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
To those calling Nightlight Innaccurate
I have compared the officially released statistics to the ones on nightlight in descending order. top ten for killer:
Official Nightlight
Surge/Jolt - 20% Pain Res - 21%
PGTW - 20% PGTW - 20%
Pain Res - 17% Surge/Jolt - 20%
Sloppy - 15% Lethal - 16%
BBQ - 12% BBQ - 16%
Lethal -12% Corrupt - 14%
STBFL - 11% STBFL - 14%
NTH - 11% Deadlock - 13%
Corrupt - 10% Sloppy - 13%
NOED - 9% NTH - 12%
The largest disparity is 4%. Ultimate weapon was released 4 days after these stat's time frame, and since its only been out for about 35 days that could shift it a decent amount, but possibly not enough to prevent a top ten.
All but deadlock and noed are present in both lists. Nightlight may not have perfect accuracy, but its at least very reliable for general perk usage.
Survivor
Official Nightlight
WoO - 28% WoO -32%
MFT - 21% MFT - 26%
Adrenaline - 20% Adrenaline - 26%
Resilience - 19% Resilience - 21%
Lithe - 16% Lithe - 17%
Self Care - 15% SB - 12%
Deja Vu - 13% Deja Vu - 10%
SB - 13% Dead Hard - 10%
Dead Hard - 10% Prove Thyself - 10%
Kindred -10% Bond - 9%
The largest disparity is 6%. The first 5 perks are exactly the same on both lists, splitting off at self care. Deja vu is in the same spot, DH has the same %.
Prove, bond, kindred, and self care are not present on both lists.
Comments
-
I've always supported Nightlight as another source of info. Before it, quite a few people believed the Steam chart was a good measuring stick of the game as a whole but the PC was only one of several ways to play it so I could never understand why people clung to this so.
Nightlight of course is reliant on player participation and what data to put in, but in order for enough data to be manipulated to reflect a bias over real results would require thousands and thousands of players to be in on it and I just can't see people doing that so cohesively.
It's a useful source to get information from.
4 -
While it's not 100% accurate, it's an effective enough weather vane to see which way the wind is blowing.
9 -
Well, it's simply in the nature of how Nightlight works that it is good practice not to ascribe too much weight to its stats. It's not a random sample because it is self-reported.
And while the stats on perk usage may seem very similar here (in their limited scope), it is important to keep in mind that with the sheer number of players, even just 1% is a massive difference. With millions upon millions of matches every month, it takes hundreds of thousands of instances of a perk being used to make for that 1%. How significant that is is easily understood when you think of the fact that for the vast majority of perks, usage percentage differences will be in multiple decimal places.
Looking at it that way, you can see that even in the 10 most popular picks where disparities can be expected to be low, differences between the stats are still quite large. Different placings, disparities of up to 6% (which again is huge in terms of player numbers, but also if you consider that 20-30% is the top most percentage perks get used at, so a disparity of 6% between them indicates a 20-30% disparity between the stats (6 of 20/30 is 20-30%)), even entire perks that appear on one list and not the other (2-4 instances of as much, in a pool of 10).
Nightlight is neat and I'm glad the community takes it upon themselves to collect stats en masse, but if BHVR's global stats that actually take everything and reflect the actual "live balance" already have to be enjoyed with caution, Nightlight's stats are to be taken with even larger grains of salt.
6 -
Those "disparities" happen because nightlights nature is geared towards mid-high mmr players.
"Official" stats include everybody, even the fresh installs who only have the basegame and perks , so they are forced into using SC , for example.
Nightlight reflects the data from the actually valuable dataset, which is mid-high mmr where people have DLCs and therefore a choice + they know how to actually optimize and play the game efficiently.
I'd argue nightlight is quite honestly better than official just because official includes all the casual / low mmr bloat stats.
4 -
Even if that were true (we have no real way of confirming this, it would ultimately be speculation and the stats might not line up with official mid to high MMR stats even if most of the players reporting on Nightlight actually were to fall within those brackets, because the people that choose to report might have similar reasons for doing so, or they may only report part of their matches rather than all of them, skewing the data), BHVR is developing the game for the global live balance experience, not just the experience in advanced MMR brackets. I will also say that it is my conviction the MMR cap is so low that even players around 100-200 hours can be and regularly are captured by what we would generously call "high MMR", so balancing for experienced and knowledgable players wouldn't even make sense from a "high MMR" perspective, not until they actually make the matchmaking more consistently pair players of comparable levels.
