MMR Experiment Reveals That The System Is Faulty
Ever since MMR came out I had always questioned if it actually was working relatively smoothly. For those who don’t know how MMR works, watch Otzdarva’s video on it; it explains it in really simple terms.
Anywho, I (like many people) noticed that ever since MMR arrived, my killer games are a lot more sweaty. More SWF running toolboxes w/ brand new part, medkits w/ styptic, Boon totems, DS, Borrowed Time, etc.
Since the bloodhunt event started, I decided to solely grind bloodpoints and not even worry about kills. And since I didn’t care about kills, I decided now would be a good time to run a little experiment: play every game as a single killer, play normally and get everyone to death hook, but then just let them finish gens and escape. I did this on Doctor just because he’s fun and I don’t mind playing a ton of games as him. So far I’ve played a little over 30 games as Doctor where I let every single survivor go.
For context, I’m a pretty average killer. Maybe slightly above average, but nothing significant. I play a big variety of killers and am pretty average at all of them. So that should give you a general idea of where my MMR would be at.
The weird thing is, though, that my games haven’t changed at all. I was expecting that after 6 or 7 straight losses that I’d start facing some potato survivors, but that never happened. 10 games, 15 games, 20 games, and eventually 30 games went by where I saw no difference. In fact, I’ve recently had some of the sweatiest games I’ve ever seen before. Yesterday I faced a 4 man SWF all running soul guard, for the people, Boon: circle of healing, and borrowed time, with 2 of them running toolboxes with brand new part, and the other 2 with medkits.
No matter how many games I lose, I keep coming up against 4 man SWF that are obviously not low MMR. Some games they may not be SWF but still have 1 or 2 teammates that are insanely good and can loop me for days, and some games every survivor is just really solid overall. Occasionally I’ll check profiles and I’ll see some of the people I go up against have 500, 1000, and even a couple times I saw 1500 hours!!!
And before you ask, no, I’m not playing at times of day where there aren’t many players, which usually causes matchmaking to pair you up with whoever is available even if they’re vastly different in MMR. I play in the evening when matchmaking is instant and you don’t ever have to wait more than 30 seconds. And no, my Doctor MMR was not particularly high before starting the 30 game losing streak. I’m not bad at doctor, but I’m definitely not winning matches all the time.
So all of this is what leads me to believe that MMR is not working how it should be. When you lose matches, the MMR value should go down, and the matchmaking system should pair you up with survivors with an average MMR similar to your own. After more than 30 consecutive losses you logically shouldn’t be going up against 4 man SWF that can CJ Tech you and pull off crazy spins, and the fact that this is the case means that Behavior didn’t actually make the MMR system the way they described it. I also think that’s why you’re not allowed to see your own MMR value.
Behavior knows that if they let you see the actual numerical MMR value, people would call them out as liars after seeing things like a killer at 600 points being paired up with a team of survivors each around 2500 points (or vice versa).
Has anyone else noticed this problem? Is this more of a killer problem than a survivor problem? Has anyone else seen or done any additional testing on this topic? I’d love to hear other people’s opinions on this.
Comments
-
Just to add even more evidence to my claim: I just played a game as Huntress to complete a daily challenge. I have played DBD for 3 years and have only played Huntress 6 times (this game being my 6th game as her). I’ve lost every single game I’ve ever played as her.
The survivor group I got paired up with was a 4 man SWF where every single one of them was iridescent (one of them was a rank 1 survivor). 3 of them were running green medkits w/ addons, one of them with a commodious toolbox w/ addons, 3 of them running Boon: Circle of Healing, 2 people running Prove Thyself, and every single one of them running Borrowed Time.
I obviously got destroyed. They knew how to run tiles, spin, and just play well in general. It’s simply not possible that they were anywhere close to my MMR. Behavior has lied to us about how the system works to match people up together.
It’s a shame that you can’t expect honesty from them in terms of game mechanics. With all this continual lying on behalf of Behavior, VHS will overtake DBD in no time.
1 -
The post is long, he can’t be wrong.
6 -
A few things (woof that was a long read, even by my usual waffling post standards).
Let's go point by point.
- I'd guess that we are at a similar MMR. Intermediate. I'm not sure what time of day you are playing at, or what region, but my games are a lot better now.
- I still get the occasional weird match. This is usually either due to smurfing (usually SWFs sandbagging their MMRs by repeatedly suiciding/AFKing to stomp at low levels), someone having a long queue time somewhere and me being the 'closest' killer or the SBMM system being seemingly offline entirely (usually characterized by absolutely instant queues in quick succession).
- 30 game losing streak...I honestly don't know if that's possible. I've had losing streaks before, but nothing close to that since ranks went away. My longest has been a 6 game streak that saw me paired up against lowbies the next day, and I was learning a brand new killer.
- That said, there does seem to be a possible 'hell' region, below standard 'low' MMR where a lot of the really dedicated smurfs congregate. I think I found it once while I was learning Demo.
- No, if you really had a 30 game loss streak you shouldn't be getting paired up with monsters. That makes me think that there's something...off here. I'm not sure what. Because I know that if I lose several games back to back, I'm almost guaranteed a string of pretty decisive victories.
