How the MMR works, and why what Patrick said makes absolute sense (from a dev POV)
Comments
-
Mmm, it would have avoided the arguments that "it doesn't reflect skill", but it wouldn't have done anything about how the current setup encourages selfish behaviours, or the backfill/bracket widening ones.
2 -
Yeah, that was the mistake they made. If they had simply said something like: "We're going to make a system based on escapes and deaths, with the belief that if you die more often you'll be put into matches with similar survivors, against killers with lower kills, and thereby hopefully improve your chance of survival" then that would be fine. As it happens, it feels like they're holding on to the word "skill" to justify a system it doesn't completely work with.
A win/loss MMR feels simpler to me.
1 -
No, and yet the argument wouldn't be as multifaceted neither, and a bigger focus on what needs to be looked at while not be swamped over by others. It's too mashed-up presently.
A clear admittance of what the system is will allow people to move on. As it happens, it's just making the crux more confusing.
Although, regarding the selfish behaviour part of your post, fortunately I'm seeing less of it and that may be because whether people go up or down the MMR, or for the simple fact nobody knows where they are, I feel people are starting to see this bizaar need to claim high MMR for what it is, which is it really doesn't matter because of how strange the matchmaking is anyway.
0 -
But... whos Patrick ? 😎🙄
0 -
The fact of the matter is you are not discussing with me the premise and actual standpoint that I have. You are simply using me as a medium for some imaginary idea of my opinion. You clearly don't want to have a discussion and an exchange of perception. You act as if I stated that the system didn't work at all while this entire time I have stated this would be an improvement. For someone that puts down so many words to say very little, you really have difficulty understanding those of others.
I think you are the one not actually discussing with me nor wanting or being able to hold an actual discussion, you still don't (want to) get that my argument wasn't about the change being an improvement, but the reasons it would be an improvement for, and that your thoughts on why it would be an improvement were based on misunderstandings of the system. It ended being a "discussion" because all I did was keep on trying to explain to you how you were making mistakes in your thinking regarding the system (and game) - the same ones I highlighted in my original reply already - and you kept saying the same things without ever actually addressing my points or substantiating yours.
I never said you made it out as if the system didn't work at all, I explicitly (with actual quotes of what you said) responded to the ways in which you made the system out to not work properly, the specific flaws it would have, highlighting that those weren't actually the flaws you mistakenly thought they were.
You already admitted that it would improve, but clearly you don't know how that is reflected in the numbers based on your response. You claim it would not do what I stated, just shows you never seen the results of such systems.
I don't know what you mean with never seeing the results of such a system, but I invite you again to show with actual maths (of a functional MMR system for DbD with values of your choosing) how the volatility index function relates to the personal and group survival metric in its effect on the mathematics of the system.
I didn't "admit" it would improve with the group survival metric as if I had to be convinced of it, I argued it would be an improvement even before ever responding to you; the reason I responded to you with regards to that was, again, that you were wrong about why it would be an improvement, claiming it would solve issues with the current system that aren't actually issues.
The fact that you also ask me to show that the current system will not create (a more volatile yet) similar result just shows that you are not engaging with me in discussion nor understand what I am speaking of.
This is more of that "gaslighting" stuff, still just as unsuccessful because the things you were "speaking of" are literally quoted in my responses to those things, and they were absolutely not about the systems creating the same end results.
I don't falsy assume that teamplay by default will lower your survival ratings, I stated it can be one of the effects of it. Once again twisting what I am stating to fit your false narrative. Some team plays and risks taken can get you killed while simulationsly providing an on par or better result for the team. The system should not punish people for playing in that manner if that is the result. It should use metrics that account for a more versatile playstyle if it is a positive for the side that they are on.
If you assume teamplayers would not "climb" in rating (which you did, it's there in your posts and in my quotes of those posts), you are stating it has to have the effect of lowering your survival rates, because with regards to the system that's literally the only effect that's relevant. And if it doesn't consistently do so, over many matches, the instances where it can have such an effect are irrelevant, because singular matches' effect on the rating are negligible by design, the rating only changes in non-irrelevant fashion over many matches.
The system does not "punish" people, you keep saying this word without ever answering what the "punishment" would be. A number change is not an actual punishment in and of itself. And the effect that number has is not either, on the contrary. Why would a player that actually dies more than half the time in their matches not want to be put into matches where they can start surviving more often?
And even if you irrationally care about the rating and consider rating changes "rewards" or "punishments", there are just as many ways for things to afftect the ratings that don't intuitively make sense to us in a group survival metric environment as there are in the personal survival metric one, such as literal anti-teamplayers (e. g. people that kill themselves) or players that contribute next to nothing to the "team" (e. g. in cases where they quickly die early on, or just hide all game and die late) being "rewarded" with rating for their ""teammates'"" survival.
Running a killer for 5 gens is teamplay, they fulfill the role of distraction while others do the gen. If the other people dont work on gens, then it won't be a 5 gen chase. The idea that teamplay still means individual actions are taking place, each at their own skill level and role. That is the whole idea of a team.
If running a killer is a "teamplay", your definition of it and indeed reason for wanting to distinguish between the metrics should reveal itself to you as absurd. A survivor running from the killer is literally running for their own life, it's not like a player that doesn't play for the "team" but only their own survival would... choose to stop running from the killer and go down.
Everything that contributes to your own survival would be a "teamplay" from that perspective, and that is a crucial realization for you to come to: If you realize that since all survivors have the same conditions that have to be met in order for them to be able to personally survive (in the context of the system, so disregarding hatch) their fight for their own personal survival is also implicitly one for everyone else's, you will come closer to understanding why personal survival correlates with group survival and vice-versa, and why distinguishing between the metrics does not lead to different end results in this regard. And the same is true for actual altruistic teamplay (rescuing others in various ways, directly helping others with heals, etc.), or so I argue, because helping your team survive is helping yourself survive, since the team helps you survive, implicitly (if they survive longer they contribute to meeting the conditions necessary for your survival even if they merely play for their own survival), and also explicitly (fellow survivors that are alive are also able to altruistically help you, rescue you, etc., and they are also more likely to do so if you helped them).
And of course, the thing is that even if there were a difference in the ratings players end up at with the different metrics, there is the next step in understanding the system, and realizing that since its only purpose is making matches happen where people can on average survive around half the time regardless of their playstyle, distinguishing between the metrics is still not relevant in this regard.
I have shown that objectively a survivor and the killer the 2 opposing forces at the end of a match can both have a winning result. It is a 4v1 game, not a 1v1. No matter how you want to twist it, the survivors are on the same team.
And it doesn't matter how much you keep stating this, it is still a failure to understand the system. The problem with MMR systems where both "sides" can win at the same time is that all players involved in a match can gain rating, which will lead to more and more players ending up at the same ratings, defeating the purpose of a matchmaking system based on such ratings. In the 1v1 math that the system works with, there can be no instance of both sides of the equation winning/gaining rating; in a 1-escape match, the 3 survivors that did not survive lose rating, their "side" doesn't gain rating. Once you understand that the system trends toward a 50% survival chance for individual players, and that a 50% survival chance of every individual player on a group of 4 players translates mathematically to a 50% survival chance of the group, you will come closer to seeing why the system even when looking at 1v1s does over many matches trend toward average 50% survival rates, per-player and per-group (2 survivors surviving, 2 dying), and that 1-escape and 1-kill matches pose no functional problem for the system whatsoever.
Short cuts means systems are not perfect. The question is whether the flaws of it are better or worse with the improvements suggested. You already admitted it would be quicker, faster and by result more accurate.
MMR systems are about winning vs. losing, in players competing to achieve diametrically opposed win conditions against each other - they are (and indeed have to be) by their very definition about the "shortcuts", insofar wins (whatever the system defines as a "win") are seen as "shortcuts", because they are about winning. People think the MMR system has to be about "skill" and matching people of similar "skill" and wins are merely being used as a "shortcut" to that end, a proxy for "skill", but that is the complete backwards understanding of the system: the system is about winning, wins are the ends in themselves, the only purpose it serves is instating a matchmaking where people have more equal chances/rates of winning in their matches on average, they are matched based on their wins. That wins indicate (are "shortcuts" or proxies for) "skill" (in this context strictly meaning: one's ability to win consistently) is merely a correlating side-effect, that can be used to make the neat assumption/statement that players matched based on their chances to win are also matched based on their "skill", but this is merely a collateral to the actual reason why players are matched based on wins, which is just wanting players to have more equal chances to suceed and more even win/loss experiences, since that is as a basic premise deemed desirable. As I've said, even if you would have to pair killers and survivors of different "skills" (or ratings) in order to be able to achieve 50% average survival/kill rates, that is what the system would want to do (although this does not seem to be the case according to BHVR, and if it were, the ideal solution would of course be to change the actual gameplay mechanics to achieve a more balanced state, not the rating system).
One "problem" could be that different killer characters are not able to reach rating levels that other killer characters are, or that their average rating levels are very disparate, and this should definitely be looked at if it is the case (and should have always been, since there are clear, stark differences in strength between the different killer characters). It's not a functional problem of the system because the players playing the killer characters will simply be matched with the respectively appropriately-rated opponents, but it is a problem in the result, which is that survivor players beyond a certain rating are much more likely to see certain killer characters in their matches. But yeah, the different killer strengths have always been something that should be brought into line more, closer to an average between them, because that way the game overall can be better balanced according to the strength of all killer characters (and the same is true to a lesser extent for perks, items, add-ons, maps, but at that point you get into game design considerations that people actually enjoy disparate gameplay experiences and don't want everything to be closer to an "average", even if that objectively makes it easier to achieve "balance").
For someone that likes to use so many words, you really have difficulty reading those of others. I have stated from the beginning that the system overly favours safe plays and doesn't account for riskier ones even if it results the team a net positive. That is why if you want to not expand the criteria it would be better to at least view it from a team perspective. Nothing you have stated has altered or informed me of anything I didn't already knew.
I mean, I have informed you of various things you objectively did not know about, such as whether the current system actually leads to more balanced overall average outcomes, and whether the devs are actually planning to change the metric to include group survival.
But regardless, you are still repeating the very same errors I have been responding to and trying to explain from the very beginning. You are the one with the difficulty reading and understanding my words (which, fair enough, it's a lot to read, but it could have ended with the very first few posts already). The system does not "favour" any plays, and it already accounts for all plays - a higher rating doesn't mean the system thinks of those players as "better", it is not intrinsically "favourable" to be at a higher rating, it just means the players there factually survive more often against other higher-rated players; and it already accounts for everything with regards to the kill/survival win condition, because everything is accounted for insofar it actually affects the rate at which players achieve or fail to achieve the win condition. For instance, it "accounts" for self-sacrifical teamplay by lowering the rating of players employing such play if they actually consistently die more often due to it, and this is not somehow not "favouring" them, since all it does is put them into matches where they can in theory and on average survive more often despite their playstyle.
Post edited by zarr on0 -
This is going to tank the quality of matches within the SBMM bracket.
It has already made some killers feel worst and it have destroyed my SoloQ experience in an effort to be inclusive with newer players.
They probably chose this method so they won't have to be bothered in balancing the SWF Meta. (cause statistics)
either way the game will soon find out its losing its playerbase as there isn't really a choice to play comfortably in this chaotic rankings.
You're either the one carrying or you're the one having to be carried, but 2 good players aren't enough to make up for the mistake of 2 bad players.
you also forgot to address the part where players farm other players for BP by hooking and slugging them multiple times but not killing them,
it's makes newbies feel miserable when they get toyed and slugged for a long time and the killer aren't facing any resistances. (it was fun as the killer)
1 -
I know the emblem system isn't perfect, but I thought it was a decent way to grade "participation" in the match.
Can't loop and die all the time? If you contributed enough to the match (healing, gens, totems, etc.) you can still rank up.
Got 8 hooks, no kills? Still considered doing an excellent job, especially if you got into lots of chases.
Tying it ONLY to escape/kills just feels like a step back to me. A bubba who facecamps the first survivor to death first down, the other 3 finish gens, and he facecamps a second survivor to death with NOED, making it a 2 hook, 2 kill game shouldn't be seen as more skillful compared to a killer who gets 8 hooks, even though nobody died. (Although I still think deaths should contribute extra compared to just hooks. Like, each death should count as an additional 2 hooks.)
Tying it to only escapes for survivors means it's all-or-nothing. Played as your typical Tapp main, banged out 3 and a half gens, got 2 hook rescued, healed survivors, did some totems, etc. and died at the end, even though your actions enabled the other survivors to escape? You're considered worse than the Blendette who literally did nothing but hide in bushes the entire game. (At least hatch escapes provide no MMR bonus, but I think they still count it as a win if the killer shuts the hatch and you escape out a door.)
1 -
My concerns have never been about whether the current system can work from a mathematical standpoint and I have stated it many times. The fact is you don't even understand the mathematical implications of the improvement you yourself speak. The fact that you state that there would be an improvement, means that the mathematical model will improve as well.
Here have a visual representation (horribly drawn in paint) of what I am speaking of, the goal is to achieve a 2 kill/escape ration for a match (this is how it looks like over many many games). Meaning that over the long run they should approach the 2 kill/escape line. The current system views this from a 0 to 1 line one per survivor, that evens out at a 0.5 ratio and applies those 4 charts in the match to the killer to from their outcome to be the sum of them all and creating the above graph in essence. While that might mathematically create a similar table to represent the whole match, it doesn't do this for the survivors. If the metrics are improved as we both claim, the line moves as displayed in the graph above as the green line regardless of which way it is approached, as the goal of the system itself doesn't change.
