"stats mean nothing" is a lie
If these stats didn't meaning nothing, why did bhvr make basekit changes according to "low Killrate" (joke in itself btw) so they can get higher kill rate stats? Make sense out of this.
Yall trying to downplay the meaning of these stats just so you can keep your comfy basekit buffs.
Reality is, these stats say a lot and it clearly shows that on average AT EVERY MMR killers are winning more than they lose.
If YOU LOSE a lot of your games it just shows you aren't that good, simple.
Comments
-
yes nurse is in fact the weakest killer. preach louder for the people in the back
31 -
You just don't know how to read stats. It means nurse is hard to play and not many can play her. Look at 5% mmr and see how she's not the weakest anymore.
Yall bad at understanding these numbers
7 -
It's not so much that they mean nothing, but more so that very broad stats like the ones we just released do not tell a full story. You cannot look at an average from all games and definitely say something like "Nurse is clearly weak, look at her kill rate!" - Of course we all know that's not true. In the right hands, she can be the deadliest Killer in the game. But the steep learning curve and the number of people trying to play her for the first time cause that average to drop.
Stats like these can give some insight in where a problem might be, but they don't tell you what the problem is or if it's a problem in the first place. You should not use them as the basis for your argument.
38 -
You mean the numbers that show that both plague and dredge are better than nurse and blight?
6 -
Stats mean nothing because they seem to imply that the game is very killer-sided, and that's not something this forum can conceive
9 -
Then explain to me why you BHVR (knowing this) still went ahead and made GENERAL BASEKIT BUFFS because you didn't like the kill rate stats you showed before? They didn't show the full story either but you still took them to do basekit changes.
You even made a post after some time saying "we happy about the basekit changes because the Killrate went up".
Make sense bhvr.
6 -
Was literally going to quote this same comment you made on Reddit yesterday. People are focusing way to much, and drawing way to many conclusions from these stats.
0 -
Another one who can't understand stats.
Plague and dredge are easier to play. Average people can get better results with them than with blight and Nurse which require more skill.
2 -
You broke out of the matrix! Congratulations
2 -
Stats only matter if i like them and they fit my narrative
15 -
I'm talking about the top 5% mmr not overall.
0 -
Doesn't change my point, it still stands.
0 -
Nurse is the most picked killer at high MMR. Nurse also has 3x the pickrate over Dredge and 2.5x over Plague. From this data alone it's very likely that Nurse is the better killer regardless of winrate. They also exclude games with DC (don't ask me why) which likely skews the stats a lot. A lot of people DC against Nurse.
The killrate (winrate) itself is not a very good indicator since MMR is supposed to bring it at 50%. It only shows that MMR does not work.
4 -
First BHVR said they wanted to increase kill rate and shake the meta, which happened so buffs did their job and are staying (hopefully)
Second, the stats can be affected by ALOT of variables, such as map, swfs(swats and casuals), add ons and so on.
Third, saying people can't read stats, when they clearly can, undermines alot of credibility for you mate
Honestly we just wanted in on the madness 😶
7 -
Friend, I'm happy to discuss things further, but the snarkiness is not necessary.
The changes we made in 6.1.0 were not solely based on stats. When looking at gameplay, it was clear that Killers were having a rough time doing what they need to do - killing, and this was in no small part due to things like generator repair times. Each of the changes we made were related to issues that many people have brought up. I would encourage you to read through the Developer Updates we posted around that time where we explained our reasoning behind each of the changes.
Beyond that, the data we have access to is far more details than averages, which allows us to do deeper dives and reveal more than just a kill rate.
Again, stats can show you where a problem might be, but they do not paint a full picture.
41 -
You made a wrong conclusion here.
You assumed that BHVR changed the things they did based on those numbers. However concurrence does not mean correlation. If you would have read the developer updates you would have had a better understanding of those things. Not to say that these statistics did not have any impact of course.
Some things are also about "internal balance" of a Killers set or the Killer role as a whole: You do not want a certain Killer to only ever pick 2 of its add-ons, you do not want all Killers to use the same perks. Same for the survivor side by the way.
