Home Dead by Daylight Forums Discussions General Discussions

SBMM doesn't work. Killer stats from 50 matches to prove.

testertester Member Posts: 617
edited December 2021 in General Discussions

Disclaimer: I'm killer main and only play survivor for Rift challenges. I played every match in sweatiest possible way, for the purposes of keeping MMR high for the test.

Disclaimer 2: This is not a complaint. I don't want to argue. The only purpose of this test is "for science" of determining effectiveness of SBMM. Please post your personal feelings and experiences. I'm not saying SBMM good/bad and not looking forward to argue about it. I'm only trying to determine SBMM works/doesn't.

It used to be different. SBMM worked... somewhat... But then, something happened after that Epic games giveaway. My killer matches became very easy. No, I didn't become spontaneously better. But I wasn't sure if I'm imagining it, so I decided to gather stats over last few days, playing variety of killers. He are my stats over 50 matches.

Average kills per match 3.02

Average Win/Loss ration is 4.33, where win is 3k or 4k and loss is 0k and 1k. 2k is draw and not counted.

Fun stat: most DC per match is on Twins. People just hate being slugged I guess.

So, my question is: Weren't we promised that SBMM will keep our average kills per match at 2? Who else is feeling that this is not the current reality?

Edit: Also, my survivor matches are super difficult. Considering I barely play one... I just got matched with killer with over 9k hours... In solo queue. Looks like there are not enough low MMR killers.

Post edited by tester on
«1

Comments

  • AkumakajiAkumakaji Member Posts: 1,528

    Interesting, you beat me to that. I am currently recording my own games and got 26 so far, and my average kill rate is 2.66

    I plan to do about 100-150 games, and I also keep taps on how many DH, BT, CoH etc meta perks I encounter.

  • testertester Member Posts: 617

    Holy crap! I'm not the only stats nerd here! Joy! Looking forward to see your stats when you get to that number of matches. But even what you have (2.66) is already not what we are told to expect from SBMM. Do you also feel that matches are easier as of last few weeks? Or do you feel no changes?

  • AkumakajiAkumakaji Member Posts: 1,528

    After a rank reset I get a few very easy baby matches, but this ones aren't in my records, because I only later decided to take notes.

    After a short while it stabilizes and I have a mix of chill, normal and sweaty games, but as a somewhat decent player, I seem to win most of them.

    After most people, who care, get into their Iri ranks, the games also seem to become more chill on the whole.

    But yeah, I guess with the influx of new players everything got a bit more chillaxed, I got the feeling that I had to sweat more and harder in the past.

  • edgarpoopedgarpoop Member Posts: 5,256
    edited December 2021

    Shhhhhhh killer is impossible due to MMR. And if you aren't losing, your hidden number can't possibly be as high as the gamers around here and reddit.

    Sarcasm aside, the matches really aren't that much more difficult on the whole. My worst game in the last two weeks is a 2k on Clown on Ormond and I play almost every night.

    I think there aren't that many players at the top end on either side, so the odds of getting a true high MMR vs high MMR match are very slim. There's always one or two very weak survivors, or the killer is well below the skill level of the 4 stack because there aren't that many high MMR killers. SBMM puts a much higher emphasis on match quantity over match quality

  • ScottJundScottJund Member Posts: 1,041

    i dont think you had to do a study for this

  • AkumakajiAkumakaji Member Posts: 1,528

    Well, you are wrong with that, my friend. Gut feeling and "I actually think it works this way" are as varied, as their are players and only numbers allow any kind of educated considerations.

    So let is all write down a few numbers and compare notes :>

  • ohheyitsbobcatohheyitsbobcat Member Posts: 1,254

    While your data is interesting it doesn't work all that well for an MMR does/doesn't work topic since every killer has different MMR. The data would be much more valuable if you had played 50 games with the same killer and got a similar result. If that was the case then you'd have some basis in claiming MMR doesn't work properly.

    Early DC's also shouldn't be counted in the stats since it skews the data a bit. A 1v3 is almost impossible to lose as killer unless you really mess up.

