The second iteration of 2v8 is now LIVE - find out more information here: https://forums.bhvr.com/dead-by-daylight/kb/articles/480-2v8-developer-update

So when is SBMM getting an update?

2»

Comments

  • AndyKuky
    AndyKuky Member Posts: 84

    I like how someone at BHVR came in to correct someone saying the SBMM only takes wins into account and ended up confirming exactly what the person claimed 😆

    ”if you play four games and win two/lose two you’ll end up around the same place where you started!”

    Well that’s the problem. I can play four solo queue games and die with 20k+ BP in all four games and no other survivor will have more than 13k, even the ones who get 5-7k for escaping. I can then follow that up with four games where I escape every match with 11k points each and the system treats them all as either escape or die, win or lose. It’s dumb.

    Basically BHVR doesn’t know how to make a competent system because their game has too much variation and RNG. So the easy solution is win/loss.

  • Peanits
    Peanits Dev Posts: 7,555

    Definitely not that much. I don't have the exact numbers handy (and our MMR wizard is still off), but to give an idea, the distribution of players across MMR is a bell curve. Very few people are at the low end of the spectrum, most people are near the middle, and very few people are in the highest end as well. The soft cap will pull in some people who are already close, but not that many.

  • Ariiawa
    Ariiawa Member Posts: 6

    it seems it's just completely random at this point. As a survivor I'm getting a game where everyone's dead at 4-3 gens and the next game I'm looping a nurse for 4 gens before she gives up

  • dugman
    dugman Member Posts: 9,713

    I’ll be curious to see the actual estimate, when we took a stab at it in another thread I think we came out to about 6% are over the softcap based on a normal distribution with mean 1000 and standard deviation 400 (which was based on assuming there were about 1% over the old cap.)

  • Timmylaw
    Timmylaw Member Posts: 227

    Are people who haven't played in a long time counted for that? Or is it only active players.

  • Ciabe
    Ciabe Member Posts: 119

    Can you escape the soft cap once entered? You talk about mmr should be self balancing, but I'm literally at 300+ loss deficit on survivor and nothing is changing. When am I going to start seeing wins again? My teammates keep becoming bigger potatoes while killers stay mediocre to pro. Does it think I'm throwing/smurfing and not letting my rating actually drop so I can get better matches?

  • usesPython
    usesPython Member Posts: 121
    edited September 2022

    If you're on a 300 loss streak as survivor and your teammates are getting worse and worse you're not over the softcap, you're in ELO hell where your bad teammates get you killed causing your MMR to go down and get matched with worse teammates.


    Also you can escape the softcap once you entered it, it just basically considers everyone with a high enough MMR as the same MMR (e.g. if the softcap is 1600 MMR then someone with 1600 MMR is the same as someone with 3000 MMR matchmaking wise, only difference is that the 3000 MMR player can lose more games before getting out of the softcap)

  • zarr
    zarr Member Posts: 994
    edited September 2022


    Would really like to see actual numbers on this as well. I have recently learned that rather than increase the MMR cap over time, you guys have supposedly lowered it even more. Something like decent players sitting at around 2000-2200MMR, really good players around 2200-2500, and gods (cheaters) reaching an effective ceiling at around 3000. Yet the matchmaking cap is apparently 1600, making all of this irrelevant. If more than 5% of players fall within the cap, that's already too much, given that even the skill ladder jump from the 0.1% to 1% of players can be huge.

    I really think that for anyone even just decent at the game, the cap is almost entirely nullifying the possible benefits of the MMR system, because it's just too wide a range. And (soft) proof for this is that decent-and-above killers/4-SWFs still win about as often as they did in the rank-based MM system, so about 80-90% of the time or more, which is ridiculous. Even more ridiculous is that really good players can have win streaks of hundreds of games in a row, which is unheard of in pretty much any competitive game with MMR. While those streaks are of course very rare in the grand scheme of things, surely you have to regard them as issues, at least insofar they are indicative of major issues with the MMR system at the upper end. The matchmaking cap really does seem like old "red ranks" now.

    I think having a ~5min. queue time window at the upper end of MMR brackets is not unreasonable, so I think for the first 5 minutes of a queue, the matchmaker should strictly look for players within a close rating range of each other (100 to at most 200MMR). Players good at and dedicated to the game won't mind sitting in queue for a few minutes. If this is not accepable in middle or low MMR brackets, just increase the range for those, or decrease the time.