Killer buffs and changes are also not always about their performance. They are also about their feeling and flair. Take Ghostface as an example: His add-ons did not get that much stronger, however now they feel much better. Or Plague, when they changed some of her interactions, something that did not outright buff her but made her more enjoyable to play.
Just an idea here: Maybe the devs did buff some baseline things for Killer so that there is no longer a 100% Killer incentive all of the time? Or maybe they did it so that in a future patch they can buff soloQ without loosing to many Killer players?
I am all in for a buff to soloQ right now and some nerfs to Killers where they belong. A nerf to Blights add-ons, the removal of hook-grabs on healthy survivors, a buff to maps (pls add brown and yellow maps...and make them do something xp), maybe even a bigger hast on the baseline BT (7%-10%) and a bit more information for survivors (seeing if somebody beside the obsession is chased as an example...).
And important question you should ask is also: What is the goal? BHVR stated that they aim for 55-60% winrate if I remember correctly. As such the numbers are actually fine.
6 -
That bias probably comes from the fact that most ppl on the forum are very experienced players with a lot of dedication to the game. So the skill level is pretty high. And high mmr games survivors tend to do very well.
3 -
The more likely conclusion from the data is that the vast majority of Blights and Nurses you encounter in your games, aren’t over performing.
Can people please look at the units of measurement on the data? The data labels are so specific that we’re sometimes using 0.5% as a separate category. If killers are off by 1% or 2%, then it’s practically unnoticeable, and it’s definitely not a big enough difference to throw these killers into separate performance tiers.
2 -
Remember that they also nerfed the twins best addons because for their statistics they were "overused" in the HIGH mmr... You heard me: Not in low mmr but in high mmr... Like the twins weren't already weak... Yep, I agree "stats means nothing" is a lie, it's a useful thing used by the devs to do senseless nerfs...
0 -
Ok, this is a fair point. Could you give us stats like escape/kill rate on a graph where MMR is the x-axis? Gen completion vs MMR? Hook states vs MMR? These graphs would tell a story, and allow for better feedback.
3 -
I haven't heard this asked for so I thought I would. Is there any statistic about the distribution of killers at each MMR? Not the kill rate, but the percentage of X killer at Y MMR.
The reason is because, if a killer is better than another one, it's easier to win games and therefore get to a higher MMR where you'll then lose games or have a lower kill rate. Then it will balance again and you'll get a higher kill rate. I'd imagine it would eventually just stabilize at a specific kill rate if you play that killer enough. But, an average killer player using a "better" killer will get boosted to a higher MMR than they would if they played a "worse" killer. Anecdotally, the survivors I play against when I use Nurse/Blight/Spirit do things that indicate a lot more skill than the survivors I play against when I play Trapper.
For a killer like Nurse, she's very strong but has a high skill floor. But, since she's strong in the right hands and likely has a higher potential, there's a higher number of Nurse's at a high MMR than killers that are traditionally weaker. But, the weaker killers at that high MMR are all cracked players because they were good enough to get there to begin with.
I suspect people are attempting to use kill rates as a proxy for how good a killer is at their maximum potential and it's just a faulty way of using the statistics.
0 -
That is interesting, any kill rate above 50% average BY DEFINITION means that MMR is not working to create fun balanced matches.
1 -
They're aiming for 60%, though.
4 -
Kill rates matched with MMR will probably just result in a straight line or something meaningless. Kill rates are used as a factor in your MMR. If you put them on the graph together, you're doing the equivalent of just measuring a line that says "f(x) = x".
I don't know if it would say anything new, but I think number of players at each MMR by killer would probably tell us more about relative strength of a killer. MMR both factors in kill rates and the skill of your opponent. Kill rates don't factor in who you killed. If I play a game with 4 people who just installed the game yesterday versus 4 people who have 8k+ hours, we wouldn't say that these are equivalent if I get a 4k. But, comparing kill rates alone does just that because matchmaking ensures that the killer with 8k+ hours isn't playing against survivors with less than 30 minutes.
0 -
That claim that "high mmr survivors tend to do well" is directly contradicted by stats. The escape rate of 39% in the stats that BHVR shared IS at high MMR.
