actually the question is whether showing MMR bracket would lead to significantly more bullying than what they currently show. It’s not immediately obvious that people who are inclined to bully someone they perceive as a weaker player won’t bully them based their Prestige or perks or loadout or any other excuse to bully them. So I think that part is in fact debatable.
1 -
Rank wouldn't matter, but there will be some matches where you effectively lose no MMR. The glicko algorithm, which they've confirmed they use for MMR, takes your opponents MMR into account when modifying your own. It's not a static value that increases or decreases. It increases or decreases relative to your opponent.
So if you play against opponents that are higher MMR than yourself, losing won't really change your rating. Going against players that are lower than your rating will change your rating, but one match doesn't really move the number very much except at the beginning. So you'd have to constantly be in matches that are against lower MMR survivors for many matches. Any match where the survivors are higher MMR, which is more likely the more you lose, will effectively not change your rating much. And the matchmaker is trying to give you a match that's close to your rating.
0 -
Because if you could, everyone would know it doesn't work.
1 -
That's not really the point I'm making - if it's revealed, it will be used that way by hostile players. Even if the absolute number of instances of toxicity doesn't go up, it is giving those toxic players another tool, and so it is still a downside that must be acknowledged before it can be argued in favour of.
I'd also personally suggest that because visible MMR would carry out beyond the matches and into things like forum discussions, as it's an easy single piece of data to demand of someone, it'd make toxicity outside the game worse too- but that is up for debate and I'll recognise that. My main point still stands either way.
0 -
I'd love to be able to see my own MMR. That'd be a nice start. Add an option to hide it to the "streamer mode" options as well, and there's literally no excuse to keep it from us. I don't care what my opponents' MMR is - I'd just like to see my own. That way, I can start questioning why it's way higher than it should be on half of the killers I barely play :')
0 -
Seeing your MMR wouldn't fix it, but just give you a better picture as to why you lost a certain game (possibly because your opponent's MMR is much higher than yours). It's simply to give context.
0 -
I think having the number though would either be a 'wake-up' call to some people who think they're much better than they are as well as be useful to those who appreciate the info. I'd certainly appreciate it.
0 -
That's basically the same as just satisfying idle curiosity, though. Seeing a higher number doesn't make the game less frustrating, nor does it solve the problem, so it's kind of just throwing fuel onto the fire of people who already think the MMR system is catastrophically failing and who are loudly, aggressively complaining about it.
I don't begrudge the desire to see it, I just don't see it having practical value that'd outweigh the downsides.
0 -
I disagree that it’s actually a downside if the number of people who bully others doesn’t change when you, say, replace Prestige on the endgame score screen with an MMR bracket indicator.
0 -
It wouldn't, unfortunately this game is so terribly balanced that its not even an accurate measurement of a player's skill, so it will always come with a big "ok but what if" at the end for most people. I dont necessarily doubt you would use the information in a more positive manner (im sure many of us would) but the ones who wont pretty much ruined it for everyone.
If you ever want examples, just check any topic the devs release any type of stats for. or even just posts referencing said occasions. They get misused and misrepresented constantly, when they're not even complete stats to begin with.
0 -
I feel that if I'm forced to play what is essentially ranked mode, I should be able to see my attendant score.
0 -
How do you mean? DBD's matchmaking system is now less of a ranked mode, as it does not have ranks attached to it and uses on the more generous, casual winrate-based matching instead of requiring you to play in a particular way.
It's just a normal quick play queue now, moreso than it was before.
0 -
Interesting... Is that only at rank 1? That would be counterproductive considering they confirmed they don't use our public ranks to use as values for MMR.
0 -
It's ranked mode with the ranks hidden. Matchmaking is uniform.
0 -
My point exactly.
People who are going to bully you in post-game, are going to use any information they have (alongside your perks).
I have literally been criticized for having "Windows of Opportunity" as a 'try hard' perk and to call me a bad player.
People have been criticized for their prestige level (if too high --> no life, if too low --> baby)
so why should we take that into account? MMR is just one more number shown.
1 -
But what makes it ranked mode? There's no like, ranks, there's no restricted rules, there's no rewards, there's no requirement to play [x] amount of the regular queue, there's no regular queue to begin with...
Seriously, I'm asking you. It's lacking literally every marker of a ranked mode and has been brought more in line with more standard matchmaking compared to where it was before- what am I missing that makes it ranked mode?
1 -
I'm not sure how it works, but I can tell you that for a period of about 10 days, I constantly died as a survivor and my MMR kept lowering to the point where I was facing almost baby killers. It took maybe, 3-4 days of escapes to get back to what I assume is my regular MMR.
0 -
It wouldn't make it less frustrating for some, sure. But instead of someone blaming a certain perk or killer and thinking it's broken, they might think, hey, maybe this perk is broken at that MMR level.
