A good Ranking System

JimmyCasket

As I hope most can agree, the current ranking system is terrible. The emblems are still broken as all heck, and the MMR is making people tryhard for some reason.

As I hope some of you have seen, Scott Jund made a review vid a few years ago, that mentions a return to BP based ranking, but upping the requirements to 20k, instead of 10. 30k would obviously be a double pip. Under 10k would be a de-pip. This would make the gen rush and leave playstyle obsolete, and incentivize interaction with the killer.

Same would go for the killer in my opinion. Would need at least 6 hooks, and 20k to pip, camping or tunneling would decrease BP gains for related actions, like injuries/downs and grabs would all be 1/4th less. This would discourage the camping and tunneling strat, and encourage a more fun playstyle that can still be effective.

dugman

I actually suspect MMR is relatively accurate, the reasons for mismatches have to do with how matches are assembled and not the MMR itself. Some examples

• If someone is a high MMR with fewer players close to their rating then they will tend to get timed out into “next available” slots with random people. Likewise if a game has a lobby with four people and one who left the lobby, it will tend to grab the next available person to fill the slot to keep those four players from waiting a long time. Essentially it’s a question of finding “best” matches versus “fast” matches, and sometimes it prioritizes making fast matches instead of making you wait even longer for close ones to come along.
• High MMR survivors often play with lower MMR friends in swfs. So you can end up with apparent mismatches where the game is trying to find a killer in between the two extremes.

I think average players who play at peak times tend to get closer matches more often than players with very high MMR and who play at off hours. Note that prolific streamers and forum posters happen to be in both categories, they will tend to have very high MMR and being prolific streamers tend to play at unusual times, so they unfortunately also would bear the brunt of a lot more mismatches of the types above than someone who is kind of middle of the road and plays at prime time. That in turn drives them to complain to their viewers about MMR being broken, and here we are. 🤷‍♂️

There’s no actual statistics though indicating MMR itself is inaccurate. Individual ratings could easily be really accurate and the mismatches are from how long matchmaking waits for good matchups or how it calculates swfs with disparate ratings, etc. There’s no evidence the old emblem system was more accurate, and in fact because those reset every month it was definitely worse in part in the weeks after a reset since it became a wild west of people all jumbled into the same brackets.

RaSavage42

Then again one side forces the other

If the Killer don't injure and hook Survivors then there's no Altruism... and if the Killer doesn't chase Survivors then there's a limit on Boldness... and vice versa... If the Survivors don't heal or unhook then again there's little Altruism... and if the Survivors don't want to be chased then there's a limit on Boldness

We all want to "win" matches so we do what we can to ensure our "win"

BubbaDredge

@dugman said:

I actually suspect MMR is relatively accurate, the reasons for mismatches have to do with how matches are assembled and not the MMR itself. Some examples

• If someone is a high MMR with fewer players close to their rating then they will tend to get timed out into “next available” slots with random people. Likewise if a game has a lobby with four people and one who left the lobby, it will tend to grab the next available person to fill the slot to keep those four players from waiting a long time. Essentially it’s a question of finding “best” matches versus “fast” matches, and sometimes it prioritizes making fast matches instead of making you wait even longer for close ones to come along.
• High MMR survivors often play with lower MMR friends in swfs. So you can end up with apparent mismatches where the game is trying to find a killer in between the two extremes.

I think average players who play at peak times tend to get closer matches more often than players with very high MMR and who play at off hours. Note that prolific streamers and forum posters happen to be in both categories, they will tend to have very high MMR and being prolific streamers tend to play at unusual times, so they unfortunately also would bear the brunt of a lot more mismatches of the types above than someone who is kind of middle of the road and plays at prime time. That in turn drives them to complain to their viewers about MMR being broken, and here we are. 🤷‍♂️

There’s no actual statistics though indicating MMR itself is inaccurate. Individual ratings could easily be really accurate and the mismatches are from how long matchmaking waits for good matchups or how it calculates swfs with disparate ratings, etc. There’s no evidence the old emblem system was more accurate, and in fact because those reset every month it was definitely worse in part in the weeks after a reset since it became a wild west of people all jumbled into the same brackets.

