What MMR Update Doesn’t Address
Just to express my concerns to either get ahead of the curve or make myself look like a fool later.
I’m not going to pretend that I was looking forward to the updates for MMR with how BHVR discussed it. I think there are three reasonable paths forward for an MMR system - go back to emblems, keep as is, or switch survivors so that they gain/lose MMR based on the total result of the group (each has pros/cons depending on what you want the game to be). I think BHVR’s approach, which they do a lot, is an overly complicated idea that won’t actually address the core issues (most likely - backfills, varying levels of sweatiness per match, and time to find a match).
We also don’t really know anymore about MMR after the livestream than we did before. We knew they were going to factor in additional things to the MMR, after the livestream we know basically the same, we just have a few pictures. We have no numbers, no specifics on weights, just the vague idea of what will be calculated.
That said, here are common complaints/issues that the MMR change won’t address:
1: I ran the killer for five gens and died so I lose MMR? - still probably true, their example (Dauntless Dwight) has a player doing everything and they ‘may’ gain MMR. The 5 gen chaser was only able to gain ‘points’ in one category (or two if they count passive generator repairs), meaning they likely get an MMR loss and everyone else an MMR gain (unless one survivor just sat around doing nothing).
-this also seems to be a larger problem with this form of measuring ‘skill’, doing a bunch of things okay is valued a lot more than doing any one thing really well.
-it's also worth noting that if this player does get downed and rescued, the other players may gain more MMR than them because they did more things.
2: Ratting - the change might impact the players who rat from the very beginning, but the much more common version, ratting once things turn (or even look like they might have turned), the player still has as much, if not more, reason, to rat for hatch. As long as they’ve done some things (some saves, some gens, maybe one okay chase), if they now rat for hatch they likely gain MMR, not lose it (presuming they don’t change hatch, but they haven’t said anything to indicate they are).
3: It’s not the emblem system - One goes up, one goes down, survivors’ and killer’s ‘points’ get compared at the end of the trial, this isn’t like the emblem system where both sides could increase. Reasonably, that feels like if both the survivors and killers have a good match, the escape/death will still be the deciding factor in who gains and losses.
4: It’s still a secret - If you get matched with a ‘bad’ player, what happened? Did MMR screw up? Was matchmaking too quick? Or did MMR work properly but that player(s) just wasn’t taking the match seriously, trying a different build, drunk/exhausted, etc.?
I don’t think the system is going to be a failure, I just don’t think it will feel very different. There’s valid concern that certain tactics on both sides might allow skilled players to maintain low MMR and beat up on weaker players. It's possible adding in additional factors will allow larger/smaller MMR swings to distinguish close and blow out trials, but they haven't indicated that being true or even a goal (or whether MMR moves enough per trial for it even to be impactful). And maybe people will just like the ‘feel’ that they had a good match, so maybe their MMR didn’t change.
But beyond that? It doesn't seem like it will actually change things with a few potential dangers to allow it to be manipulated.
Comments
I think that's a fair assessment of how the MMR update will go, given what we know about how it works. I think a lot of players have similar concerns. But I do think it will be an improvement for somewhat experienced SoloQ players who find themselves in matches with relatively inexperienced teammates. I'm sure we've all had matches when we've done 2 or 3 gens, saved everyone from hook and had at least one decent chase but died in the end because nobody else was pulling their weight. So it might give players who find themselves in low MMR hell an opportunity to get better future teammates simply by progressing the game.
I feel like the direction of travel for the game has been towards making it a more competitive experience overall. It's not what I personally think is good for the game, but it does seem like that's were the majority of the player base is heading these days (which I think is a generational shift as much as anything else. Kids these days don't have the history that older gamers have of having to sit next to the guys you're playing multiplayer games with. I think that's given them a stronger competitive drive on average than past generations because in the old days if you stomped your opponent too hard, you'd quickly run out of people willing to play with you (especially so if you opted for the cheesiest strategies. Anyway, I digress…).
