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!
If the devs are so horny on using MMR why are they unable to use it properly?
What I mean is simple conditions that should alter your MMR with lets say:
- - If you dont use meta perks or use meta perks (Each slot equipped with one meta perk would increase your MMR for that search)
- -Same goes to addons and items, if you dont equip one you should be matched with ppl that either dont use any or weak stuff, same geos to killer addons ofc
- -If your character is low level then your MMR should also be recued and try to match you with ppl close to you level, doesnt has to be perfect.
You could say "Well maybe it is already doing that but there are simply not enough players online so you dont get matched with them." Thats bullshit cuz I see often enough games or get with my FULL SWF WITH FULL META killers that are low level or even play their first game ever with that Killer. Its not just in the last few weeks, it has been always like that.
I usually dont ######### on the devs that easy cuz I am mostly very neutral to most things however I dont see ANY possible explanation or reason why they should not do that.
Comments
-
By my understanding the main conditions fo that alter your MMR are simple.
Survivor gains or loses MMR if they escape or die
Killer gains or loses MMR if they kill survivors or let them escape
For better or for worse, you can't deny it's simple.
1 -
That is how you lose and gain mmr but my point here is your MMR should alter according to your loadout. This is nothing new, many other games that use mmr also do that exact thing.
4 -
The killer cube will not go down without a fight.
Really the original pip system was easy to pip on, it just took ~40 hours of gameplay to get to Rank 1. In its simplicity however, I think it did what it needed to do in order to keep newbies and pros apart. Well, at least until the 13th hit.
If you played a lot you were probably good at the game so you'd red ranks quickly. If you played occasionally and weren't too worried about it or couldn't get a 7k point average each game...you probably didn't get that high in ranks. You played with people like you.
Then you had green/purple ranks with people like me that couldn't decide if they wanted easy matches or to be challenged. It was a simpler time.
6 -
Agree I really miss the old pip system, it wasnt perfect for sure but it did its job WAY better then MMR and it never got constant changes, tests or attention from the devs and still was better without support for years.
11 -
I love these posts about MMR from people who don't know anything about basic algorithms but presume to think that they can immediately design a much better system than people who've been working on it for well over a year with all the data at hand. It's not arrogant at all. 😂
1 -
They are using it properly, the way MMR functions in DBD is how the majority of MMR systems work from my research.
There are a bunch of reasons why this is a better decision than having the system pay super close attention to how a single match went or what variables went into that match, but what I really want to hit on right now is that this is just how MMR systems work. They take your overall winrate and base the number on that.
1 -
I was so excited by the idea of MMR (back when I played this game consistently) because I thought they would separate all Killer MMR from each other. I really wanted to experiment and learn a killer at lower ranks so I wasn’t abused harder then normal. Nothing changed though. Either play viable killers+perks+add-ons or get washed 😂
Oh BHVR.
1 -
Ye lets ignore the suggestions I made there trying to improve it but instead try to make fun about that, how old are you?
1 -
Thanks for explaining however I am already aware of that.
0 -
MMR doesnt even matter. If you've been waiting for a match it'll just pair you up with anyone - regardless of MMR.
MMR could be good - if it was ever actually used...
3 -
Then... what's the point of this thread?
I'm not trying to get a 'gotcha' here, if you understand that this is how MMR functions and that it's the better option than looking too closely at a single match, what is your overall point...?
1 -
In a perfect world an MMR which takes into account everything mentioned (SWF, Solos, Add-Ons, Items, Perks) would exist. In a perfect World, Maps would also be calculated, including everything what happened due to RNG (the Survivors who died on a Map with the minimum amount of Pallets are not worse than those who survived on the same map with the maximum amount of Pallets).
However, something like this will never be possible.
This is most likely the reason why they went for a super-simple MMR - no MMR in this game can be accurate, because there are way too many factors.
3 -
I agree with your 3 points. Perhaps they're concerned about locking player's loadout, people weren't that happy when they locked the killer at the lobby.
0 -
Yeah... just got a good team and we got a one perk new Ghostie... who was completely overwhelmed.
0 -
exactly
1 -
I wouldnt call addons or perks as "closely looking". Many other games also have that exact same function. My point here is that ppl who are new to the game OR if you get a new character which barely any perks shouldnt get matched with other full equipped players. Atleast if you have a low level character you shouldnt be paired with other that are already full.