    I do think MMR should be tweaked and looked at more, I just don't think this data proves anything.

  • AkumakajiAkumakaji Member Posts: 1,528

    Your objection is valid, just one thought: didn't BHVR say that MMR is just split between Killers and Survivors, but consistent within that group? I think they are somewhat adjusting for perk slots, as I noticed that a lvl 5 killer got a higher chance to play against other babies, but beyond lvl 15, you should be back to your normal MMR.

  • testertester Member Posts: 617

    "The data would be much more valuable if you had played 50 games with the same killer and got a similar result. If that was the case then you'd have some basis in claiming MMR doesn't work properly." - Please explain. I never play same killer. Are you saying that SBMM is not supposed to work for people who always switch killers? I don't understand your reasoning.

  • testertester Member Posts: 617

    Wow! It's like we have same mind :) I also get the same weird feeling when reading reddit/forums. And I also had same suspicion that SBMM prioritizes matching speed. My match difficulty is like on the clock. Time after dinner is more difficult than any other times of day. Could be due to SWF or could be due to different player composition.

  • Barbarossa2020Barbarossa2020 Member Posts: 1,029

    Well it only proves that in a region that had a free giveaway the sbmm maybe be skewed.

  • testertester Member Posts: 617

    That's my feeling, but I can't test in other regions easily to compare. Wait... I have VPN. Would that work for being matched with other regions? Which regions didn't have free give away?

  • Barbarossa2020Barbarossa2020 Member Posts: 1,029

    Im in Australia, and bought dbd via steam, but i'd wager that Epic games has little or no presence here.

  • ohheyitsbobcatohheyitsbobcat Member Posts: 1,254

    Depending on how much you play, kind of. Every killer has a separate MMR value tied to them meaning that for some killers your MMR will be higher or lower then others. Your Spirit could have close to 1900 while your Trapper might be 1500, we don't know. As such, every killer needs to be treated separately for their W/L averages and your data which shows only 1-2 matches for the majority isn't very good for showing anything, other then you played multiple killers and won a lot.

    It'd be like if I took a match or two where I 2k'd and tried saying MMR works great just because of that. 50 matches with the same killer is a lot more valuable as far as data is concerned.

  • ScottJundScottJund Member Posts: 1,041

    Your data doesn't matter, unfortunately. No one knows their MMR and yet will continue to justify all their losses due to their MMR. It doesn't matter what you show them.

  • testertester Member Posts: 617

    I still don't get it. I played over a hundred hours from the start of MMR. Are you saying that even that amount of time is not enough to normalize MMR if someone plays different killers? This data is from last 5 days. My match difficulty is consistently same for the last few weeks.

  • AkumakajiAkumakaji Member Posts: 1,528

    Some data is better then no data or gut feeling.

    You are right, I don't know my MMR score, frankly no one does. But the claim was that the system should be balanced around 2K, and the available data seems to indicate that this doesn't work as intended.

    Yet, most of us experience a change in the quality of their matches, ie way more sweaty and (BAD WORD) to the walls kind of matches and a very loud majority (I assume here) wants the old system back, as they can't cope with the constant stress of sweaty game after sweaty game.

    And the only condolence are nasty comments the like of "hur dur that's on you. You are too good, should have played worse."

    By collecting any kind of date (assuming that we don't lie) at least gives us some way of interpreting what's going on and quantifying it.

  • testertester Member Posts: 617

    There is no intention to prove anything. My intention is to determine if SBMM matches us with people of equal skill. My data suggests it doesn't. But it's just 1 person. Therefore intention of the post is to ask others to contribute, so we have more data.

  • DemonDaddyDemonDaddy Member Posts: 3,330

    Mmr only effectively works for survivors, killers are placed wherever mm needs them to be whether it be up or down.

