Feels like the devs don't understand what's wrong with the MMR system
I read through the official news topic here: https://forum.deadbydaylight.com/en/discussion/315687
All it does is talk about how some people favor fast queue times while other people favor fair matchmaking,The way they describe how matchmaking tries to find people close to your MMR and then expand over time to find people further away from your MMR is normal for any kind of matchmaking system, and it sounds based on that post that they'll just experiment with these numbers.
But it doesn't adress the issue that the way people's MMR is decided to begin with is the core of the issue. Based on interviews I've read, MMR for killers seems to be purely based on whether or not you get kills or let survivors escape, and for survivors it seems to be based on whether or not you escape or die.
Do the devs understand that there's more to the game than living or dying? If you're able to play like a god (on either side) and make the match miserable for whoever you're facing, and then be able to lower your MMR anyways by choosing to die at the end or let the survivors escape... Then you can keep facing newbies forever.
And that's the problem... Not everybody plays to win, personally I enjoy dying to the killer at the end, and after reading the interviews about how MMR is calculated it made so much sense why I keep matching with newbies despite having over 1500 hours.
To fix the MMR system, they need to look at more than just whether or not survivors live or die.
the MMR system needs to look at more variables than that. For example, how long a chase lasts before the survivor goes down, percentage of missed skill checks, percentage of missed swings, how fast a killer is able to locate survivors, how much time a survivor spends idling without progressing the objective. These are just examples, the most important thing is probably how long chases go on for before a down, since that's the most skillful interaction on both the survivor's and killer's side.
I apologize if this is very long, it's just concerning that nothing about the MMR calculation itself is mentioned when they talk about changes, and that they only mention how these numbers are planned to be used while finding matches.
Comments
-
That is certainly a thing that a lot of people have said over the months that MMR has been active, but that doesn't actually make it true.
I've never heard a good argument for why the way the MMR is calculated is wrong, it usually seems to boil down to people misunderstanding what the system is for- a focus on measuring individual skill within a specific match, instead of sorting into general skill brackets.
The devs do understand that there's more to the game than living or dying, and their responses make sense overall. Those things matter in a single match, but if you're looking at aggregate data, winrate is a good proxy for also measuring not only what plays are being attempted, but also how many pay off.
6 -
"I've never heard a good argument for why the way the MMR is calculated is wrong, it usually seems to boil down to people misunderstanding what the system is for- a focus on measuring individual skill within a specific match, instead of sorting into general skill brackets."
I feel like I had a pretty good argument there. You can literally play like a god against people out of your league, giving them a miserable time, but then choose to lower your MMR just to give them the "win" at the end, even tho the match itself was terrible.
Having more variables doesn't "sort it into general skill brackets". You use all the different variables to calculate the player's overall skill. You don't go "oh this person is good at looping killers, match him with a killer who's good at catching survivors". But it's still a variable that determines skill. Living or dying or getting kills does not determine this at all, as skillful players can make the choice to avoid it.
14 -
Sorting into general skill brackets is the current goal. The point isn't to take a hyper-specific look at a single player and ask how skilfully they played each and every match, the point is to have a set of skill brackets for all players based on how skilled they are at winning their matches.
The system only needs to be a general matchmaker, it doesn't need to function like a ranked ladder or sort for a particularly intensely competitive game, it just needs to separate people into broad categories. I do think it has gaps - I'd like to see some adjustments for good loopers that die at the endgame as the only kill, so those players don't lower too far and make games miserable for less skilled killers - but the core design is not a problem.
Players who know how the system works will always have ways of lowering themselves intentionally. There's only so much you can do to compensate for intentionally bad actors.
2 -
Too much effort,I do how one day they gonna put the effort in making a decent matchmaking system that's not flawed at its core.
0 -
Seeing as they compared their game to hockey, I don't hold out much hope that they'll realize why they're current MMR system is terrible.
18 -
Really hope all this testing means they're looking at the kills/escapes system and are willing to change it.
The whole community wants it to change,why not change it?
2 -
Honestly, this is getting so boring that people don't understand how algorithms work.
Do the devs understand that there's more to the game than living or dying?
Yes, the devs understand this. Do YOU understand that you can't put all that into a number (or numbers) that reflect your rating??
That's literally the core of the issue. How do you assign a value - however the value is composed - that accurately reflects all the nuances you want it to reflect. Please explain. I'll wait.
