Am I reading this right? BHVR essentially flat out lied about SBMM?
I supported SBMM from the get go. From the latest testing, my experience was overall positive. I loved the idea of it, too. I thought one of the main points of it was to more accurately assess your skill by taking multiple factors into consideration aside from kills/escapes/emblems
Iirc, a developer themselves even said they tested mmr purely by kills/escapes in the past and it sucked. So.....what's different now? I thought this new algorithm was supposed to be smarter. I thought that the game would finally be able to better assess the skills of a 0K killer and a first-dead Survivor but guess not.
I'm willing to not use the "L" word regarding Pinhead because that is likely out of BHVR's control(shame, they go above and beyond once and somebody else has to come in ruin all that hard work and effort) but this doesn't have that same excuse.
What happened to the SBMM logic from the last test? Because I'd be willing to bet two Double Cheeseburgers that the SBMM we have now is completely different, more specifically they trashed everything at the last second and replaced it all with a burning tire.
I really thought lazy, low effort play styles would be punished with MMR but I suppose I was the real sucker this whole time?
Welp, now I'm feeling kinda spiteful, not gonna lie. This whole time I've been going "easy" by playing full aggro Legion and Nemesis with 0 gen regression. I guess after 7PM I'm just gonna play Demogorgon only because he is IMO the best Anti-SWF killer in the game with the upcoming Spirit changes and is an easy 4k machine.
Comments
It looks like it. I supported it, and still do because my games are better, but yeah.
One of the reasons I might have more chill games than others is because I usually let one (or more) survivors go especially if they didn't do very well. And if it's through the gate...I might not have actually done them any favours :/
Technically, the system is using match duration to factor how much a person gains or loses. Not that it means much since the end result is still the same, if you escape you go up and if you die you go down.
Considering the old rank based matchmaking, it was a realy low bar to improve upon. So I'm not really surprised that basing MMR off kills/escapes is still a better system.
After all, if you're consistently getting kills/escapes, then you're obviously doing something right, and your opponents are obviously below your skill level.
Yeah. I also thought this. They did say it wouldn’t be only escapes/kills but now..
i defended their take on SBMM as well. My matches did get better or at least not worse. But.. having these contradicting statements and having such a basic system after such a long development time.. this feels underwhelming.
Have to agree with this.
The new system is still better than the old one, but jesus, did it have to take so long?
-post deleted-
In the end of the day my matches got better so I don't really care.
I do hope that everything they said about SBMM being death/escape based was a lie to keep us from finding how it actually works. Even going so far as to hide fake data in the files to be found. I really do hope it is like that....but I doubt it.
I just deleted my post because I’m pretty certain they’re actually being real…
When you put it like that, this is kind of embarrassing for them. So many years to come up with this. Ha ha ha...
Remember that the 'old rank based matchmaking' was actually 'no matchmaking' for killers, for pretty much the entire last 9 months before SBMM. It was causing endless queue timers, and a developer confirmed that they had indefinitely disabled killer matchmaking back in...February? I believe.
SBMM has some annoying features, but man - that was miserable.
I certainly hoped that it was a bit more sophisticated than 'escape = MMR increase, death = MMR decrease' as this works completely contrary to how genuinely skilled players actually behave. If you lead the killer on a 2 minute chase leading to the last 3 gens being popped, but then get clobbered by NOED, your MMR should certainly not be penalized.
All this is does is encourage selfish behavior from survivors, and frustrating 'hide and seek' play against killers.
I knew the system was broken when all my survivor matches were just awful. I usually sacrifice myself for my teammates and it makes sense now that my teammates are just consistently extremely bad. I guess I gotta play selfish now lol
The lie I'm least pleased with was that SBMM was supposed to set you at a low level for new killers so you could try out the new guy's power without getting stomped over and over again. Some killers like Blight, Nurse, Pin Head, and Oni have REALLY steep learning curves, and you're not learning how to use their power if every time you try to use it, the SWF sweat team evades you easily and finishes the game in 3 minutes.
Putting brand new, never-before-touched killers at an AVERAGE of all your other characters is miserable. It flat out fails to do the one thing that Killer-specific Matchmaking was supposed to fix. At this point, you may as well just turn killer-switching back on, because every killer's going to be more or less the same level of Survivors teabagging your every missed move.
The new system seems better than the old system, which is a plus to me. I do remember them saying they wanted it to be more complicated, but, just because the version they ended up with is simple doesn't mean it's wrong. Sometimes you try something complicated and it doesn't work.
I also think there was an element of struggling to define what would constitute a good match, and it seems like the definition they settled on is that you have roughly a 50/50 change of surviving as survivor and roughly a 50/50 chance to kill each individual survivor as killer. That's a really clear outcome to build a system around, and I bet it took a while to define it and then to run tests to determine, for example, that you had to remove hatch from the calculations and account for changing probabilities as the number of survivors goes down.