0 -
I agree with this, however I am not asking for complicated fucntions. Atleast if you que as an lvl 10 trapper you shouldnt get 4 survivors p3 with all perks lvl 3. I met few days ago with my swf squad a lvl 14 onryo that just bought the DLC. This guy had 1,3k hours which is fine in that sense, however had had only 2 lvl 1 perks against our full stacked squad, dont see any reason for that. This happens rather frequently, I also meet occasionally low level survivor with my full leveled p3 killers.
0 -
hes gotta get better at hockey!
1 -
The only way to set this up would be to prevent changing of perks or loadout once match making begins. I'd personally be for it, but I can just imagine the amount of complaining in the forums such a change would generate.
0 -
there are ppl asking for Artist reworks who just came out a few months ago, I wouldnt worry about such complains much tbh.
0 -
Are you certain about that? Because all of the games I could find information on very explicitly didn't have that kind of function, they very explicitly did just look at a player's winrate to calculate their MMR. This is for two main reasons- one, it's just much harder on its face to objectively measure how powerful/skilful/etc any of those things are in a way that isn't context dependant, and two, it'd throw the system completely out of whack if those things ever changed. For instance, basing your MMR partially on whether you have Dead Hard equipped would then require the system be changed to accommodate for Dead Hard ever getting nerfed- then snowball that out to every perk, item, and addon in the game, and you see why this is a monumental amount of work.
That's why it's based on winrate. If your game is balanced, anything that'd be deemed more powerful or more skilful by something more involved will reliably lead to more wins, and if your game isn't balanced, your matchmaking system can't be built around whatever's unbalanced about it. It's simpler, it's cleaner, it doesn't require mountains of maintenance for any changed aspect of the game, and it works just as well.
Now, obviously you shouldn't be matched against new players; we know that the system is broken in how it matches players, and that's what these recent tests have been trying to fix to my knowledge. To the point about a low-level killer without all equipment slots unlocked... I don't think that's a matchmaking problem, I think that's a problem with how perk and addon dependent a lot of killers are. This is less applicable to survivors since they all share a perfectly serviceable basekit, but it applies a little- and either way, that's a balance issue and not a matchmaking issue.
0 -
How can you improve an algorithmic system unless you have some fundamental understanding of how it works in the first place? Your "suggestions" only really make sense to those on the far left of the Dunning-Kruger effect... i.e. they don't, to anyone with some knowledge of how it actually works.
0 -
Hearthstone and world of warcraft both had this for example and it worked very wellt for both games. I wouldnt do it that complicated about which perk and which not but more about if a perk is lvl 1-2 or 3 and how many you have equipped.
0 -
If it is that hard to do why do other games also can do that? Just asking
0 -
Those things are still balance issues and not matchmaking issues, though? The fact that you have to dump a bunch of bloodpoints into a character before you have all perk slots unlocked or even usable perks to fill them with isn't a problem with the matchmaking, and especially not the MMR.
(Does WoW have PVP? I legit didn't know that, I haven't played it before.)
0 -
Their equation is wrong. Kills do not equal skill nor does pure escaping. Until they realize they have to include chase time, killer stuns and the killer hooking the system will always fail.
0 -
lol, sorry but what a horrible idea
thank god you're not on the team lololol
0 -
can you imagine making a system that takes all of that in consideration actually work?
they would break the game for sure and you'd still be here complaining
what you guys want can't be done, not without making you wait 10 minutes in line to fine the ACTUAL person who plays ACTUALLY the same as you so you be pleased.
I don't know in what world you guys live.
0 -
You don't make rational decisions while horny.
1 -
ye wow has pvp xD, if you have a low equipped character and que for arenas you get matched with opponents with a similar equipment level to yours.
What do you mean with balance issue? Its just the progression of the game, you start small and have to level up which is totally fine, you just shouldnt meet ppl that are way above your own level (Progression wise).
0 -
WoW does, yes. Has since launch in 2004. Not only that, but in typical Blizzard asinine fashion, they even turned it into an eSport.
3 -
The balance issue is that you have to dump a bunch of bloodpoints into a character before they're usable at your level. The solution to that problem is either to have characters start with at least three perk slots and their teachables already at level three, or to fundamentally change how loadouts work at all- changing the matchmaking is just more points of potential failure, which I don't think is going to help in the long run. It's just another thing the devs would have to juggle to make sure it's working consistently across all tiers of play, to say nothing of characters that start with wildly stronger teachable perks than others.