2 -
Current MMR does not measure individual skill because escaping doesn’t demonstrate skill. To use the dev’s awful team analogy: a player can be individually great while still being on a bad team. Some of the greatest players in NFL history have been on historically bad teams (see the Detroit Lions.) The team being bad doesn’t make the individual player bad.
4 -
It measures individual skill over a large batch of games. You're right, there are too many variables for escape/kill to be considered an indication of skill in one match. But overall, across a large batch of games, if you're winning more games than you're losing, you're clearly doing something right- and if you're not, that's a balance issue, not an MMR issue.
3 -
The actual MMR doesnt focus on measuring individual skill at all. Its just boiled down to one basic indicator, and its factors are completely ignored.
2 -
As I said above, it does measure individual skill in general. It isn't concerned with whether you played 'right' in one game, it's concerned about your overall winrate.
0 -
What's so hard about it?
It's not harder than designing any kind of point counting system. You just need to know what gameplay features are dependent on skill.
A big part of that is chase length. Short length means killer did good, survivor did bad. The game already keeps track of chase length so should not be hard to adapt. As mentioned earlier, percentage of missed skill checks could be used as well, since experienced players are less likely to miss them.
These systems are not as hard to design as you seem to think. I am a programmer and could easily make a system based on gameplay actions. The problem here is that the first step is to have the right idea of what actions are considered skillful and what actions are considered less skillful, and it feels like the developers are stuck at the idea phase.
2 -
Did you not read what i read? Overall win rate is not indicative of skill. Plenty of athletes are great on teams with losing records over a season. Survivors rely on each other to escape, just as an athlete relies on the team to win.
4 -
Okay, but what does that have to do with placing them in an MMR system? It's not about how secretly good they are, or how much potential they have, it's about their chance of winning. That's the point of an MMR system, to match players with a roughly equal chance of winning.
2 -
This game's MMR punishes you for doing well, while simultaneously enforcing the player to take advantage of every single thing they can. It's both counter-intuitive and basically boils down to "if you want to win, you better sweat your ass off, that way you'll get even more sweaty matches as a reward." That type of thing works in games that are on even playing field (like most competitive games) but not when the 1v1 balance isnt there. With the absence of any coordination in rank/mmr/pips/challenges/etc you're trying to squeeze too many metrics into simple binary results, and we can use the bad hockey example to illustrate, though not in the usual way.
People like to focus on that example for showing very skilled play being both unrewarded and even punished, right? It actually affects even more facets: The players who avoid interaction and get gens done in <5 minutes aren't bad players, oftentimes thats the right thing for a good player to do. By finishing the gens so quickly however, they not only push the killer to play more lame (as they have to get at least 2 kills as well) but vastly limit the strategies both the killer and their own survivors have to alter the game. It basically devolves into a flowchart.
Meanwhile, the more they're one of the escapees, the more they're going to get 4 other players with the same mentality. Solo survivor is generally bad at most ranks simply because most survivors are bad. There are a very large number of amazing survivors, and a good deal more that are intermediate/above average, but it seems like survivors always want their MMR high so they can rely on their teammates more, and end up with swf tier efficiency. Then once they get it, they only see Blight and Nurse, NOED, tunneling 1-2 out, facecamping, etc... Because its the only tools killers generally have at that high of a level. So once again, being punished for good MMR.
A lot of MMRs problems would be lessened if the players had any indication of what their standing is, even just in a non-sharable way. While you can often tell an approximation of where you're at based on who you get matched with, backfill/swf/queue times/etc can all make things a lot less consistent to judge until suddenly there's 1 gen left, or 2 teammates dead.
I've embraced the "i honestly don't care about the outcome of the match" mentality as pretty much the only way I can play anymore, especially killer. "losing means i'll get easier matches" is a recursive fallacy (since winning said easier matches would then give harder ones) so people can't even use that mindset anymore. Its extremely easy for people to get frustrated with the way MMR and matchmaking in general is currently, but tbh its always been. just harder to know what you're actually mad at vs raninbow ranks in the past.
8 -
you're confusing "skill" with "winning".
My overall winrate is trash, because I have all the perks in the game and don't care about escaping for that extra 5k blood points.
So I usually let killers get those points at the very end, i by letting them kill me, even though I could have escaped.
