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!
The problem with MMR: DBD is not hockey
Comments
-
Don't get me wrong. That is absolutely ######### up and MMR needs some serious work.
That ain't my gripe, it's the people who look for meme quotes like the hockey and "was that really a skilled play?"
The mental gymnastics people go through on this forum is out of control.
0 -
Regardless of what sport you compare DbD to, MMR needs to have about 10x more nuance than it currently does.
If you get dunked on, the game shouldn't automatically try to dump you against some 0 hours survivors. And if you happen to get a 4k, it shouldn't rocket you up to teams with 90% 4-man escape rates. The point of a rating system is to make micro adjustments as you play one match to the next, so things shake out into a comfortable middle ground, where all the best players play each other, the bad/new players play each other, and everybody else lands somewhere in between, tweaking up or down slightly as they improve (or worsen). The current system is basically an On/Off switch of Pro/Potato.
To use a sports metaphor: you don't get a shot at the Stanley Cup, just because you won your first game. It should take time, hard work, and tons of skill before you're in the finals, to play against the best the game has to offer.
Of course, that's just part of the problem. The REAL problem is that matchmaking just throws out MMR more often than not. If they were going to prioritize matchmaking time so much more than MMR, then there was no reason to waste all the time, money, and effort making a system to replace the grade-matching one. Current, MMR is a DISTANT second-though, compared to getting games going. If you want to treat the game like a competitive entity instead of a party game, you have to make MMR matching the absolute priority. And that means (in the current, survivor-buffing & killer-nerfing development cycle) if you're in a team of unkillable survivors, you're going to have to eat 20-30 minute queues because there just aren't enough superstar killer players to match up with.
To use another sports metaphor: If you don't have enough teams in the Major League, you don't match them up against High School teams just to get a game going. You play fewer games. Nobody wants to see the Detroit Redwings obliterate the local Boy's Varsity Ice Hockey team.
NOTE: Yes, I'm aware they want to tweak the 'back-fill' thing. But that's just one of the "queue time is more important than game balance" problems. They need to fix the whole system, top to bottom. And when the find that it takes a hour or more to get a killer capable of handling hyper-pro SWF sweat-squad teams, maybe THEN they'll have the precious data they so crave.
3 -
Yeah I mean I find them funny and somewhat bad statements. It's like the 'pretty good job so far' line.
I don't base my dislike for SBMM/MMR purely on those quotes, that would also be pretty silly
3 -
I'm not really trying to hone in on the quote itself, but more so on the general competitive mindset that attitude pushes. By having a system that only cares about the end result and does not reward people for playing with actual depth beyond that, it detracts from the general gameplay experience. This system relegates killers to tunneling and survivors to selfish escapes, which generally makes the game less fun. Why should I go for 12 hooks if that win is directly equivalent to one with 4 hooks? Why should a survivor run to unhook their teammate with one gen left to go when leaving them to die and finishing the last gen is the objectively optimal way to rank up?
1 -
Right, I'm the one that wants to argue. 🙄
0 -
What you're really saying is you want a different meta, as everything you described are just efficient outcomes.
If we're going to be competitive, we're
Edit driving and half post was deleted
0 -
I mean, yeah. Clearly. Your response to his original post was just a disagreement with really nothing to offer. Then when the OP further clarified himself you shifted the goal posts hard as if his whole reason for posting this was to say "slugging is bad".
I mean, you don't have to agree with the guy, but at least add something to the Convo besides steering the conversation to another us vs them post. Like, Imagine reading this first two posts on this thread and only gleaming "slugging bad". Weird stuff.
2 -
So here is my stance, if you're going to add a system that only exists to functions as a more in-depth calculation of skill then that system should do that function effectively based on how the game is actually played and actually scored. I'm not saying DBD needs to rise to an e-sports level game standard, but it is making use of system from that realm poorly. Not only is it making use of said system poorly but also very obfuscative. There is another game that has MMR and is a bit closer to DBD's style of game called Spacelords. It used to be called Raiders of the Broken Planet. It's a more casual game with a more casual approach, but even in that game people still want more specificity put into their MMR system.
Why? Because it simply helps sort other issues such accurate matchmaking, in Spacelords case end of game scoring, and can generally just make for a better experience when you inevitably need to make adjusts to said system.
As to pips not being a win/loss condition, that's the only mechanical metric that determines a win or loss for a killer and realistically if you escape as survivor and don't pip you've not really progressed in any meaningful way. This of course has its own flaws as you can have a game where 5 people pip which technically shouldn't be a thing to begin with, but it is.
