Read more Dead by Daylight ➜ https://deadbydaylight.mgn.gg
I’m live right now! get on over here – https://www.twitch.tv/tru3ta1ent
FOLLOW EVERYTHING!
www.youtube.com/tru3ta1ent |
https://www.twitch.tv/tru3ta1ent
https://twitter.com/TrU3Ta1ent |
https://www.instagram.com/tru3ta1entreal |
https://discordapp.com/invite/thetru3mmunity |
https://www.tru3ta1ent.store |
https://www.tiktok.com/@tru3ta1ent |
#tru3ta1ent #dbd #deadbydaylight
source
You gotta love this guy…
"I am a an analytical person who analysis data smartly."
a few moments later…
"guys.. what does it mean to have an average kill rate of 60%? So you tie 6 out of 10 games?" omegalul
Here is the formular for 10 games: (40*0.61 – 2*N_draw)/4 = N_win, now trial and error N_draw until N_draw + N_win = 10 (approx.) and you know how many draws and wins (4K win) you would have in 10 matches for a given average kill rate (here: 0.61 = 61%)
examples:
Kill rate = 50%: N_draw = 10, N_win = 0
Kill rate = 61.3%: N_draw = 8, N_win = 2.13 (exact: N_draw = 7.74, N_win = 2.26)
Kill rate = 75%: N_draw = 5, N_win = 5
Kill rate = 100%: N_draw = 0, N_win = 10
61.3% would be around 24 – 25 kills
I think you will get 1 kill in the end just by standing next to a hook. So 10 kills disappear. Then is 14 left as a bonus. Usually when game goes well it goes well and you will get 4 kills. In which case you would get 4 games with 4k. Then 1 or 2 games with draw and 4 games with a loss. You also need to account for throwing the game, dc etc. So the actual win rate might be even lower than that.
Everyone if you want to see a breakdown on this go watch Otzdarva
I'm a survivor main and play killer only for challenges why am I playing 2,000 hour swf for 1 and for 2 like I said I'm a survivor main and they need to stop with having all the perks where survivors take 3 or 4 hits to go down it's crazy.
I'm starting to bealive that devs are survivor mains and when they die from a killer or a killers simply destroys them they nerf him
Today on: the devs show data in a very specific way to make it look like they did something.
I wish this game encouraged hooks more, torturing the survivors, make them all have fun.
Apply debuffs for each fresh hook, for each stage 2, so that the game stalls enough that you can actually hook everyone twice. And so that I can run my builds without having two gens pop at the start of the game lol.
The devs don't seem to understand the most fun on the game is the interactions between survivors/killers.
MMR should be based on chase length, hooks and saves.
But oh well one can dream, I still really enjoy the game, but I think it could be much better.
7/10 times you draw, at best, using the strongest killer.
The remaining 3/10 times you either win or lose.
Which means, 3/10 is the best odds you get of winning using the strongest killers.
Perfectly balanced.
for me all this graph shit just said what was the most played killers in DBD on the month of september, thats how meaningless this is
I still find the chart very weird since and i could be wrong the previous charts have had a MUCH LARGER pool to pool from for data where this is just the past month of data not including factors of how many people are actually playing the game compared to the time period of the last chart and since this chart only has 1 month the other chart had months most likely? like you really cant take a months period of time and data and try to say "hey its better than the previous data we had" since the previous set had so much more time and other variables that are not present in this new one besides of course "KiLlS" is the measurement
Tru3's voice is like Morgan Freeman's in terms of ASMR, its perfectly crips
Can someone tell me if Survivor speed got buffed? I tried to play Spirit it is impossible. I hit one survivor and they run over the whole map.
Or they always make it to a save spot even if i should have already gotten them. I play both survivor and killer so I know the survivor speed i know where i can get them and where not because i am in the same situation.
I had to quit three games because of that and I think these people are cheating.
First game: I use BBQ and see just one person i phase and someone drops a pallet on me. Where did they come from??? Gen was also done way to fast
Second game: they press forward they drop a pallet i break it and phase and they are already way to far away.
Third game: I hit her and she runs across the whole map. There is this perk that makes this speed 2 seconds longer but my phasing just got ignored. She was also looking back and running in no objects, very sus. Another survivor aways knew when i was coming and run exactly in the other direction without having any perks for that.
