Visit the Kill Switch Master List for more information on these and other current known issues: https://forums.bhvr.com/dead-by-daylight/kb/articles/299-kill-switch-master-list
We encourage you to be as honest as possible in letting us know how you feel about the game. The information and answers provided are anonymous, not shared with any third-party, and will not be used for purposes other than survey analysis.
Access the survey HERE!
Any Plans to Reinstate MMR?
I know the old mmr system was far from perfect, but is there any plans to reinstate some sort of rhyme or reason to matchmaking? It would make for a better game if the two sides had some sort of equal footing so there was a chance for a good fun game.
If MMR is impossible how about a golf handicap scenario where the side which is the underdog gets some sort of buff so gameplay is a bit more evenly weighed?
Comments
-
Would you be willing to wait longer for an MMR match rather than being matched with 3 level 1 Survivors against a level 100 Killer?
I guess I do not understand why the idea of matching similar skilled players together is not part of the design of the game.
4 -
I think there may be some misconceptions here, so I'll do my best to clear them up.
The game does use an MMR system for matchmaking and it has not been removed. That said, not every match will be perfect. Matchmaking systems need to balance match quality and queue times carefully. If it prioritizes making the absolute perfect matches every time, you're going to wind up waiting a very long time for a match. While some may be okay with that wait, others don't care as much and would rather play sooner. It's a tricky thing to perfect and something we're always tweaking to improve. The vast majority of matches are already within a very close margin where your odds of winning vary by a few percent.
There are some extreme cases where this is not possible, though. For example, if there aren't enough players near your rating playing in your area at the time you're playing. In cases like this, the matchmaking system has to gradually expand its search range until it can find you a match, otherwise you may be waiting a very long time (possibly forever). This scenario is fairly uncommon, and while it's unfortunate, there's not much we can do there. The matchmaking system can only work with the pool of players that are available to it.
But as far as levels go, they mean absolutely nothing and have no impact on matchmaking. If an expert survivor happened to pick a new character this match, they aren't suddenly a terrible player. Especially since the progression overhaul, they even have access to most perks right away. The same goes for grades and prestige; they have absolutely zero impact on matchmaking and do not reflect someone's skill in any way.
The exception to this is the PTB. Matchmaking ratings are not used on the PTB since there are not enough players on the test build to create quality matches. Therefore, it's disabled and you can be paired with anyone.
14 -
Lobby dodging, people transitioning from console with 1000s of hours to PC or people making Smurf accounts could also be a determining factor as to why games feel off balance.
10 -
What about people learning new killers? I remember how you guys talked about the mmr getting lowered to incentivize learning new killers. So someone who has 1k hours on Trapper alone but never touched Nurse, gets easier survivors until the player learns to play Nurse. This isn´t the case right now. Which makes learning new killers a chore. Especially killers with a high learning curve.
6 -
When the matchmaking system "gradually expand its search range," does it prioritize the outer ends of the population bell curve? For example, if the MMR ranges from 0 to 100, does it give priorities for 0s and 100s (with the least numbers of players) over 10s and 90s so that 0s will much more likely to be matched with 5s and 10s, and 100s will be matched with 95s and 90s, instead of the least number of players being stuck with each other (0s and 100s being matched together because of the difficulty in finding others with similar ratings) From the complaints sometime heard on the forum (I don't know if it's true or not) it sometimes sounds like brand new players are going against players with thousands of hours - to the point where it's hard to believe that there is absolutely no other players in between those skill levels during that time in that region. Or do all of those cases occur only because of random backfills? ...any progress made on the backfill issues?
2 -
Not sure I follow, but it works something like this:
- You search for a match.
- The game tries to find you a match with other people close to your rating.
- If a certain amount of time has passed and it can't find you a match, it slowly starts to expand the search range.
- The range expands until you find a match OR until it has expanded to the maximum range.
It doesn't prioritize anyone in particular, but the people who would notice it most are the people on the very low or very high end of the spectrum since there tend to be fewer players at the extremes. Since there's far more people around the average skill level, it's not common for the matchmaking system to expand much- if at all- before they find a match. But that said, the maximum expanded range means that players on the super low end and on the super high end would never be matched together, no matter how long they wait.
8 -
I guess what I was wondering is when the matchmaking expands, it should be much easier for the middle 80% of the ratings curve to immediately find a game with each other than for the top and bottom 10% to find a game with players within that 80%. Just wanted to make sure that top and bottom players are not being matched with each other because everyone in the 80% had already found a match amongst each other. It sounds like it's not - thanks.
