Shrine Discrepancies
Here is a list I compiled as to Shrine perk frequency between killers vs. survivors. I chose 10 perks for each side that I would deem relatively meta (killer really only has 8 imo so I had to add STBFL and Thrilling Tremors, though I cut Detective's on the survivor side for an even 10) and did the math on the average frequency of appearance since either perk release or October 2016 (when Shrine became a thing), whichever comes later. My findings are as follows:
Given that there are more survivor teachables than killer teachables, meta killer teachables should in theory show up more often. As we can see, that isn't really the case. On the general, killers seem to get meta perks in rotation far less often, which in my eyes is indicative of a particular bias toward one side. This isn't based off pure conjecture or the typical "moris got nerfed now where are keys" argument. This is actual data.
And I'm not saying the devs are doing this intentionally, either. I think they're just choosing perks they like to be put in Shrine, which is probably why Enduring has shown up 20 times, more than any other perk in the game. I'm typically rank 1 and I already have all these killer teachables, so Shrine doesn't really affect me. But as someone who had to do the grind once as well as paying for all these characters, it was a LOT of time and money. I just think the devs need to take a step back and decide they gotta throw the newer killer mains a bone more often, since they're the one thing this community desperately needs to save our survivor queue times.
EDIT 1: I understand the argument that it is randomized. To a certain extent, that is probably true. That said, we are looking at a broader set of data here than many would consider. I'm not just cherrypicking perks like BBQ and Dead Hard to do a single perk comparison. This is a far larger sample size of perks scattered across more than 4 1/2 years of Shrine data that, in my opinion, demonstrates a general trend that either their pseudo-random algorithm is inherently flawed, or that they need to better balance what human intervention is there to determine Shrine perks.
EDIT 2: A friend and I were continuing on this discussion and we've reached a slightly different conclusion. We think it entirely boils down to the Shrine algorithm favoring older perks, not meta ones. In going over every single perk in the game, even in the last 2-3 years base game teachables appear somewhat more often than even some of the other older teachables in the game (pre-Shrine). In coming to this conclusion we realized the algorithm used to calculate Shrine perks may return older perks on the list far more frequently, even multiple years after their release. It just so happens that the meta survivor perks happen to be older than the meta killer perks. For example, in 2019 and 2020 Up The Ante showed up 2 and 4 times respectively, compared to Ruin showing up 3 and 2 times. Ruin in that case is still relatively constant compared to all-time results. Same goes for Sprint Burst. Interesting phenomenon, still likely due to a flawed algorithm.
Comments
has undying really not appeared at all lmao
Wait, Undying has never been in the shrine? I was saving for that one... I don't want to buy Blight.
I would strongly recommend buying him. I main no recovery GF, but Blight is always a really fun change of pace. I've had tremendous amounts of fun with him.
The best part of current shrine is BBQ which was there whole week before event started....where is my WGLF bhvr?
Probability does not work the way you think it should. The fact that Enduring has appeared 20 times is a statistical abnormality, no different from when you get 5 heads in a row when flipping a coin. It's unlikely (about 3%), but possible. Getting struck by lightning has about a 0.0003% chance of happening, yet around 24000 people are still hit and killed by lightning every year.
PS: The odds of any one perk appearing actually decrease over time due to the increase in possible perks.
WGLF has appeared 9 times. The last was in 29/01/2020.
BBQ has appeared 6 times, the last being in 23/06/2021.
BHVR has never said Shrine Of Secrets is a random selection. The fact that it is not is very clear given the fact that Object and DS both appeared immediately surrounding their recent reworks.
They have, actually, with the exception that 2 perks are guaranteed to be survivor teachables and 2 are guaranteed to be killer teachables.
EDIT: If memory serves, any perks that are needed for tome challenges are also guaranteed to appear when the challenges are released, but I'm not 100% sure if that's what they said.
Find me one singular quote from a dev that suggests that each set of teachables (being the 2 for each role) are fully randomized and not at all influenced by developer decision. Even Zanshin Tactics appeared right around the time it got buffed recently. The amount of times that happens seems to indicate the assumption that the devs have at least some influence is correct.
See my post count? I've been here a lot longer than most regular users. I know what I'm talking about.
EDIT: In case you want something more recent, see below.
I am playing since may 2020 so I never seen WGLF in the shrine yet. It would make sense if they put it with BBQ into the event shrine first day.
I'm pretty sure the devs can override options in the Shrine of Secrets if they feel it necessary. They have put up perks in the shrine when challenges for those perks were active in the past.
1: Peanits is a Community Manager, not a Developer. He may well be under the assumption that that is the case, but the underlying system may not reflect that.
2: It could very well be down to their algorithm being flawed, though the amount of times we've seen a perk being changed then appearing in shrine (regardless of whether it is in Tome, like with Zanshin Tactics) seems to indicate there is a noticeable amount of human intervention.
It's absolutely intentional.