I don’t think you learn much from either graph. They didn’t mention all the survivors and killers losing on purpose. Or survivors just letting killers kill them.
Pig secretly A tier
People will argue these statistics help, but can’t tell you what exactly it tells you
Stats rant: This is a misuse of averages. Not every kill is "equal". In particular, going from 0K to 1K in a match is usually trivial. Each additional kill in the same match requires increasingly more effort. So two 3Ks is not equivalent to six 1Ks, and you can't just math them together.
Instead, let's see the distributions. Especially when there are five neat categories: 0K, 1K, 2K, 3K, 4K.
It's a little silly they only have the rates for this month, because the newest killer (wesker) has a disproportionately high play rate right now
Played 7 games as solo survivor last night: 4/7 matches the first survivor to go down dc’d. 2 other games the survivor tried to kill themselves on hook. I escaped two games out of 7. If people just didn’t dc I’d like to think would escape more.
sound like solo q need more nerfs to mE….
gens should be 180 seconds slower
so low mobility killers like Myers and pig have sufficient time to beat the shit out you.
If you enjoyed please give it a like! It allows more people to see it! Also, If you want to see more Dead by Daylight content, SUBSCRIBE to the youtube, follow my twitch @ http://www.twitch.tv/tru3ta1ent and follow me on tiktok https://vm.tiktok.com/ZMLsW8u8p/ for exclusive content ❤
Displaying the data in terms of "average kills per game" rather than a percentage would make it more intuitive to understand. For example, (as has been stated a few time below), the over all kills average for top 5% MMR would be 2.45 kills per game. Wesker at top 5% MMR is getting 2.6 kills per game, and trapper is getting 2.24 kills per game. Generally, the percentages make it seem like the differences are larger than they are.
Really, the most useful statistic would be the standard deviations. If we know how much variablility there was in these means, that would tell us how reliably these menas really are. We can see a lot of movement in Nurse (likely indicating a large SD) and little movement on Dredge (small SD). DIsplaying only the means can be misleading.
Unfortunately, there are infinitely many possiblities to calculate 61.3%, which makes this very sad (see below for an example). This makes this data extremely useless. I study engineering and those are highschool level of graphs, no (computer-/data-) scientist would ever take those seriously. It's been made by community managers, not by any data analyst for sure.
One possible example is:
4x4k (100% kill rate), 4x3k (75% kill rate), 4x2k (50% kill rate), 1x1k (25% kill rate), 2x0k (0% kill rate) equaling to 61.67% mean kill rate. You can vary those numbers like crazy.
If you play 25 times, you confront 100 survivors. Wesker kills 63 survivors in 25 matches. Freddy kills 62 survivors in 25 matches. Pig kills 61 survivors in 25 matches. As you say, the killers are very close to each other, what means that there isnt big differences between killers. 63 is almost the same distant to 50 than is to 75. So Wesker just sometimes gets 2 kilss and sometimes do 3 kills. He is in between those two results.
tbh i think a nurse is easier to juke around than wesker. But that might only be me i guess.
the depth for just a game is crazy lol
This is a better system. When the Emblem System was the mmr (aka Rank) the game was an absolute joke for most of the roster. Where as certain killers could benefit from it (true chase mobility killers) most of them(non mobile/territory killers) could not whatsoever. This gave us all horrible statistics and would also give us an era of many many unwinnable situations for all killers. Survivors splitting gens only to be able to do all bt saves endgame and the killer having to eat a ds on the last survivor who was farmed by their team. This game was way too easy for survivors. Also a system where veteran players could face a set of babies or a baby player.
(My opinion)Why anybody missed this is beyond me, no one should ever EVER have to face a baby in dbd especially witn a veteran. Fuck the emblem system and for anyone who smurfs or enjoys that kind of play is a piece of shit. (Opinion over.)
When you go off of kills compared to hooks you recieve proper data since the game was based off rng in the first place. This is an unpopular fact within an asymmetrical game. You cannot mathematically caculate hooks into this situation because of oak offerings, sabo, etc. With the alternative of certain killers ( the ones that had no chance to win the emblem system in the first place) when played to the extreme can make saving, tunneling chases, and traversing certain parts of the maps downright unbearable for an individual survivor while the team is trying to protect their teammate from being killed. At the higher levels of dbd these are the only ways to obtain a set of kills with those types of killers where as going for hooks (catch and release) would be a loss no matter what.