I just wish the devs were more transparent in terms of what they take into consideration when they're balancing the game. They've mentioned in the past how Nurse always ranks low on the % of killers that are played and therefore they don't think she's a problem, but then again they adjusted a perk for the upcoming chapter because of her.
Having "prove thyself" at low MMR/ solo q would not nearly be as broken or effective as in high MMR/swf... so MMR must be taken into consideration and I believe revealing those values would help the players formulate those arguments better.
0 -
When u said getting to rank 1 as surv without escaping u lost me.
Can you please do a video of no stop no escaping and having pips.
0 -
it's essentially ranked mode without everything else you mentioned. Which is what makes it a bit 'unfair' and 'not fun'.
If it wasn't ranked, they'd just match you with the fastest random players and your matches would vary greatly.
0 -
?
Have you never played games where you died in the end but still got +1 or +2 pips? The pips are not solely based on your escape.
0 -
Two pips without escaping? Yeah... vs friendly killer and later give him a kill.
A pip yeah but really situational 4 unhooks and killer being bad and doing gens and totems.
Right now im on the place were i get tunneled out of game when 2 gens done at most 1 gen done vs killers with the most broken ######### ever
0 -
But that's not how other games function, so why would it be how it functions here?
Other games use matchmaking too without being ranked mode. Even the ones that have ranked mode still use matchmaking for their quick play, it's just looser (I assume) and typically the match rules are different too.
The things I listed are the things that set ranked mode apart from a quick play mode. MMR is not one of the things that sets them apart.
1 -
It doesn't have visible rank anymore, but its still a "ranked" system as long as MMR exists. The outcome of each game contributes toward what the game recognizes as your "skill" and thus, attempts to match you with similar opponents. Quickplay would be if MMR was off entirely, as it would not be using those types of considerations in matchmaking and instead focusing on speed. Any system that uses any type of MMR is ranked in all but name.
0 -
Yeah, it depends how long you last in the game. From experience, the longer you last in the game, the more pips you'll have. (Unless of course you're doing nothing but hiding all game)
But yeah, I meant that it's easy to lower your MMR by dying at the end of every match, but still getting pips to increase your rank.
0 -
I mean, sure, but doesn't that mean "quick play" basically doesn't exist? Other games use their MMR systems for both queues, the difference is in how strict it is.
Also, that's how this game worked before as well. Emblem based matchmaking actually was a series of ranks used to match you with other players that you were expected to engage with actively in order to make them increase. That's way closer to ranked than the current system, which is totally passive and based off how you normally play.
0 -
Appreciate the honesty 👍
0 -
It's definitely not. There wouldn't be a softcap and acceptable range (350) that accounts 40% of the range between a fresh account (900) and a soft cap (1700) in a ranked mode. If anything, what we have now is the opposite. The best players often play against very average players, and it shows.
0 -
That's true, they designed it this way so the best players don't have long queue times (or have to go up against the same people all the time) but then it hurts the average player IMO.
0 -
A lot of games don't show you your MMR though.
1 -
True, but in a lot of those games, you don't feel like you're getting stomped on every game or every other game. I don't know if their MMR is better or what, but I can't think of another game where you wait to load into a match, start the match and then within 2-3 minutes, you're basically held hostage until you're kicked out of the game.
Well that's not true, I guess there's fortnite like that, but at least the other players are more or less on your level.
0 -
I'd rather hear this now and then over the numbers complaints that showing it would lead to
1 -
many games do not have ranked progress affect unranked modes. They exist entirely separate, with the unranked mode often just having connection based matchmaking. In this case its all ranked, but stays ranked and tries to switch over to quickplay style matchmaking while still keeping ranked tracking. Its probably the worst possible way to handle it, but they know how much separating the two would splinter matchmaking entirely.
Also the only reason its based off of how you play is because you have to play within its expectations? You can play well and die and play poor and get a 4k, the current system's only results matter approach can be anathema to good play more often than not.
0 -
That is not how I'm given to understand most games function. Unranked modes still use MMR, it's just a different MMR score to your ranked one, and while one doesn't affect the other that fact is basically irrelevant to the topic of DBD's matchmaking as DBD doesn't have separate queues to begin with.
DBD doesn't try to "switch over to quickplay style" matchmaking, it just broadens the range of acceptable matches when queue times are long, which I'm pretty sure happens elsewhere too? Nothing about DBD's matchmaking is actually unique, or somehow analogous to ranked modes specifically, as far as my understanding goes.
As for your last point, I don't see how. Playing poorly and winning is a pretty uncommon scenario, for either side, so in general your winrate is indicative of how well you normally play. I guess if you broaden it to "playing averagely but the other side makes more mistakes", sure, but you're still winning those matches if you capitalise on mistakes and so pushing you up to higher lobbies is still the right move if it happens consistently.