0 -
I just want to start by saying that none of these are bad ideas. On paper, they all make sense. Actually incorporating them into a matchmaking system is where things start to get tricky.
For starters, let's look at just perks. Someone has to subjectively pick which perks are meta, we cannot base it solely on how often they're used. A perk like BBQ & Chili is incredibly popular, but I think we can all agree that it's not because slapping it on will carry you to victory. Applying a penalty to it wouldn't make a whole lot of sense. So already, you need to selectively apply penalties to perks and also determine how much of a penalty each one should carry. This is already highly subjective and I strongly doubt that you could ever find a value for each one that everyone considers fair. There are bound to be loads of disagreements over penalties (e.g. This perk has too much of a penalty. It's not even that good. Now I get super tough matches just for using it.)
This also does not factor in the fact that not all perks are equal. Some perks are stronger when combined with other perks/add-ons. Take Dead Man's Switch, for example. On its own, it's okay. Put in on a Killer like The Artist that can easily force people off of generators and it becomes much stronger. How do you apply a single value to it that doesn't discourage using it on other Killers while also fairly penalizing it on ones that can make full use of it?
All that aside, these penalties/bonuses are redundant when it comes to how a matchmaking system actually works. By design, using strong things inherently increases your chances of winning. When you win, your rating goes up and you'll end up against tougher opponents. Put simply, the matchmaker does not care how you got there. It doesn't care if you're a perkless legend, it doesn't care if you run meta every match with ultra rare add-ons. What it cares about is your chances of winning, and its' only goal it to pair you with people where your strategy (whatever that may be) works about half of the time. When it stops working more than half the time, you plateau and stop climbing. This doesn't mean that you're a great player because you have the best stuff, it just means that the matchmaker has successfully put you in a place where you're no longer pummeling new players.
Adding another layer of prediction on top of that would only make this worse, and would only be of benefit in a scenario where you can see your rating (and it actually mattered). Say you're a 1,500 player, but you're running really good stuff so the game treats you as a 1,700 player. If you lose your match, you only get shoved further down into matches with less experienced players who are not prepared to deal with the meta you're running. In a scenario where you can see your rating and it means something, sure, this matters. This stops someone from artificially boosting their rating by running all the best stuff. This is not, however, the case. Your rating means nothing beyond placing you in matches where you've got an equal chance of winning and losing, and it already does that without predictive penalties & bonuses.
21 -
Your rating means nothing beyond placing you in matches where you've got an equal chance of winning and losing, and it already does that without predictive penalties & bonuses.
Maybe it creates matches with an equal chance of winning or losing in some alternate reality or parallel Earth, but that's a whole different thread topic.
5 -
There are some edge cases that can create some matches that should not happen. There's a few things we're working on fixing to prevent them, but you can find more info on that here.
7 -
because the most base line requirement you need to fullfill to have a functioning MMR system is a balanced game and we already massively fail at that.
the game being imbalanced is not necessarily a bad thing by the way - there are a couple of very good reasons for that (most importantly that a balanced DbD is a boring DbD) - but there is no point in trying to push a matchmaking system which overall goal is to balance matches by having every player be on the same skill level, when one of the two competing teams gets heavily favoured by the game to the point where the team that isnt being favoured gets obliterated in response.
It always baffled me how the Devs stated that they "balance around fun" on one hand, but then decide a matchmaking system like that was a good idea on the other.
3 -
I undersand the part about picking specific perks, however my main point was here about ppl that generally dont even have the perks to begin with. It was mentioned that if you play on a new killer your MMR should be low to begin with since you are learning that specific character, however I have seen more then enough cases where ppl got matched with ppl way above their levels even with their main killers (without survivors that fill an empty slot after someone left)
0 -
Just a thought on using character level as basis for mmr, and correct me if I'm misinterpreting your idea but:
Let's say hypothetically I only play wraith. I am good at him and am high mmr with wraith. But, I'm getting bored and I want to learn sadako. I level them up to 15, get 4 slots perk slots and maybe a whatever build but it will do.
If we use character level, I am mostly now going to get baby survivors on sadako, who plays almost very similarly as wraith (as all m1 killers do at their core). My skills as an m1 killer in chase still carry over so all the mindgames I'm doing end up with downs, even more so than usual because they are baby survivors.
On the flip side, what if I start leveling up blight, get his perks unlocked and eventually max him out. Now I'm at 50+ character level on him, using some good addons, and I'm getting really really good survivors and struggling to do anything.