Nothing so strange about that, most multiplayer games are inherently competitive. But what I do find strange is BHVR's continued insistence that we still shouldn't know what our relative skill level is. It's still completely hidden from us in the new system, making it difficult to really know if it's working as intended or not. I don't think they will ever do a deep dive on how the system really works. They seem to prefer that players don't know an MMR system even exists at all.
That doesn't say much for BHVR's confidence in quality of the updated system.
I think what you need to do is separate out MMR and Matchmaking - both of those things are not the same.
These things are examples of what is covered with Matchmaking - except perhaps varying levels of sweatiness which falls into a different category on its own tbf.
With kills and escapes, only those were the outcome for the current MMR which was missing a massive part of the game, hence we looked at the system overall and looked at what were the important factors in the game.
The point of it keeping hidden is ultimately the number is meaningless to the vast majority of players, who play the game casually and do not care if they're high MMR or medium MMR (there's only a very small portion of players in that high mmr bracket after all). We don't want players to only focus on MMR and play in a way to maximise it, as its a number that's used for matchmaking only.
Are there any plans to look at the lobby itself, in particular dodging and shopping around? Imo that too need to be addressed at the same time, as it too often leads to the rough backfills that causes lots of problems.
I think what you need to do is separate out MMR and Matchmaking - both of those things are not the same.I believe that's what I was doing, saying things other than MMR are the potential core issues.
And while they're different under the hood, the end result to the player is the same. If one of those is the best designed system in the world, and the other is broken, the end result doesn't work (and players have no way to know what went wrong).
If the problem is something else, fiddling with MMR is just window dressing to the problems.
With kills and escapes, only those were the outcome for the current MMR which was missing a massive part of the game, hence we looked at the system overall and looked at what were the important factors in the game.Just for clarity, is BHVR walking back the logic that was done for the old system? That the reasons given for why just tracking the end result (hockey analogy) are no longer believed to be true?
Because I'm one of those who thought it made a lot of sense and have defended it, but if it was wrong, I'll take the loss if BHVR's evidence says it wasn't working.
The point of it keeping hidden is ultimately the number is meaningless to the vast majority of players, who play the game casually and do not care if they're high MMR or medium MMR (there's only a very small portion of players in that high mmr bracket after all). We don't want players to only focus on MMR and play in a way to maximise it, as its a number that's used for matchmaking only.1: Are we going to learn the exact impacts of various MMR factors, or will those be a secret? Like what percent comes down to the escape/death?
2: I understand why MMR is not publicly shown, I don't even disagree, but its still the same issue that when a match up doesn't feel right people will blame MMR. The new system isn't going to fix that.
Its one of the things on my list, but that would be when we look at matchmaking overall
I agree for your final point there, players will still blame MMR - that's an unfortunate fact that happens in pretty much all games. There's no system that's perfect and its especially hard in an asymetrical game.
Kills and Escapes will still matter, its just that now they're not the only thing - it should address other things in a match, and I think that's been a lot of what players have said over the last few years. There are definitely big parts of that which do make sense - kills and escapes are one of the main constants in the game - so I don't think we're disagreeing here.
The game has changed a lot from what it was a few years ago, as have the players. We recognised that we could improve the system on how we calculate players MMR and we've worked to do that. Its been running in the background for a while as a kind of test (not impacting games, not used for matchmaking etc) and the data is looking good on it so far. It might need tweaking when its Live, we'll be keeping a close eye on it and making any adjustments as necessary.
Does the team have any intent to modify the stat tracker page to include team escape rate or any information from the new 'tally' MMR system, now that performance is team based rather than individual on the survivor side? Assuming I've even understood that correctly, I may have gotten that wrong.
I think it will still be calculated per individual survivor. And the previous MMR did take into account your performance compared to teammates if you were in a party (i.e. less MMR gain if your whole party escapes, since you have more of a chance they will rescue you)
Hm very good points, that means you will get bad MMR if you dont farm points from now on?
Performance is still not team based but I would love to see team escape rate. Let's actually treat survivor abandons as a death while we're at it.
I think many of us would love to see team escape rate.
Its not a team based MMR, though some would prefer that as well. Doing more things, such as rescues and heals, will count toward the points earned, but whether those players escape or not I don't 'think' has any impact.