If an individual improves and starts over/under performing compared to those lines, their MMR will drop or rise accordingly to try and match that once again, so the quicker and more efficient it is as detecting that the better and more accurate your rating will be determined. The lines do not hit the 2 mark, as people consistently fluctuate between the upper and lower boundaries (volatility) as they keep winning and losing, a range based on the accuracy of the system that it deems acceptable.
You love throwing the math at me, but I understand how it functions. You did not explain any of this, you did not adjust my perception on how the system functions. All you did was tell me that the devs are seeking to adjust the metrics to improve the system, while considering a more hybrid approach to make the team and sum of the results play more of a role.
You claim that my perspective is wrong, but from a game design perspective all that I have stated is simply true and please feel free to illustrate exactly would be wrong with the above graph. I look beyond the system in a vacuum and its mathematical model, clearly something you are incapable of and don't even try to attempt. Therefore we are having very different conversations, as I am not disputing the fact how the MMR system functions from a mathematical standpoint. The current system punishes risk plays that might result in the overall result improving or staying equal and it only favours safety ones from a personal perspective for survivors. This is what I consider a dire flaw with it and this would be addressed by viewing it from a team perspective, the sum of the total.
From a math point of view it is simply this:
If you escape you get +1 a win, if you die you hit are a 0 and lost.
- If you make a risky play for the teams overall outcome and are successful, you personally were already set to escape so it remains a +1 and all that changes is someone else however also now achieves a +1. Yet as this is a random stranger, why would anyone care under the current one?
- If you make a risky play for the teams overall outcome and make a 1 for 1 change, now you personally are now set to 0 instead of +1, while someone else is set to +1. While this does not affect the overall numbers of the graphs, it does affect yours negatively.
- If you make a risky play and fail, now the teams overall outcome goes down, your +1 changed to 0 and that of the other remained a 0.
This is how it functions in the current system. There is no incentive to put your own neck on the line to improve the teams overall outcome. It even incentivizes you to ditch anyone that does in any situation where you can see a line of play where you personally can escape. Nothing you have stated has addressed this concern, just tried to excuse it as irrelevant. While that is how the system promotes and favours certain playstyles. They mathematically achieve the same goal, but what it promotes are very different achievements and actions.
Naturally as a result if one person is taking a risk, means that everyone on that side is affected and should be invested in supporting and making it a success or at least a neutral outcome. You might consider this undesirable, but my argument is that it is better than the current system that promotes no risky play to be done instead or concern for the survival of others. It is a game after all where you are supposed to be 4 against 1, wouldn't it make sense of it to promote the 4 to work together to make their side win - at the moment it just goes for personal victories and that is the main complaint? It is about winning and losing, but the detail lies in what is to be considered a win or a loss! If both the killer and survivor can win in the same match, why would a survivor need to care about their side winning - why take any risks, a 1 escape is all you need to win?
The system as it is now in promotes to not work together as long as you can escape. If someone is taking a risk for the sake of someone else, it promotes just ditching them and lowering their chances to succeed or trade evenly for your own gain. Which in turn reflects positively on the killer and make them overperform more often and raising in rank, they get an extra kill making their losses less negative, they get the risk takers more often to die and boosting them positively. Regardless of the path taken it can cause friction within the group of survivors, there is no set path to take and yet the system removes the higher risk/reward scenario for the people that would need to take the risk to begin with. The guy on the hook isn't the one making the choice after all, it is the others that are. They can still secure the X escape by running out, but they could also consider the +1 escape at the risk of a bigger loss. It is a game, it should reward those that take the risk and succeed, not just even out at best.
If you view the data that was mined, it showed that killers raise significantly faster in rank to the soft cap than the survivors did, this might contribute to that? If killers disproportionally are raising their MMR compared to survivors, as they are far more likely to benefit of the selfish behavior the survivor metrics promote. This results in matching similar MMR to create a more balanced match becomes more difficult and results in more uneven match ups to avoid extreme queue times, the longer you wait the wider the search becomes speed over quality. There are significantly more survivors complaining about long queue times than for killers, which means that it widens and they are more likely to be paired with a killer at higher ranking then themselves as there are more high ranked killers awaiting a match themselves - since there aren't enough survivors up there with them.
Now, where is my wrong perception here? People are playing a game, many people will take unnecessary risks because as you stated, it gets the adrenaline pumping, it is fun, it is a game... yet the system doesn't promote or reward this at all for the individuals and I would argue at the detriment of its own quality. You keep telling me to prove my point with numbers and the math behind it, well here you go.
It is about more than just does the system make it a 50% chance to escape or not, it is about the ripple effects it has. Like if this doesn't make you understand what the hell I am talking about... you are just a hopeless case.
Post edited by Kalinikta on2 -
Talk about brick walls.
No, it doesn't make me understand, because it doesn't make much of any sense, due to the reasons I've been pointing out all along. Let's start once again...
The fact that you state that there would be an improvement, means that the mathematical model will improve as well.
I never said the mathematics of the system would not improve. I explicitly highlighted the way in which I assume they would improve - the equations would more quickly lead to the ratings where survivors actually settle around 50% escape rates. Less matches to arrive at the same ratings. Of course this is an improvement of the mathematical model, but our disagreements pertained stuff like you saying people would end up at different ratings with the different metrics altogether, that the current metric would lead to problems of "both sides winning", and other things, that I quoted verbatim whenever I responded to them.
While that might mathematically create a similar table to represent the whole match, it doesn't do this for the survivors. If the metrics are improved as we both claim, the line moves as displayed in the graph above as the green line regardless of which way it is approached, as the goal of the system itself doesn't change.
This is bogus. It doesn't matter whether you look at a match from the "perspective" of the killer or the survivors, if each of the survivors in a match has a 50% chance of surviving, on average 50% of them will survive. As long as the metrics arrive at that 50%, there is literally no difference in the achieved end result of the graphs you look at over time, no matter whether you looked at individual survivors with a 50% chance to escape themselves, or at whether 50% of the survivors overall escaped in their matches. I argue (and have argued from the very beginning) that a graph would more quickly arrive at those same ratings that yield a 50% chance per-survivor (automatically also meaning per-group/match), have a steeper incline. But your argument went the way of saying such graphs would look different between the metrics altogether, lead to different players settling at different ratings, lead to issues of "both sides winning" and as such more and more players ending up at high ratings, which would skew the results more, and more stuff.
If an individual improves and starts over/under performing compared to those lines, their MMR will drop or rise accordingly to try and match that once again, so the quicker and more efficient it is as detecting that the better and more accurate your rating will be determined. The lines do not hit the 2 mark, as people consistently fluctuate between the upper and lower boundaries (volatility) as they keep winning and losing, a range based on the accuracy of the system that it deems acceptable.
Yes, indeed, it will be quicker, there will be less matches needed to arrive at the appropriate ratings (the same ratings we arrive at with the current metric), less potentially uneven, fluctuating match results to get there... which is what I've been arguing from the get-go. That is not what I contested at any point. I contested things such as you claiming different people would end up at different ratings altogether.
If you make a risky play for the teams overall outcome and are successful, you personally were already set to escape so it remains a +1 and all that changes is someone else however also now achieves a +1. Yet as this is a random stranger, why would anyone care under the current one?
Why would anyone care about the "+1" is the question you should instead be asking yourself in order to realize the flaws in your thinking. The only reasons for it would be irrationally caring about one's invisible rating, whereas the reasons to go back into the map to help someone else in a scenario where you could otherwise just escape are plenty and more rational, such as seeking gameplay, thrill, the feeling of success if you manage to rescue them, the feeling of solidarity, and on.
If you make a risky play for the teams overall outcome and make a 1 for 1 change, now you personally are now set to 0 instead of +1, while someone else is set to +1. While this does not affect the overall numbers of the graphs, it does affect yours negatively.
But you still have no reason to care about your "graph", there is no negative effect no matter how it behaves if it actually behaves according to your survival over many matches, if your actual chosen playstyle actually consistently leads to those graph changes. If you personally actually die due to sacrificing yourself for others and do so so consistently that your graph is affected to trend more toward 0 than it otherwise would, you want the system to reflect this with a lower rating because why would you not want to be matched such that you can survive more often rather than constantly keep dying at the ratings where you have shown you will keep dying with your playstyle? Even disregarding that in matches with a 50% overall survival chance any survivor is likely to escape around 50% of the time on average and that no one player is likely to consistently more often be among the two that die, even if that were the case, the system seeks more even personal success chances for individual players, and this is desirable for pretty much everyone, because most of everyone wants to succeed and not constantly die, not least because players in this game most of the time aren't in actual "teams" but queue up individually. And for the few people that actually prefer to die in order for their teammates to survive, and don't mind constantly dying, falling in rating to a point where they can survive while still helping their teammates to survive in the same ways is still not a negative effect, unless they are a "masochist" and for some reason want to actually constantly die helping their teammates rather than only die/survive half the time helping their teammates. At that point we'd basically be arguing the system insofar it "encourages" playstyles due to people caring irrationally much about ratings, "hurts" people that have yet other psychological oddities on top of this, such as outright wanting to die. And even on multiple levels of absurdity of this, it doesn't really matter, as I'm sure such players would still find reckless risks to take for their teammates that lead to their demise even at the lowest levels (hell, BOT Trapper from the tutorial can kill you if you take silly risks).
If you make a risky play and fail, now the teams overall outcome goes down, your +1 changed to 0 and that of the other remained a 0.
And this is how even with the group survival metric, even if you irrationally care about rating changes, risky altruistic plays are not encouraged. In any case where you die for the survival of 1 other survivor in your stead, the risk was actually not worth taking from the group survival perspective, because you risked 25% of the team (yourself) for a net-neutral return, in what could have also ended in a net-negative return. In these scenarios the group survival metric would also regularly dictate simply leaving because the risk is not worth it (if you irrationally care about rating changes).
This is how it functions in the current system. There is no incentive to put your own neck on the line to improve the teams overall outcome. It even incentivizes you to ditch anyone that does in any situation where you can see a line of play where you personally can escape. Nothing you have stated has addressed this concern, just tried to excuse it as irrelevant. While that is how the system promotes and favours certain playstyles. They mathematically achieve the same goal, but what it promotes are very different achievements and actions.
First of all, you have to manufacture very specific scenarios where you can actually see a "line of play" that leads to your own survival without helping other survivors. Such as when you are literally in the exit gates and would have to decide to go back into the map for someone else. In most of any other case, you helping fellow survivors increases your personal chances of survival. And again, in scenarios where you can actually secure your own survival without risking anything for the survival of others, the group metric would still regularly "encourage" (if you irrationally care about rating changes) simply leaving, because that is the most secure play from the group rating perspective too. Only in scenarios where you can very safely rescue at least 1 other survivor with it actually not leading to you yourself dying is it at all "worth it" to go for such plays from the "I care about rating" group metric perspective, and I guarantee that people do not refrain from making such plays with the current metric either, even if they know about the metric and care about their rating.
On the other hand, you can manufacture specific scenarios where people that like to go for risky plays are discouraged from doing so from the group metric "I care about rating" perspective. Say someone is on a hook and 3 other survivors are in an exit gate. Now one of those 3 players might be inclined to go and try to rescue that survivor... but they've already won, all they need to do is leave. Them going back into the map is risking their rating increase, regularly at best leading to the same outcome, while at worst leading to a decreased rating for everyone due to them dying. These risks are simply not "worth it" from the group metric perspective either, that's why you see actual teams in tournaments leave the last survivor to die 99.9% of the time. If people go for rescue attempts in these scenarios, they have to do so in basically throwing away rational risk calculations to begin with, in either metric and without caring about ratings at all, and do so for the gameplay, the fun, the thrill, etc.
Naturally as a result if one person is taking a risk, means that everyone on that side is affected and should be invested in supporting and making it a success or at least a neutral outcome. You might consider this undesirable, but my argument is that it is better than the current system that promotes no risky play to be done instead or concern for the survival of others. It is a game after all where you are supposed to be 4 against 1, wouldn't it make sense of it to promote the 4 to work together to make their side win - at the moment it just goes for personal victories and that is the main complaint? It is about winning and losing, but the detail lies in what is to be considered a win or a loss!
You might consider it desirable for survivors to be concerned about the survival of others just as much as their own, or even more, but you do not have to consider it as such. I have repeatedly pointed out that even if it were true that the people that irrationally care about their rating play in more selfish ways than they would otherwise, that is not objectively a flaw of the system from a game design perspective either, because the game is not strictly a team game and may very well be seen as a solo survival game.
Regardless, I can only repeat that playing for one's own survival automatically leads to playing for that of others, and that playing for the survival of others automatically leads to playing for that of oneself. Sure there are exceptions to this, just like there are exceptions where selfish play can increase group survival and altruistic play can decrease group survival, but overall, the correlations here are such that there is no difference in distinguishing between personal and group survival. Absolutely not in the results of the functioning of the system to achieve its intended goal because 50% per-survivor is 50% per-4-survivors, but also not by the nature of the game or indeed the "I irrationally care about ratings" perspective, because not only is teamplay (and at times risky teamplay) that increases group survival necessary for one's own survival, but for every case where it is not, there are cases where avoiding teamplay (and especially risky teamplay) is necessary to increase group survival.
Even if you irrationally care about rating, and even if you accept the argument that playing for the team is desirable and how the game should be played, there are only exceptions where the personal metric would actually "encourage" different plays from the group metric, and these exceptions exist in both directions, as there are instances where the group metric would "discourage" teamplays that come at a risk to oneself.