Few more questions:
1. You said:
"The vast majority of matches are already within a very close margin where your odds of winning vary by a few percent."
What percentage of current matches require the game to expand the search for players beyond what BHVR would consider a fair match for all players?
2. How well does MMR predict the escape outcome for SWF games vs solo-que games?
3. Any progress made on the backfill issues caused by lobby dodging? Mandy had suggested locking the lobby - could we at least test that out to see if it would improve the quality of matches?
2 -
#3 would be great to atleast test for a week.
4 -
So then how is it possible that one of the most well known Survivor player that should by all reasonable accounts be well over the MMR cap got matched against a killer with 97 hours in the game? I understand that playtime doesn't equate to skill, but I find it impossible to accept that a multiple competition winning survivor was pitted against someone who only recently picked up the game and matchmaking was doing a pretty good job so far.
2 -
I think the elephant in the room is the backfill system and lobby dodging at this point. The initial accuracy of the lobby is a moot point when the overwhelming majority of my lobbies have at least one backfill. A survivor or killer dodges in at least 70% of my lobbies.
9 -
It’s because they are again showing that they don’t care about their most loyal, veteran players. He straight up said that this happens more often at the extremes of the MMR brackets. It’s the same reason they balance the game for scrubs who occasionally play. They don’t care.
1 -
This clarification brings me, and I'm sure many others, joy. I hope many eyes read take a gander at this post.
0 -
...because it's only natural that the extreme ends of a bell curve would have fewer closer relations...
This is a mathematical truth, not a conspiracy.
2 -
I always find it amusing when people say they should balance DBD around high level gameplay... which is literally nothing but facecamping, tunneling, gen rushing, discord SWF, and comp dropping pallets. Aka the things people whine about the most in DBD.
5 -
Thank you for always explaining the MMR system when it come up, it is very helpful ad clears many misconceptions.
I am curious, would it be feasible to have a 2 system matchmaking where players can control their matchmaking preferences in settings.
Setting 1: Prioritize Time - Literally the present system, unchanged other than the "always improving updates" you mentioned.
Setting 2: Prioritize Rating - A system where time is not a factor and you are only matched once players at your rating have been found. Due to the way lobbies and backfill mechanic work, if a player left the lobby, the whole lobby would need to break and the search would begin again.
This way players can have their cake and eat it too. Personally, I'll always choose the present system. But, this wuld allow players who state "I'll wait however long, as long as it's a well matches game" to actually do that. But, it won't force it on any player. Additionally, if the player is in control, they should have nothing to complain about.
I do realize this could be unrealistic if you invest the resources into an implementation/system such as this and +90% of the players stick to the present system.
2 -
Is there any data around how common or uncommon it is to have what you would define as a "fair match" of "close mmr"
Because matchmaking feels like a giant pendulum to me where i'll get one of 3 scenarios.
I'll get 1 match that is a SWF kill squad that probably plays comp, teabags me, trolls me, and overall destroys me. Then my next game is the complete opposite, a bunch of survivors trolling each other, and i 4k at 5 gens. Then my next match is relatively normal. Then i might get a SWF kill squad again.
It feels like matchmaking is balancing me around a 50% win rate., but the way it does it is by just throwing some 10k hour squad on coms at me, and then throws a bunch of babies at me that i stomp. Great 50% win rate, balanced.
The other scenario i get is when i play killer and i just win 10-20 games in a row by a bunch of people who seem like they are barely trying. Then i'll get that SWF kill squad, and then another 10-20 games of a bunch of people who i stomp easily.
Then further still, on killers i don't normally play, i feel like i get thrown that SWF 10k hour squad every time. Like, i generally don't play trapper for example, but if i get a challenge or something where i need to play him, i'll get that death squad almost every time. But then on my nurse main that i play where i legitimately probably haven't lost a game in months with, i'll get babies constantly. Now sure you might say that is due to experience, i play nurse a lot so i'm really good with her, but i don't play trapper much so i'm bad with him. But i can say for sure that you can TELL the different between these groups of survivors by how they play, and how fast gens go, and all the other factors. It is just something that you "know" from playing them.
6 -
Balance for people who play the game, not the Claudette crouching around the map and not helping the team. It seems like an obvious choice to me.
3 -
Conspiracy theory time:
Many perceive slightly better players than themselves as a comp-SWF-squad, and slightly worse players as troll baby-survivors. Maybe because of the game's snowbally nature, maybe because we can't imagine how much better and how much worse players are out there besides those we usually face.