0 -
Wait, they actually used glicko? Lol, that's what I suggested multiple times on Reddit.
0 -
No thanks, players don't need a reason to play sweatier than they already do.
0 -
Because some players have inferiority complexes.
0 -
I would really like to see my killer MMRs, just to see which killer I'm best and worse with. Of course I know from experience who I like and who not, but I would bet there are some surprises. Killers I really like to play and have bad MMR with, and killers I hate to play but have a high MMR with.
0 -
To my knowledge, yes. Or some other algorithm that's based on glicko but I think it's specifically glicko 2. It was mentioned in a Q&A. I think it was the same one with the hockey analogy. What Patrick said is that the algorithm mostly works as a typical glicko system would for survivor where each match gets fed into the algorithm to produce a new rating (and I assume a ratings deviation based on the algorithm).
On the other hand, for killer, it uses a batching mechanism where all 4 matches are fed in as a single batch for a single moment in time. This is because your MMR score isn't expected to be different within the same match. So the batching gives a more reliable number on the other end. I think, if they didn't do the batching, you'd get a different MMR number depending on the order it was fed into the algorithm even within the same match.
When MMR was first revealed in a leak, I remember the leak said something about kills/escapes being the metric plus amount of time in a match and then it said something like "your score changes by 20 going up or 20 going down". After it was confirmed the kills/escapes thing was true, the second one became just a kind of normal assumption even among some large streamers. I've seen Otz not necessarily say it, but seem to base his assumptions about the MMR cap off of what I assume is that leak. Patrick's words in the Q&A seem to directly contradict that portion of the leak and it's likely the leakers saw something like "in one match, this happened and the MMR score changed by 20 so that must be the value." Then they insinuated that it was always true no matter the reason.
But we'll likely never learn more because this community is a bunch of children and preferred to make hockey jokes. I'll admit I'm a bit sad about that.
It would be different for different people. I don't know too much about it or how many matches the previous person played, but 2 weeks could be playing 10 matches or so. If they let survivors escape or they were survivor and died, it's not necessarily the case that it would change especially if they had a very low ratings deviation, their opponent had a high ratings deviation, and their opponent's MMR happened to be higher than theirs. I don't have too much of an understanding of the algorithm as my expertise isn't statistics, but based on the Wikipedia entry, there's a score and a ratings deviation. As you play more games, the ratings deviation goes down and your score starts to go to a specific value. As you don't play, certain "rating periods". If enough rating periods pass without data, the ratings deviation goes back up and your MMR would start floating more again.
If you have a high ratings deviation, as a new player likely would, you'll also have a lot more deviation in your matches. If I have a rating of 1200 and a ratings deviation of 100, then the system will believe my MMR is anywhere between 1100-1300. If I'm really at the 1100 side of that, I could get matched with someone who has a score of 1500 especially if their rating deviation is also very high or matchmaking is a bit slow that day and it widens the search enough.
We also know from the MMR tests they did that they have a "bidirectional matchmaking" switch. When searching for a match, it starts at a narrow range close to your MMR and gradually expands as time goes on. When it sees two sides that need a match and the MMR range overlaps, it will match them together presumably in some kind of queue. With bidirectional matchmaking, it requires both sides to have their match search range overlap with each other. We know that system is off and unidirectional matchmaking is on. That means that only one side needs to overlap with the other side to get matched. So if you go into a queue at a time when matchmaking times are imbalanced, you might have a 2 second queue, but the other side was waiting 5 minutes. If they were waiting 5 minutes, you may get too big of a deviation where it'll just feel like matchmaking is off.
If you look on Wikipedia, there's solid math for the system, but there's also a lot of "derive magic number from data and input here" that would be very difficult to figure out. It's not much of a surprise that it took a long time to do this when you consider that they have to build in the data collection capabilities, run data analysis, tweak the numbers, and they have to account for the fact that one side is a team but also not a team. I feel like if people listened to Patrick when he was trying to explain it rather than mock him for hockey analogies, we could have learned a lot about the struggles and tradeoffs they had when developing the system rather than just mocking a guy for trying his best and doing his job. It was a rare moment of incredibly transparency that you wouldn't hear from a game developer because that information would be considered "trade secrets" and the community just kind of ######### on him for it.
1 -
Good info, and agreed all around.
I think one of the biggest misunderstandings people have is when they cry "MMR isn't working." MMR is a different component from matchmaking.
MMR is working if, given a particular matchup, it can correctly predict the outcome of the match. That's it. If a really good killer faces really bad survivors, and MMR says it's going to be a 4k, then MMR works. If the prediction was incorrect, updating players' ratings post-match is the attempt to "fix" it.
The problem right now is that matchmaking is effectively indistinguishable from matching players up randomly. The most accurate MMR system in the world won't matter if you're not effectively using the information it produces to build quality matches.
0