If you don't believe their stats, bad news for you, I calculated the empirical winrate of popular survivor streamers, and it seems to be around that number, if not lower: https://forum.deadbydaylight.com/en/discussion/339868/massive-discrepancy-between-bhvr-stats-and-reality-solo-queue-seems-to-be-30-winrate (Note: unlike BHVR I did not discard matches with DCs and counted hatch as 0.5)
The reality is, escape rate is just that low. The stats that BHVR shared ARE representative of what you'll get if you queue as a survivor.
0 -
That seems like a jumped conclusion for me. Why should 50% be the optimal kill-rate?
It should be slightly higher to imply a balanced game, because you have to add a margin. If one survivor gets killed early on the game is very likely to end in a 4k or 3k+hatch. A 4k and even a 3people escape is more likely than a 2:2 which would be the perfect outcome of every match because of the momentum a single kill creates even in a balanced game.
This one kill early can happen due to bad matchmaking (wider MMR range after a while of searching), people playing as a stack (good players with their newby friend) or other unfortunate events like suicide on hook or archive/achievement challenges. As such the margin has to stay and now we could argue about how big it is supposed to be...
2 -
Source? I don't think they have said this anywhere, but if true that would make sense.
0 -
Looking at gameplay? What gameplay please? I would like to know what gameplays you were watching which lead to these decisions.
And why are you not revealing these more detailed datas to the community? Why give us 2 charts instead of all the data you used to make those decision?
We are all smart and capable of analysing huge and complex datas. Just share them and we can make our own conclusions and "paint a full picture" as you said
1 -
Can we also get winrates by perk? I think that will also add to the whole story. Like obviously if someone is playing optimal Nurse is gonna have a very different game than someone playing "nice" Nurse. Or someone who plays killers with no perks or not full perks. Don't those also affect the killrate? I understand that they are not the majority, my point is just. Can you track which builds per killer have the highest and lowest killrate and if yes, would you be able to share them?
0 -
The problem is people always move the goalposts. First it was "let them show us data for high MMR and then we'll talk". Now they've shown the data for high MMR and it's the same people acting like stats don't matter at all since the data didn't give them the answer they wanted. People only see what they want to see and stats will never convince some people.
6 -
Again, I'd suggest reading the Developer Updates for more explanations, but to use generator times as an example, there was a very clear trend of slowdown perks being used. Watching countless streams, videos, and even playing the game, it was clear that generators could often fly by unless the Killer had some way to slow them down. Even among the feedback we received, it was very common for people to ask for slower repairs. On the data side of things, we can see average match lengths, even compare matches where a perk is used compared to matches where it is not. All these sources play a role in those decisions.
As for the more detailed data, the issue isn't whether or not you'd be able to comprehend it, it's displaying the data in a way that a person can actually read. We have all sorts of interactive dashboards that can show us things like how a Killer perks across skill ranges along with their popularity. The issue is: You cannot read it as a simple JPG. Even if we're just looking at Killers, you'll have around 30 lines crossing over each other, making it impossible to make any sense of. Our tools allow us to focus on a single (or handful) of data points because they are interactive, but a JPG does not.
All that said, even then data does not show a full picture. I cannot stress this enough, just because your experiences with something don't line up with the numbers doesn't mean it's invalid. Numbers can show you where a problem might be, but they are absolutely not the only thing you should use to backup your arguments.
17 -
Stats mean nothing because they seem to imply Nurse and Blight aren't broken and that's not something survivors can conceive.
3 -
Stats are still not as accurate as they seem I think.
0 -
I read then Dev update back then already.
The thing is, I'm still very curious what streamers bhvr was watching to make those decisions. I would love to hear some names to get a picture on the decisionmaking. Not every streamer is the same.
Also, is bhvr currently still watching streamers for balancing?
Why is bhvr not creating some kind "leaderboard" then? Showing all kinds of statics everyone can pick and chose? Like the leaderboards some fans created. Live updated, showing kill rates of the day/week/month most used perks of the day/week/month.
And then let the community create their own statistics to focus on and share.