Let's be real for a moment, a lot of killers don't have a high skill floor to pick up and use if you got the basics to killer down. All 115% killers can play the exact same way - I can play the doctor the same way I can play the artist if we ignore their specific powers. Your skills as a 115% m1 killer absolutely transfer over through out all the killers in that category. While learning artist's crows and zoning with them have a higher skill floor than learning doctor's zaps, at the end of the day you can still just m1 without your power. Like, if I ever get fed up in a chase as huntress or run out of hatchets, I'll m1 a survivor, it's not too far off of a concept.
As far as how it works now? Is it perfect to learn a new character based on your highest mmr? I think it's close. I think there could be a tweek or two made. I'd like to sorta see an average of all your killers - if I'm tip top mmr nurse, average mmr spirit, and now I'm learning billy, I'm going to get survivors on the same level as my nurse which is gonna be rough.
And on an engineering standpoint (I don't talk about it often, I'm a software engineer) - this algorithm of basing it factors such as character levels, addons and types of perks chosen would be disgusting to write, test, and would inherently break when perks are changed. I think we should remember that the meta in this game are essentially bandaid fixes for various game mechanics people perceive as unfair (Unbreakable = slugging, DS = tunneling, etc, corrupt = survivors spawing on gens together) and it shouldn't be looked at as static - as perks get nerfed, buffed, reworked the meta changes (with the exception of the survivor meta, the killer meta in regards to perks and addons have a little bit more variety). That is a lot to factor in and when a perk becomes meta, it's not immediately obvious at first until people play around with it and find out it works great (like: pain res/dms). It's not impossible, nothing really is but it's just sort of a bad idea for this game imo.
0 -
For new characters that you don't have perks on, I feel that's the one that's arguably the most doable. Though if it's not handled very carefully, you could end up with the inverse problem (you're treated lower than you are, but you end up winning and get pushed higher).
In any case, I think @jesterkind hit the nail on the head here:
This would be an issue with progression in the first place, not necessarily the matchmaking system. Balance and progression definitely have an effect on matchmaking, but attempting to address an issue with one by fixing another is likely not the best way to go about it. Ideally the progression system works in a way where this isn't a problem in the first place rather than needing a bandaid fix on the matchmaker's side.
8 -
I agree with a lot that has been said. As a software developer, I can imagine the nightmare it must be to balance an asymmetrical game around so much subjectiveness. I can certainly see where establishing the *win* conditions a while back has helped with MMR logic. I'm sure there's more to it than I realize; I consider myself a casual DbD player.
Personally, I think matchmaking ought to be based upon player mastery, but how mastery is determined should be revisited. I will say, however, if you define a win for survivors as a unit, instead of individually, I could see some merit with that, but defining a win individually for survivor is fundamentally flawed since a Killer is one unit and 4 Survivors are one unit.
As survivor, my chances of winning the game (escaping) depends significantly upon the mastery of the other 3 survivors. Thus, if a survivor team wins a match, all of the survivors who participated should share in that win in some way - even the ones who died. The same with respect to losses. You play as a unit and you share the outcome as a unit just like with every other known competition.
Mastery ought to be calculated based upon individual points and team outcomes, against some difficulty metric. You'd want to weight the individual points scored higher than team outcomes but the team outcome should still be a factor in determining mastery. The difficulty metric is important as you will need it to determine whether mastery is gained, maintained or lost.
The challenge that I see with this approach is with respect to whether Killer mastery will scale with Survivor mastery, if Survivor queues are loaded with the same or similar mastery levels and other important factors like Survivor team advantages (SWF). I realize this is a controversial topic.
I'm sure your team is well aware of the scaling problem. At the end of the day, you'd have to decide if competitive matchmaking is really the best route for the majority of your customers. Perhaps you could introduce a Competitive Arena along side a Casual Trial. If you go down that road, it's easier for you to establish player expectations upfront and deal with scaling issues with handicaps.
Otherwise, remove MMR and make matchmaking more random and mysterious.
~ Steven
1 -
Your second-last paragraph is basically the hockey thing all over again. Thanks for nothing
Post edited by egg_ on2 -
Thank you Peanits for the response, it is fantastic to get more insight into the thought process of MMR.
2 -
Is sad to see that you guys are so stubborn in keeping this Sbmm when the community doesn't want it at all, last two test of MMR were based on grades imo and they felt so good.