I see what you're saying, mismatched trials are better than waiting forever for the perfect match. The system is working to keep queues down.

The problem with MMR is that kills/escapes are (from what I understand) all it looks at. An easy 4K doesn't make me a good killer, a 2k against a really good team does. The best survivor isn't always the ones that escape, often it's the opposite.

The emblems recognize different aspects of play, including kills and escapes. The best players are the ones that accomplish the most, all we're looking at are results. I didn't learn to play this game thinking "kills are all that matter" and I don't really want to.

Tiufal

Whats should a BP based ranking change? As survivor its not in your power how much BP you get, its under full control of the killer player. So a ranking system of this kind is completely impossible to establish.

The actual MMR system doesnt reflect the playstyle of DbD. All that matters is 4k and Exit-Escape. Thats just a sooo small part of how games can end, that its inaccurate by default. If youre a teamplayer and want to get out your mates, your doomed to fall in MMR to the bottom if you sacrifice yourself for your team. On killer side, if you just play with all those "legit strategies" getting 4k is pretty easy, but doesnt make you a skilled player and you will just rise more and more until you get extremely outskilled. You can just compensate much skill with those "strategies", that are dumb to no return.

The emblem system was the easiest way to make a good base for a multilevel matchmaking system. Tweaks to them would have been badly needed, but shouldnt been so hard to implement, mostly. Then you would just add playtime/player level/devotion to ensure new players arent matched with veterans, and you basically have a full working matchmaking (if you think of 4 solos being matched). For SWF the matchmaking would just have been returned to the good old average, thats what worked best the past years.

dugman

@BubbaDredge said:

https://forum.deadbydaylight.com/en/discussion/comment/2724241#Comment_2724241

I see what you're saying, mismatched trials are better than waiting forever for the perfect match. The system is working to keep queues down.

The problem with MMR is that kills/escapes are (from what I understand) all it looks at. An easy 4K doesn't make me a good killer, a 2k against a really good team does. The best survivor isn't always the ones that escape, often it's the opposite.

The emblems recognize different aspects of play, including kills and escapes. The best players are the ones that accomplish the most, all we're looking at are results. I didn't learn to play this game thinking "kills are all that matter" and I don't really want to.

I know the devs have said that the system looks at more than just kills and escapes. To clarify, if a survivor is killed then the killer's MMR goes up and the survivors goes down, and vice versa, that much is true. However, if I recall correctly, how much the MMRs go up or down depend on other factors, including

• the relative MMRs of the players
• whether it was the first or second or third or fourth kill in the match (i.e. once one or two survivors are killed is becomes significantly easier to get another kill, so the MMR doesn't change as much on later kills as it does on that first one)
• I seem to recall mention that how long it takes to get a kill is also a factor in how much MMR goes up or down. I think a very fast first kill increases killer MMR more than a kill that took a long time to get (maybe? Hopefully I remember that right.)
• Every killer character also has its own MMR. So as above, depending on your relative MMR against a survivors, if you get a kill on a killer you're not very good with that has a lower MMR it will increase more than if you get a kill on a killer you are great with that already has a higher MMR

In the end kills and escapes are the ultimate goals, so that's why they determine if MMR goes up or down. Yeah, playing for hooks or getting in chases or whatever is fun, but those are tools to help you get the goals of escaping or killing. For instance the emblem system also gives killers extra points if they kick a ton of generators and pallets and walls, but I think you'd be hard pressed to argue that a killer who kicks all the generators is "playing well".

So to sum up, kills and escapes determine if MMR goes up or down, but how much it goes up or down depends on other things. Emblems are a fun side note but they're not really a better indication than the MMR of what constitutes a "good player" and the old matchmaking that relied on them didn't really produce better results than the current one in my opinion.