My suggestion on reworking survivor MMR
So here's a thing that I half-agree with Patrick: the endresults matter. However, how those endresults affect MMR, also matter.
In terms of ELO, the better performing players in a team tend to gain more MMR than their teammates if they are underperforming. Or to even go back to Hockey Players, even if a very high valued player loses the game, their contract price might still increase because of their overall performance.
So here is my suggestion to MMR:
Include the escapes and deaths, BUT spread out the MMR over players based on how well they distracted the killer. That way, if there is a survivor looping the killer for 5 gens, and 3 teammates escape(the 5 gen looper dying in this scenario), the survivors who distracted the killer the least wont gain much, if any at all(or even lose some), while the dead survivor gains most of it. This might be a bit harder to implement than the current MMR system. The math shouldnt be too hard tho, as you can ratio it across survivors, the hardest part is ofcourse applying the MMR after the match has ended since a survivor can die very early on and find another match before their old game is over. But in general, the survivor with the lowest chase time should lose the most/gain the least mmr, rather than the first one to die, and the survivor with the highest chase time should gain the most/lose the least mmr.
Any good working MMR system for an asymmetrical game should never be an easy implementation of "kills vs escapes" and just be done with it, because its not a 4v4, its not a 1v1. Maybe a useful implication would even be to apply the build up MMR everyday at the same time, or apply everytime a user logs out, or both. In the long run, it's not gonna matter if you get MMR on a match by match basis. If your MMR changes over time at a fairly accurate manner, that, on long term would create better matches, and more accurate MMR representation than the current system.
I'm not saying its a horrible system, it just has to compensate for the fact that it's asymmetrical too.
Take Smash Brothers Melee tournaments for example. Losing a game doesnt mean you can no longer win the tournament, because there are things like RNG affecting the results, while in normal tournaments, losing a game often means you cant get 1st place anymore. They HAVE to adjust the tournament rules because of those factors. DBD isnt a standard game, so why is the standard MMR considered fine? Yes, its gonna take more work to fully fledge out the mmr system, yes, its gonna take some headscratching, but the fact that players have been jumping ship because of MMR alone kinda indicates that MMR isnt working as expected. If MMR worked as expected, you wouldnt see a drop of 20%. It might still be a drop due to seasonal fluctuations, but it's never been 20% because of that.
Comments
On their Q&A stream they showed they don't intend to change MMR system ever in the future, because the lead game designer thinks the current system is perfect and flawless, and not a single change can ever make it better. So you are suggesting ideas to the wall.
You're wrong in interpretation there. They dont want to change the MMR system, that being, to another system. That doesnt mean they dont want to improve the current system here.
Context matters. When talking about not wanting to change the system, they used the hockey teams as an analogy for why they think endresults matter more than chases and hooks. Which is true, a killer getting 4 kills in 4 hooks often means that survivors heavily misplayed or the killer did really good. Either way, its an indication of difference in skill regardless of the hooks. As for survivors, it doesnt matter if 1 player got chased for 5 gens, if 3 survivors die, they still lost.
The problem with their MMR system, is that it doesnt take asymmetry in consideration. Where 1 survivor could be mismatched for whatever reason, and losing a lot of MMR because his teammates dont help. This is where the MMR system fails.
You could argue more hooks equals more skill, which could very well be true, but in either scenario, if you were able to get 8 hooks, but no kills, you clearly werent skilled enough to kill someone, or you didnt kill someone on purpose.
So no, my suggestion isnt to a wall, as their biggest issue with MMR is players leaving the match because they died early and MMR currently being coded to update once you leave the match. Because there is a chance that 1 player finishes 3 matches, loses 2 very early and wins 1 quickly, while the 2 losses havent finished their game yet. Making matchmaking a literal clusterf*ck if you apply it like that. So applying MMR changes later, or at a set time, wouldnt really screw with it.
So to reiterate again: My suggestion doesnt change the MMR system, it is just meant to improve the current one.