Visit the Kill Switch Master List for more information on this and other current known issues: https://forums.bhvr.com/dead-by-daylight/kb/articles/299-kill-switch-master-list
100 SoloQ Match Statistics, An SBMM Test
Hello there everyone!
After a recent exchange with Peanits where we briefly touched upon the issue of balanced games with the Skill Based Match Making and the statistics that are derived of the trials by the dev team ( https://forums.bhvr.com/dead-by-daylight/discussion/comment/3591596#Comment_3591596 ), I decided to do a little informal experiment and see how balanced my matches felt and what "objective" results I got in regards to the SBMM.
I believe this gives a rough estimate of what matches are like in the 1600+ MMR skill bracket, which is the second highest one, followed by the soft cap that was recently raised. I believe that based on my opponents and teammates in the past, many of whom I cannot possibly fathom to be under the 1600 mark (Otz, Hens333, True, KnightLight and these are just the names I recognize).
I usually play anonymous mode, but for this experiment, I made my profile public and played as Yui, who for the most part, doesn't seem to come with a lot of perceived stereotypes as I wanted my matches to be as general as I could possibly make them. I played all games with the same loadout (Lithe, OTR, Prove, We'll Make It) + a brown medkit. I never dodged and I played to win, which meant that if I could, I prioritized my escape, but was willing to throw myself at the killer if that resulted in the rest of my team successfully escaping. I also set out to only give up a match if 2 people DC-ed or went next on hook.
All matches were played a week after Chucky was released and before the Chill event began to sample the "usual" games that are played. This coincided with flashlights being disabled so results might be skewed in that regard.
Match results:
4k: 28 matches
3k: 12 matches
2k: 11 matches
1k: 24 matches
0k: 25 matches
I actually only died in 48 of the trials which resulted in a significantly higher escape rate than usual, which I consider to be the result of the flashlights being disabled and survivors being less likely to waste time trailing the killer. Still, for the original purpose of this experiment, I feel like it supports my perception of the SBMM system, most games are heavily favoring one side and "balanced" trials (however you define them) are exceedingly rare.
I even rated all matches subjectively on a scale of 1-5 where 1 was the P100 Nurse 12 hooking the team before a single gen was done and 5 being the one where the killer couldn't get a single hook by the time the exit gates were opened.
5/5: 14 matches
4/5:: 30 matches
3/5: 19 matches
2/5: 22 matches
1/5: 15 matches
Very similar results, although on this occasion, the balanced matches were not the least common. Obviously, this being completely subjective means that how accurate of a picture they paint is questionable.
Hook states are tracked in the spreadsheet document at the end of this post.
Tunneling
One statistic that I kept track of was tunneling. I had 3 categories here. Hard tunneling, which meant that the killer would focus on the first survivor they hooked exclusively, until they were either dead, escaped through the gates or the killer lost sight of them and had to pick another target. Soft tunneling were cases where somebody was tunneled immediately off the hook, but not as the first survivor that got hooked. For the purposes of this experiment, I did not consider tunneling after the last gen popped to be tunneling. Tally was as follows:
Hard tunneling in 47 matches.
Soft tunneling in 15 matches.
No tunneling at all or until end game: 38 matches.
Please bring DS and OTR to your matches!
Top 10 Maps
6 matches: Azarov's Resting place, Bedham, Haddonfield, Game, RPD,
5 matches: Ormond, Yamaoka Estate
4 matches: Crotus Prenn, Pale Rose, Rancid Abbatoir, Shelter Woods, Sanctum of Wrath
Top 10 Killers encountered
14 Wesker
9 Blight
8 Nurse
6 Chucky and Oni
4 Doctor, Hillbilly, Huntress, Legion, Nemesis, Spirit
Top 10 Killer perks
44 times: Pain Res and Pop
31 times: Corrupt
19 times: Bamboozle
18 times: Deadlock and Shadowborn
16 times: Nowhere to hide
15 times: Sloppy Butcher
13 times: Lethal Pursuer and Save The Best For Last
Top 10 Survivors encountered
24 times: Feng
23 times: Nea
20 times: Claudette
18 times: Nancy
17 times: Ace
15 times: Yui
14 times: Kate and Mikhaela
12 times: Dwight and Meg
Top 10 Survivor perks
130 times: Windows of Opportunity
91 times: Adrenaline
87 times: Lithe
73 times: Resilience
49 times: Off The Record
46 times: Dead Hard
44 times: Balanced Landing
42 times: Deja Vu
36 times: Decisive Strike and Sprint Burst
For the full list of matches, maps, hook states, killers, perks etc, here's a spreadsheet document:
And here you can see every end game screenshot I took, with all names removed:
Hope you found this interesting, let me know if there are any questions!
Comments
-
Interesting and kudos for the effort.
However one thing I find is always missing from these sort of things, is backfill.
You say you never dodged, but you don't say if (and half the time you can't tell so it's probably fair that you can't guess) you were ever backfilled into a lobby that effectively has NO matchmaking taking place.
There's no easy way to track this, but it's a limitation that is always going to skew these sort of stats. How many of the 'extreme' results actually fall outside of the SBMM system, rather than being symptomatic of it.
