@EryxApexI agree. Just to be clear I do not suggest using level to determine skill with accuracy, I just used it as an example where I am fairly new to the game and playing with veterans, that should never be the case.
For me, 1/10 games is enjoyable, 1/50 is good. You are always going to have bad games, so the target for design should be 9/10 enjoyable, 1/20 good. That is very far off atm.
But where do we draw the lines? Keep in mind that "skill" is a continuous variable.. How can we design a SBMM that makes ALL players, at ALL skill levels, NOT have to play against someone much better than him? Its not possible unless we have a super narrow SBMM, but that is not practically possible since the queue times would be too long for both the high elo and low elo players since skill distribution is most likely normal distributed.
And I think the current SBMM will sort out the really bad players in a separate queue where new players or just really noobish players are allowed to learn the basics. But when you hit a certain skill level, probably measured in K/D, you will have to fight with the big boys.
I honestly feel that most SBMM complainers want public games to be like ranked games. So a much easier and better solution would be for Respawn to FIX RANKED by making it super hard to smurf in ranked. Then, people who want to play vs players his own skill could do that in ranked, regardless if he is in silver, or master.
I think the problem is that the player skill distribution most likely is normal distributed with a mean surprisingly low and a standard deviation probably also low. I wouldn't be surprised if the mean K/D of active players <1.
So lets say, for the sake of the argument, that they would create public lobbies of all players >2.5 k/d
Here are just some problems of the top of my head:
1) Super long queue times. Last season, only 4.68% of the active ranked players were Diamond or better, and you dont even need to have 2.5 k/d to reach diamond. So from that, maybe we can say that there are actually FAR LESS players than 4.68 of the player base with 2.5 K/D, and creating public matches with less than 4.68 % of players will probably not even be possible in the smaller regions (20 min queue time?). Is our goal to make it impossible for good players to play this game?
2) And even if these "high elo" public lobbies could be created; ALL players in them would drastically drop in K/D. We must remember than skill is relative. Take me for example. I have 2.4 overall k/d now but only 1.63 K/D last season I played ranked (reached D2) and this number is including my silver-platinum games (started in silver that season). My "diamond" k/d was probably far less than <1.
So how would this work in this SBMM you are suggesting? Bottom line is that if we have a narrow SBMM, EVERY player, regardless of skill level, would have a k/d going towards the same value (theoretically, the value would be 1 in super narrow SBMM) because, once again; skill is relative. So after a while, how would this sbmm know what player was good or bad, since good or bad is determined what opponent we are playing against.
But on the subject of ranked, I totally agree. Ranked is a mess and is designed to make people come back and play (ebmm?) , not to provide a competitive game mode. Remove the smurfs, remove or reduce soft reset, enable both demotions between divisions and fast promotions. When someone clearly is playing on a much higher skill level than his corresponding rank, there is nothing wrong with moving him up several ranks. If demotions where enabled, anyone who is moved up too far will eventually be demoted back down to the rank better corresponding to his actual skill.
Community ManagerCommunity Manager