The idea of looking at this subject in whether or not the added metrics soften the loss is so alien to me. Like couching everything in kills and escapes puts the system at its softest level. This stuff works both ways, the more data you take into account the more exacting scoring becomes so if you're someone who sits by a locker pointing to dodge crows and do nothing of value for the entire game then get out the door you're going get credit for escaping, but you're MMR would be penalized because you didn't actually engage in meaningful gameplay.
Even if you just want to have the system purely to even out matchmaking kills and escapes alone do a poor job of that. Also, kills and escapes are the only thing currently taken into account. I'm not sure as to why you think there is more being calculated as this has been stated in a Q&A and reported on.
2 -
The comment you're replying to is directly in line with the point I originally made. All I did was add in a couple paragraphs for clarification and modify my subsequent wording to make it fit and/or fix any redundant statements the additional paragraphs created. The end result is it reads better and the intention of my original post is slightly better articulated. Not sure what you're trying to get at.
1 -
"if you escape as survivor and don't pip you've not really progressed in any meaningful way"
I agree! But you did win. Winning the game and escalating the game's progression system are separate, and they should be separate. A progression system needs to take more into account than a pure matchmaking system, and the progression system we have does; you need to play a very specific way in order to max out your emblems, to the point that honestly I think it's a flaw in that system. The progression system gets to demand more of you than the basic rules of the game, because it's there to test your skill, not to simply decide if you win or not.
To the rest of your points, I think that's the core thing that's getting in the way here- the matchmaking does only exist to facilitate matchmaking, it's not a want situation. That's all it is- it's not a progression system, it isn't there to reward players, it isn't there to give feedback or even really calculate exactly how skilled you were within a specific match. There are other systems in the game to fill those gaps and we've seen how poorly it goes when those are used for matchmaking too.
(Also, no- it only takes kills and escapes into account but it's not a simple binary. If as survivor you die first, you lose more MMR than someone who dies third, and there's similar changes to your MMR gain/loss based on how different your MMR score is to your opponent's going in.)
The bottom line is: This is a generally casual game that has a matchmaking system wholly divorced from its progression system. It does not need to be all that specific because it's not there to test player skill, just push people with comparable winrates together. Anything that gets in the way of that is a balance issue, and needs to be addressed separately.
0 -
Alright then, if it only exists to facilitate matchmaking then why use the MMR format at all? Why call it "Skill Based Matchmaking" if it isn't meant to in anyway account for your actual ability to play the game?
They could have sorted matchmaking out by simply setting up matchmaking gates where new players could only play with new players and based the entire system off total time spent playing the game. We know they have that number and even have information to such a granular level that they can figure out who actually played Bubba so they could pass out iri shards.
There were more effective ways to get the same or slightly better results without even remotely crossing into a territory where you needed to make a background mmr system nor did it need to be called Skill Based anything. I know this gonna edge into pedantic territory, but what has been conveyed by their messaging does not match either the definition of the words uses or reflect the reality of the implementation of the system in question.
If you want skill based matchmaking then by default your progression system needs to be an integral part in that. If you just want new players not to get curb stomped you can simply let new players play against each other till they get enough game time in to push out of the gate into the wider playing field.
1 -
...Because that's what MMR is for...?
It's literally Match Making Rating. The MMR format exists to facilitate matchmaking, that is its entire purpose. It is often paired with a progression system in the form of a competitive ladder, but it's not even slightly synonymous with that sort of system.
It also is still skill-based, just a different definition of skill than you're used to. It doesn't exist to figure out how skilfully you played each match, it exists to assign you a general skill bracket based on your winrate. Both of those are accurate definitions of skill in this context, so they haven't strictly done anything wrong in calling it that- though I'd agree it is misleading to the layperson, I didn't know this at first either.
2 -
I never "steered" anything anywhere. And when I want to actually "argue" about something I typically give examples and explain my position in a bit of detail.
I'm not "arguing" with the original poster, the only thing I did was point out 1) that the hockey kerfuffle is overdone; and 2) that his side comment about slugging was entirely dismissive of a perfectly valid way to win the game.
Now YOU I'm arguing with. See the difference?
0 -
"Chess factors in not only flat wins and losses, but time spent in game and number of moves. Fighting games that used the TrueSkill rating system judged you not only on wins and losses, but factors such as offensiveness, defensiveness, execution of combinations."
Please cite sources.
Chess usually uses ELO or Glicko rating systems, and neither one uses "time spent in game" or "number of moves" to determine players' ratings.