I also remember them doing a survey where they asked people what a win was, and I'm willing to bet that the results validated this system -- I'm pretty sure most people said "win" means you escape or get a certain number of kills.
TL;DR it's simpler than what they initially proposed, but I still think it's okay.
Look, I get it if you like the current system, that really isn't something we can or should argue about. But don't defend the blatant lies we've been told. SBMMR was made out to be this complex system that would do its best to calculate your skill, it isn't, it never was and we've been fed nonsense for months and months.
SBMM is Survival based matchmaking, not skill based.
Also, they said:
I don't consider this a blatant lie, though (and I don't know why it would matter if it were, but that's a separate issue). They slightly overstated how involved the calculation is, but it's true that they don't just use kills and escapes to determine the rating, and it takes several factors into account (we know death order and hatch are two of those factors), the goal is to have you win half of your matches, and, from some points of view, the calculation is complex (we don't know exactly how it calculates the killer MMR, for example, because it has to factor in four 1v1s).
If you're disappointed by what the system turned out to be, or if you imagined something different when you read that statement, that's fair. But I don't think it's a lie.
If you view this entire situation as them "slightly overstating how involved the calculation is", then there really isn't much to talk about. The way I see this, is yet another Corporation playing the "how do you do fellow gamers" shtick left and right, while aiming for as little transparency as possible, because that makes them impervious to any negative feedback. And the tone in which this was explained on the stream really pushed my buttons.
Don't get me wrong, it would be incredibly naive to even assume that Behaviour would be any different, but I suppose it's the fact that it's never been made more obvious. Am I a little bit salty about the entire situation? Yes, yes I am.
My guess as to what happened: When they set out to make MMR, they fully intended on factoring in a number of different things pertaining to skill. I imagine they realized at some point that survivor especially is extremely nuanced, and to accurately reflect a survivor's performance is to delve into Sabermetrics while at the same time confront the map RNG elephant in the room.
Not only that, but your success as a survivor is very dependent on your teammates, and to accurately grade one survivor's performance, you need to take all survivors from that match into account. That creates an issue where your MMR from the last match is still being calculated while you're already dead and queued up for another match.
We can quantify some things to a degree. Efficiency can be loosely quantified in chase, generators, etc. But then you have things that are harder to assign a numeric value. A player lurked near a hook, didn't save, got found, and also got downed near the hook. The game is completely lost. How do we quantify that numerically in regards to the other players? The other players are dying because one player threw, but that's an obvious one.
So here's a harder one: a 3v1 with one gen left, and two survivors choose to double up on a gen but the chase doesn't last long enough to finish it. Match ends in a 4k. Was the chase an issue, or was the problem the decision to double up on the gen instead of splitting two separate gens and guaranteeing one pops? How do we quantify decision-making in that scenario?
It would be awesome if MMR could grade survivor efficiency vs region average. But I think a lot of people don't realize what they're asking. That's an absurd statistical formula to develop. Don't get me wrong, I think just doing escapes and kills is too simple, but a perfect system would be a wild ask.
Me too support it since my games are better because I play casual on both side.
Survivors: I play with only Lithe as 2nd chance, happens to die more. And for that I vs Killers who happens to kill less for playing fair (I had many matches where Killer Mori first survivor at 8th hook (each had double hook already). The game is far more enjoyable for me even if I die, than escape vs tunnel/camp killer
Killers: I dont play for kill, but to feel power trip with End game build. Kill rate is inconsistent so I keep go up & down.
Because escape & kill are goals for many players, the system is stupidly work well. I find its unfair of Killers who play that fair in my games should get higher "skill base rank" than Killers who do everything to have 4k such as tunnel,camp, slug for the last.
While this is probably the case, the frustrating thing is that people were saying this from literally day 1. You can't have a functioning MMR system in this game without taking into a billion different variables: killer, killer perks, survivor perks, survivor items, survivor addons, killer addons, map, map RNG, map strength for killer, pallet usage, etc. Its not that we're disappointed they couldn't do it. It's that we told them they couldn't, then laboured away for almost 2 years to then tell us they couldn't.
Sure.
But the system should absolutely not be increasing the MMR of a survivor who potatoed on gens for the entire game while penalizing their teammates who ran interference, looped and saved expertly but happened to get mumped by NOED at the end.
This is why it being a hidden score was a big red flag to me. If it was a well put together system that legitimately rewarded and at least semi accurately determined skill then it being known to at least yourself would not have led to any major abuse. Also they just straight up lied about many things concerning mmr the past few years so that's cool.