1 -
I don't believe mmr is legitimately used. I won't believe until I see it. So many times I'll check after a match because the other survivor couldn't last long in chase, missed multiple skill checks, hides across the map when chase is nowhere near them, etc. Everytime it's a new player. Same with a killer that obviously makes dumb mistakes and gets smoked. I spend maybe 10 seconds waiting for a match and the lobby is filled within a few seconds, so I don't really believe these replies by Peanits.
2 -
I suggest yall look into matchmaking right now and see if something has changed since as of late I feel my matches have been way too one-sided. I very rarely see balance matches anymore, there are either me(the killer) stomping the survivors at 4 gens or the survivors competely owning me and I walk away with 2 to 3 hooks and no kills. I know I'm not the only one who feels the matchmaking last few months have been really off. The streamer Hen333 made a post about it on Twitter a few weeks ago also.
2 -
To clarify what Peanits said, they don't want to make the majority of players feel they need to wait longer for a more fair match as many of those same players will get bored of waiting and just go play another game like Civ V, so until further notice, we're stuck with the faster Q times and worse match-ups. :/
1 -
I can definitely tell the difference. There are what i call "normal" survivors. And then there are the "babies" and then the "efficient" ones. And you can definitely definitely tell the difference just in how they move and what they do. The "comp squad" is an exaggeration sure. But it was meant to illustrate the point rather than try to explain details.
The point is, my matches swing wildly back and forth in terms of quality and fairness for seemingly random reasons.
2 -
If it were balanced around the Claudette crouching around the map there wouldn't be so many complaints about solo queue. It would be awful for playing killer as it would be balanced around one or two people. A player to do gens and one to run the killer for 5 of them.
It would be like balancing the game for killers that don't know how to chase. They try to bloodlust every safe pallet on a map, refuse to break them, and take turns around obstacles as wide as possible. They'd have to remove 90% of the pallets and allow killers to phase through the objects to balance for that.
0 -
Well you simply cannot fix that if that is true.
1 -
I think a big part of what has alot of people convinced that DBD matchmaking is near to non-existant is a combination of two main things
The first thing is that there is a possibility that the maximum range is a bit too wide at the extreme ends. The main complaint I hear is from cases of extremely skilled players with thousands and thousands of hours matching into people who have maybe a few hundred- and such a fact shows well before the match ends, as such games tend to be terribly onesided in the favor of the veteran. I hear and see from alot of high end players that something in the park of 90% of their matches are basically trivial because they match into people who don't know what they're doing.
The second thing, and the thing that much easier to fix, is the fact that MMR/Ranking is not visible or apparent to players. The way you've phrased your post here implies that the MMR system is very granular with a wide range of tiers or brackets for skill; but that is about as far from common consensus on MMR as it gets from the perspective of the players (as propagated by the game's content creation community). Most players are convinced that DBD MMR consists of a very small number of brackets, the absolute highest end of which is extraordinarily wide and encompasses the vast majority of the playerbase- therefore resulting in these skewed matches being the norm.
If that is in fact the case, then the fault is clearly a flawed system for calculating MMR which does not accurately reflect the skill of the players and is falsely placing too many people in the highest bracket. If it is not the case, and the system is in fact granular as your post seems to imply, then the easiest fix would be to simply make MMRs visible to players. Their own, and those of their opponents post-match; including MMR changes from match results. This would give everyone a much better idea of how the MMR is actually working and do absolute wonders to mitigate complaints about the system- as if people can clearly see their relative granular rankings, they may be much more inclined to believe that the other players simply made too many mistakes and had a bad match.
Of course, this would also happen to expose any potential flaws in the MMR calculation since it would be glaringly obvious if the system is giving players too much MMR for wins relative to losses- or if the game is even correctly identifying what constitutes a win or a loss in the first place.
3 -
^ This almost sounds like a roundabout way of saying that the whole MMR system just doesn't work, period...
9 -
For me it feels the majority of the playerbase is at least above the soft cap at this point. And for how this MMR works, it means it will be still a mixed bag of average, good and really good players. Maybe it's time to higher the soft cap a bit?
1 -
It's not far from perfect its useless you get potatoes with you so call things as they are
1 -
Did you know that modern MMR is broken? That's why there's been a lot of discussions about MMR on the forums lately. No matter how much you listen to the explanation in words, no one can believe you if you can't see the rating.
Consider making casual and ranked matches as others have suggested. The reason is simple. This is because there are players who value "time" and players who value "quality". And make sure that only you can see the rating. This is for credit. What to do is very simple.