- I have a theory that MMR does not always update correctly and can take a few hours to kick changes in. This seems to be the case, because I'll get stomped three times before bed, win once (I always try to go out on a win), get stomped 3 more times in the morning, go to work and the second evening rolls around I'm flattening people for several games in a row.
- That all said - yes. Sometimes I get paired up against people I shouldn't be - either people substantially below or above my skill level. That said, it happens a lot less frequently than it did before MMR.
0 -
I get the feeling that rank, while kind of meaningless now, plays a role in your MMR and what survivor kits you run into in the later ranks. Seems to hold true. My rank never goes down, because of the grade changes, and the survivors I run into are all using the same kits. Either that or there are other hidden effects at play for sure.
I'll say that ever since I hit high-ish MMR I have seen the same builds over and over with so little variance that it isn't funny. Doesn't matter if I get 4 games in a row where all 4 survivors leave. Nothing changes. So the only thing I can imagine is the grade system still plays a role in our MMR.
0 -
MMR doesnt work because the game isnt balanced. MMR will never work because the game isn't balanced. This isn't a hard concept
3 -
That...actually wouldn't surprise me. I was literally thinking this today,
1 -
Your guys are not taking something into consideration- Que times. Que times supersede MMR. If there are not enough High MMR killers to go around, it doesn’t matter how many games in a row you lose- you and every other mid MMR killer will get placed with a sweaty 1% Swf for that evening. Pay close attention to how fast your lobbies Que up when you’re killer. Especially if there are already 4 in there when you join or 4 immediately join after your lobby is created.
0 -
Losing 30 games isn't that many. Considering how long MMRs been going 30 games of losses probably offsets like... 2 weeks of MMR? They said there's no cap to how high you can go so it's not like you can be sitting at the top cap and lose x games to be at the bottom.
0 -
He just said he's playing during the middle of the day and que times are instant
I hate when people don't read the whole post
1 -
I've been saying this, I'm high mmr on Trapper currently
Yesterday I played pig who I rarely played and just farmed with survivors idk 10-15 games
Hopped on today first match was a all beamers all bt all deadhead all spinechill all circle of healing match
Love to see it but ofc nobody believes me and I typically get some guy telling me my games are made up
But ik they aren't cause I'm the one playing them lol
1 -
He just said he rarely plays Doctor so he shouldve already started at a lower mmr
And the devs even said for killers it's not 1v4 its 1v1v1v1
So therefore that's 30 games of losing 4 times the normal amount of mmr
Umm if it's working how they say it should he definitely should've lost a ton of mmr don't be silly
1 -
I just wanted to start playing after long break and got destroyed super hard and it looks I would have to play over hundred games to get normal survivors again... ######### that
My main killers are just unplayable for me atm.
0 -
The devs also said that killer-specific MMR hasn't been implemented.
There's also been no confirmation by the devs on any sort of values. Yes, they said each survivor is considered an individual win or loss, but they never said that that means a killer gains 4 times as much MMR for a 4k as a survivor does for an escape.
If 30 matches were enough to jump a significant amount then you'd be in high ranks one week and low the next just out of sheer luck.
30 games is not enough to expect a significant difference in an MMR system.
Also "losing 4 times the normal amount of MMR" is just wrong, since you ALWAYS have that same amount of MMR for grabs in a killer match.
0 -
"The devs also said that killer-specific MMR hasn't been implemented."
When the devs said that?
From what I know killer-specific MMR was promoted from start and I am pretty sure it is active.
1 -
Can't find the post but I'm fairly sure right after it launched they said something along the lines of "it's something we will look at in the future."
0 -
I've never heard then say anything about it not being implemented
I'm not saying they gain or lose 4 times a survivor wut I'm saying is if one guy escaped is -30 then four man escape is -30+-30+-30+-30 idk about you but to me 30 games of that sounds like a lot
30 matches definitely is enough to change mmr significantly cause inwoukd say it took me maybe 50 to 60 straight 3ks or 4ks to reach high mmr on Trapper so yea dude idk where your getting your numbers from
Also technically no if someone escapes through hatch then it's considered a draw and nobody loses or gains mmr
0 -
I don't think it's physically possible to treat every killer you haven't played as a brand new killer with zero MMR. That would lead to utter destruction and chaos for new players. More likely that they instead average your MMR rating for a killer you haven't played yet, then increase or decrease from there. So even when playing a killer you haven't played, if you're already in high killer MMR games, your average MMR will be slightly higher than the true average for this game. So the match will still be somewhat engaging and not completely one sided where you stomp newbies for hours on end.
0 -
I'm pretty sure it just takes your highest MMR Killer and then assigns that as your MMR for any new characters.
It worked that way when I first bought Freddy. Zero MMR, didn't even own him.
Played a Nemesis game prior to that, a good 4-man. Not god-tier, but definitely well above the average.
Level 1 Freddy queues up and is instantly matched with the exact same 4 man.
1 -
Well that kinda confirms something for me too because you said they weren't exactly top notch but definitely above average. And being a 4 man swf just obfuscates MMR altogether because most people in that party are going to climb... statistically speaking
0