While matches are a lot more intense and not as easy anymore for both sides (as it should be in a slasher game), with this system you actually can continue to balance all of the roster and survivors without having to listen to just the survivor community(emblem system days).
Kills and escapes will forever be the only way to balance this game since both sides exist, not just the survivors. While the math is fuzzy right now, it is so much better now because both sides have a chance now.
Nurse is the less cause a small quantity of player can play good whit her, so the majority of the community lose whit her
It's pointless data
they should just release more specific data but for each individual killer
The data is incomplete, and they are aware of that but release it "for fun" anwyay. Its actually a big part of why they have that disclaimer in there about considering other factors, they're saying to take it with a massive grain of salt. They do use more data than this in their decisions, but like you said, these stats are entirely pointless from an analytical standpoint as they are. The only thing they can be used for is to create hypothesis in order to use more accurate data to then corroborate or disprove.
Freddy 62%? If I’m being completely honest I haven’t played against a freddy at least 5 months. (I play everyday)
BHVR are “winning”. 2 kills truly is the standard.
Pig used to be the best killer and now she is top 3 tier. Nurse used to be bellow 50% but I bet because of the “auto-aim” patches she actually got stealth buffed for noobs since it’s only blinking rather than old Blink+ aim (at least that’s what I had to do pre-Crossplay consoles, absolute pain to master back then).
Hooks mean nothing, Death means everything.
The idea behind this chart isnt to show how many kills or hooks but how killers are right now. So if you look to the overhaul you re goin to see every killer above average kill. They are basically doin well. They can take a look at wesker and figure out why hes too strong right now and maybe nerf something and also look at trapper and see why hes so weak right now and buff something. and of course this chart confirm what behaviour wanted: buff killer so they can get more kills. They managed a way to do it.
The issue with this data, is it lacks critical context. Honestly the data is kind of pointless without the context. It just shows who gets the most kills and the average.
having it as % is just wrong anyway they should show in numbers…. since 50% of 10 is only 5 and what is the % out of we need the number to be % out of
Idk, I still think the perfectly balanced outcome is when one survivor is sacrificed with only 1 or 2 gens remaining, then two more survivors are sacrificed in an intense endgame, with only one survivor managing to escape. So ideally the kill rate should hover at around 75%. Seeing the top killers just barely squeak out a 2k on average… that's not balanced. If you, as a killer, are only getting TWO KILLS on average, you are getting absolutely dabbed on, mate.
As someone who does a lot of data science in engineering, these graphs are a bit less than the absolute bare minimum you could possibly do. It gives you numbers but doesn't tell you anything. It tells you the ending of the story, without telling you anything about the plot, which is strange because they obviously collect that information: number of hooks for each kill, the gen progress at each kill, and the amount of time it took to get each hook or kill. I'm quite the amateur but it only takes me a few seconds to dream up a couple graphs telling some really educational stories with only this small set of data. Why aren't they doing this? Possibly because the story it would tell, is not the story they want to believe.
All they’d need to do is include Standard Deviations (or some other similar measure of variability) with these averages. That would answer a lot of what Tru3’s asking. But regardless, I completely agree that it’s unclear what (if anything) they’re trying to SHOW with these data. No good scientific paper ever shows a figure like this unless the authors also DESCRIBE what they think the figure is actually showing. These guys just vomited some bar graphs on a page and said “look what we maked”. Why? What are we to glean?
from this data, no one can say a killer is OP, because a 61.3% kill rate tells us that no killer can consistently get a 4k every game to qualify the killer as being OP….
thus how can anyone justify statements of "Killer is OP" or "Solo Q is weak", because the data tells us that no killer can consistently 4K every time the game is played?
No matter how this game has been updated, I've always been able to count on Tru3 for an honest, and well thought out opinion. Keep up the good work
10 games has 40 survivors. 60% of 40 survivors is 24 survivors. Try to split 24 kills evenly ends up with 20 kills per game and remainder 4 so you could arbitrarily get 2 kills for 7 games and 3 kills for 2 games where a survivor got the hatch and 4 kills for one game. This is out of 10 games. Or it could be 3 kills for 4 games and 2 kills for 6 games.
They want to push a kill rate of 60% on average for all killers and they seemingly only use kills as their determinate factor (no maps, hooks, chase time, perks, addons, etc.), which is unfortunate from an analytical POV.