2 -
Some random quirks i noticed with matchmaking
- if you have an escape streak, you get randomly assined to a new player lobby
- if you play a high prestige char, you get in a lobby with high prestige chars. if you play a low prestige char you get in a lobby with low prestige chars. the time played on steam profils stays in the same range.
- when killer is 100% on EU server for me it feels like there is no matchmaking on the survivor side. just random skill levels (from 1 hour to 10k in one team) together.
- mmr is ignored when someone dodges the lobby and there are a lot of dodged lobbies
5 -
Since there have been several suggestions in this thread that promoted separating those who want quality matchmaking vs fast matchmaking, I'm going to bring up why BHVR has probably not done that.
By separating the queue into "quality" lobbies and "fast" lobbies, you are actually dividing up the pool of available players. So even if you want "fast" matchmaking, half of the people in the current queue may be unavailable because they want "quality" matchmaking, and therefore you may be forced to wait twice as long - it is no longer a "fast" matchmaking system. Players who want "quality" matchmaking may also need to wait e.g. 20-30 minutes instead of 5-10 minutes also - this would be because even though there may currently be others who are at a similar MMR, they may get immediately pulled into the "fast" lobbies because that's what they had picked as their preference. (although some folks, like me, may be ok with this.)
This game already divides players by 1.MMR, 2. eight server locations around the world, 3. and role (survivor or killer) that they play. (And remember, they want the killer:survivor ratio to always be 1:4 as much as possible.) It is also influenced by the time of day that players are able to play. BHVR may be hesitant to separate the queue further because of that.
4 -
"Fast" Robbies should not be subject to MMR as they are now. In other words, the lobby will be created as soon as 5 people gather, so if the queue is split and the number of players is halved, it is expected that the matching time will actually be faster. Also, I think that there are few players who care about matching time because players who emphasize "quality" literally gather in "Quality" Robbies. Finally, by having "fast" lobbies and "quality" lobbies, both players who want "fast" and players who want "quality" can enjoy it (currently neither player can enjoy it), so the player population I think that will increase.
2 -
The last time they asked this community in a survey whether they favor queue times over quality matchmaking the overwhelming response was for the queue times.
1 -
This game isn’t casual. It was overwhelming when I started playing in 2018, and is even more so now due to all of the content. This is a common talking point that I hear and it is completely wrong. It takes a ton of time to learn how to play properly. The devs should embrace the fact that the game is obtuse and balance around those who stick around. If they are balancing to make the game casual, then they are failing miserably.
2 -
Getting the blasted Cross Progression done can solve your second point.
1 -
Except again, nothing about the game is casual. I think when people say they want it balanced to be casual, they are actually saying that they want easier wins.
4 -
I'm surprised so many people say this game is a casual game. If it's a casual game, why are so many people commenting on balance changes? I think it's because people want the game to be fair. So what does the game look like when we seek fairness? The answer is simple. Makes the game more competitive. In other words, it can be said that many people are really looking for competitiveness in their hearts. Fairness⇄Competitiveness
4 -
They said they routinely move the soft cap to ensure that roughly the top 5% of the playerbase is above it.
I'm certainly not above it, and I intend to keep it that way.
0 -
It has not been a casual game in a long time. Every change is made to make it highly competitive . MMR does not work. If you play Survivor you may get one person in your MMR but the other two will usually be way way way lower. There needs to be a change to make it fun for all I can not get friends to play because they will not grind 1000 hours to get a decent game.
2 -
Rank based match making was 10000000 times better.
5 -
I think a lot of this is true and goes to human psychology. If MMR works players should get to a point where they feel like they hit top tier competitors, regardless of whether it is true or not.
In addition to the snowballing, I think people don't really grasp the probabilities. Example: let's say I'm using huntress and I have a 50% chance to hit the survivors with a hatcher based on our skill levels. One out of four times I'll hit the first two shots and be doing great, one out of four times I'll miss the first two and potentially be way behind. But people tend to think in the first scenario that the survivor was just really bad and in the second that the survivor was really good, when its more of a case of randomness. Given how important the early game is, early good luck/bad luck can be a deciding factor.
I'm surprised so many people say this game is a casual game. If it's a casual game, why are so many people commenting on balance changes?
If you took everyone on the forums, reddit, etc. posting about the game, we'd still make up well under 1% of the player base.
Also, a casual game can still have balance changes, things can be broken, those two aren't really related.
2 -
This. I value quality infinitely more than time. I will wait for a lobby so long as the subsequent trial has balance. It's not like there's so little to do on the internet during the wait. Plus if things are more balanced the amount of DCs/suicides will naturally reduce, as the painful and obvious imbalance is one of the primary drivers of DC/suicide.