I have a feeling this is something that can't be answered, but: will we the players, even if we don't know our own MMRs, actually know how the MMR breaks down (weighs) the various activities. Or will the exact specifics be a mystery, just that its more than escapes/deaths?
there's no plans to give the exact maths, no current plans at least. I can't speak about future as that could change, but for right now its not even in discussion.
Thank you, I appreciate it. I think its better for the community that everyone knows that for discussions going forward.
As discussed above, we won't know the exact breakdowns. Potentially? It really depends on how they weigh things because its NOT the emblem system where both sides could go up.
People have already voiced the worry about slugging killers and gen rushing survivors staying at lower MMRs. As I mentioned above, on the survivor side I think doing a lot of different things okay will be given more MMR than doing one thing really well, but whether that is an appreciable difference is impossible to say if things run in the background. That does create the potential there will be really amazing chase survivors (as an example) who never do anything but chase who stay at lower MMR than they would under the current system.
On the killer side it feels like a tunneling /end game build could get a lot of 3/4ks, but also not pick up much under the Gatekeeper/Malicious 'emblems'. Again without weighs impossible to know the degree of impacts.
There's going to be a lot of wait and see, which will be followed by many, many guesses.
There are plenty issues with balancing the game. If we make killers balanced around high mmr swf then why would you ever want to queue as solo survivor if it'd be just a kill simulator. On the other hand if we balance killers around low mmr survivors then there will be no chance to win against high mmr swfs. That's why later killers are designed well. The balancing point should be Jason/Dracula level where the killers are fairly good but still have reasonable counterplays. Killers like Judgment or Vecna those who dictate the outcome by missing or hitting no matter the skill of other side. Then there is group of isnanely boring killers like Knight or Trickster who are so unfun to go against yet they are quite weak.
From what i understand this wont ever happen, so any change to mmr is meaningless because ignored by matchmaking so streamers get their 1000+ win in row against newbies, rarely encountering any good player. This also means the majority of the playerbase should be low mmr, and the game balances around the top 10% survivor players with communication and being premade, with thousand of hours.
They have tightened matchmaking before and, predictably, experienced players hate playing against other experienced players. They prefer weaker opponents. So there were loud complaints and YouTube videos. They also had a testing phase once and the preferred level/s of matchmaking for a lot of folks unfortunately increased queue times for others (ie low and high MMR).
I think this new system might be even worse, maybe now even if you did a bare minimum and didnt escape, you still wont lose mmr and easily gain. Why is this a problem?
Because according to new mmr everybody outside of the absolute worst will be "high mmr". Either you gain easily, or you dont lose mmr easily, same thing.
I dont even understand why people care about the invisible mmr, so the system kill or escape was good enough.
The problem was always that streamers hated when had to wait games for say 10 minutes to get according players to skill level, so mmr ignored completely, just so the top few percent get instant or fast matchmaking, and then there is dodging, which destroys matchmaking anyway.
So what will the be impact?
None. Or getting worse.
I think this new system might be even worse, maybe now even if you did a bare minimum and didnt escape, you still wont lose mmr and easily gain. Why is this a problem?That I don't think is correct. MMR still works off one side up, one side down, and (I think on this part) a 1 to 1 survivor to killer comparison. So if the survivor goes up 10, the killer goes down 10. The issue is whether what they have is actually anymore reflective of skill than what existed.
I dont even understand why people care about the invisible mmr, so the system kill or escape was good enough.This is true. I wouldn't even be making this post if BHVR had never promoted they were doing it, built up hype around it, and had a livestream about it. It's made players expect something better, but we aren't actually going to learn how it works, and things will still probably feel the same. If it had been an off handed change - "hey, we're tweaking MMR to incorporate some additional factors into the rating at the end of the match, we think this will yield a slight improvement" - I wouldn't care. But they've made players expect something that I think will leave them sorely disappointed.
But I don't think queue time is a valid reason to matchmake weak players with strong players.I don't think any of us do, but BHVR seems to disagree.