And what fighting games use the TrueSkill system, but also take into account all of these other variables for matchmaking? The entire idea behind using the TrueSkill system is that the use of other "performance" variables would make the matchmaking LESS accurate so they purposefully avoided adding them to its algorithm. What game developers would first pick TrueSkill as its ratings system and then proceed to muck it up with other variables?
2 -
I'm just here vibing. Have a great night.
1 -
Problem with MMR is just that it's ignored, nothing else.
0 -
I think in the way SBMMR is... does lead to boring games (one way or the other)
Kills Vs Escapes
But yes there are many ways to show skill in the game
I think Kills V Escapes should trump them... Like ( in your metaphors) getting 10 Hooks with one Kill would mean that you credit for those 10 Hooks but you get more credit for that 1 Kill... And the same for getting 4 Hooks and 4 Kills would mean that you get credit for those 4 Hooks but more for the 4 Kills
And on the flip side... If I were to do 3 Gens and die that would mean I do get credit for completing 3 Gen but not as much seeing how I died. Or I could be that Survivor saving everyone but end up dead I would get some credit for all of those saves but it doesn't mean that much seeing how I died
Plus it would be hard to implement something like that into the game (especially with how the game is coded)
0 -
It doesn't make the game any more boring than what it is now, people have no reason to farm MMR, they can't even see it for reason.
Intention of this system is to separate sweaty "boring" players to less-sweaty casual players, which does work nicely even with kills/escapes as long as there is enough players.
0 -
Off the top of my head TrueSkill was used in the x360 release of BlazBlue and IIRC Street Fighter 4 both had Trueskill sections of their leaderboard. Saint's Row has TrueSkill ranking and its not remotely a competitive game for what its worth. TrueSkill is patented by Microsoft and used in a lot of the games that came out on x360. TrueSkill2 was released around the same time frame as Gears of War 4 and is baked into Halo 5 as well.
To the chess ELO statement, I was wrong, I mistook a proposal for an activity bonus being added to score as a rule and made an assumption that was clearly incorrect.
1 -
I tried to find your claims online about SFIV and BlazBlue developers using Trueskill and then deciding to add other variable metrics such as "offensiveness, defensiveness, execution of combinations" (your words), but I haven't had any success - please cite your sources. It's absolutely counterintuitive to me, given that like I stated earlier, creators of TrueSkill avoided using metrics like that on purpose.
"It is interesting to note that Elo’s update equation depends only on the win/loss outcome. Similarly, TrueSkill’s update equations take into account only the finishing order of the players/teams involved. None of these systems incorporates the actual final score, say, the number of kills in a shooting game or the finishing time in a racing game. This is a deliberate choice.
First, taking into account only the finishing order makes these rating systems universal because in almost any game a finishing order can be computed from the detailed game outcomes.
Second, it is crucial that the purpose of the game and the behavior of the rating system be aligned: people striving for high ratings should be forced to play in accordance with the spirit of the game. Taking the margin into account by which a game was won can be very misleading. As an example, consider a player who is trying to catch up with the opponent before the time is up, thereby taking into account the risk of falling further behind due to counter chances of the opponent. Should the rating system be designed to discourage this fighting spirit?
For the updates of the individual players’ scores in team games it is crucial to consider the finishing order of the teams only. If individual performance metrics are used in the rating system players will aim at maximizing those metrics instead of teaming up with their team members and try to win the game. For example, if you choose to reward the number of flags scored in a capture the flag game, do not be surprised if you find everyone rushing for the flag at once, but also some dead bodies who died with the flag in their hands betrayed by an overly ambitious team mate."
Sirlin, most famous for his thoughts on game designs and balance, also expressed his thoughts about TrueSkill in this manner when addressing Overwatch's matchmaking system years ago:
https://www.sirlin.net/posts/tag/Overwatch
TrueSkill intentionally and explicitly does NOT use any individual performance metrics. Their argument is that no matter what game you're talking about and no matter what metrics you measure to determine how well a given player did, it's necessarily imperfect compared to using only win/loss. The point of trying to guess if a player did well or not is how much they contributed win/loss, but the win/loss stat is the most accurate measure, they say. You'd introducing error by adding ANY other metric.
In addition to introducing error, you're warping incentives. For example, if you measure "damage done" as one metric, then it means players will attempt to maximize "winning AND damage done" rather than just "winning," which is not great. You can also very easily accidentally do a lot worse: you might accidentally give incentive not to play support heroes in a game where you really need support heroes on your team. (It seems this is already true in Overwatch.)