Shrine perks tend to be things that are useful in the upcoming rift challenges. For example in the few weeks before the previous tome, that featured a lot of Yui and Oni, their perks were in the shrine, every single one. Plus there's the perk combos, Stridor and Iron Will were both in the shrine together, that's no coincidence.
Perks that are more sought after, particularly Undying, are therefore going to come up less, so that they can encourage players to buy the character they come with instead.
Blight isn't even a year old yet, so I'm not at all surprised Undying hasn't been in the shrine yet. But this event, with one day per shrine, would be the perfect opportunity for them to put it in the shrine, for one day only.
@cluxdx Is this better?
https://forum.deadbydaylight.com/en/kb/articles/163-the-shrine-of-secrets#which-perks-can-appear-in-the-shrine
Bestie please tell me you see the large "DEV" tag next to Peanits' name. Please please tell me that you don't think that someone working for BHVR doesn't have communication with other employees of BHVR??!!
You just want to ignore any facts to fit your narrative of "BHVR HATES KILLER PLAYERS" when that's literally not the case. Orion has already provided you with the proof so stop being stubborn lmao
Please it's so sad it's almost funny
All of the survivor perks came to the game in 2017 or earlier, 5 of the Killer perks are from 2018 or later, so I'm not sure if that is a good comparison.
Undying appear right after the nerf
I never said BHVR hates killer players. "And I'm not saying the devs are doing this intentionally" is a pretty big indicator of that. And in my eyes the dev tag on the forums is not an accurate representation of every BHVR employee's job description. To me a developer is someone directly responsible for coding the game. For example, a web developer is not a guy the company pays to communicate ideas to the guy who creates the website; the one coding the actual website is. I would not consider Mathieu Cote as a developer either because that's not in his job description. And I'm not selling them short, either. Someone like Peanits is equally important as the ones modeling, coding, balancing, etc. because their job is to take community feedback and communicate that to the rest of the team so they know what to focus on and fix. That said, since he wasn't responsible for coding the algorithm they use to pseudo-randomly select shrine perks, unless he himself has seen the code and understands how it works he cannot and will not know exactly how it chooses the perks. All he would know is what he was told. And based off the devs' recent history with the awful spaghetti code of DBD, I think it's fair to assume their Shrine algorithm (assuming there is zero human intervention, which I doubt) is equally as flawed as the rest of their game, which is very much so.
I'm still waiting for undying Sadge
And so please share, in your infinite wisdom, the sources for your claims that the Shrine isn't randomly selected, preferably complete with the code that makes it so. Because, rather unfortunately, we actually have a source for our claim that isn't "Well most of DBD is buggy". Because unless you've seen the code, how would you know?
I assume you know the analogy about monkeys on the typewriter? That applies here too. Probability can do what is thought to be near-impossible, because if it isn't actually impossible then it can happen.
The pattern in my eyes has far too large a sample size to not have even remotely evened out over time. I mean we're talking about more than 4 1/2 years of data. Sure, it's possible to flip a coin 500 times and land heads on all of them, but it isn't feasible to do that without doing something to deliberately make it happen, like weighting one side of the coin. Based on the amount of times perks that are not in Tome challenges appear right after major changes to those perks (like with Zanshin, Object, DS, and BT), I think it's a safe assumption there's more human intervention than they're letting on. BHVR has never exactly been forthcoming in the inner workings of their game or even their team; that much is indicated by the sheer amount of things they've said they were doing or thinking about that we magically never heard of again. I'm drawing a conclusion based off a statistical pattern; the Infinite Monkey Theorem doesn't typically apply in the real world because of the insane amount of trials that it would likely take to make such a pattern occur. Speedrunning mod teams used probability to determine that Dream was cheating, and as it turns out in most extraordinary cases, assuming something is up is a more than safe assumption when you've got that large a sample size. For example, a man has been struck by lightning seven times. You would think that's insanely unlikely until you look deeper and realize that he was a park ranger typically perched in a tall watchtower in an empty field with no lightning rods.
THE SHRINE OF SECRETS IS NOW ALSO A CONSPIRACY AGAINST KILLERS HOW COULD I HAVE BEEN SO BLIND
"This game is too easy for [X] even a new [X] player can win against any [Y] player no matter how experienced the [Y] player is. [X] players have all the strongest perks, the devs favour them and they get buffed all the time while [Y] gets nothing but nerfs. And I'm actually a [X] main before you all start."
Okay so, firstly, have you heard of paragraphs? It makes your writing significantly easier to read.
The pattern in my eyes has far too large a sample size to not have even remotely evened out over time. I mean we're talking about more than 4 1/2 years of data.
Are we now? Because if we take a look at the perks you're on about most of them barely break through the 4 year mark, with the exception of perks from the original characters.
Also if we take a look at the perks on the survivor side, every single perk except for Dead Hard was released within a year of the game's release. Dead Hard was released on 27th July 2017, so just over a year and a month after the release. Looking at the killer side, Tinkerer, Ruin, Devour, and STBFL are the only perks released in this timeframe (BBQ was 14th September 2017). PGTW, Discordance, and Corrupt were Year 2 perks, while Thrilling was a Year 3 perk and Undying was a Year 4 perk.