This ends up being a downwards spiral where my matches are so easy because i get killed a lot, and then I feel bad for looping the killer the whole match so I feel even more inclined to give them the satisfaction of killing that annoying looper at the end.
Me dying while everyone else escapes doesn't make me lack skill, but the game registers my result as being bad at the game simply because I died.
the whole point of an MMR system is to match people with people in a similar skill level to keep matches fun and "fair" to a certain degree. So there's no point in having an MMR system at all if they're not properly trying to detect people's ability to play well in the game. Currently, I would have better matches if matchmaking was completely random.
4 -
...I'm not sure why you intentionally choosing to throw your games at the last minute translates to the MMR system being bad...?
1 -
They're 2 separate things, and I understand not everyone does this, but I've met a lot of people on both sides give away the "win" just as a way to be nice to other players after playing well. I'm good at the game, but I also want to be a good sport and not feel like a complete douchebag, and I want everyone to have a fun game, not just myself. That's why I do this, unrelated to the MMR system... Yet the MMR system sees this as my gameplay being weak and lowers my rating.
The reason why the MMR system is bad is because the purpose of MMR matchmaking is to take skill into account to match people into fair matches, yet people's MMR ratings have NOTHING to do will skill level.
The current system only works for players with the mindset that they always want to WIN, but a lot of experienced players keep their sanity by caring about more than just winning when playing this game. Having fun is more important than winning to a lot of us.
3 -
another way to think about it is this:
Imagine an adult playing a game of chess against a kid. The adult plays real strategies until the kid only has a couple of chess pieces left, and then starts losing on purpose.
Now... Would you judge this adult as being bad at chess, and enter him into a chess contest for kids... oooor would you recognize that he's good at chess but chose to lose?
See, the MMR system does the former. It fails to recognize any skillful gameplay and just says the adult sucks because he lost.
2 -
Those changes are just going to make it harder for killers to de-mmr. You wouldn't be able to farm and 8 hook then let everyone go anymore, you'll pretty much be forced to the "Insidious Lightborn AFK in the basement" strat which survivors hate so much since it leads to them depipping.
0 -
Theres not enough people playing killer for the MMR to work correctly in the first place. The pool isn't big enough to accommodate survivors across all MMR range brackets and keep the queue times the same for all players. High MMR survivors would probably suffer the most if they were to put a cap on what range you could play in. They would probably have 15-20 mins queues waiting for a killer within that range bracket.
0 -
Remember, this is Just a test. They said that feedbacks will be read
0 -
That's a conundrum to say the least
Do you go all in for faster Queue times or for a "balanced" match
Cause at this point that's what I'm seeing
0 -
...What? The point of the MMR system is that you're not supposed to be getting games that wildly unbalanced to begin with, but even if you do get a game like that, the system doesn't say "you suck", the system puts one tally point in the loss column. One game doesn't swing your MMR all that much.
Even if you are getting matches like that all the time, you're still intentionally throwing, and it's a very strange take to say that the MMR system is bad because it can't account for bad actors purposefully losing their games. Any system is going to have that flaw.
1 -
What's so hard about it?
Are you having a laugh? I literally spelled it out. How do you compose a set of numbers, or any other objectively comparable criteria (objective being the operative word, being as the comparisons is done by a machine), that adequately encapsulates everything you guys want
"Dood doood.... did you see how I jumped through a window and tea-bagged the killer... I need that in my MMR!!!!!!!!"
Please.
What's so hard about it? You tell me... if it's not hard, then please design it for us. I've love to see what you can come up with. Just the composition of the ranking, not how it's reached, etc. I just want the data of what you would see if you wanted to see anyone's MMR, according to the imaginary way you guys think it ought to work.
It's not harder than designing any kind of point counting system. You just need to know what gameplay features are dependent on skill.
See, you say this... but don't actually offer anything by way of demonstration. I can just as easily make absurd statements too.
What's so hard about flying? Birds do it. Insects do it. Heck, even planes do it. So I don't see why people shouldn't be able too. It's not so hard. Just attach wings.
A big part of that is chase length. Short length means killer did good, survivor did bad. The game already keeps track of chase length so should not be hard to adapt. As mentioned earlier, percentage of missed skill checks could be used as well, since experienced players are less likely to miss them.