The issue ultimately however is that clearly, BHVR's balance decisions are more or less completely at their own whim, rationale and reasoning. Even insofar they are based on stats we have no real insight into just how much that informs their decisions because they share very little and only very generalized stats. There are so many completely disparate "balance aspects" in the game (think Blight add-ons being buffed versus Demo add-ons being nerfed, for just one extreme example out of hundreds such), we have no reason to believe that any set of however representative stats would change their... "philosophy", let's say, when it comes to these things.
2 -
Self Care is heavily run in the asian servers and always has been. This is why its usage charts so high, not because it is freely available. Gameplay there is more individual based rather than teamwork based, with SWF less common.
Nightlight lacks a global view of the game.
5 -
The vast majority of the playerbase is in the low to mid MMR area, and that's where they mostly balance for, and what they keep in mind for changes mainly.
3 -
That kind of balance perspective is what leads to blatantly busted perk designs.
Balance should inherently be done with the most efficient/ best players in mind.
A perk balanced in high mmr is a balanced perk for any mmr, due to the nature of trickle down Balance.
It's plain obvious to see that if they buffed a perk like hyperfocus because "low-mid mmr players struggle with their skillchecks" , it would lead to absolutely horrible gen slam matches in the high MMRs because people actually hit their skillchecks.
---
Ironically people call mft balanced because of this exact problem. "Well MFT in low ranks has minimal effect, it's only the good loopers that make it seem busted"
Imo, MFT is long overdue for a ol yeller treatment.
0 -
This is also incorrect. Things that are balanced at high MMR can be completely, abysmally unplayable at lower MMR, because of the inherent skill gap.
What you actually need to do as a game dev is to balance for everyone, and skew towards the average. You don't ignore what's causing problems higher up, but you do the sensible thing and apply your finite focus more consistently towards the majority. The game needs its newer and lesser skilled players because they are by far a majority in most games, I see no reason to assume it would be different for DBD.
To the wider topic: it's not that Nighlight is always inherently inaccurate, it's that you can't guarantee that the numbers you're seeing are valid. It's usually a good general indicator for top picked things in EU/NA servers, but it's clear that it has a bias that trends towards those two things; it trends towards the top picked stuff and it trends towards EU/NA servers, because of who is uploading their data there.
It's fine to refer to it, but you can't treat its numbers as gospel. You can't say "this specific perk has X% pickrate, that's clearly a problem!" when you have no real way of knowing if that specific number is actually accurate.
6 -
Usually you're supposed to provide a valid counter argument with some kind of evidence, not just go "well you're wrong lol".
Perks that are balanced at high mmr are balanced at lower mmrs this is defacto.
A perk *looking* like its good because the other side has severe skill issue is not a *balance problem* it would be like claiming any means necessary is overpowered when you refuse to kick pallets / do chase cleanup.
0 -
"Perks that are balanced at high mmr are balanced at lower mmrs this is defacto."
LMAO, this is the most blatantly wrong thing I've heard in a long time.
Old Ruin Undying was balanced at high MMR because people would split up and power through it.
The perk was NOT balanced at average MMR because you couldn't reasonably cleanse 5 totems without throwing.
Ancient Ruin was balanced at high MMR because players had totem spots memorized and because they could consistently hit Great Skill Checks.
The perk was not balanced at average MMR because new players didn't know where the totems where and couldn't hit Greats.
I can keep going, but I hope you get the point.
7 -
Original ruin wasn't balanced at high mmr to begin with.
It takes AI level of "skill" to pull off greats 100% of the time just to *not lose progress*.
And i highly doubt people had "totem spots memorized". Like sure you can remember a couple predetermined spots per map, but remembering 5 totem spots per map + every single tile permutation on said map is downright impossible.
But sure, keep throwing up perks.
0 -
I did? Things that are balanced at higher MMR will not be balanced at lower MMR because of the inherent skill gap. It's also possible for things to be the other way around, too; something that a low skill player cannot utilise well, but a higher skill player can make absolutely unbearable.