  • edgarpoopedgarpoop Member Posts: 5,256

    Which kind of brings up a more baffling question: what purpose is it serving for BHVR to hide MMR? We're pretty much all assuming it doesn't work or sucks anyway at this point. We can't give effective feedback on something we can't really prove exists. If they at least showed us our MMR, we could tell them "hey, these matchmaking conditions don't really work when this range is used" etc. But they have it set up so we can't tell if it's even on half the time and everyone is pissed off anyway.

  • testertester Member Posts: 617

    My guess is the purpose of hiding it is that they know it doesn't work and they can't fix it, so they just don't want you to know and complain that it doesn't.

  • testertester Member Posts: 617

    Doesn't even work as survivor. Just tried one match and got matched with Spirit with 9200 hours of experience... In solo queue. And I'm barely able to play survivor.

  • ohheyitsbobcatohheyitsbobcat Member Posts: 1,254

    Every killer has their own separate MMR value. If I took two back to back games where I 2k'd and tried arguing MMR works would you agree? Of course not which is the point here since most of your games are only with each killer 1 or 2 times.

    Every single killer has to be taken as a separate value for W/L. You can't mix them together in a single pool and then say MMR does or doesn't work since they're not meant to be taken together.

  • PulsarPulsar Member Posts: 12,792

    Separate MMR is basically meaningless, I've pretty much confirmed it.

  • ohheyitsbobcatohheyitsbobcat Member Posts: 1,254

    Maybe but it doesn't change the initial point that having a mixed pool of killer data means just as little. Unless you're trying to say there isn't separate MMR in which case I heavily disagree.

  • PulsarPulsar Member Posts: 12,792

    If there is, it matters very little.

    I played a game as Nemesis, someone who I have put a considerable amount of time into and I got a good 4 man. They definitely pushed me to play my best.

    I BOUGHT Freddy after that game. I had never played him before. I didn't even level him up. I queued for a match and was nearly instantly put into a lobby with the same 4 people.

    I also did this with Pig on a separate day.

  • testertester Member Posts: 617

    Yep. I can confirm this. Same experience. I consistently matched with people of same skill with my best and worst killers.

  • ohheyitsbobcatohheyitsbobcat Member Posts: 1,254
    edited December 2021

    I'm pretty sure they've said or its been mentioned at least that when you play a new killer you get a sort of average of all your killers MMR instead of at the beginner level. I know the devs had said they don't want experienced players to go up against new players even if they're playing a killer they've never played before. I think they mentioned that it can take a bit for ones MMR on a killer to settle as it throws you back and forth to try and see what level your at. I might be misremembering though.

    On my end, I can say separate MMR works. As Huntress I played around 30-40 games and I either farmed or lost on purpose in all of them as I was working on the skilled Huntress trophy and was not going to do that against good players. By the end, I was getting some of the worst players I've ever seen. After getting the trophy I switched back to Oni who I play the most and got ######### on. I remember it specifically because it was such a shift from just messing around and losing on purpose.

    This was back in the beginning of November or so. I have also since played Huntress a couple days ago for a daily and they were still pretty dang bad while the match I just had as Legion today was rough as hell.

    It's all anecdotal of course but in my experience separate MMR works well enough. It could be better when you start a new killer, I give you that. My first couple matches as Artist I got destroyed before it settled to normal groups.

    Edit: Here's a screenshot of what I was trying to say in my first paragraph. I'll see if I can find more where they've talked about it.

    Edit 2: Another screenshot talking about the same thing. I wanted to find one where they talked about it bouncing you back and forth when you start but was unable. That one I might just be misremembering.

  • LoneSlingerLoneSlinger Member Posts: 500

    Did you miss the part where he said that he got matched with the same exact people

    Even if there was a average that either means one of 2 things

    1: the brackets are large asf and mass people get grouped together

    Or 2: by average they mean it's within 50-200 of your best killer which still sucks ass

    I wouldn't take anything they say about this mmr too seriously or as a fact

    They went on for years saying this was gonna be a grand system that was more complex then kills and escapes

    Then boom 2 years later and what is it based on oh that's right kills and escapes truly revolutionary

  • ScottJundScottJund Member Posts: 1,041
    edited December 2021
Sign In or Register to comment.