In many cases, it's almost hopeless to even devise a metric. If a character's role has to do with healing, you can't actually use how much they healed as a measure of much. If you did, it would penalize a healer on a team that played so well they didn't need as much healing. Or even worse is a character like Mei. Her ice walls can do a lot, her slow and freeze effects can do a lot too. But to actually quantify that into a metric correlated to win-rate? That's a huge error effect waiting to happen. My friend suggested the best metric how effective you were with her is to monitor the opponent's chat to detect how much they are cursing about Mei.
Yet another issue is that it's easy to accidentally create competition within a team for no real reason. For example, if number of kills is a metric that affects your rating, then your teammate killing an enemy that you could have killed essentially "stole" ranking points from you. That's clearly a bad dynamic.
I think Microsoft TrueSkill's reasoning makes sense here. It's a good case against ever using any individual performance metric when adjusting ranking points after a win or a loss.
2 -
I don't have sources to cite. I was working off I could A) Remember from my past engagement with the games in question, B) Believed based on conversations had with others, and C) What I've read over the years.
1 -
Just like how others on this forum could have read your original post about chess and (possibly) some fighting games using TrueSkill with random additional variables? And then they'll continue to spread these false information on the internet, convinced that they're true, just like you were?
If you aren't sure of something, please fact-check first.
0 -
I was sure, I was also wrong. I am only a human doing their level best. I fact check when I can and other times I put my trust in the words of others.
1 -
As humans doing our very best, let's strive together to learn from our mistakes and not repeat them. None of us are perfect, but we can always try to be better. I do appreciate you admitting that you were wrong about that subject - I don't think most people on this forum has the decency to do that.
1 -
Takes too energy trying to fight against the fact that at some point I'll be wrong. Learning I'm wrong means I can have the right answers later assuming I remember things.
2 -
"playing super altruistic and trying to get as many people out as possible inherently increases your chances of dying, but it's typically the most fun way to play"
'playing altruistic' having no mean in this game unless you rescued someone with bt at egc and die because of it,
even never rescueing first hooked person yourself could be most altruistic play for your team
some people pretend to rescue from 60meters away from hook because they just get bore to gen
0 -
Matchmaking based on time playing the game would be awful. I play with a group that plays for fun and doesn't really try to go and get better at the game. They just like getting chased, getting scared, and having funny things happen. This is a very common way that people play this game. Why would this be a good metric to send us against harder killers? Experience can be correlated with a person's skill, but it also doesn't mean a person is skilled because experience has to be paired with an active desire to learn and improve. If you don't care to improve but just like playing the game at your current level, why should you be penalized? Seems like a good way to cause players to want to leave the game.
The usage of "player play time is more indicative of skill" came around because being rank 1 was considered such a joke that DBD veterans had to find some other way to rank themselves and play time was just about the only metric they could easily see. It's not actually a good metric for player skill. It actively discourages people from playing the game more.
0 -
I'm pretty sure I mentioned the time gate in relation to new players being able to play against each other rather than the current setup where new players can end up playing against someone with hundreds or thousands of hours of play time. Even at your most casual level of play you're still going to learn the game to decent degree. The concept of gated matchmaking isn't that uncommon and at least the time metric deals with the biggest complaint I've personally seen in regards to matchmaking.
People with zero to ten hours of gameplay being matched up with people with hundreds to thousands of hours of gameplay.
0 -
I agree with pretty much all of this... except when you get to this point:
The more data points you use in your MMR system up to a point gives a better understanding of how each player approaches the game and better indicates their level of skill.
You referred to the end result as "their level of skill", which indicates it's a single parameter (or in some cases, two, like in TrueSkill which used a delta and mu to indicate skill level and confidence of that level, i.e. it indicated your skill was within a particular range rather than assigning a single number).
Once you've reduced the "skill" down to a single number (or even a range), you've literally lost all the data as to how you got there. That single number doesn't include any of the pip data. There's no way to extract that data from your number.
Which then begs the question... why bother collecting and using that data in the first instance when, literally, it just makes the maths far more complicated and its lost/discarded in the end anyway? Particularly, when there's an easier proxy to achieve the same result, i.e. kills/escapes, which, when you look at the figures overall (which the devs can do, and have done), there's a close parallel... at least close enough that the overall kill/escape ratio sits at approx 50/50 which is what they were targetting.
You could argue the extra data will be fine-tuning, and this would be agreeable. But there's a point of diminishing returns, and it wouldn't change anything to any effect such that you'd see a difference. The bigger problem is the holes in the MMR process (rather than the metrics) which is the back-filling, etc. Resolving that would likely make the MMR better and have a bigger effect than implementing PIP data (once you've worked out how to attribute those pips accordingly)
0