That's the easy stuff. You're just looking at what happened in ONE match. The system then needs to distil that down into what becomes your ranking at that point. And that's what I want to see. What is the format.
Just to save you the trouble of the wrong answer, you might say "the rank should just be a number, and it goes up by X, or down by Y, depending on what happened in each specific game". Great. Except with this arbitrary number, there's nothing to differentiate someone who escaped 20 matches by doing nothing (and their rank went up for the escapes), or someone who did a lot of chases and escaped once against a bad killer.
Do you see the problem now? Everything you want measured has to get abstracted away at some point.
These systems are not as hard to design as you seem to think. I am a programmer and could easily make a system based on gameplay actions. The problem here is that the first step is to have the right idea of what actions are considered skillful and what actions are considered less skillful, and it feels like the developers are stuck at the idea phase.
I'm gonna call BS... not least because, as a s/w developer of over 20 years myself I've not heard of anyone refer to themselves as a programmer (but hey, maybe it's a regional thing), but mainly because you literally talk what looks like the talk, but you don't walk at all. Some people, myself included, would call this sophistry.
So, as per the post of mine which you quoted, I specifically asked "How do you assign a value - however the value is composed - that accurately reflects all the nuances you want it to reflect. Please explain. I'll wait."
You've not answered this. You've just made vague statements like "it's easy", etc, without any explanation of how the data would even be composed, let alone compared.
0 -
i think your definition of "throwing" is different from mine.
I'm not throwing the game at all.
I am playing to the best of my ability, and making sure to help other survivors and playing like a good team mate.
Throwing implies playing like garbage on purpose, or sabotaging the game for others, or overall just not participating in normal gameplay.
Simply choosing to not walk out of a doorway at the end is not the same as throwing a game, at least not what i think most people would consider "throwing a game" means.
I feel like you have too much of a "win/lose" mindset to understand this issue. you've agreed that the point of an MMR system is to prevent people from getting wildly unbalanced games, yet you keep talking as if winning or losing is the only thing that matters. Winning or losing at the VERY END has nothing to do with how balanced or unbalanced the ENTIRETY of a match is.
0 -
MMR is just a number to match people. Its not a GPA or evaluation. Its not a comment on you as a player.
3 -
"Throwing" just means "losing on purpose", you don't have to be playing intentionally garbage for the entire match to throw the game at the very end.
I'm not making a moral judgement here, that's just what it is that's being described- what's being described is intentionally subverting the win/loss condition of the game, if you're making the choice to do that then you should also be accepting that there are consequences for that decision. If you do it enough, you're going to end up in lower-MMR games, because you've intentionally been losing.
The only way to avoid that situation would be if the MMR system allowed players to rise in MMR if they lost the game, and that seems pretty obviously unintuitive.
0 -
That's your definition. In my experience throwing involves sabotaging matches in general. Talking about different games in general here, not just dbd... People who "throw matches" often sabotage their team mates too, or refuse to participate in normal gameplay or perform actions that lowers the chances of the team winning.
In DBD, you also have to ask yourself... What IS a win, really?
It's a team game, survivors VS killers. I consider it a win for the survivors if a majority escape. Killer wins if a majority dies. It's a draw if half escape and half die.
"Throwing" is short for "throwing the game". So if I make sure all the survivors escape and then I choose to die at the end, I am not throwing the game. I helped all the other team mates survive, and took one for the team. That's not throwing.
Throwing would be, if I got downed first, and then kill myself on the hook and leave the rest of the team to play the match on their own. Or, if I spend the whole match crouching around the edges and refusing to help the team or touch any generators. That's also throwing.
Either way, that's not what this topic was meant to be about. The point of this topic is that the purpose of MMR systems is to get decent matchmaking, and calculating MMR based on escapes and deaths is an awful way to determine skill level in order to find fair matches. It feels pointless to talk to you because you ignore all the points i make about the MMR system being bad, and just talk about the definition of throwing. So i'm gonna leave this here as I've already made my point in this thread, and I hope developers find this and read some of the points about the MMR system.
0 -
Ironically during an early Q&A the devs said that basing the mmr on kills is a terrible idea.
I really wonder why they changed their opinion and basically turned this into the Victory Cube instead of keeping something like the Emblem system.
4 -
I am perfectly content to stop using the term "throwing" if you like, that isn't really the important part; the important part is that you're choosing to lose at the last second. The choice is being made, and that choice will have ramifications.
