Interesting Stuff in the Stats

[Mors_Indecepta]

I've been using nightlight.gg a lot lately, and it's pretty interesting to see how things actually play out in real matches (that aren't my own).
I find it particularly interesting that the top 3 deadliest killers, at this pre-patch moment, are not Nurse, Blight and Spirit. According to the thousands of matches they've compiled, the 3 deadliest killers are (in order of kill-rate) Skull Merchant, Plague, and Sadako (with my boy, the Dredge, coming in at #4).

Of course, kill rate doesn't tell the entire story...which killer(s) have the highest rate of 4-escapes and 4-kills? That's also important, I think.
Skull Merchant has the highest percentage of 4-kills (~37.88% of matches).
The killer who most commonly has to see all 4 survivors run out the gate is Ghostface, with ~27.4% of his matches ending that way.

What about survivors, though? Obviously, they all play with the same mechanics, but I learned a lot by checking out the most popular perks and builds. Primarily, I learned that a lot of survivors are pretty much all running the same perk builds. All the top builds listed for survivor are just variations of the same build...Windows of Opportunity, Adrenaline, Dead Hard/Lithe, and Resilience/Prove Thyself.
31% of the survivor community is using Windows of Opportunity alone. That's not a bad thing inherently...just can't help but think if that many killers were using the same perk, we'd all have to take shelter from the radioactive fallout that would result from how hard BHVR would nuke it.

[Probably afk tbh]

Originally posted by TheMetalOne:

I've been using nightlight.gg a lot lately, and it's pretty interesting to see how things actually play out in real matches (that aren't my own).
I find it particularly interesting that the top 3 deadliest killers, at this pre-patch moment, are not Nurse, Blight and Spirit. According to the thousands of matches they've compiled, the 3 deadliest killers are (in order of kill-rate) Skull Merchant, Plague, and Sadako (with my boy, the Dredge, coming in at #4).

Of course, kill rate doesn't tell the entire story...which killer(s) have the highest rate of 4-escapes and 4-kills? That's also important, I think.
Skull Merchant has the highest percentage of 4-kills (~37.88% of matches).
The killer who most commonly has to see all 4 survivors run out the gate is Ghostface, with ~27.4% of his matches ending that way.

What about survivors, though? Obviously, they all play with the same mechanics, but I learned a lot by checking out the most popular perks and builds. Primarily, I learned that a lot of survivors are pretty much all running the same perk builds. All the top builds listed for survivor are just variations of the same build...Windows of Opportunity, Adrenaline, Dead Hard/Lithe, and Resilience/Prove Thyself.
31% of the survivor community is using Windows of Opportunity alone. That's not a bad thing inherently...just can't help but think if that many killers were using the same perk, we'd all have to take shelter from the radioactive fallout that would result from how hard BHVR would nuke it.

I kinda disagree, if there are 200 perks and a single one is at 31% it is a problem. Either the perk is being used as a bandaid for a core flaw of the game or it's overpowered. That said I really wouldn't trust nightlite, it's interesting but with all the data being self selected by the community I find it hard to draw conclusions from.

[Tragopan]

Survivor perks don't see much variety for three reasons:

1. The vast majority of Survivors are solo que and thus bring in perks that help alleviate some solo que woes (WoO letting people know what is available/been dropped)

2. Survivors do not have character specific powers that define their kit, thus the meta for Survivors will be repeated more so than with Killers.

3. Most Survivor perks SUCK. Like, truly and utterly suck. Most of them are memes. The same is not true with Killer perks.

[Voidlines]

Originally posted by Tragopan:

Survivor perks don't see much variety for three reasons:
3. Most Survivor perks SUCK. Like, truly and utterly suck. Most of them are memes. The same is not true with Killer perks.

*Laughs in Hangman's Trick, Zanshin Tactics, Grim Embrace, Furtive Chase, Predator, Beast of Prey, Bloodhound*

I still agree with a lot of survivor perks sucking but saying it's not true for killer perks? Lmao.

[Tragopan]

Originally posted by Chaos:

Originally posted by Tragopan:

Survivor perks don't see much variety for three reasons:
3. Most Survivor perks SUCK. Like, truly and utterly suck. Most of them are memes. The same is not true with Killer perks.

*Laughs in Hangman's Trick, Zanshin Tactics, Grim Embrace, Furtive Chase, Predator, Beast of Prey, Bloodhound*

I still agree with a lot of survivor perks sucking but saying it's not true for killer perks? Lmao.

Bloodhound is actually a perk I enjoy. Zanshin Tactics is really useful on Nemesis or Huntress. Predator is also a perk I get a lot of use out of. Furitive Chase is amazing on insta down Killers.

Killer still has its useless perks but no where near as many as Survivor.

[State Line Crosser]

Originally posted by Probably afk tbh:

That said I really wouldn't trust nightlite

I wish bhvr allowed public access to their statistical data like other multiplayer games.

Originally posted by Tragopan:

Killer still has its useless perks but no where near as many as Survivor.

Maybe you ought to do a list and compare them? I'd be curious to know who has more. Although I think you may be trying to amuse us with your promotion of predator.

[￭ Yuri ￭]

Perks little used by killers do not necessarily mean they are useless, since each killer is unique in terms of its power and each perk can end up having good synergy if have a good consistency (the effect just needs consistency)

I.E: Spirit make an insane use of perks like stridor that is currently little used according to the site you extract the information from perks, and there's nothing wrong with that, the perk already has a solid and good effect. Same with Zanshin Tactics that is god with Pyramid, artist, etc . So it's just not used much because other perks are more advantageous and better used by all killers.