The system as it is now in promotes to not work together as long as you can escape. If someone is taking a risk for the sake of someone else, it promotes just ditching them and lowering their chances to succeed or trade evenly for your own gain. Which in turn reflects positively on the killer and make them overperform more often and raising in rank, they get an extra kill making their losses less negative, they get the risk takers more often to die and boosting them positively. Regardless of the path taken it can cause friction within the group of survivors, there is no set path to take and yet the system removes the higher risk/reward scenario for the people that would need to take the risk to begin with. The guy on the hook isn't the one making the choice after all, it is the others that are. They can still secure the X escape by running out, but they could also consider the +1 escape at the risk of a bigger loss. It is a game, it should reward those that take the risk and succeed, not just even out at best.
First, your assumption that this would overall make killers overperform is again based on not understanding that the system works out to instate 50% survival rates in both cases, and that it doesn't matter whether you look at individual survivors or groups of survivors for this, as it is the same average outcome of 2 kills/2 escapes. It does not "positively reflect on the killer".
Secondly, even if we do say teamplay is desirable (which is not objectively the case, players and designers may well prefer more "selfish" approaches to the game) and assume people irrationally care a lot about their ratings, the group metric does not regularly "encourage" risky play any more than the personal metric, and there are instances where it outright discourages it, especially once we consider that people now can also be affected by irrationally caring for their teammates' ratings, which risk-taking behaviour can put in peril since one's own death now directly negatively affects everyone's ratings.
If anything I could argue that if the group metric would actually "encourage" risky altruism more often where the personal survival metric does not, that would actually play into the killer's hand because that gives them way more opportunities to get more kills than is otherwise the case. I could argue risky teamplays (and particularly in the endgame where players can see an actual "selfish" line of play to their personal survival) are way more likely to result in decreased overall survival than they are to result in increased overall survival. (This would of course still not matter for the functionality of the system since it trends toward 50% overall in the matchups it creates in either case, with simply the people taking the right risks and avoiding the wrong ones climbing ratings either way... but even if you do go into "irrational" arguments of the system encouraging certain playstyles in players that otherwise would play differently, it is not as clear-cut as you assume it is.)
If you view the data that was mined, it showed that killers raise significantly faster in rank to the soft cap than the survivors did, this might contribute to that? If killers disproportionally are raising their MMR compared to survivors, as they are far more likely to benefit of the selfish behavior the survivor metrics promote. This results in matching similar MMR to create a more balanced match becomes more difficult and results in more uneven match ups to avoid extreme queue times, the longer you wait the wider the search becomes speed over quality. There are significantly more survivors complaining about long queue times than for killers, which means that it widens and they are more likely to be paired with a killer at higher ranking then themselves as there are more high ranked killers awaiting a match themselves - since there aren't enough survivors up there with them.
The data was mined at the very beginning of the system's implementation, which means the numbers were a result of the old matchmaking system where killers were matched against survivors based on rank. It is obvious why there were disproportionally more killers at high MMR at that point because we know the ranking system led to high kill rates on average, and especially in red ranks.
Both the assumption that this is still the case and the queue time assumption are merely that, assumptions. We have no idea what the distribution now is, what queue times globally are actually like, and even if we did, that could indicate issues with a ton of things rather than the system, the system would most likely just be highlighting certain issues, which it would in either case no matter which metric.
And again, functionally the system would trend toward 50% kill/survival rates regardless of which of the two metrics is used, which is its purpose, and so the metric switch would definitely not help alleviate any issues that may lead to killers rising in rating faster, or the like. If such problems actually do exist due to the system, they can only be found at other parts of it, the metric change would not change anything about it.
Now, where is my wrong perception here? People are playing a game, many people will take unnecessary risks because as you stated, it gets the adrenaline pumping, it is fun, it is a game... yet the system doesn't promote or reward this at all for the individuals and I would argue at the detriment of its own quality. You keep telling me to prove my point with numbers and the math behind it, well here you go.
Disregarding that you used no numbers and math to actually prove your points, your perception is wrong in the following regards:
- You assume the game is strictly a team game and people should be encouraged to play it as such. This is not objectively the case, and it is also not objectively the case that self-sacrifical teamplay makes for more desirable gameplay. Subjective opinions on this can and do differ.
- You assume people care that much about their ratings. I highly doubt this, because not only is it most likely only a portion of the playerbase that even knows about the rating and personal survival metric, but of those that do, they still have a ton of other, much more tangible things to care about that inform their plays. How often since MMR implementation have you seen a player come across an open hatch but actively refuse to jump into it and run to the gates instead? I basically never see that. This should be an indicator that people do not care about the rating much at all.
- You assume teamplay that leads to increased group survival chances/rates leads to the teamplayers themselves having reduced personal survival chances/rates, which is wrong because of the correlations between the two. Likewise, you assume the group survival metric would encourage risk-taking/self-sacrificing teamplay where the personal would not, but it can discourage such play just as much, and there are instances where it if anything discourages it even more so.
- You assume rating changes are a "punishment" or "reward", not understanding that getting a decreased rating is not a punishment and does not lead to ill effects. All it does is put players into matches where they stand a better chance to survive if they did not stand a 50% chance to do so on average at higher ratings.
- You assume the different metrics lead to different results, such as different players ending up in different matchups and killers rising in rating disproportional to survivors. This is wrong because regardless of which of the two metrics you apply, kill/survival rates average out to approaching 50%. It may do this faster with the group metric (I think it will), but it will not ultimately end up creating different matchups.
- You assume 1-escape and 1-kill matches stand in the way of the proper functioning of the system, by way of creating instances where "both sides win", not understanding that the mathematical logic of 1v1 creates no such instances where both sides of the equation gain rating, and leads to none of the issues that "both sides winning" would (more and more players gaining rating and all ending up at the same levels, or the "soft" cap).
- Many more issues in your perception, such as whether the system selects for "skill", what that "skill" is, whether the system has to select for "skill", misperceptions (or deliberate misrepresentations) regarding what arguments I am making...
It is about more than just does the system make it a 50% chance to escape or not, it is about the ripple effects it has. Like if this doesn't make you understand what the hell I am talking about... you are just a hopeless case.
The rating system itself literally only is about creating more even success chances/rates overall, approaching 50% for "everyone" on average. If it works to do that (which it does, notably more so than the old system, and despite matchmaker issues), it works, and the metric change would change nothing about this (other than probably make players approach those 50% more quickly, which is my assumption). The "ripple effect" it may have of "encouraging" more selfish play of people that irrationally care about their rating is not objectively a negative effect, and can even be seen as a positive effect. And regardless, it does not stand in the way of the proper functioning of the system, and teamplayers even if they would consistently die and fall in rating despite having higher group survival rates in their matches (which I'm convinced is not at all the case, those players consistently survive more often themselves too), suffer no ill consequences from their lower rating - the only thing it does is put them into matches where despite their playstyle, they have higher chances to stop dying that often, eventually arriving at a place where they too should be able to survive as often as they die playing in this way.
And I've explained all of this in various ways various times, and would likely have to keep doing so indefinitely if we keep going.
1 -
It's helpful to remember that the point of the matchmaking system isn't to determine who's more skilful- it was a mistake of them to use the word "skill" in the name, even if it is still accurate when you're looking at aggregate batches of MMR brackets instead of individual games.
What the matchmaking system asks of a player when the match ends isn't "how skilful were you", it's "should you go against stronger or weaker opponents". If you won that match, it's a point towards needing stronger opponents, and if you lost then it's a point away from needing stronger opponents. Amass enough points, you move up the bracket.
The facecamping Bubba should go against stronger opponents. They shouldn't be stomping noobs who can't deal with them, they should be going against teams that know what to do. Same for the immersed Claudette- she should be escalated to being the weak link against a good killer because the cheese strat should stop working eventually.
It's also worth mentioning that the MMR system should not account for problems with the game's balancing, it should very explicitly be built as though those problems don't exist so it's future proofed for when the problems are fixed.
2 -
I have now proven and showcased my point using a mathematical model in a more clear fashion, you don't care about the graph... but that is the math model you speak of. You ask why I state it as 1 or 0 for the survivors as that is the definition of a win or loss in a mathematical or computer science situation. You do not care about the graph, because you are disingenuous in actually caring about the math behind the system. You complain about not using numbers, while I did not run any simulations or numbers and simply showed the theoretical outcome of a system like this. I can state over 10 billion games played, but it doesn't really matter as they are numbers that I pull out of a hat. The hypocrisy is that you have not once used any math yourself to strengthen you argument on why it would be better, while even admitting that a group view would likely result in a faster result. Which I showcased within the math analysis of what that means.
Additionally even with my drawing it out for you and explaining it from a math model point of view, you still make inaccurate statements about what I am talking about. There is no use talking to you, as you clearly are not interested in listening or understanding what I am stating and just want to uphold some inaccurate imaginary idea of what you believe I believe... instead of speaking about what I actually believe and state. No matter how I try to clarify and rectify the assumptions you made about what I wrote, you just tell me that you know better what I believe than myself. You might not believe me, but I do think having a MMR system was a good decision for the developers to invest in. I just don't think they are using the best metrics to make the best game.
You claim why would someone care about their invisible rating, while I am not stating that they do. I state that people care about winning and the wins defined by the developers are as I stated. What is defined as a win or a loss influences how people play and acting like people don't want to win or shouldn't is so unrealistic. Therefore by proxy they do care about their rating, as it is defined by how often you win against how often you lose.
As you accurately point out if you are in the exit gate with 3, the safe choice would still be to leave and in many cases would still be done with the improvements I stated and I am not trying to deny people from making that choice. Their team can make the more risky move and go for the 4 man out, try to save them however a 3 escape would be considered a win and therefore the risk isn't really worth it. It could also be that they are with 2 in the exit gate with 1 person on the hook. Then the decision would be to go for the safe draw or risk it all for the win, maybe it is now worth the risk? If they manage to pull it off, the team that went back are the ones that should be rewarded, but as it currently stands that isn't the case.
A game design and the underlying systems should aim to reward those that take risks and succeed. If you cannot even comprehend that this results in a better overall game and system you simply don't want to understand my perspective. My goal is to provide feedback to have the game be the best version of itself, not to excuse it because it works if you ignore everything else and the ripple effect it has within the game.
These are not my assumptions, these are the assumptions you impose on me and while some are accurate just because they are objectively true not all of them are.
Assumptions 1: This is a team game.
This is a 4 versus 1 game according to the game description, therefore it is a team game and a solo game based on which side you play. Claiming the survivors are not on the same team is simply inaccurate. You don't have to be in a pre-made to be on the same team, the game allows you to join a team and the MMR system seeks one for you. Those on the same side are on the same team, they simply are even if they are made up of individual solo players. It is a game of Survivors Versus Killer.
Assumption 2: People care about their rating.
The rating is the result of winning or losing, people that play games try to win and therefore as a result care about their rating, as explained before it is a derivative of the fact whether you win or lose.
Assumption 3: Teamplay leads to increased survival rates
This is a well established fact, the more coordinated a team is the more powerful and higher likelihood they will win. There is a reason why SWF's are the strongest type of team within the game, as they play together to get the match win. You are to focused on the individuals survival rate, while the game should look at the groups survival rate to establish a win for all.
Assumption 4: Rating changes are a "punishment" or "reward"
It is not about rating, if you read the actual post you would have seen that it is about rewarding those that take a risk.
Assumption 5: Different metrics lead to different results, such as different players ending up in different matchups and killers rising in rating disproportional to survivors
This is based on the data mined information that was made available and multiple people discussed and showcased. This is not an assumption, this is directly pulled out of the game by a specific group. I indicate one of the factors to why that is the case and how different metrics lead to different decision making.
Assumption 6: You assume 1-escape and 1-kill matches stand in the way of the proper functioning of the system
Once again, just showcases that you don't want to actually understand what I am stating regardless of what I stated and are purely reflecting on some imaginary belief. Not once have I stated that the system isn't functioning, I am stating that it can be improved and be adjusted to improve the overall game.
Assumption 7: Many more issues in your perception, such as whether the system selects for "skill", what that "skill" is, whether the system has to select for "skill", misperceptions (or deliberate misrepresentations) regarding what arguments I am making...
The use of " marks to indicate that "skill" is some arbitrary aspect that one can fill in. The fact of the matter is that skill can be very well established if one would expand on the premise of the metrics used to account for that. If you read my posts you would see that I barely touch on it and speak far more on playstyles, overall decision making and their potential outcomes and not the skill that is involved in achieving them.
There is literally nothing wrong with my perception, you simply don't want to understand it. I look beyond the does the system function in a vacuum as I have stated many many many times now. You ignore that aspect completely and state that it is the only thing that matters, but it does not. The MMR system is a single piece of the overall product namely the game, my goal is to provide feedback to create the best game! What is the ripple effect that the MMR system does by defining a win and loss in the game and what impact does that have within the game itself.
I have laid out clear examples on what effects occur based on the current system, how the suggestions would reflect and change. You however have not once showcased any consideration for any of the concerns others bring up that they have with the system and what it promotes. The math works, what is the issue? Well, the gameplay it promotes and how it defines a win condition! I am not one of those that advocates for a total different mindset either, I still use the same metrics of kills and escapes just in a different manner.
Post edited by Kalinikta on2 -
Nerf Twins some more. That should do it.
0 -
3 games in a row I decided to use AFK Pig strat. So with video tape addon, i went stealth and body blocked one of the box.
Match 1: 3K 1E
Two head pops, downed third which did a suicide on hook. Next to 4th hatch spawned.
Match 2: 4K
3x head pop, 4th gave up
Match 3: 4K
2x head pop. Downed and hooked third. Coincidentally gates opened so with the hook bloodwarden procced. Went toward the 4th guy, downed him via noed.