You also don't have to ask yourself what a win is, since the win condition has been relatively clearly defined for a while. The only point of ambiguity that's ever existed is whether a 3k counts for a killer win or if it has to be a 4k, everything else has been established since the game's release.
But yes, talking about throwing wasn't really the point here, that's something you brought up and it confused me enough to derail the conversation. The bottom line is, the way the MMR number is calculated seems perfectly logical to me; if you're winning more matches than you're losing, you're doing something right. If you're not doing something right but still winning more than you're losing, then that's a balance issue, not a matchmaking issue.
0 -
Part of the problem is that most people here aren't able to define "skill" in regard to pvp games - they can give examples on what they *think* are skills, but can't give a specific definition nor can they explain why they should be considered as being "skillful".
It doesn't matter if a chess player can juggle all 16 chess pieces in the air - it doesn't make them a skilled "chess" player. Would it take a lot of effort and time to learn how to juggle 16 chess pieces at the same time? Sure. Would only a very small fraction of chess players be able to even do that? Sure. But unless that particular ability correlates with actually winning at chess over multitudes of matches, juggling during chess matches don't make them "skilled" at chess. The definition of "skill" in pvp has nothing to do with effort put in by the players or its difficulty. It may make them a skilled juggler, but certainly not a skilled chess player. Because the ultimate objective of chess is to checkmate the opponent's King, just as it is to escape as a Survivor, or kill as many as possible as the Killer in DBD. That's the whole point. And if they can do so consistently over many matches, it shows that they know the intricacy of how to win in that particular game - that's what it means to have "skill" in a game. Each move within the game should be made in consideration of whether it raises their probability to win against that particular opponent. Well, how about if a player is able to always have more captured pieces than their opponent at any given time? Same thing - does it give them a higher win rate against their opponents over many matches? If yes, then sure, it can be considered a skill. If not, then no, it's not a skill. (And even if it can be considered a skill in their current pool of opponents, if the quality of their opponents rises to the point where that same ability/strategy/tactic no longer works, then it's not a good skill to have anymore.) Ultimately, the only thing that matters in determining whether something is a "skill" in a specific game is their ability to win over course of time. Only the *results* over multitudes of games can accurately measure whether the actions that players took was skilled or not.
Same thing with any other pvp game. Any game. I don't know much about hockey, but let's take basketball as an example. Say this one particular player is usually able to score more points than the other nine players on the court. Is that player skilled at "shooting"? Sure. But the point of playing the game "basketball" is to score more points *than your opponents* by the endgame. If that player is always only concerned about scoring points and therefore standing by the opponents' goal like an idiot waiting for teammates' passes, and neglecting the fact that in their current environment, it may be more beneficial to help out with defending and getting the rebounds, then no, that player shouldn't be considered "skilled" at the game of basketball, but only with the "shooting" aspect. If that player is only tunnel-visioning one small aspect of the game to excel, and not understanding what the game needs or require as a whole, then why should they be considered "skillful" at basketball? If that player's only forte is shooting baskets, but continually makes bad judgments on other aspects of playing the game of basketball (passing when it was better to dribble past the opponent, or dribbling when it was better to have passed) why would anyone consider him "skilled" at basketball as a whole? If that player continually fails to learn from losing multiple matches due to those mistakes and keeps losing because of their failure to learn, why should they face more skillful opponents?
Many on this forum mistakingly believe that basing matchmaking simply on the win condition of a game only reveal a part of what players do. It's the opposite. Matchmaking through using only the win condition actually considers every aspect of every player, every interaction, and every moment and every environmental circumstances of each game, and judges whether the players made the right calls regarding the match. If the players made enough right calls (at least, moreso than their opponents), then they would have won. If they made less, then they would have lost. That's why the biggest matchmaking systems like ELO, Glicko, and TrueSkill work so well - it sums up everything that happened within the match, and grade them by win/loss.