The data given out is not enough apparently, so give us the possibility to access more data
0 -
The stats that BHVR shared DO imply that Nurse is broken at top MMR. She has the highest pickrate and Blight follows not too far behind. Both have 60% winrate wich is high considering their pickrate, notice the other top pickrates (Huntress and Legion) have lower winrates. They also do not count games with DC which I think impact Nurse games way more than other killers. I actually don't think Nurse is broken but you are insane to think the stats align with this. What the stats seem to tell is a story along the lines of "Nurse is broken but hard to play".
2 -
But from what i read, pre 6.1.0 killers would still have an average rate of 52-53% killrate which means they would still kill more on average than survivors would escape. I dont truly believe they were having such a "rough time".
8 -
-
They had no rough time. It's just made up. People mix balance with skill sometimes.
2 -
You're forgetting that they supposedly don't even count DCs AND a lot of killer mains, especially those that decimate their opponents, let the last survivor go in order to "be nice" and it's safe to assume some do so purposefully to lower their kill rates (mostly people who know they're wrecking survivors and don't want their killer nerfed). So, if the stats can't be trusted it' s because the killer is probably doing BETTER than what we're seeing. Definitely not worse.
As far as nurse goes, it's logical why she's lower than some other killers. The average person can't play her. Yeah, she's deadly in the right hands (a special few) but most people aren't going to waste their time trying to "get gud" when there are so many other enjoyable options. Nurse is basically unplayable on console anyway, so that excludes a ton of people right there.
In the end, stats do matter. They are undoubtedly higher than what we're being shown.
1 -
So what about Wesker? Highest pick rate by far, should have an even lower kill rate, right?
The stats dont tell that story. You are infering that story from other factors outside of what the stats show, and using that to explain the stats.
That doesn't mean it's invalid, but it's not direct statistics.
3 -
A possible reason for this is because of the 4v1 dynamic and what it means for player engagement.
I've played games where the killer got a 2k, but they really struggled the entire game and mostly just got the kills due to endgame shenanigans. In some of those, one or two people just get chased for the entire match. For the other 2-3 people, the match is kind of... boring? You're just on gen duty the entire game and escape easily with no effort.
The game is more than just an end result but mostly how you got there. Killers might have been getting an average of 2k, but maybe it was happening in an unhealthy way. It's generally more fun for killers to get 8 hooks and kill nobody than to get 1 hook the entire game and then camp that hook out because the 5th generator already popped.
3 -
They imply that nurse and blight are harder to learn and most killers aren't actually as good at the game as they think, thats why the whine about survivors so much.
2 -
'The stats' don't imply that at all.
You're inferring that from other factors.
The stats don't provide enough data to draw that conclusion, that's merely your attempt to explain the stats.
0 -
'The stats' don't imply that at all.
You're inferring that from other factors.
The stats don't provide enough data to draw that conclusion, that's merely your attempt to explain the stats.
0 -
It sounds like you are trying really hard to accuse the devs of either not knowing how to read their own data or of lying.
4 -
Wesker is indeed totally broken on a statistical basis. I excluded him because he's new and survivors have seemingly not learned how to play against him yet.
You are welcome to provide a similar explanation for Nurse. I don't see any that could make sense. In fact, DCs being discarded makes me think her true killrate might be higher.
Also notice, I didn't claim that "Stats show that Nurse is broken" but instead that "if stats show something, it's more likely that she is broken rather than not broken"
0 -
0
-
But that is a very inconsistent way of justification. I mean, i agree some games the killer would get the kills on endgame, but that is one type of outcome out of many.
Some killers get gen rushed and get 0k.
Some killers stomp the surv and get 4k at 5 gens.
Some killer struggle to get 1k.
Some killer camp to get 3k and the last surv got hatch.
Some killers get their kills on endgame because their perk build was made for that stage.
I dont think it's fair to say the 52-3% was done in a unhealthy way, nor i think that bhvr can accurately gather percentage data on "healthy" matches and "unhealthy" matches.
0 -
You know, when you try to reverse someone's facetious, sarcastic remark back on itself, you just look stupid.
2 -
Don't worry, stats will matter again when they will show that escape rate is too high. That's basically how the selfish killer-main mentality works.
3