Ive put 0 effort and by the BHVR definition of my gameplay I deserve to be in highest MMR
1 -
Not quite- by BHVR's definition of your gameplay, what you deserve is to be going against stronger survivors who'll do a better job against cheese strats.
Even if they can't with this strat specifically, surely you see how that's a problem with Pig's addons and not with the MMR system, right?
0 -
Is it? Please explain me how you would nerf Pig, so I cannot do cheesy tactic ?
Please also explain how after all nerfs, the survivors will have to be nerfed.
Please explain me, how SWF who prefer to sweat gens using such builds and brand new parts is not a problematic at getting a balance?
The definition of MMR should be the same way treated as in a hockey you determine MVP and not by shots and goals determine who won.
SBMM is not applicable in this game. Balance will never be achieved.
SWF will always have advantage and will have more escapes then deaths.
As long as you can spread 2by2 on gens. As long as you can pressure gens on one side, not bother much on healing as you can rely on Adrenaline. It will always be unbalanced game.
Killer cannot down as fast as gens can be done.
It is asymetric game. How can if everybody have 1000 MMR the game be balanced with 1000:4000 MMR + balancing it around 2K 2E.
The ratio of 1:4 is far from achieved. The ratio between time and balance of lobbies will never be optimal.
BHVR is delusional, where they look in wrong direction of where the development of the game should be focused.
SBMM is Patricks baby. And ofc he will defend it until the last drop
1 -
I'm not... entirely sure what your point is? It seems like you want the matchmaking system to take current balance problems into account, which is a pretty bad idea- if you base it on things like that, you'd have to keep updating the MMR system to accommodate whenever things are fixed.
Unless you're saying that a matchmaking system can never work because the game won't ever be truly balanced, in which case... sure, but surely at that point it doesn't matter what matchmaking system they use if nothing could ever work...?
0 -
I love your wording with a pinch of political spiciness.
Yes. Nobody enjoys the current system nor how the mmr is calculated. People want to play game for fun, to chill. Not to have over and over a sweaty matches, where after a few games you get exhausted and need something else.
Have a look at steam charts and the active player base drop.
And SBMM has nothing to do with it?
Killer pick & kill rates were suddenly for whole playerbase, why is classical high mmr list missing?
Question around K:S ratio was dodged. For what reason there wasnt at least average on daily basis - different timezones and prime times.
SBMM is least important thing that dbd needs.
If you do not see any other problems and the adaptation of SBMM is the best thing ever.
Then I feel sorry for you honestly
I dont care about future SBMM updates as there is a valid bhvrs TM of “soon”.
But the current SBMM + lack of killers (queue times says all) will never work as intended.
1 -
...Genuinely not sure what "political spiciness" is in this context, and I'm not being facetious there.
So, people still can play the game to chill- and in the cases where they can't, MMR has nothing to do with it, most of the time it's just the game's current balance problems and those existed before the MMR did. Regarding the Steam charts, I'm pretty sure the game's still going up- last time I checked we're still above the playercount of this time last year. Playercounts go up and down all the time- and while a massive drop would be semi relevant to this conversation, it's not really the case here, at least to my understanding.
The killer pick rate chart also seems kind of irrelevant? I don't know why they didn't show that this time around, there could be any number of reasons. I can say that the question about Killer/Survivor ratio was not dodged, though, it was answered very conclusively and very concretely- there isn't a static K/S ratio because it changes depending on time of day, region, and a whole host of other factors. There's no reason not to believe that.
Like I don't mean to be rude but I really don't see how anything you just said is relevant, beyond a general "people don't like it", which is pretty circumstantial at best. Most people don't like MMR, it's true, but a whole heck of a lot of those people also don't understand the MMR, so I don't think their opinion really counts for much if we're talking about how the system functions. The only other point you're making is that the game is unbalanced, which... has nothing to do with MMR?
0 -
ok
1 -
I have now proven and showcased my point using a mathematical model in a more clear fashion, you don't care about the graph... but that is the math model you speak of.
What point did you prove and showcase, specifically? The graph shows little of real value, it's just 4 curves randomly approaching 0 and 4 kill results. Your explanation of it truthfully didn't make much sense to me, and I'd be glad if you could try and explain more clearly, but the conclusion you came to is still wrong. It doesn't matter whether you attach a 50% survival chance to every single survivor or put together 4 survivors each of which have a 50% "group" survival chance for survivors in their matches, it doesn't matter how you "approach" it, it will still be the same outcome, coalescing around 50% survival rates per-player and per-group in matches on average.
Either way, I asked you to showcase how the volatility index function relates to an MMR system. You know, the function that measures volatility on stock markets, using actual math? And I only asked you to do this because you brought it up and I was interested, it's not like it's making any point we have been discussing - if it shows that the group metric would reduce "volatility", that'd be dandy, I would like to see that, but I already said from the beginning that I expect the group metric will reduce fluctuations, that was never subject of discussion.
You ask why I state it as 1 or 0 for the survivors as that is the definition of a win or loss in a mathematical or computer science situation. You do not care about the graph, because you are disingenuous in actually caring about the math behind the system. You complain about not using numbers, while I did not run any simulations or numbers and simply showed the theoretical outcome of a system like this. I can state over 10 billion games played, but it doesn't really matter as they are numbers that I pull out of a hat. The hypocrisy is that you have not once used any math yourself to strengthen you argument on why it would be better, while even admitting that a group view would likely result in a faster result. Which I showcased within the math analysis of what that means.
Again, I only asked you to show the actual math because you were the one bringing up the volatility index and I was interested whether you were actually able to demonstrate how that is relevant, but you weren't. I used math in an earlier post modelling a rudimentary MMR system and showed that the group metric would probably lead to some better results in terms of stabilizing ratings more quickly. But that is not relevant to the argument, the mathematical parts relevant to the argument were: you falsely assuming the system would skew its math due to "both sides winning" with the 1v1 rating calculations, and you thinking singular matches are significant for a system that by design works over many matches, where singular match outcomes are statistically and functionally insignificant.
Additionally even with my drawing it out for you and explaining it from a math model point of view, you still make inaccurate statements about what I am talking about. There is no use talking to you, as you clearly are not interested in listening or understanding what I am stating and just want to uphold some inaccurate imaginary idea of what you believe I believe... instead of speaking about what I actually believe and state. No matter how I try to clarify and rectify the assumptions you made about what I wrote, you just tell me that you know better what I believe than myself. You might not believe me, but I do think having a MMR system was a good decision for the developers to invest in. I just don't think they are using the best metrics to make the best game.
I always quoted what I argued against. You claiming you did not make the arguments you did is still not leading anywhere.
You claim why would someone care about their invisible rating, while I am not stating that they do. I state that people care about winning and the wins defined by the developers are as I stated. What is defined as a win or a loss influences how people play and acting like people don't want to win or shouldn't is so unrealistic. Therefore by proxy they do care about their rating, as it is defined by how often you win against how often you lose.
Surviving and killing have always been what people consider winning in this game, it's literally in the game description. People at large have always cared about their survival more than anything else, that didn't change just because now an invisible rating used for matchmaking purposes looks at these things.
As you accurately point out if you are in the exit gate with 3, the safe choice would still be to leave and in many cases would still be done with the improvements I stated and I am not trying to deny people from making that choice. Their team can make the more risky move and go for the 4 man out, try to save them however a 3 escape would be considered a win and therefore the risk isn't really worth it. It could also be that they are with 2 in the exit gate with 1 person on the hook. Then the decision would be to go for the safe draw or risk it all for the win, maybe it is now worth the risk? If they manage to pull it off, the team that went back are the ones that should be rewarded, but as it currently stands that isn't the case.
Again, there are scenarios where the group metric could "encourage" certain risky plays more, but there's also others where it can do the opposite. Take the 2 in exit 1 on hook case - if they just get out, they will secure at least a draw for everyone's ratings, risking a rescue attempt is not regularly worth it here because endgame saves that don't result in a hook trade (or worse) are fairly rare. But regardless, now if we look at it with the personal metric, a player might very well be more willing to go for the risk, because if they die doing so, the worst that happens is that their own rating decreases, they are not risking anyone else's; the second survivor is free to leave if they want to secure their own rating increase, and the survivor on hook would have lost rating either way, so the worst case scenario is the same for them. In the group metric scenario on the other hand, you are "responsible" for everyone's ratings and might not be comfortable taking risks putting that on the line in scenarios such as these. If the survivor dies taking a risk in such a case, with the group metric they will have caused not only their own but everyone's ratings to drop.
And this is all from the perspective that a non-negligible amount of players actually irrationally care about their rating and base decisions and actions on it, which I still can't see to begin with.
A game design and the underlying systems should aim to reward those that take risks and succeed. If you cannot even comprehend that this results in a better overall game and system you simply don't want to understand my perspective. My goal is to provide feedback to have the game be the best version of itself, not to excuse it because it works if you ignore everything else and the ripple effect it has within the game.
It does not "reward" anything, the actual rewards for taking risks and succeeding are blatant and plenty, and have been long before we've had invisible ratings.
It's fair enough if your feedback is "with the group metric more players will feel more like a team, which is nice because I like teamplay and consider a more teamplay-oriented gameplay experience to be desirable", and I would agree and would even want for BHVR to outright tell people in the game that they are being rated based on group survival, but the arguments were about the proper functioning of the system, not subjective preferences, the things I argued against were criticisms of the system such as it leading to the "wrong" players "climbing ranks", the 1v1 metric not leading to the right results from the 1v4 perspective, the 1v1 rating adjustments skewing the math because "both sides can win", all of which is just objectively false.
Assumptions 1: This is a team game.
This is a 4 versus 1 game according to the game description, therefore it is a team game and a solo game based on which side you play. Claiming the survivors are not on the same team is simply inaccurate. You don't have to be in a pre-made to be on the same team, the game allows you to join a team and the MMR system seeks one for you. Those on the same side are on the same team, they simply are even if they are made up of individual solo players. It is a game of Survivors Versus Killer.
It's not strictly a team game. People can care about and play for their own survival more than anything else, and that is not a "wrong" way to play the game. The basic game premise is just playing for survival, most of all your own, there is no implication that you should risk your own life for that of others. I agree that it is better if played as a team, and would even welcome if the game told players something along the lines of "you won" or "you tied" if 2/3 other survivors escape despite the respective player dying, but that's just not the reality of the game, and BHVR is making a deliberate design choice not treating the game strictly as a team game, not implementing voice chat, etc.
Assumption 2: People care about their rating.
The rating is the result of winning or losing, people that play games try to win and therefore as a result care about their rating, as explained before it is a derivative of the fact whether you win or lose.
That's backwards logic. People care about winning and always have, they don't need an invisible rating to care about winning. Again, people don't feel like they've "not survived" if they take the hatch, and I have yet to see anyone refuse to jump into the hatch in order to get a gate escape instead, even if the hatch was literally in front of the gate. People care to survive, they don't care about the rating.
But sure, there are definitely some people that care about the rating somewhat, I just highly doubt it has much of an actual effect on the game overall.
Assumption 3: Teamplay leads to increased survival rates
This is a well established fact, the more coordinated a team is the more powerful and higher likelihood they will win. There is a reason why SWF's are the strongest type of team within the game, as they play together to get the match win. You are to focused on the individuals survival rate, while the game should look at the groups survival rate to establish a win for all.
The point is that teamplay that leads to increased group survival rates also leads to increased personal survival rates of the players employing such teamplay. You said "safe" players would climb the "ranks", not teamplayers, implying teamplay would consistently lead to the teamplayers dying.
Assumption 4: Rating changes are a "punishment" or "reward"
It is not about rating, if you read the actual post you would have seen that it is about rewarding those that take a risk.
But rewarding how? The rating is not a reward in and of itself. And the effect of the rating is simply that players are matched such that they have more equal chances to succeed. Why would a teamplayer that actually consistently dies (loses rating) not want to be put into matches where they can survive more often. They want to help their team survive, they don't want to necessarily die doing so.
Assumption 5: Different metrics lead to different results, such as different players ending up in different matchups and killers rising in rating disproportional to survivors
This is based on the data mined information that was made available and multiple people discussed and showcased. This is not an assumption, this is directly pulled out of the game by a specific group. I indicate one of the factors to why that is the case and how different metrics lead to different decision making.
Again, the mined data was not actually based on the MMR matchmaking system, there were a lot of killers with inflated ratings due to the game having been doing calculations in the background while the rank-based MM was still active. And even if there were now, that would not be due to the metric used for survivors. As I've said, if anything encouraging risky plays (and particularly in the endgame) would lead to inflated kill rates.
The metrics leading to different decision-making of players that care about their rating is well and all, but even to whatever extent that actually happens, it's still not a flaw of the system, not least because "selfish" decisions if it were to encourage them are not objectively negative. But the point of contention here wasn't about that, it was about whether team-players and "skilled" players would end up at the same ratings, because you argued they wouldn't, that the current metric would select for "safe" and "selfish" players, that team-players and even generally "skilled" players would not climb "ranks". I argued against that from the very beginning.
Assumption 6: You assume 1-escape and 1-kill matches stand in the way of the proper functioning of the system
Once again, just showcases that you don't want to actually understand what I am stating regardless of what I stated and are purely reflecting on some imaginary belief. Not once have I stated that the system isn't functioning, I am stating that it can be improved and be adjusted to improve the overall game.