Let's take "saving someone from the hook" as an example. What do you have to consider in regards to deciding whether it raises your probability to win in the end? How far away is the hook? How much time will you lose by heading over to the hook? What if Hex Ruins in effect? Is there anyone closer? Is the killer camping or proxy camping? Would it be better to continue to work on a generator until they're closer to the next hook stage, or unhook them immediately so that you have more people working on generators earlier? Do I have "We'll Make It?" Do I have "Borrowed Time?" Or is it better to bet on someone else who may have those perks in saving them? Does my opponent have "Pop Goes the Weasel?" and/or "Call of Brine?" and therefore better to finish off this generator I'm working on? Is someone else already going for the save, or have they've been found and are already in chase? Am I already injured and on second stage, making extremely risky for me to go for the save and better have someone else do it? All these questions regarding "who" "what" "when" "where" and "why" are only tiny fragments of consideration that one should be making while playing the game. And the only way you know you've made the right decisions at each moment is to look at the final outcome of the game. The outcome of the game (who won, who lost) is the ultimate presentation of the accumulation of everything that has happened in the game for all players and decisions they've made, every interaction and synergies between different elements like killer powers and perks, all environmental factors like size of map and its randomization - and condenses them to the outcome.
But why not "weigh" these different elements that you personally believe is important in winning? Because the numbering of these "weights" constantly changes, depending on all those circumstances that I've just mentioned above. It's subjective, and creates inaccuracy. And that's just if there are no changes being made to the game. When there are constant updates like in DBD, there's no way to weigh certain actions with a specific number when things always get shifted around. Some things/actions that are "powerful/important" now may not be so powerful/important later, and therefore have less relevance in one's ability to win. Sure, they can assign a specific number to "length of chase" but then what happens if they make changes that makes the "hiding" aspect more important to escape on certain new maps? What if they add new mechanics like "fighting" that are alternatives to "chasing" or "hiding" and thereby making them both less important? Nevermind how synergies between these gameplay mechanics, hundreds of perks, maps, killer powers, add-on, all can change and shift around after an update - who would keep making sure all of these are constantly assigned a correct number and weight so that vast majority of the players have a fair chance of winning after each match? Not BHVR, and certainly not its players. But sure, go ahead and try to assign them all of those variables some numbers - I'll wait.
2 -
Literally had bad game after bad game as survivor, usually get a few good games and wins in but not had one. Seriously bad matchmaking, hacking killers, high tier killers against low tier survs.
At this rate I’d be surprised if anyone is willing to play anymore, it’s pointless even trying survivor if you don’t play SWF, which is annoying
0 -
Given that I felt some differences between the 28th and the 29th, I guess they at least know how to tune it.
When all is said and done, one issue is the random factor.
A big map? Usually bad for killers.
A map with chain-able powerful tiles? Bad for killers.
A map with loads of pallets, including a few "god" ones? Bad for killers.
...
There are a few maps that are less favorable for survivors but besides the ones that are good for Nurse (but not for most others), I can't recall at the moment.
Until they make the maps smaller and more balanced, the MMR score will be wacky.
0 -
The only way to fix the SBMM system is to get more killer players into the game. It will never work with the current shortage of killers because queue times for survivors would be way too long. Offer BP/shard rewards for killers, make quality of life improvements for killers, nerf some of the most OP survivors perks, etc will incentivize more people to play killer. If you have more killers in the pool then SBMM will work better.
0 -
What was victory cube?
0 -
A win for killer is a 3 or 4k and a win for survivor should be a 3 or 4 man out. The problem is they treat survivor as a 1v1v1v1 while killer is a 4v1... If they actually changed survivors goal to get as many out as possible then the game would be better.
0 -
Honestly, if they just favored teamplay over selfishness, the game would play better. Forget about win conditions. People just fixate on those. You should gain or lose mmr based on your contribution.
Killer is fine with kill based mmr in my opinion. I know that one is not particularly popular, however.
1 -
What we had as piping requirement before emblems.
Only hooks and kills mattered. Killer required 9 points to pip. Hooks gave 1 point and kills 3 points.
So a killer who face camped 2 survivors and got 1 additional hook got a guaranteed pip. While a killer who got 7 hooks but no kill, depiped.
The prevalent tactic to rank up was to camp hard.
3 -
Weird
0 -
Here is the wiki on it
0 -
Oh ok dbd pre April 2019 is weird to me because that's when I started playing
0 -
Let me make MMR real simple for everyone.
Devs: Let's move away from the original pip system based on points. Introducing the killer cube! The killer must now get 2 kills and a hook to pip! Wait. Nobody likes this. Let's move away from it. Introducing Emblems!