0 -
For what it's worth, I'm impartial. Like it if you want, hate it if you want- at the end of the day I'm just here to take the feedback. I'm sharing this information not to convince you, but because a better understanding of how the system works will only lead to more productive discussions.
That said, regardless of what the last test was, I implore you to check the feedback section for those two tests. While they may have been better for you, there are many people who disagree. It's important to remember that no matchmaking system is a one-size-fits-all solution.
9 -
It’s sad that you really think only your opinion or the one of the vocal minority matters. Also the tests really show that the devs are not that stubborn but are willing to improve the SBMM system.
Also you don’t know what the last two tests were, but even if it was grade based, the current grade system is just playtime for the month, it can’t match people properly when they haven’t played for a short while. And there were many people disliking those tests as well (and there is literally no way possible where everyone will be happy)
4 -
Thank you Peanits, it's good to see the thought process laid out in a concise manner.
However, if I can ask as I don't know, of course, what's under discussion, has rebalancing game mechanics been discussed?
IMO, the issue with MMR on the killer's side, in a macro overview, isn't a roughly 50% win rate, but that playing, as you advance in MMR your games become less fun and more stressful.
This is due, once again imo, in the fact that tunneling, hard camping and gen rushing become more and more staples of high MMR play. Those playstyles are not fun.
When there are multiple hooks, chases, etc, as was rewarded by the Emblem system DbD is a very engaging, fun game. I purposely keep myself at that level by setting artificial limits on when I hook survivors so that I stay at an MMR where I can enjoy my games. Playing too efficiently leads to a playstyle where preparing for killer kind of feels like you need to skip rope, chug energy drinks, and play 'Eye Of The Tiger' before logging in.
This frustration was alleviated more by the ranking system as you could get a really difficult match followed by easier matches and it didn't feel like you have to be hyper focused all the time.
Tunneling, camping and gen rushing would still exist without MMR but there were other rewards to not doing them (red ranks were the visible achievements and you couldn't 4 hook 4K and get red ranks before). If those strategies could, in the future, be made less effective so that more prolonged games of Dbd would become more of a norm that might alleviate some of the concerns about MMR.
1 -
MMR works best if both sides want to play the game but as there has been a mass exodus of killers for whatever reason (coh? boon totems? that entire month where survivors could hop into a snowman? who can know or speculate as to why) it doesn't actually work out well since the asymmetry is now lopsided.
People obviously still want to play survivor enough that they'll wait 10 minutes in line like at disneyworld just to get a killer, any killer but it wasn't always that way. The game had problems, but throwing in competitive numbers and stuff exposed them and now you either quit playing killer side or you adapt and get meaner like some of us have >:)
0 -
Because of reasons.
If you're going to ask an extraordinary vague question, you should expect nothing better than an extraordinarily vague answer. You're not even trying to compare apples and oranges, but rather you've just gone straight to apples vs fruit in general whilst expecting a serious answer 😂
0 -
"Put simply, the matchmaker does not care how you got there."
This right here is a problem though. Matchmaking not caring how someone got there makes solo q a miserable time. The matches aren't fun because all that matters is the escape/kill, not if the gameplay was actually fun or not. I want matchmaking to take into account how my teammates played in a previous match because in the end if I escape/die it doesn't matter if the match itself wasn't fun.
Did the teammate escape because they were a team player that worked on gens and went for safe unhooks/heals when not in a chase? OR were they hiding/doing nothing to avoid being caught, when they were in a chase would lead killer to other survivors, only unhooks were unsafe in front of killer where the killer would instantly start chasing the unhooked survivor, etc. and they escaped either by lucking out opening exit after killer closes hatch or teammates were able to complete objectives even though they had a teammate that was working against them?
Playing with teammates that are team players, even if I die, is generally a fun time. Playing with teammates that do nothing/are basically spending the match sabotaging the team even if I escape isn't fun, it's annoying/frustrating - especially with current SBMM knowing if they escape too I'll be matched with them again or they'll be ruining some other survivors matches.
Have matchmaking take into account how people are playing because an escape doesn't make up for an unfun match.
2 -
Hey as long as you guys keep queue times quick then I'm down for whatever. While I enjoyed reminiscing about the old matchmaking system, I sidestepped the 30 minute red rank killer queues. Let's not do that again.
However if you bring back headless Myers or chainsaw Myers I won't be upset.
0