You said with the per-survivor survival metric, 1-escape matches would be instances of "both sides winning", which is simply an error in thinking and not understanding the system's functioning. Both sides of the equation cannot win, and the issue that non-diametrically opposed win conditions would lead to by the way of "both sides winning" does not exist with this metric, it does not lead to everyone gradually climbing ratings.
Assumption 7: Many more issues in your perception, such as whether the system selects for "skill", what that "skill" is, whether the system has to select for "skill", misperceptions (or deliberate misrepresentations) regarding what arguments I am making...
The use of " marks to indicate that "skill" is some arbitrary aspect that one can fill in. The fact of the matter is that skill can be very well established if one would expand on the premise of the metrics used to account for that. If you read my posts you would see that I barely touch on it and speak far more on playstyles, overall decision making and their potential outcomes and not the skill that is involved in achieving them.
The "skill" argument was a more minor point of contention, but it was still just something you were wrong about, starting from the first post and the false assumption that skilled players ("5-gen runners") would not climb ratings based on singular exceptional match scenarios, and later developing into you saying the devs are wrong to use the word "skill" in the way that they are, but they aren't, it's not a wrong use of the word to refer to players good at consistently winning.
There is literally nothing wrong with my perception, you simply don't want to understand it. I look beyond the does the system function in a vacuum as I have stated many many many times now. You ignore that aspect completely and state that it is the only thing that matters, but it does not. The MMR system is a single piece of the overall product namely the game, my goal is to provide feedback to create the best game! What is the ripple effect that the MMR system does by defining a win and loss in the game and what impact does that have within the game itself.
I have admitted at every step of the way that the argument of the ratings being soft (irrational/emotional) incentives for people's playstyles/decisions/actions has merit. I have doubted whether it actually has a notable impact on the game, but I have not said it doesn't matter what side-effects the system may have as long as it functions. But I have also pointed out that even if it were to encourage more selfish playstyles/decisions/actions for every single player, that would not objectively be a flaw, since that is preference.
As long as the system functions to instate more even win/loss ratios for players, approaching 50% (which it does), it functions just as well in principle with the per-survivor metric as it will with the group metric, only that I suspect the latter will make players approach those ratings that lead to more even ratios faster, and/or stabilize them around those ratings more. Either way, everyone is still free to play however they want, and if their chosen playstyle preference (and even if it would be "informed" by the metric) leads to them dying often enough to fall in rating, that only means they will be more likely to be able to survive more often as a result at those lower ratings. There's nothing bad about it, it's not a viable criticism of the system.
I have laid out clear examples on what effects occur based on the current system, how the suggestions would reflect and change. You however have not once showcased any consideration for any of the concerns others bring up that they have with the system and what it promotes. The math works, what is the issue? Well, the gameplay it promotes and how it defines a win condition! I am not one of those that advocates for a total different mindset either, I still use the same metrics of kills and escapes just in a different manner.
I did address your main argument of the personal survival metric "promoting" certain gameplay, and I acknowledged that you as opposed to others are not entirely unreasonable/unfair in your treatment of the system. I also admitted that I came on a little harshly at first, since you weren't actually making some of the ridiculous arguments others are. But that doesn't change anything about the flaws in thinking that I did explicitly reply to and try to explain.
0 -
You just proven that you have no idea what you are talking about while acting like some kind of expert. You keep trying to tell me that my reasoning and manner of thinking is incorrect, yet when I throw at you the actual manner in which the system functions from the theoretical math model you try to hide behind you state you don't know what you are looking at.
The graph that I show you is how a mathematical model of a MMR system would look like and is known as an average volatility graph. You have the minimum end return value and the highest and then run a ton of games over it, within the industry the standards are like simulations of 1,000,000,000 games played to determine a volatility accuracy. The desired result if everything works accordingly is what we then call in the mathematical model the nominal. Therefore the reduction in fluctuation that you speak of is pretty much the area between the upper bounds and the lower bounds. The lines you see in my quickly drawn graph pretty much represent the two boundaries which one could experience to eventually end up at the correct nominal (50% escape rate).
I cannot put in the numbers, because we don't have the numbers and to actually calculate these properly you either need to put in a bunch of data into it either by simulation or having actual data on it. Therefore I cannot tell you whether the angles I showcased are accurate and me making up numbers doesn't change the manner in which it responds compared to me just winging it like I did, as I would be the one making up the numbers. Applying the calculations would take far more time and that is totally not something I am going to do to prove a point on a forum. You state yourself that the effect on the system isn't up for debate, so me showing this on the hand of actual numbers is quite pointless and time intensive, while an illustration does the exact same.
You state that my logic is backwards by stating that the developers defining a win influences the players. That the hatch is a perfect example of it. Yet in my model the hatch would be considered far more a win, corresponding with what the game pretty much indicates as a win than it currently does. In the case that the game ends up in an overall loss, the final player still has a chance to get a draw just for themselves. The killer is hardly affected by this, as they are still judged on the overall games result. It is why I don't try and address the MMR system from their perspective, while they are naturally influenced by the manner in which survivors play. The whole goal of my suggestion is to have both sides of the versus to be judged on the same criteria.
I have stated time and time again that the game design implications and the metrics used in the system aren't aligned to create the best game and experience. Looking at the overall outcome of the side that you play as is done with killer would result in a better system and you have conceded this fact multiple times.
Frankly speaking, you can believe what you want. You have admitted that it has an effect and yet because we cannot quantify how big it is you don't care. The quality and how good the system is as long as it works is irrelevant according to you. While that is exactly what I am addressing, the quality of the system and I have never claimed the current system doesn't work. I am discussing improvements to a system, while you are trying to have my prove that the current system doesn't work while making no such claims. We aren't having the same discussion, because you don't want to understand where I actually stand.
Post edited by Kalinikta on0 -
You just proven that you have no idea what you are talking about while acting like some kind of expert. You keep trying to tell me that my reasoning and manner of thinking is incorrect, yet when I throw at you the actual manner in which the system functions from the theoretical math model you try to hide behind you state you don't know what you are looking at.
What part of your reasoning and manner of thinking does the graph support, in any way? I didn't say the graph made no sense, it's simple enough, just that the curves you input are random, and that the graph doesn't actually provide anything supporting any arguments you've made. How does it show that the group metric would reduce volatility, and how does it relate to the volatility index?
The lines you see in my quickly drawn graph pretty much represent the two boundaries which one could experience to eventually end up at the correct nominal (50% escape rate).
Yes, and this would true regardless of which metric you look at results from. The graphs would just be expected to be somewhat more extreme in their aligning at the nominal with the group metric, or so I suspect, which is obvious if it is true, but not mathematically proven to be true with your graph.
You state yourself that the effect on the system isn't up for debate, so me showing this on the hand of actual numbers is quite pointless and time intensive, while an illustration does the exact same.
I do think it has that effect, but I've explained why I think it does (using numbers, I showed that with the per-survivor survival metric, killers could lose rating in "draws" against mismatchedly-rated groups, or gain a lot of rating if they kill all survivors due to weak links in such groups; if you look at the overall result (and use group average ratings), you can much more easily stabilize players around balanced outcomes by simply leaving everyone's ratings unchanged in them, and not have rating disparities within survivor groups cause as much of a potential shift on both ends - there are still things that skew ratings with that metric too, you'll still need many iterations to arrive at stable ratings and more even matches, but it should take less iterations). I just don't see how your graph/explanation of it demonstrates this.
You state that my logic is backwards by stating that the developers defining a win influences the players. That the hatch is a perfect example of it. Yet in my model the hatch would be considered far more a win, corresponding with what the game pretty much indicates as a win than it currently does. In the case that the game ends up in an overall loss, the final player still has a chance to get a draw just for themselves. The killer is hardly affected by this, as they are still judged on the overall games result. It is why I don't try and address the MMR system from their perspective, while they are naturally influenced by the manner in which survivors play. The whole goal of my suggestion is to have both sides of the versus to be judged on the same criteria.
Currently, the hatch doesn't factor into MMR at all, it is the only "true" draw state we have because neither the killer nor survivor rating is affected by it. I only used this to showcase that people care more about surviving (like they've always done) than they do about ratings, because I haven't seen anyone refrain from taking the hatch in favour of leaving through an exit, even in cases where they were completely free to choose between the two. I also haven't heard of players saying they feel less good if they escape through hatch now, or as if they hadn't truly "survived".
Either way, you bring up an interesting point, namely how to deal with the hatch in a group survival metric environment. Should a hatch escape count toward everyone's rating, making all survivors lose less rating than they would in a 4-kill match, or should it treat the match as a 4k for everyone but the survivor that did escape through hatch? I'm honestly not sure, but here I do feel like your argument that people care about rating shows itself to be relevant: if players knew that their hatch escape can "benefit" the entire team's rating, they might be more thrilled to play for and succeed in getting a hatch escape, and who knows, this might also lead to dead players sticking around more often spectating, rooting for remaining players.
I have stated time and time again that the game design implications and the metrics used in the system aren't aligned to create the best game and experience. Looking at the overall outcome of the side that you play as is done with killer would result in a better system and you have conceded this fact multiple times.
I didn't "concede" that fact, I purported it myself, and agreed with you about that aspect once it came up. It was the other stuff I argued against.
Frankly speaking, you can believe what you want. You have admitted that it has an effect and yet because we cannot quantify how big it is you don't care. The quality and how good the system is as long as it works is irrelevant according to you. While that is exactly what I am addressing, the quality of the system and I have never claimed the current system doesn't work. I am discussing improvements to a system, while you are trying to have my prove that the current system doesn't work while making no such claims. We aren't having the same discussion, because you don't want to understand where I actually stand.
My arguments weren't about your entire standpoint, certainly not after the first few replies, but about the specific things I addressed. I even agreed early on that the soft factor of the group metric perhaps kinda-sorta encouraging more self-sacrificing or at least risk-taking teamplay is an upside and another reason why switching would be good (although far from the primary reason, for me anyway). I just later pointed out that just because we both consider that to be a desirable effect, not everyone has to or does. Maybe one of the reasons why the devs are going for a hybrid approach specifically is that they are afraid that with an exclusively group-based metric, there could be more frequent edge-cases where players can go on dying streaks, and they want to still factor in personal survival to avoid this because they speculate/know that ultimately, players on average care much more about their own survival than that of others.
0 -
Au contraire...
"I love how that dude zarr and plays by shady keep defending the system when I can list two flaws with the system rn that can't be argued against
1:it's kills and escapes based matchmaking we all know this everyone knows this, that means it's easy asf to abuse all I have to do is die on purpose every game or let everyone escape every game and all of a sudden boom I'm smurfing"
Err... no dude. That's not a flaw with "the system" (presumably you mean the MMR system), that's a flaw with anything and everything that doesn't know or understand your intent. Any system can only match you according to your measurable progress... so if you deliberately cripple your progress, obviously that's going to reflect in your standing.
To use the Kobe example, what are you expecting if Kobe just stands on the side of the court for the whole game and throws it? For the system to reward him "oh, he's a good player normally, he's just making a point right now?" No dude, his standing will go down.
That's not a flaw of "the system", that's literally how rankings work... and you've not explained why it shouldn't work like that, and therefore why it's a flaw.
Love it when people thing they've got these killer points without realising the fundamental flaws in it, lol
0 -
"It's helpful to remember that the point of the matchmaking system isn't to determine who's more skilful- it was a mistake of them to use the word "skill" in the name, even if it is still accurate when you're looking at aggregate batches of MMR brackets instead of individual games."
I don't think it was a mistake to use the word "skill" - it's an accurate enough term that's been in use for decades (I referenced TrueSkill in my OP which is 20+ years old).
The mistake was not recognising that the general community would understand something else from the term "skill", and not catering the language accordingly.
It's all well and good for BHVR to refer to "skill" internally... the devs know what they're talking about, and in fairness, it's the most accurate short-hand term (anyone who disagrees, please suggest an alternative). But if you do any PR - including live-streams - then there's a degree of responsibility on you to understand your audience and understand how they'll interpret your words. I think this was the mistake Patrick fell into.
As a dev, everything Patrick said made complete sense to me. But to my non-dev friends, he was full of contradictions and making because they didn't understand the background... and unfortunately as a community, the masses are too immature to recognise that it might be them who's not understanding. It's much easier to say "developer = dumb" which is the trap too many fall into.
1 -
You clearly haven't a clue what you are talking about.
1. Me making up numbers can result in anything I want it to, as it has no foundation for it. Therefore doing so and acting like it would is completely delusional. Go do it yourself and prove me wrong, you claim to understand it so much better than I do... so do what you ask from others else simply stop demanding numbers.
2. Killers even out on 2ks affecting them, you love throwing it at big numbers and the diversity of going up and dow goes both ways. The current system is more volatile because of it as the min and max are further apart.
3. Hatch does affect MMR as it removes a loss and a win from the equation, either the killer or survivor. It is literally the only way a survivor can draw at this point in time. Making the sum for the killer reliant on 3 outcomes instead of 4.
You try to tell me that I don't understand the system and use it against me... yet somehow when it comes to the math you want to limit it to a single scenario, while if anyone uses it to showcase risk/reward balance in the game you go... but that will level out. The hypocritical arguments, the personal deflection and trying to unjustified act like an expert is so transparent.
You act like I am presenting my suggestion as gospel, but I already admitted they have more options to improve the system and it needs to consider group performance. You state they want to make a hybrid approach, literally meaning that all the points I brought forth are aspects they are considering... because guess what that is the reason people dislike the current. You are the one white knighting for the current system, while the developers by your own accord are not sticking with it. They have the numbers, and are adding team performance into the MMR!
1 -
Yeah, to be clear that is what I meant- it wasn't wrong to use the term because it still makes sense in context, but using it for public-facing descriptions was imo a mistake because of the loaded associations that it has.