*Time passes*
New Devs: Let's move away from the Emblem system and ranks. Introducing MMR! The killer must now get 2 kills and a hook to pip! Wait. Nobody likes this. Let's move away from it. Introducing -whatever comes next-
That's all there is to it.
6 -
Idk what they were doing these past few days with matchmaking tests. I can't wait for the announcement to say what was favored each day, but so far.... is ok for survivor, but as killer I get sweats on a killer I never played... All the games I played last night were like that.
I think mmr rn is... better compared to the game comparing skill to your rank (and then because it took to long it paired you with whoever).... I thought what if they changed the "win" condition from escapes and kills to points you earned (if that's even possible and convert the points to then 1-4 points. -2 if you died or something idk). But what about those games where one survivor runs the killer for the whole game and you just barerly like get 15k from gens and a totem ? You'll barerly rank up ! (To be fair that's how it is anyway.)
Idk I think unless some revelation happens after those matchmaking tests, keeping the current mmr but showing you your mmr rating before finding a game lobby, and then after the match is over is a solution.
0 -
I literally offered examples of multiple things that could be used to count MMr, yet you completely ignored all of it.
Tbh i didnt even read your entire message after it was apparent you didn't read mine.
0 -
I just dont agree with you with what a "win" is.
It's a team game for the survivors, and for me it's a win if i can help all other survivors escape by keeping the killer busy the whole match. Dying at the end really doesn't make a difference in this game, I just miss out on 5k bloodpointsz that's it. I reach rank 1 easily without needing to escape.
I dont always let the killer kill me, by the way.
I do however enjoy interacting with the killer more than i like doing gens, so i put myself in danger to keep others safe and we "win" a lot of matches with all of us escaping, but if someone dies it's usually me because killers get pissed after i distract them for 5 full gens and i often end up getting face camped.
This is another example of the MMR calculation being bad. If you can loop the killer for 5 gens, you basically won yhe match for your team. Yet you rank down in yhe MMR aystem because the killer got mad and face canped you.
Idk why anybody would think this is a good system. Clearly there's more to the game than whether or not you, personally, lived or died
1 -
The irony here is, I did read your message. You just can't comprehend my response.
Contrary to what you say, you literally offering nothing, other than vague ideas which cannot be translated into a number. Yes, you offered ideas as to what can be measured on a match-by-match basis, which is fine. But the trick is condensing that into a number which carries forward. You offered nothing here, and this is what MMR needs... some kind of measurable metric that's based on your history of games, not just a single one-off game. You need to work out how to format all those metrics you want into this objectively comparable dataset. You offered nothing in this regard.
0 -
I mean, that's not up for debate? The win condition has been the same for the entirety of the game's existence, whether you personally care to fulfil it or not. I don't personally tie my enjoyment of a match to whether I won or not, but I also don't pretend that means it's up for discussion what a win is.
For the record, I do think a potential gap in the system is someone who can run the killer for five gens but dies in the endgame as the only kill, but that'd still have to be happening in more games than it doesn't for it to be statistically relevant for the MMR system, so all that really needs to change is to lower those players slightly less to compensate and it won't cause any issues. It's also not a gap because that player would be "punished" by being ranked down, MMR isn't a reward or punishment, but rather because lowering them to the point where they bully newer killers wouldn't be particularly fun for those killers.
But again: In order for a situation like that to matter in MMR, it has to happen repeatedly, more often than it doesn't.
Regarding your last point: That's true of any game that has an MMR system, though, and it clearly works okay for those games.
1 -
Wdym, I've met plenty of people saying that a survivor win is when 3 or more survivors escape.
Im not losing just because i die if i die getting everyone else out.
Just like how I dont lose a round of counter strike just because i die, if my team does well enough I still win.
In any team game, if you perform actions that cause your team to win and they do win, you also win regardless of whether you lived or not.
I feel like your mindset on this is very narrow
1 -
The win condition for playing survivor has always been to escape, and if your team gets out that's even better. We can argue back and forth about whether designing the game as a semi-team based one was a good idea, but that's the game that it is. It's not "my mindset", it's just what the game is.
That's not to say that team-based adjustments for the MMR system we have would be a bad idea, mind - I think it's unproven that it's necessary but if MMR needs more adjusting after its current flaws are fixed then that'd be a great first place to look - but the win condition doesn't actually have anything to do with the MMR, it's been the case forever.
0