It turns out that the same conclusion cannot be drawn from the survivors' perks. Like all are "skins" by default, so in fact little used perks is because it doesn't have a consistent enough effect (bad) or because there are better alternatives.

As for the killers in the stats in top, I don't think it's wrong, either they have track or insta down skills and can perfectly secure 3-gen insanely, like plague and SM. Sadako has track, but her best build is around slug and the ghost face is the only one that is more aggressive and can chase, without a more defensive play style

[feral]

there is no need to draw far-reaching conclusions from the informations of a site that records the data of 12-14k matches per month. more matches are played PER HOUR than that.
it's a cute little hobby to look at the site sometimes, but the accuracy of these statistics is nowhere near the accuracy of the developers' own statistics for internal use.
i don't even mention that the data can be very different from region to region, which is not mentioned anywhere on the site.

[Tragopan]

Originally posted by State Line Crosser:

Originally posted by Probably afk tbh:

That said I really wouldn't trust nightlite

I wish bhvr allowed public access to their statistical data like other multiplayer games.

Originally posted by Tragopan:

Killer still has its useless perks but no where near as many as Survivor.

Maybe you ought to do a list and compare them? I'd be curious to know who has more. Although I think you may be trying to amuse us with your promotion of predator.

I definitely don't have the patience or motivation to go through all those perks right now... but I'll tell you what I like to do with Predator. I usually don't feel like I need meta perks to win as Killer so I use comfort picks instead (which are still strong). Shadowborne, Predator, Bloodhound (all ironically Wraith perks) are things I use quite often on practically every Killer. Shadowborne is OP for me, personally, because I get very dizzy in first person view and the larger FOV help combat that. Wish I didn't need to give up a perk slot for it but there it is.

Predator is fantastic on high mobility Killers. Scratch marks are the primary way we find Survivors who are not next to objectives. If the Killer gets even a small hint to where a Survivor was that's usually all we need to find them and if you're good at your chases then finding Survivors is all you need. It's a perk that does not require a condition to trigger like aura reading perks. I use it a lot on Wraith, Blight, Billy, Oni, Deathslinger, Huntress, sometimes even Bubba.

[State Line Crosser]

Originally posted by Tragopan:

Predator is fantastic on high mobility Killers.

I mixed predator up with beast of prey. It's a lot more understandable now.

[Mors_Indecepta]

Originally posted by feral:

there is no need to draw far-reaching conclusions from the informations of a site that records the data of 12-14k matches per month. more matches are played PER HOUR than that.
it's a cute little hobby to look at the site sometimes, but the accuracy of these statistics is nowhere near the accuracy of the developers' own statistics for internal use.
i don't even mention that the data can be very different from region to region, which is not mentioned anywhere on the site.

Do you want to edit this and see if you can fit a little more condescension in there?
As anyone with education in statistics will tell you, using a smaller sample to extrapolate the approximate numbers of the whole is a legitimate method of obtaining statistical data.
If BHVR would make the data available to the public at all times, I would be looking at that, but they don't.

[Reckall]

Originally posted by TheMetalOne:

Originally posted by feral:

there is no need to draw far-reaching conclusions from the informations of a site that records the data of 12-14k matches per month. more matches are played PER HOUR than that.
it's a cute little hobby to look at the site sometimes, but the accuracy of these statistics is nowhere near the accuracy of the developers' own statistics for internal use.
i don't even mention that the data can be very different from region to region, which is not mentioned anywhere on the site.

Do you want to edit this and see if you can fit a little more condescension in there?
As anyone with education in statistics will tell you, using a smaller sample to extrapolate the approximate numbers of the whole is a legitimate method of obtaining statistical data. The data certainly isn't exact, but it doesn't have to be to draw conclusions from it.

I do have an education in Statistics and the above is true. The bigger the pool, the smaller the sample needs to be. If the same Killer appears twice in ten matches, examined out of 100 played, you can't say that 20% of players like that Killer. But if the same Killer appears 2,000 times in 10,000 matches, examined out of one million... let's say that the span can be 15%&-25%. And this range becomes smaller with even bigger numbers.

However, when human beings are involved statistics go right out of the window. This is why "Exit Polls" are a farce. "OF COURSE I voted for the CURRENT THING!" says a guy to a stranger with a microphone. Alone in the poll he actually voted... what he wanted to vote.

I like Nightlight.gg, but I'm the first who, after a 0K, sometimes forgets to upload the match. And mine is a honest mistake, usually made because I'm a bit depressed (or disgusted after bullies/SFWs). Now, add to people like me all those players who rage quit, who generally rage because they turned out to be the worst in the match, and who generally rage full stop.

Anyway, as a Sadako main I can tell two things: People underestimate her A LOT, and against a band of SWFs the best you can hope for is to cause some heart attacks before the unavoidable 0K.

[feral]

distortion of statistics, and their inaccuracy can be traced back to countless reasons, but the most common of which are sampling procedures and errors in method selection.
all i said is that ALL information is available to the developers, and if they interpret this data correctly, it is impossible for a site that only has access to a fraction of this information to reflect a more accurate picture of reality than all the data possessed by the developers.
it is another question whether the developers correctly interpret the data at their disposal or not.

by the way, i didn't say it condescendingly. i check the site myself from time to time.
and yes, of course, we would all be happier if the developers shared statistics with us more often (and not incompletely), but unfortunately this is not the case.