Admittedly this is probably a greater issue for Killers than it is for Survivors as their own matchmaking result isn't offset with any teammates. A backfilled killer is either a high MMR player getting an easy game or a low MMR killer getting stomped.
So you could for example track if any killers dodged your lobby and were replaced, as this could be the cause for many of your 0K results.
4 -
Right, I actually forgot to say this!
At first, I fully intended to keep track of dodging, but as you said, there was no way for me to know if anybody had dodged before I could see it, so I decided to just ignore it.
As far as I can remember, at least 3 or 4 times when more than 2 killers dodged, we got absolutely massacred and only one time was it either a draw or a win for us. Tiny sample size, but backfill seems to be a huge problem for SoloQ, giving higher MMR killers.
1 -
If backfills were such a small occurrence, that's actually fairly reassuring, that players aren't dodging quite so much any more. Which probably lines up with my more recent experience come to think of it.
Although just 1 dodge is enough to negate matchmaking, it needn't be 2 or more dodges in a single lobby.
0 -
I was going to ask this, but then I realized I could do it myself. I was curious how many streaks you went on as that seems to be my experience, I'll hit a bunch of games were the survivors get crushed, then hit a streak of games where everything feels fine. After doing the streaks I checked a couple of other things.
Your longest streak of dying: 4
Your longest streak of escapes: 9
Your longest streak of team losses: 4
Your longest streak of team wins: 5
Times your team won but you died (eg 1k and you were the one dead): 4
Your average rank: 2.66 (with 1.00 being 1 in all 100 matches and 2.50 being a hypothetically perfectly average score)
-I'm a little surprised your escapes are above the average, but point wise the score's have you being a slightly below average teammate.
Apologies if I made any errors, did counting and math while tired.
Question: What time of day did you roughly play during or was it varied?
Thanks for the data, is always interesting to see.
0 -
I love seeing this sort of data so thanks for sharing!
I tracked my own games for a good 6 months and it's interesting, I think by the end I saw a tunneler an average of 1/10 games. But I'm definitely not in high MMR, as you must be, so I imagine people try harder in that bracket than my own. My escape rate settled on 35%, which surprised me tbh. You had a pretty impressive result I must say.
1 -
I think I have an idea as to why my points were that low. Since I was trying to play pretty optimally, I assumed that SoloQ being SoloQ, I had to be cranking out the gens, so repairing took my number one priority. If a teammate's HUD was empty, I assumed they were going for the save and I kept on holding M1. I also didn't cleanse totems, search chests etc, trying to minimalize my time wasted, resulting in doing less and accumulating less bloodpoints. Of course I admit it might just be because I'm bad lol!
Matches were played on the weekdays 14:00-18/19:00 until 100% bonus was on Survivor. On weekends, 12-18:00 until 100& bonus was on Survivor. EU server.
0 -
Yeah, there's also the fact that if you look at my match stats, in the beginning, there were more matches were no tunneling took place. But as I kept on escaping, the tunneling instances went up and up and up.
Then again, I truly believe flashlights being disabled was a huge part in my escape rate, I escape significantly less often usually. Might repeat this experiment between events to see how much changed.
0 -
"I believe this gives a rough estimate of what matches are like in the 1600+ MMR skill bracket, which is the second highest one, followed by the soft cap that was recently raised."
Do you have data on the brackets (names or values) and is it possible to check where I'm situated in? I haven't finished reading your post but here I got stuck. Where do you get this number?
0 -
Is this the EU region? Given some of the names you mentioned I would think so.
I know there's significant differences between the regions, with playstyles in the US, South America and Asia ones vastly different. Such as in Asia perks like Self Care are still highly run, and EU being notoriously more "sweaty" and more heavily PC than other regions.
0 -
This was fun and interesting to read. Thanks!
1 -
I actually don't know how much we are allowed to discuss the way the matchmaking works, since they intentionally don't let us see the numbers and we have to make educated guesses. If you look on google, you should find some answers though!
Yes, all these matches were played in the EU region.
1 -
Interesting results.
Could you go into more detail on how you rated each of these matches? It's kind of hard to tell how balanced a match is, if you're not using some specific norm (in form of a list of criteria).
0 -
I tried to take into account as many variables as humanly possible to judge how likely it was that the other side could have won if things played out just a little differently.
Did the Killer have short or long chases? Did he tunnel, camp, slug or 3 gen? What was his perk and addon loadout? How efficient were my teammates on gens? How often did they drop shack pallet at 5 gens left against an M1 killer without being injured? Obviously, this is all very subjective and hard to judge, but I couldn't think of a system to score the matches on a more objective standpoint. Keeping track of hooks is basically a proxy to that. A killer getting 8 hooks without tunneling and slowdown performed better than a killer who tunneled one survivor and couldn't even get 3 hooks. If anyone has any ideas for an accurate system for the future I'm all ears!
0 -
That's good to hear. I remember your post about what we considered a balanced match and it was definitely thought provoking. It's not always easy to tell, how balanced the match was just by how many gens were done / hooks were scored.
0
