SBMM will never be effective in this game as long as match format is asymmetrical

Just because it's an asymmetrical game does not mean you can't use metrics to try and measure skill and then match players without a big skillgap. It just isn't as easy as in a symmetrical game. Currently "skill" is measured by the outcome of a trial, instead of taking the plenty events into consideration that happen during a trial, that's where changes to skill measurement could begin (which then would be utilized by SBMM).

No it actually could work pretty well

For Killer it’s really simple, the more hooks you get and the more Survivors you kill the more MMR you gain, you could also add stuff like how fast your chases are, how well you kept gens stalled, etc.

For Survivor it’s also fairly simple, how long did you keep the Killer busy in chase, how many heals did you provide for your team, how many gens did you do, etc.

And also showing people their MMR and how playing in certain ways affects it, that would also be really nice.

However, the current MMR system doesn’t factor in any of this, for Killer it only factors in kills and for Survivor it only factors in escapes, its purpose is to get people into matches faster, not accurately put people of similar skill levels together, that’s why everyone complains about it, it’s a random number generator wearing the skin of a SBMM system.

I don't know that your examples here really prove much, definitely on the killer end. Tackling that one first, the question that needs to be asked is "how should a matchmaker treat these players", and we've got some fairly clear answers for most of them.

Killers one and five, the sweaty Nurse and the sweaty Trapper, are trying very hard to win with multiple shortcuts. A matchmaker must assume reasonable balance, so it must assume those shortcuts get easier to deal with higher up the skill curve, so those players should climb until they face proper resistance for their tactics. This is what should happen with the current system (setting aside situations where the matching portion hiccups, anyway), so no issues there.

Killer two, our cracked Billy, clearly wants to practice their power more than they want to outright win, and their results reflect that. A player like that should be able to maintain a reasonably stable position in the matchmaker but they are ultimately getting subpar results, so a slow decline until they actually improve or level out would be appropriate. Again, that's what should happen here.

Killer three, the memeing Myers, has basically the same argument to be made, save for that they care about memeing around more than winning. Their results reflect a stable position, so they'll just keep trucking on. Annoying for people who face it sometimes, sure, but that'd be the case wherever in the matchmaker they land.

Killer four is the biggest potential outlier, but at a certain point, you have to accept that players who lose on purpose will get to lower their matchmaking level. Trying to stop this would be a fool's errand, even if a few reasonable safeguards could be put in place, hypothetically.
At least our example here is someone who lets the players go, so not too disruptive to the ecosystem, so to speak.

For survivor, sure, it's a bit messier… but they're all engaging with a clear playstyle that gives them clear results. Sable will drop until they can improve in chase, and survivor 4 will continue to drop because they're just not really contributing much, but that's mostly appropriate- the thing being measured is how consistently they can win the game and therefore what kind of resistance they'd put up to the opponent, and none of these examples really show someone who's slipping through the cracks in either direction.
Sure, the last example is an annoying player, but that's not a matchmaker problem- in this example they aren't climbing, they're just sitting at a stable position.

I could see survivor 4 maybe being an outlier, but at a certain point you should be expected to engage with the basic objective progression and the basic gameplay loop before being considered skilled even if we move to only measuring performance in one game, right?

I prefer'd the emblem matchmaking, contribution based matchmaking like this without a cap would probably do much more than Escapes/Kills. People would be playing against players who play more efficiently for points and get out/sacrifice, and be punished for tunneling/camping and genrushing/avoiding chase too much.

that way the toxic playstyles stay in the middle, and those who play fair go to the top, and people have to literally outright do nothing on purpose to fall down/die early.

but that's just me being idealistic it probably wont work like that

I like how many different reasons people came up with to highlight the problems with matchmaking in DbD.

Anyway, I'll throw on a couple:

Having one side get MMR increasing events ~60% of the time, and the other ~40% of the time. This could balance out, but it creates more complexity.

Having one side with unclear win conditions - play as a team, or play for yourself, BHVR is cool with both, but the MMR system only likes one.

I think that's generally what they should do (though average survivors MMR is tough, if you had hypothetical MMRs of a 9//8/8/3 hitting a 7 killer, the 3 is likely to drag the survivors down way more than the other survivors outranking the killer).

I've always thought the idea of designing a VS game without a win condition a fool's errand, but have to admit DbD has been incredibly successful with that formula. I think BHVR wants to keep the idea that sometimes the game will be a 4 person SWF going for a 3 out, and sometimes 4 solos each playing for themselves, so balance is always going to be an issue (and that they prize long term balance over individual trials). I think awarding MMR via overall team result would help, but do worry this could create an unforgiving low MMR hell where players who aren't very good are trapped with rats as teammates.