More to the point, though, I am not specifically talking about perks here, I'm talking about the game as a whole. Setups that might be balanced at high MMR might be unbearable for one side lower down the latter; that's the entire reason the term "noobstomper" even exists.
Obviously to some degree lower skilled players should be expected to improve their skill, but there is a point at which it becomes unreasonable to expect a sharp increase in skill in a very short period of time, if something is genuinely unbalanced for lower levels.
Like I said, you balance for everyone and you skew for the average. That just makes sense.
6 -
Yes... lower mmr players are indeed expected to improve overtime instead of complaining , thats the whole point.
If low mmrs struggle to hit skillchecks we don't go and make skillchecks easier to hit because that will fundamentally break high mmr and gut skillcheck perks, you just teach them how to listen to the sound cue and practice in order to get good.
If a low mmr claims ghostface is overpowered because of his stealth we don't nerf ghostface, you tell the low mmr to turn his neck and be vigilant.
Perks are a problem when no matter how good you are they are still opressively strong and you can't work around them.
FTP+BU are a great example of something that regardless of how good a killer is, he cannot outplay or work around the insta pickup+10s endurance.
1 -
Yes, and, there is a line where something can actually be unfair against lower skilled players.
Original Ruin would be a good example. Higher skilled players hit Great skillchecks much more frequently, meaning they had a pretty decent chance against that perk. Lower skilled players, though, can range from only being able to hit Good skillchecks to not even being able to consistently do that. Ergo, Ruin was a perk that was unfair to lower skilled players even though they were capable of improving their own skill to aid in countering it.
You have to find a middle ground. This goes for any game, not just DBD. Ergo, you balance for everyone and skew towards the average, to make sure that lower skilled players are being given fair challenges without doing too much harm to the experience of higher skilled players. That is the only way of balancing live service games like this that will ever stand a chance of working.
5 -
You didn't need to hit all of them, just enough to get the gen done. Worse players couldn't hit ANY without either missing or blowing up the gen.
Likewise, yes, we did have totem spots memorized. There weren't THAT many maps and the RNG has been the same for 7 years.
2 -
I've already adressed ruin in a previous post, the perk isn't a valid example because it fundamentally wasn't even balanced in high mmr.
0 -
They do not balance that way. They've consistently made changes this way forever. If players actually want their suggestions seriously considered they need to keep the entire playerbase and all platforms in mind. It's how they do things.
2 -
Yup yup
0 -
I think it's also notable that nightlight is a self reported stat tool. Which most likely means that a lot of the stats are centered around bettter players since i really doubt a newer player cares or even knows about nightlight.
3 -
How do the killer pick numbers compare?
0 -
Stats for kill rates are also taken from submitted survivor matches exclusively. That means the stats represent the average kills random killers get against survivors on the more experienced side. Even if we were to assume that people reporting on NL don't have certain motivations for doing so (such as wanting to submit games in which they did well), this still skews the stats very artificially. Not least because more experienced and dedicated players that would even know about NL and be motivated to submit matches can also be expected to be playing SWF with other (experienced) players much more often than the average random player.
Fairly well. Of the top 10 killers between August and September, 9 are present in both stats (Xeno, Wesker, Huntress, Blight, Legion, Wraith, Myers, Nurse, Ghostface), the only difference being that NL has Nemesis and BHVR Trapper. The pick rate % difference is also not too great, with both having Xeno at 14% and the biggest discrepancy between the stats being 2% (Wesker NL 7% and Wraith NL 4% vs. Wesker BHVR 5% and Wraith BHVR 6%), average deviation more so being around 1%.
Although it again has to be kept in mind that given how large the numbers the stats are derived from are, even small percentual differences are quite significant. And also that a, say, 2% difference from 6% is a 33% deviation. Pick rates on NL are also based on submitted survivor matches exclusively, which I think actually makes them more representative of the global average. If they were to include reported killer stats, I am pretty sure that the pick rates would change significantly given that it would then skew toward killers more experienced and dedicated players prefer to play.
0 -
I think kill rates would be a more important comparison to make. A perk being used 18% vs 24% of the time isn't going to matter much (though it's still a statistically significant difference) but a 54% kill rate vs 60% is.
0