Simply calling it an MMR instead of SBMM would've alleviated some of the strain- not all of it, but a little bit.
1 -
- Dude, you were the one starting to talk about volatility indexes, I said feel free to use a functioning MMR system with example values because we don't know the real ones to actually showcase what you mean. That graph you drew isn't actual math, it's just... results randomly trending towards a mean. How does it relate to the actual math of an MMR system, let alone to a difference between personal and group survival metrics in an MMR system for DbD? Mind you, I never said the group survival metric would not have the effect of reducing fluctuations within the rating developments, I pointed out as much myself, I just wanted you to show what you mathematically mean with the volatility index function relating to this.
- The min. and max. in both metrics are 0ks and 4ks, 0% or 100% survival, both per-player and per-group... But the group metric allows for more granularity, since 25%/75% group survival results have a more predictable impact. What do you mean "even out on 2ks affecting them" - that they can be rated +/-0 in draws if using the group survival metric? Literally my point, and not one you brought up before.
- Not the point. Escaping by hatch gives +0 rating for the survivor, as opposed to escaping by gate, so if they cared about rating, they would always want to go out of their way to avoid hatch if they can get a gate escape. This is not behaviour I have noticed, which I used indicatively to say players don't even seem to care enough about rating to simply not jump into hatch.
I specifically pointed out flaws in your understanding of the system. Like, you can't just say stuff like "both sides win, same problem as if it were based on hooks!", or "1v1 metric doesn't work from the 1v4 perspective", or "star players don't 'climb in ranking'!" and then act as if you didn't make objectively wrong assertions. What do you mean I want to limit the math to a single scenario? What scenario? Which part of the math specifically?
I've always said I think switching to the group metric would be an improvement, what is your point? I didn't contest what you presented on the basis of it being gospel, I contested the things I did contest on the basis of them being false. The OP of the thread that has professionally worked with such systems had already told you you have no clue what you are talking about in your original posts, and pointed to the flaws in your thinking, I merely went and further tried to explain the flaws in your thinking to you. Doesn't mean everything you've said is wrong, or that you ever said the system would not work at all, you didn't... that was never the point.
Post edited by zarr on0 -
- You actually talked about it, but frankly with your lack of understanding of the math model I am not surprised that you missed it. When you talk about how impactful things are on the numbers and claiming that they are insignificant, you are technically speaking of the volatility. All I did was use the correct verbiage.
- You speak of the killer being able to lose MMR while 2king, yet they can also gain MMR while 2king as it all is about the relative level of the survivors against their own. Additionally the system can still form the killers MMR to account for all of this even if a group metric is applied to the survivors or balance it out more easily by using averages and making that outcome less volatile.
- The fact that you do not understand that the ability to negate a negative and positive score for the other party based on whether you do escape or not by the door, impacts the mathematical outcome of the system. Taking the hatch isn't always an alternative to the gates, it might be the alternative to dieing. In either case this reflects back on the opponent you are facing as well as your own rating. These thing have a bigger impact on the numbers than you might think.
You like hammering on the so called flaw of thinking that both sides can win in the current system. It is a 4v1 game, the meaning of versus is that they compete against each other. If at the end of a match a victor emerges on both sides of that v sign - they are versus, both sides shouldn't be able to have a victor if they are versus each other and that is not flawed logic, that is quite simple actually. This isn't a bad line of thinking, it is viewing the entire match and looking at the result of it. Then claiming that the MMR should look at who hit the win condition and who didn't.
To bring it back into the team sport analysis that was used, it is stating that the highest scorer of the opponents team won, even though their team lost. In sports they don't care who scored or not, all they care about is which side scored more... nobody cares whether it was 0-1 or 25-300 in the end one side wins. In DBD this equivalent it shouldn't matter if it was a 0-4, a 1-3, a 4-0 or a 3-1 there is a clear victor at the end and the only draw is a 2-2. The team that scored 25 don't have an individual win, they lost. It is literally the same issue as comparing it to hooks, whether or not the killer got many hooks the survivors escaped. It doesn't matter that the opponent made 25 points, you made 300 and won. It doesn't matter how the survivors or killer played, 3 of them are dead.
You even by admitting that switching to the group metric it would be an improvement. Therefore all that hiding you do behind the math model is invalidated as we agree it would become better. We both state that the current system mathematically still makes sense which means an improvement is about the accuracy and quality of the result (which is reflected in the volatility). Now you can argue how significant of an improvement it would be and we simply do not have the answer to that and I never claimed we did. You however want me to make up numbers to showcase this, while you refuse to do it yourself and I argue that it doesn't matter as those would be all fictional anyway.
Then we go into the game design aspect and you have even conceded on this front:
"I even agreed early on that the soft factor of the group metric perhaps kinda-sorta encouraging more self-sacrificing or at least risk-taking teamplay is an upside and another reason why switching would be good (although far from the primary reason, for me anyway)."
Whether this is a primary reason for you or not, that is literally purely subjective. It however does also implicate that all that I have been stating is accurate and correct. Therefore my line of thinking again isn't flawed. The reason why this is my primary motivator is actually because the mathematical improvements are one thing and I don't believe it is that significant and is why for the majority I focused on this aspect. You don't care about it, yet I believe the main outcry against this system is exactly due to the consequences of these soft factors.
Sorry, but you might want to call on your companion here to step in, because frankly you are out of touch with the actual discussion that I am having and the feedback I am giving. You pretty much conceded on each front without noticing, while we might value them differently on their importance yet that is a pure subjective take on it anyway and judging someone for it as you do is flat out arrogant and out of line. You can try and hide behind another persons professional experience, but he isn't the only one with that type of experience. You think I can draw up a graph like I did without any experience? If you want to act like an authority, show your own credentials as those of someone else don't actually mean you know what you are talking about.
1 -
- It wasn't about the "verbiage" of volatility, it was specifically about the volatility index. Here https://www.macroption.com/vix-calculation is a simple step-by-step guide on the math of that.
- Ah, so you did mean the thing I have shown with numbers and you never before stated. Gotcha.
- You are great at avoiding talking about actual points. The point of me bringing up hatch was only, simply, to indicatively highlight that I don't see that people care about rating as you claim. Again, even in cases where the hatch was right by a gate, I have yet to see anyone choose to take the gate escape over the hatch escape.
You like hammering on the so called flaw of thinking that both sides can win in the current system. It is a 4v1 game, the meaning of versus is that they compete against each other. If at the end of a match a victor emerges on both sides of that v sign - they are versus, both sides shouldn't be able to have a victor if they are versus each other and that is not flawed logic, that is quite simple actually. This isn't a bad line of thinking, it is viewing the entire match and looking at the result of it. Then claiming that the MMR should look at who hit the win condition and who didn't.
If you look at it in the way that the system actually does, there can be no instance of both sides of the equation gaining rating. You simply aren't understanding that doing it 1v1 leads to no such issues you were claiming it does. The only issue "both sides winning" leads to is more and more people climbing ratings with time, invalidating them. The 1-escape scenario does not lead to this, because it is statistically irrelevant overall, and statistically even more irrelevant for a given player. And for any player it would not be statistically irrelevant for, they should be gaining rating, because otherwise they would be able to do what they do, somehow consistently escape through gates the sole survivor, exceed the 50% success rate target, but yet get worse and worse opponents, getting less and less likely to ever approach the target that the system solely exists to seek.
But even regardless, even when looking at it 1v4, how do you not get that both "sides" do not win if 3 survivors do not gain rating, and that a 50% survival rate is based on the 1v4 perspective because from the 1v1 perspective the only sensible survival rate would be 0%?
To bring it back into the team sport analysis that was used, it is stating that the highest scorer of the opponents team won, even though their team lost. In sports they don't care who scored or not, all they care about is which side scored more... nobody cares whether it was 0-1 or 25-300 in the end one side wins. In DBD this equivalent it shouldn't matter if it was a 0-4, a 1-3, a 4-0 or a 3-1 there is a clear victor at the end and the only draw is a 2-2. The team that scored 25 don't have an individual win, they lost. It is literally the same issue as comparing it to hooks, whether or not the killer got many hooks the survivors escaped. It doesn't matter that the opponent made 25 points, you made 300 and won. It doesn't matter how the survivors or killer played, 3 of them are dead.
DbD is not a team game in the sense sports are. Even outside of the context of you overblowing the analogy that was used and what it was used for, using team sports to apply to DbD logic is flawed. I've explained this at length (little reminders: people regularly play alone, having actual set teams is the rarest exception; everyone in DbD may only play for their own survival; everyone in DbD "scores goals", shooting themselves over the exit gate line, so to say), and the system may well be designed like this because as opposed to you, BHVR understands that most people care about their personal survival more than any stranger's, and that most players would not be happy dying 10 times in a row just because 2+ other players survived each time. Not least because they often may not even be around to see this outcome, having left the trial after their death. That is likely big part of the reason why they'll keep looking at individual survival rates too.
And it is still not the same issue that looking at hooks would lead to. You have to think about what it systemically means that the 1 lone survivor that lives gains rating, and realize that as long as it does not lead to more and more players gradually climbing rating by consistently doing so, it poses no such issues. And nobody is consistently escaping through gates in matches where 3 people died. And if they were, they should be gaining rating, because they must be doing something that leads to this consistently, it is statistically impossible for someone to be consistently more lucky than other players. And they should not be individually escaping constantly either way, not even the sole survivor, but approach 50% escape rate, as that is the system's only purpose, and increasing the rating of such players as such necessary for its functioning toward achieving its goal.
You even by admitting that switching to the group metric it would be an improvement. Therefore all that hiding you do behind the math model is invalidated as we agree it would become better. We both state that the current system mathematically still makes sense which means an improvement is about the accuracy and quality of the result (which is reflected in the volatility). Now you can argue how significant of an improvement it would be and we simply do not have the answer to that and I never claimed we did. You however want me to make up numbers to showcase this, while you refuse to do it yourself and I argue that it doesn't matter as those would be all fictional anyway.
I only asked for actual math because you brought up the volatility index, saying it would show how the group metric mathematically reflects on the MMR system. And I only did because I was interested, not because I was claiming the group metric would have no such effect of decreasing "volatility", as I have from the beginning myself said and later explained with numbers, stabilizing ratings more around desired outcomes, cutting away some fluctuations.
The things I argued were not about whether the group metric would improve the system in this way, but whether it would in the other ways you claimed it would, the ways in which you claimed the current system is flawed or not functioning desirably, such as "both sides winning", "star and team-players not climbing in rating as opposed to safe and selfish players", "the system not working from the 1v4 perspective", all of which is still just as wrong as it was then: both sides of the equation are mathematically unable to win; if you look at more than singular exceptional matches like you have to with the system, star players and team-players that have increased group survival rates (i. e. are actually good at what they do) will also have increased personal survival rates, and as such climb rating (but even if this weren't the case, it would pose no problems since it would still lead to all players eventually settling at ratings where they can personally survive more often, so what if these or any players aren't at the highest ratings); 50% survival rate works out to be balanced both from the personal and group perspective, and is indeed based on the 1v4 perspective because no player could ever realistically survive a 1v1 in DbD. If DbD were a 1v5 game instead, you could easily understand that not 2.5 survivors can escape in any given match, and that therefore a balanced 50% survival rate can never lead to balanced results in singular matches, that you would always have "wins" for either side with at best 2 kills vs. 3 escapes or vice-versa, yet still the 50% balance target would work on average over many matches and when looking at players individually. You just fail to understand this for the reality of the game where a 50% balance is possible from both perspectives and can in any given match lead to a balanced overall result too (2 kills, 2 escapes), but likewise also works with all other combinations over many matches (0, 1, 3, 4 kills vs. 4, 3, 1, 0 escapes). Well, conveniently it doesn't matter, it comes out to the same end result, the system will work to achieve its desired and desirable goal with 1v1 math just the same (only I suspect it will do so faster with 1v4 math).
Whether this is a primary reason for you or not, that is literally purely subjective. It however does also implicate that all that I have been stating is accurate and correct. Therefore my line of thinking again isn't flawed. The reason why this is my primary motivator is actually because the mathematical improvements are one thing and I don't believe it is that significant and is why for the majority I focused on this aspect. You don't care about it, yet I believe the main outcry against this system is exactly due to the consequences of these soft factors.
I did not say your thinking is flawed in the sense that such side-effects of the system can and do exist to extents (otherwise I would not have agreed with it...), but that it is flawed in the sense of making it out to be something objectively desirable for the system to encourage such self-sacrificing/risk-taking teamplay behaviour. Because opinions may very well and do differ regarding whether it is desirable to risk and give one's own life for that of others in the game rather than mostly play for one's own, whether the gameplay it yields is desirable. That means if and insofar the personal survival metric "encourages" "selfish" play, and the group survival metric "selfless" play, using either of the two is not objectively an improvement over the other, but a game design consideration that has to be informed by subjective considerations. And I for one do very much assume BHVR still thinks of the game more as a solo survival horror type, where 4 individual players happen to have the same objectives and may work together (in rudimentary ways; there's a reason BHVR is rejecting the idea of implementing in-game voice chat that most actual team games these day have), but are not strictly a team and may play "selfishly" and for their own survival above all too, and do not personally "lose" if they survive just because the other survivors may have died. And again, this is also because they probably realize that most people mostly care about their own personal survival, are satisfied most of all (or in many cases indeed only at all satisfied) if they themselves survive.