So, it's hardly a fair comparison.
Furthermore, look at how many times Calm Spirit, Urban Evasion, or Botany Knowledge has appeared in the Shrine.
Sure, it's possible to flip a coin 500 times and land heads on all of them, but it isn't feasible to do that without doing something to deliberately make it happen, like weighting one side of the coin
Yes, actually, it is. That's how probability works- regardless of how small the chance, it is still entirely possible.
Based on the amount of times perks that are not in Tome challenges appear right after major changes to those perks (like with Zanshin, Object, DS, and BT), I think it's a safe assumption there's more human intervention than they're letting on.
You're seeing patterns where there are none. It's fine, it's what humans do (we look for patterns everywhere) but it's just random chance.
BHVR has never exactly been forthcoming in the inner workings of their game or even their team; that much is indicated by the sheer amount of things they've said they were doing or thinking about that we magically never heard of again.
Actually, they have. Because things go through the consultation phases, which is when all they can say is "We're working on this", and are then either developed further (OOO rework) or completely scrapped (Early-game Collapse)
I'm drawing a conclusion based off a statistical pattern; the Infinite Monkey Theorem doesn't typically apply in the real world because of the insane amount of trials that it would likely take to make such a pattern occur.
Still creating a pattern where there is none. Random chance doesn't have a pattern, which is what makes it random.
Speedrunning mod teams used probability to determine that Dream was cheating, and as it turns out in most extraordinary cases, assuming something is up is a more than safe assumption when you've got that large a sample size.
Except we have proof that the developers have confirmed the Shrine is randomised. People have been digging through the code of DBD- it's how we know it's such a mess- and they haven't raised anything. So, again, the burden of proof falls upon you since you're the one accusing it of not being randomised.
The Shrine of Secrets is randomised unless one of three criteria are met:
For one, that comment is one paragraph. If I shortened it into two sentence paragraphs it would also lack proper cohesion as the subject matter of each paragraph would remain the same. Hardly an argument.
Secondly, all months before October 2016 were discounted because Shrine wasn't a thing before then. So 2017 perks aren't as far away as you think they are. They still account for 4 years of data out of a total of just over 4 1/2 years, so the perk pool was still relatively small during their release and as such their average frequency should be somewhat similar. It is not.
Also, I agree Undying should be discounted; it is far too new to draw a conclusion on it alone because of a now larger perk pool. I just needed 10 meta perks, and to be quite frank, killers have fewer perks that rotate through the meta so not including it would've left my data with a gap. Plus, people would've mentioned it anyway. It was more important that my data included it than not.
Lastly, you're assuming true random, which anyone worth their salt will tell you does not exist in programming. Every random simulation requires an algorithm created by humans, and therefore is subject to error. RNG can also be exploited, such as how you can guarantee specific patterns of random characters in Melee. There are nearly always things in games that influence RNG factors. In older Pokemon games, so much as specific characters in your name could influence your chances of encountering specific Pokemon. As time has gone on RNG programming has improved, but it is not perfect and there is almost always a predictable pattern in typical implementations once the underlying algorithm is understood.
This whole conversation doesn't make sense anyway because all survivors are the same, killers need a bigger variety of perks because powers work better with some than others, let me give some examples not on your "meta killer perks".
Infectious is meta on fast down killers like Oni, Billy, Nurse, Spirit & Leatherface
Sloppy / Brutal Strength are meta on M1 killers like Wraith, Trapper, Hag, Doc, Myers, Legion etc
Iron Maiden is meta on Huntress / Trickster
and besides, this benefits killers even more because they're more likely to unlock all teachables without levelling each chartacter, unlike survivors who will have to spend more bp and in game time to get all the teachables because the same perks show up more often and they also have 11 more perks than killer. (That's 33 more bloodwebs to max a survivor btw)
There is no denying that there are perks that are on the general deemed as "meta" for killers on the whole. For example, Tinkerer and Discordance are most definitely meta tracking perks. And slowdown is a bare minimum requirement for competitive killers, therefore perks like Pop and Ruin are definitely meta. Corrupt Intervention is also the single most universal killer perk in the game, and as such is definitely meta. There is without question a general trend of killer perks that are seen at high level far more often than others. Are you gonna tell me that Ruin/Undying/Tinkerer wasn't the meta because killers like myself still ran Pop and BBQ in some builds? One-off killer perk picks like Huntress with Iron Maiden is kind of a moot point because that still isn't the general trend for the vast majority of killers.
Side note: Sloppy is not good on Legion. Frenzy attacks are not basic attacks.
Regardless, I'm leaving this thread here. I simply aimed to bring forth stats, and I feel I've said all there is to be said for my side of the argument.
Thank you, thank you, thank you.
People's understanding of math on these forums just drives me crazy 🤪!!
side note: I never said it has anything to do with his power, he's still an M1 killer that benefits from sloppy.