I understand there are those that do want to get lobbies as fast as possible, regardless of getting stomped every time. That's fine. They can use the fast lobby setting while those who prefer a more balanced game can use the quality lobby setting.
The important thing is that there IS a choice, and that is actually works properly. Not like the trial feedback rating above the scoreboard, not like the player kudos feedback, not like the current lag icons. NOT a placebo.
2 -
Thing is, BHVR had a test week last year where they messed with their SBMM every day to see how people reacted. On the more accurate MMR days, people overwhelmingly did not like the longer waits. In fact, at its most strictest there were people waiting for upwards of an hour. Lots of people say they'll be happy to wait longer yet when that time arrived, it was recieved negatively.
1 -
The issue with that was everyone was in the same boat, so it was an all or nothing situation. What I'm proposing is letting the players decide which they want. If you want a fast lobby, you can turn that setting on. If you want a balanced/quality lobby, you can use that setting. This way each type of player can get what they want, or as close as reasonably possible.
2 -
The problem with this is that you're dividing the playerbase, which itself effects matchmaking negatively.
With the people who select 'long queue/closer matches' out of the 'fast queue/looser matches' running, both options will end up taking longer.
Fast queues will end up having no matchmaking at all, and long queues will take a ridiculously long time to the point where you'll probably give up waiting around the 30 minute mark. Everyone loses.
1 -
Wouldn't it be a good idea then to implement a competetive and a casual mode?
Competetive would be for the players that focus on balance and that are ok with waiting a bit longer for their matches and casual would focus on the players that want a fast match making experience.
2 -
I'm against anything that splits the playerbase sigificantly.
I don't want to sit in lobbys for 5 minutes plus. Sure once it's a minor inconvienence. But 200 times it's a reason I stop playing your game. It sucks in Apex, it sucks extra hard in Overwatch, and it would suck here too.
Absolutely no.
0 -
Let's say there is a queue (in the same region, around the same time) that looks like below with its role, MMR, and preferred matchmaking type.
- Surv 500 Speed
- Surv 500 Speed
- Surv 500 Quality
- Surv 500 Quality
- Kill 500 Speed
- Surv 1000 Speed
- Surv 1000 Quality
- Surv 1000 Quality
- Surv 1000 Quality
- Kill 1000 Quality
Now, if there was no preferred matchmaking type, these players would immediately form two lobbies - one with MMR around 500 (top five), and another lobby with MMR around 1000 (bottom five).
However, when you add another variable to matchmaking (in this case, Speed vs Quality of matches), now you will end up with three incomplete lobbies:
- Surv 500 Speed
- Surv 500 Speed
- Surv 1000 Speed
- Kill 500 Speed
- Surv 500 Quality
- Surv 500 Quality
- Surv 1000 Quality
- Surv 1000 Quality
- Surv 1000 Quality
- Kill 1000 Quality
So now you would not only require 5 more players, but those 5 extra players would need to fulfill very precise roles, MMR, and matchmaking preference so that their lobby would accept them. And this is just a simple scenario in which there are only two ratings (500, 1000) - realistically the MMR range would be much more diverse and varied, and if those players had picked "Quality" for those specific MMRs, each one of those would have to have a separate lobby, and would not able to share any of the "Speed" lobbies even if few of them have similar ratings. Additionally, in the example above, it has the ideal 1:4 ratio of killers to survivors. Realistically it may be much more lopsided, dependent on the time of the day, which will also affect the number of accessible lobbies.
Now add "different region/servers" and "SWF" (besides the issues of adding SWF with wide MMR gaps, remember that duo SWFs can never be in the same lobby with 3-man or 4-man, and 3-man teams can never be in the same lobby with duo or 4-man, and that issue is exacerbated if one of the SWF prefers speed while another prefers quality) and you have even more additional variables to consider when forming lobbies.
So in a perfect world with a perfect ratio of players, adding this new variable (of preferred matchmaking type) may benefit the formation of lobbies to the satisfaction of majority of players - but I'm not sure how realistic that really is - and as I said, this may be what concerns BHVR.
1 -
Because long time players don't want competitive matches, they want quick easy wins. So all the best and sweatiest players would just play Casual, with an even bigger chance of being paired with newbies they can bully.
2 -
Most games know that queue quality and time cannot coexist. So it seems that you are trying to measure the coexistence of "quality" and "time" with vague MMR, but please understand that it is impossible. It's best to split the queue in two.
0