Sorry, but you might want to call on your companion here to step in, because frankly you are out of touch with the actual discussion that I am having and the feedback I am giving. You pretty much conceded on each front without noticing, while we might value them differently on their importance yet that is a pure subjective take on it anyway and judging someone for it as you do is flat out arrogant and out of line. You can try and hide behind another persons professional experience, but he isn't the only one with that type of experience. You think I can draw up a graph like I did without any experience? If you want to act like an authority, show your own credentials as those of someone else don't actually mean you know what you are talking about.
The nerve to call me "arrogant" after saying something like "call on your companion here to step in". You are pretty hilarious dude.
Here I am repeating the very basic arguments that we have been circling around from the very start. This absolutely would not ever lead anywhere. Go ahead and have the last word, I'm ditching out of this "discussion" with you.
Post edited by zarr on0 -
You can google a math explanation Congratulations! Now apply it, get the average volatility graph and see what roles out. You wanted to know what it would look like, well that is what I drew out for you
You act like you are proving me wrong and repeating that I dont understand how the system works throughout this conversation. All while admitting that the aspects I bring up have an impact, would be an improvement, etc. Yet dismiss it for some unknown reason based on nothing than the assumption it wouldn't have a big impact. Yet somehow you aren't arrogant? You even hide behind someone else's professional life, you brought that up.
The fact that even with the graph I gave, you still state I don't see how it can produce the desired result of a 2k is mindboggling. It is literally what I drew... but oh well. You live in an imaginary world.
If the developers want to create a situation where survivors act like they are on the same team, as a 4v1 game indicates... maybe the system should promote that. You stated and pointed out they are moving in the direction of team based outcome, but sure it isn't desirable or impactful or their view of the game.
You now have the arrogance to fill in how the developers see their game. Claim that if anyone has a different perspective they are wrong and don't understand. It is the key element of the complaints about the system, that it doesn't account for those playstyles. I have from the get go have stated it is a different angle. Yet you kept coming at me with huge explanations on why that was wrong... now once again you simply yielded.
This is indeed done, you have literally admitted to each and every one of my points and agreed. While still having an incorrect view of where I stand, even though I drew it out for you. You never had an interest in understanding what I was stating and just wanted to respond to an imaginary opinion.
1 -
Just a reminder to everyone here that DBD was never intended to be a "team" game.
Mathieu: "It is an asymmetrical multiplayer game where you can live in the world of these 70s and 80s horror movies. It’s anti-social multiplayer where survivors have to collaborate and cooperate up to a certain point. The old saying is that you don’t need to be fast than the killer, you just need to be faster than your friends. Whereas the people playing as the killer, they just need to straight up murder everyone."
https://www.mmorpg.com/mobile/features.cfm?read=13188&game=1501&ismb=1
At best, it's always been about forming a temporary alliance. Part of the "skill" came from figuring out how much to help others to increase your own chance of escaping, and knowing when to cut them off.
After all, the winning condition - the objective - has always been about your own escape, not the overall numbers of escape per match.
2 -
i love the “anti social” but also let us introduce you a SWF where you can be social and using comms to be more competitive
1 -
And 3., it incentivizes unhealthy gameplay.
If a killer knows that the only thing that matters in terms of them being considered a "skilled" player is getting kills and that killer values being viewed as skilled, they have every incentive to be that Bubba camping the hook. If he's facing a team anywhere near his skill level, that's an almost guaranteed 2k with the potential for more if survivors play altruistically. I took a break from the game for a few months in the fall, but since I came back I have encountered an unbelievable amount of killers playing like this. I had literally 8 matches in a row the other night in which the first person on the hook was camped to death. Didn't matter if the survivors were in the area or not, trying to save or not, or pounding gens or not. In every match the killer did not leave the hook. This would sometimes happen back in the day, especially after a long first chase, but in ~1300 hours I'd never experienced camping anywhere close to that consistent before. I've noticed an increase in killers who slug for the 4k since I came back as well as well. Why risk one survivor escaping through the hatch when what used to be a +1 regardless is now the difference between a 3k and 4k? The lack of fun for the one survivor left to afk on the ground certainly isn't enough of a deterrent.
Similarly unfun behavior appears to be showing up on the survivor side too. For example, let's say someone is hooked just as the gens are powered. Before you would be incentivized to go back for the save, since even if you die you'd gain emblems for the unhook and would be punished if you left your teammate to die. For those running WGLF, you could get two stacks and an extra 50% BP just for trading places with the person on the hook. That more than offsets the loss of your 5000 BP for escaping (minus late unhook, protection hit, struggle points, etc. on the way to getting the save, of course). Nowadays, though, it seems like more people than ever just leave without even trying. Why risk having your MMR drop and getting matched up with potatoes? I'm sure the increase in hook camping isn't exactly a catalyst for altruism, but still. Players on both sides seem to be playing differently than six months ago and I can't think of a good reason for that besides the change in ranking system.
I admit that there could be other compounding factors, though. Maybe others aren't haven't this experience and I've just been outrageously unlucky. Maybe my survivor MMR is high and I'm facing stronger, more competitive killers who are more likely to slug for the 4k. Maybe my survivors MMR is low and I'm facing more killers who feel their only chance to win is to get one hook and camp it like their life depends on it, and being paired with more baby survivors who are too terrified to play altruistically. My money's on the ranking system itself changing player behavior, though.
1 -
But why have MMR at all if it's not meant to reflect player skill? Is the goal not to have better killers play against better survivors and vice versa? At that point the old system of rank-based matchmaking, while flawed, is still a better system to base matchmaking on. Even if you want to use kills and escapes as a proxy for everything else in the game, it should absolutely consider kills and escapes for the whole team, not the individual player. @SilentShepherd makes a great analogy to +/- in basketball and I completely agree that this is what MMR should be based on. I'm fine with keeping it to kills/escapes if so, though.
For example, let's say you're the only survivor to die in a 1k. The system should consider that a +1 for every member of the survivor team. Conversely, let's say you're the only survivor to escape in a 3k. The system should consider that a -1 for every member of the survivor team. This would make matchmaking more directly based on player skill while also incentivizing good behavior for survivors. For example, let's say you're in a situation with three survivors left alive and the gates open. One survivor is hooked, one is injured and getting chased, and one is healthy. In the current system, if the healthy survivor just leaves, MMR increases their rating. In my proposed system, if the healthy survivor just leaves, they're punished; they'd almost certainly be the only survivor, meaning everyone on the team would get a -1. Let's say they went back, got the unhook, and took a hit for the injured survivor in the chase so that they could escape. Even if they died as a result, then, they'd get a 0, which would be better than the rating for just leaving (and rightly so).
This would even be a simpler system and one that is more representative of what the actual game is. It's not four individual survivors each having a 1v1 battle with a killer. It's a team of four survivors having a 4v1 battle with the killer. Each team member has the ability to make or break the outcome for the rest of the team. This should be reflected in MMR.
1 -
Thank you for your acknowledgement and for the excellent post.
We are not even being paid for this, and we can come up with better solutions than what we have. It's just so disheartening to see the choices that DBD devs make.
Why are they incentivizing UNHEALTHY/UNFUN gameplay...
Why don't they use a +/- system to more accurately rate players? This would create so much more HEALTHY gameplay and a better overall experience for every player.. because dying does not mean you need to "git gud"... it means you took the killer on and did risky things to HELP YOUR TEAM ESCAPE. And depending on how successful you were... how many generators you did, how long your chases were, how many unhooks, whatever.... all of this is taken into account. People would feel much better at playing a team game, and not feel so bad if they die.
2 -
This perspective is too individualistic. A useless survivor should not be given a higher skill rating on the grounds that now they'll be more likely to die. This is ignoring the fact that survivors like that will forever have a boosted MMR, which destroys the experience of their teammates at their MMR.
In the current system, rather than each 50th percentile MMR lobby containing four survivors of average skill, it could just as easily contain a mix of survivors with 10th percentile actual skill who prioritize their own survival at all costs and survivors with 90th percentile actual skill who are extremely altruistic and are willing to die to save their teammates. Boosting the skill rating of useless, counterproductive survivors is an awful idea because it impacts the in-game experience of those more skilled teammates. They will have endless strings of matches where they run the killer for 5 minutes and then get left to die on their first hook, and the like. It makes no sense to build a matchmaking system that punishes skilled and altruistic players by saddling them with a disproportionate amount of useless and selfish teammates.
In a +/- style MMR system, even one just based on team escapes and nothing else, a useless player's ranking would be hurt if they were only survivor to escape a trial. This is a good thing. Now they're playing with teammates that won't be able to carry them and against killers that aren't very good. This will incentivize them to come out of the bush and actually play the game, for one. It also doesn't punish the killer, since this survivor will only stay in low MMR so long as they're not helping the team. In other words, even if they escape a lot, they'll be basically paying off that escape by gift-wrapping the other three kills for the killer. If this survivor actually starts working to help the team, their match outcomes will start to improve, and their MMR will increase. They'll then be matched against stronger killers who are able to counter their increased productivity.
See my previous comment in this thread for more about this system, but it would wipe the floor with the current system. There is no reason to have an MMR system for matchmaking if the ratings don't do a good job of showing player skill. Currently they don't. Players are being evaluated as individuals in a team game, which makes no sense.
1 -
? What does SWF have to do with the main objective of the game design? Sure, SWF will have more information and therefore would have more awareness of when they should help or cut off other survivors. That doesn't change the win condition.
0 -
Sure thing! Completely agree. The old pip system was flawed, but it at least incentivized good behavior. It had a bit of that +/- principle in it too because everyone's rank would suffer if survivors were left to die on the hook, and survivors who did little to help the team would derank. I wish they had still considered the importance of team play when building the MMR system.
1 -
the quote of what cote said compared to the state of dbd
1 -
? My point is that DBD isn't a team game. Existence of SWF doesn't change that. Players in SWF can choose to play like they are in a team, but that is more of problem with players, not the game.
0 -
You mean this win condition as defined by Mathieu in the same article and interview that you quote?
One of the difficulties is that the game does not have a clear-cut win/lose condition. It's more of a gradient of did I have a good game, do I feel like I’ve accomplished much, was it fun?Your argument is this is not a team game, yet a Survive With Friends was introduced the year after that interview. It is not a team according to you, but the definition of with is together. There is only one game mode and it most definitely has teams on the survivor end. The system doesn't care whether you are a pre-made or not, the introduction of SWF is literally the introduction of the concept of survivors being on a team. They threw away the anti-social approach of survivors when they added the feature.
Like you can literally use the same interview to claim that SWF and MMR have no place in the game and yet they are now integral parts of the game and aren't going to go anywhere. The game evolved and the features and systems that it added are direct products of the changes in what the game is. You can cherry pick, but the game isn't anymore what it was back then. In that interview they stated they were going to re-invent the game and they did... it isn't a hide and seek game anymore either.
2 -
*Your argument is this is not a team game, yet a Survive With Friends was introduced the year after that interview.
?
The interview was posted on November of 2018. SWF arrived in Patch 1.0.3 in July of 2016. What in the world are you talking about?
*The system doesn't care whether you are a pre-made or not, the introduction of SWF is literally the introduction of the concept of survivors being on a team. They threw away the anti-social approach of survivors when they added the feature.
As I told the other forum goer, SWF simply allows friends to play together - nowhere does it imply that they have to be a "team." Just like in solo-que, they must form a temporary alliance, but that isn't actually a "team." Of course, players can choose to pretend that they are in a team, and proceed to play the game that way, but that is simply a player problem, not the game's. Sharing of information among SWF does allow for better awareness of when to help each other (so that they themselves can escape), and when to give up on them (so that they themselves can escape) - it's definitely a huge advantage, but that doesn't make them automatically a "team" nor does it change the win condition.
For good measure, let's review what the survivor's win condition is again:
Under the current in-game tutorial: "Survivors: How To Win - The primary objective of a Survivor in a trial is to escape."
On DBD's official webpage (under game objectives - how to win): "As a Survivor, your main objective is to escape the map and live another day."
In either case, nowhere does the game's win condition for survivors indicate that they are a team (e.g. "Survivors win if 2 or more survivor escape the trial in a match.")
*You can cherry pick, but the game isn't anymore what it was back then. In that interview they stated they were going to re-invent the game and they did... it isn't a hide and seek game anymore either.
Sure, they can change the mechanics within the game - sometimes strengthening the chase-aspects of the game, and other times the "hide and seek" aspects. Hell, if they want to, they can add "search for a generator parts" mechanics or "fight the killer" mechanics - but all of those mechanics won't change the overall goal of the game, which for survivors, is simply to escape themselves, and has nothing to do with helping other players survive.
All I've done is give you exactly what DBD's game director has said on the matter - if you have a quote (and not just some inferences) from an equally reliable source that states that the game's win condition for survivors rely on them being a "team", you are more than welcome to do that. Otherwise, it just sounds like you are arguing for what you personally want DBD to be designed as, not how it actually is.
1 -
This is about the statement win condition not changing, the quote of the same game developer within the same interview:
"One of the difficulties is that the game does not have a clear-cut win/lose condition. It's more of a gradient of did I have a good game, do I feel like I’ve accomplished much, was it fun?"
Are you going to use only the cherry picked aspect of the same interview that falls in line with your argument or look at the entirety what was said at the time? You use it as an undisputable point, while ignoring literally the rest of the interview.
The fact of the matter is that the developers have avoided for years to create a clear message on what a win/lose condition is. They now state it is escapes and kill based and compare it to a team sport. Their choice of a comparison, there are enough solo sports that it could be used and yet they picked a team sport. There is a reason the emblem system was created to look at the different aspects to let you climb, it was about more than just kills/escapes and tried to embody the gradient.
The definition of the feature of Survive With Friends is a direct statement of them trying to survive together, therefore your pretend to be a team is simply inaccurate it is an active choice of the survivors, one the game explicitly offers to the players. They join as a group, a party, a team to face off against a killer, to try and survive with this group. Also, simply read up on the perks;
- Strengthens the potential in your and your team's Aura-reading abilities.
- We have to work as a team, I need you to survive so that I can survive!
- You are able to organise a team to cooperate more efficiently.
There are tons of references that survivors are/can be part of a team, claiming that you are all lone wolves going at it alone is simply not true. Can you decide to be more selfish and a lone wolf, sure. Yet survivors can choose to cooperate and work as a team. Does the current system hold any regard to treating the survivors as a team, it is an option that the game presents to us after all. Yet none of the plays, risks and actions that correspond to working together as a team are equivalently rewarded when successful. The win solely looks at the lone wolf approach.
People talk about teammates, not other sole survivors and you can dismiss this all you want. Claim it is a misconception of the player base, but these are the people that play the game. These are the people that for a large part have been extremely negative about the MMR system and are the people that the company relies on to keep playing. They listen to the community, according to the article that you yourself mention and that the product is no longer just theirs, but all of ours.
We’ve learned that this is not just our game anymore, there are a million people playing every month and these people are part of what the game is and what the game should be. And they should have a say in what the future of the game is. So we pay quite a lot of close attention to what comes from our community.From the developer himself. Well guess what... people aren't happy, people are complaining that if they do well and feel that they had a good game, that they have accomplished much and their team as a whole has more escapes than those that were killed... if they are the sole survivor that died, they are not granted the win like the rest, they are told they lost. The vast majority of players will look at the end result, a 4k, a 3k, a 2k, 1k, 0k and understand whether their side won or lost. Just like with hockey, people get the concept of an overall score.
The win condition didn't change right? To many people it has changed, significantly punishing those that play in the other survivor style; as a team. People say they cannot remove SWF because, that is their money maker, that is what the game runs on... guess what these for the vast majority see themselves as a team. I am not disputing any of these claims, SWFs are a vital reason why the game is where it is today.
Look at it from a business, how many people approach the survivor end of the game as the lone wolf choice versus how many approach it from the team choice? Who exactly do you want to keep happy and have the system account for? You can support both, but that would require expanding the parameters, considering solo performance on top of just the kills/escapes at the end of a match (in my opinion the better option). Yet if the developers want the easy clear cut win/lose condition of kills vs escapes, the team approach would be the better and smarter move.
2 -
DBD is absolutely a team game. Survivors all have the same objective and they need to show some level of coordination and altruism in order to succeed. That's not to say there aren't selfish players, but the core concept is a 4v1, not four concurrent 1v1s. This is true for solos and SWF alike.
2 -
*Are you going to use only the cherry picked aspect of the same interview that falls in line with your argument or look at the entirety what was said at the time? You use it as an undisputable point, while ignoring literally the rest of the interview.
If you want to look at the full context of the interview without cherry picking it, that answer was provided specifically in regards to its failure in the e-sports scene. And yes, when you look at the Killer's win condition, unlike the Survivor's win condition, it doesn't have a clear-cut win/lose condition. (DBD website - "The Killer's main objective is to please The Entity by sacrificing Survivors on Hooks found in the environment." - it doesn't specifically mention how many sacrifices are required for a win.) As such, it doesn't really work out in regards to esports/tournament scene. But Survivor's win condition has always been very clear cut. And it never had to do with the total number of survivors being able to escape.
*The fact of the matter is that the developers have avoided for years to create a clear message on what a win/lose condition is.
Actually, win condition has never changed. In its official webpage, as well as in the tutorial, it has always been straight forward and clear cut. However, you are absolutely right that developers have always muddled this fact - it's a reflection of the lack of confidence they've had in this game design from the very beginning. The only reason why they have become more honest about its win condition now is only because they found out that the wincon of the game must be directly correlated with its matchmaking system for it to have any worth. Otherwise they would still be hiding it under the rug - disgusting.
*They now state it is escapes and kill based and compare it to a team sport. Their choice of a comparison, there are enough solo sports that it could be used and yet they picked a team sport.
Haha, you are reading way too much into it. They could have picked ANY pvp game in the world and the analogy would still have worked - hocky was simply the first thing it probably popped up in his head. The message is simply about how the meaning of "skill" in games are directly related to its win condition, and nothing else.
*There is a reason the emblem system was created to look at the different aspects to let you climb, it was about more than just kills/escapes and tried to embody the gradient.
If you believe that, then you must also understand that there is also a reason emblem system is no longer used for matchmaking.
*The definition of the feature of Survive With Friends is a direct statement of them trying to survive together, therefore your pretend to be a team is simply inaccurate it is an active choice of the survivors, one the game explicitly offers to the players. They join as a group, a party, a team to face off against a killer, to try and survive with this group.
Sure, you can aspire to survive together - but it has nothing to do with the win condition, which is for each individual to find a way to escape. Like you said, it's just a feature - not a win condition.
*Also, simply read up on the perks
You are right - some of these perk description probably should be fixed to more accurately reflect the game design, and not the lore - since it's probably confusing to people who are desperately looking for a reason to believe that it's a "team" game.
*Can you decide to be more selfish and a lone wolf, sure. Yet survivors can choose to cooperate and work as a team. Does the current system hold any regard to treating the survivors as a team, it is an option that the game presents to us after all.
That's just it. Players can be selfish and a lone wolf - they can choose to cooperate and work as a team. There are plenty of options within the game. But at the end of the day, what is the win condition? What is the ultimate goal? What's the objective? Simply to escape.
*Yet none of the plays, risks and actions that correspond to working together as a team are equivalently rewarded when successful. The win solely looks at the lone wolf approach.
Of course not. The cooperation and the alliance in the game are definitely rewarded - if done correctly: it's rewarded with the player's personal escape - because that's the entire point of playing. And the lone wolf approach? - again, if done correctly, it's rewarded with his/her escape - again, because that's the entire goal. For most players, it's not one or the other - they have to figure when it's the correct approach within the match, and when it is not.
*People talk about teammates, not other sole survivors and you can dismiss this all you want. Claim it is a misconception of the player base, but these are the people that play the game. These are the people that for a large part have been extremely negative about the MMR system and are the people that the company relies on to keep playing. They listen to the community, according to the article that you yourself mention and that the product is no longer just theirs, but all of ours.
And that's why it has become such a problematic game - instead of being true to its game design and sticking to it, the developers started relying on its players to define the game for them. Wincon is the most basic tenets of any game design - yet they compromised its existence for years. No wonder players are unhappy - they are being told "no" for the first time.
*The win condition didn't change right? To many people it has changed, significantly punishing those that play in the other survivor style; as a team. People say they cannot remove SWF because, that is their money maker, that is what the game runs on... guess what these for the vast majority see themselves as a team. I am not disputing any of these claims, SWFs are a vital reason why the game is where it is today.
Look at it from a business, how many people approach the survivor end of the game as the lone wolf choice versus how many approach it from the team choice? Who exactly do you want to keep happy and have the system account for? You can support both, but that would require expanding the parameters, considering solo performance on top of just the kills/escapes at the end of a match (in my opinion the better option). Yet if the developers want the easy clear cut win/lose condition of kills vs escapes, the team approach would be the better and smarter move.
To me, that's a player base problem. Or rather, the quality of the player-base. Those who never truly loved or accepted this game for what it is. But yeah, you are right; I don't really look at it from the business perspective. Nor do I care about its player-base. The only thing I care about is being true to its game design - and its win condition has never been changed.
...so no quotes from the developers that would back your opinion that this game's win condition have to do with Survivors being a "team", eh? Bunch of inference and guesses, eh? I see. Like I said, it sounds like this is more about what YOU want, and not the game actually is.
1 -
Survivors each have the same objective of individually escaping. And to do that, they need to cooperate and show altruism to a certain point until their own survival is ensured. What is your point?
0 -
The game does not frame the survivor's objective as an individual one. It remains agnostic.
Many people, including myself and the others replying to you in this thread, treat this as a team game. Even if you disagree, you should agree that those who play selfishly should statistically hurt the overall survival rates in their lobbies, meaning that their presence in a match increases the chance of the killer completing their objective regardless of whether or not they survive. I feel they should be punished for that from an MMR standpoint.
Designing the MMR system as a 1v1 rewards selfish play as much as it rewards skill. An MMR system designed as a 4v1 would do a much better job of putting similarly skilled players (i.e. players that do more gens, chase longer, save their teammates more reliably, etc.) in higher skill brackets. So, even if you're the guy leaving me to die on first hook every match, you should want an MMR system that doesn't reward that. You'll just end up going against killers who are better than you.
1 -
It does reflect player skill, it just isn't evaluating your personal skill based on one game. Skill in this instance isn't being used to refer to how skilfully you evaded the killer, or even how much skill you show in the macro game of going for saves at the right time and working on the right generators, but rather your skill at consistently winning.
Any one play you make could be a fluke. The system isn't interested in that, it wants to know where you're at overall, and that's where the controversial statement from the devs come in- if your skilful plays aren't consistently getting you to win, then the system doesn't view them as skilful because the point is to achieve a roughly 50% chance of winning the game it puts you in.
Regarding your other points about how the system could work, it seems a bit complicated for not much reward? We don't have any reason to disbelieve the devs when they say the number they get from MMR is accurate, we know which part is broken and it's not the kill/escape based number, so why bother putting in all that work when what we have seems to work?
2 -
*The game does not frame the survivor's objective as an individual one. It remains agnostic.
What is agnostic about:
- Under the current in-game tutorial: "Survivors: How To Win - The primary objective of a Survivor in a trial is to escape."
- On DBD's official webpage (under game objectives - how to win): "As a Survivor, your main objective is to escape the map and live another day."
Nowhere does it indicate helping other survivors escape. Or perhaps you are concluding that because it doesn't say anything about helping other survivors escape, we should make random assumptions about it? Maybe because it doesn't say anything about tea-bagging the killer ten times to win, we should all just assume that it's also part of the win condition?
*Many people, including myself and the others replying to you in this thread, treat this as a team game.
You treat it how you want; the game allows all kinds of behavior, bad plays as well as good plays. It doesn't have anything to do with what the win condition is though.
*Even if you disagree, you should agree that those who play selfishly should statistically hurt the overall survival rates in their lobbies, meaning that their presence in a match increases the chance of the killer completing their objective regardless of whether or not they survive. I feel they should be punished for that from an MMR standpoint.
Why do I care about how the opponent does in the matchup? The objective of the game is for me to escape, not to make sure the killer gets least number of kills as possible.
*Designing the MMR system as a 1v1 rewards selfish play as much as it rewards skill.
MMR system will only reward selfish plays if the selfish plays lead to the player's escape. If the selfish plays end up in their own death (or they have to take the hatch), then it's not rewarded in any way. Part of "skill" is in knowing when selfish plays will help them reach the goal of escaping, and knowing when cooperative plays will help them instead.
* An MMR system designed as a 4v1 would do a much better job of putting similarly skilled players (i.e. players that do more gens, chase longer, save their teammates more reliably, etc.) in higher skill brackets.
Why should this "selfish" player who consistently is able to achieve his goal of escaping be stuck with players who constantly do gens/chases/saves but always end up losing/dying?
*You'll just end up going against killers who are better than you.
As I should be, until I reach the level of killer in which that same tactic won't work on anymore. What's the problem?
1 -
The current system is evaluating "skill" and, hell, even "winning" based on a definition of skill that most players would not accept. Would you consider it a "win" if the killer 3ks but you escape? I certainly would not. I'd say "we got stomped, but at least I got out". The current system is also saying that the teammate who gives up in a winnable match and instead hides in a locker until the hatch spawns is actually more skillful than the people actually trying to win chases and do gens. It's saying that the teammate who never goes for saves during the EGC is actually more skilled, no matter how free the save is or how many teammates they could have helped escape.
I'm sure the number the devs are getting from MMR is accurate enough, but only at predicting whether a given survivor will survive the trial. It does nothing to predict how. As a result, you'll see "skilled" survivors who happen to play altruistically with the same MMR as much less skilled survivors who play selfishly. By "skilled" here, I mean a pretty generic definition that an average player might use; survivors with good game sense, who do objectives, don't waste time, and chase well.
The result of this that these skilled, altruistic players get consistently screwed over by their teammates. MMR creates frequent situations where skilled, altruistic players run over-matched killers for five minutes and then get left to die on the first hook, because their teammate was too scared and/or selfish to go for an easy save. That's one of the main problems with the current system. If you play DBD as a team game, then, your reward is frequently getting paired with selfish potatoes against weaker killers. It's a much worse experience than the old pip system was, where the only concern was "are they good (but actually good; not just selfish), do they play a lot, or both"?
My other comments get more into the unfun behaviors that the current system incentivizes: leaving your teammates to die the second you know you can get out, slugging for the 4k, camping the first hook, etc.
---
I don't think it would be very hard to change the algorithm to consider team survival instead of individual survival. Just add/deduct a certain amount of points from the killer based on the average MMR of the survivor team, then proportionally add/deduct those points to/from the survivors based on their individual MMRs. The only thing I can think of that could be tricky is that survivors can theoretically requeue and even finish a new match before the result of the previous match is final, but it shouldn't be that big of a deal to work this out. One solution would be keeping the MMR completely normalized around a set number, for example, and basing your analytics on percentiles rather than hard numbers. Not ideal, but an acceptable floor. Lots of room for improvement over my off-the-cuff brainstorming, especially when doing nothing is hurting gameplay.
1