Kills ≠ Skill, some arbitrary factors play into unrewarded kills/escapes. Unwarranted kills happen more, hence the extra 10% from the 50% killrate which would equal 2 kills per match that they want.
However, these do not specify what perks and addons are available to each killer and the usage rate that plays into said kill rates and how many hook states.
So we know… Basically nothing, since only an averaged percentage result is shown and not the raw data used to make the calculations. Essentially, we see the result and can only theorize how it got to that point.
What we know :
– The kill rate for each killer at 2 specified (but still vague) ranges (as a percentage, not a decimal)
– The pickrate… "" …(as a percentage, not an integer)
What we don't know :
– Matches played for each killer across ranges (Determines Pickrate)
– Total Matches played for all killers across ranges (Solidifies data pool as a base comparison point)
– The number of Kills for each game of each killer and their average [as integers and decimals respectively] (Arbitrary, but is used as a comparison with other data, also helps determine which killers can defend hooks better)
– The number of manual hook states (which should be refered to as : 'Hook Actions') achieved in each game by each killer (Assuming which map each killer played on for each match + perks and addons were also factored, this helps determine which killers actually win chases more than others)
– Which addons were used the most for each killer and in each match (Helps determine probable cause for a higher/lower placement)
– Which perks were used the most for each killer and in each match (Helps determine probable cause for a higher/lower placement)
With this data, graphs with numerical value can be charted which determines these :
– Killer killrate (on each map)
– Killer Hookrate (on each map)
– Killer killrate (overall)
– Killer Hookrate (overall)
* Perks and addons should be factored in manually to determine probable increase or decrease to numerical value
Time taken chasing as a total time (in seconds) for each match may also be used to support the above statistics and determine chase shutdown ability for each killer. (Nurse ends chases faster than Trapper, etc)
I believe the 5%–15% MMR range, using the above values is the best indicator of actual killer strength.
It would have had much more sense to show the most frequent outcome for every killer …. Average means nothing.
Source: I’m a statistician.
Strangely, the killers that obtain the most kills are paid killers … what a coincidence … in this data I only see marketing, unluckily, not analytics for sure
Yeah I wanna see how many books happen per game cuz RN what I see is a lot of ppl getting one hooked cuz their left to die, that's my experience at least since the meta shift and especially the halloween event
Nurse gets 2,1 (rounded up from 2,08 for convenience) kills per game, what does that mean?
it means that in 10 games, a "perfectly average" nurse player has:
between 1 and 7 wins.
between 0 and 6 losses.
between 0 and 9 draws.
between 21 and 101 hooks.
Meanwhile a "perfectly average" Wesker/Pinhead/Sadako have 2,5 kills per game(rounded down from 2,54):
between 3 and 8 wins.
between 0 and 5 losses.
between 0 and 7 draws.
between 25 and 105 hooks
Lastly, here is what a theoretical "ultra-balanced killer" with 50% killrate (2kpg) does in 10 games:
between 0 and 6 wins.
between 0 and 6 losses.
between 0 and 10 draws.
between 20 and 100 hooks.
here is some interesting data I can pull from these numbers:
Being the best killer in the game or being the worst killer in the game differs as little as 0,4 hooks per game, but "identical performance" can differ as much as 8 hooks per game. (this is highly illustrative of just how much slugging/tunneling etc. matter; saving a single hook state per gam is the difference between overperforming compared to the best killer or underperforming compared to the worst)
a "perfectly balanced" killer has a win rate of anywhere beteen 40% and 60%; the more 4k potential a killer has, the more games that killer has to lose in order to be considered "balanced" by a killrate system; this means any killer that is comparatively good at getting 4k's (for example: killers who are good at slugging) have this feature "balanced out" by having a lower win rate compared to killers who struggle to get the 4k. (this might play a part in the "perceived performance vsstatistical performance" discrepancy between the community and BHVR's balance choices; a killer can have a below average win rate and still be considered "overperforming" by BHVR)
the top 5% playerbase kill rate is 2,45, which rounds up to the same 2,5 the Wesker/Pinhead/Sadako calculation round down to; you'd have to count 100 games for these percentages to be statistically significant.
All this does is pushing surv mains in their Killer op crusade.. even tho this data is telling nothing, the entiteled ones will use it to base their argumentation on it….
Idk wtf the devs wanna tell us
As someone who had to study research for psychology, I'm no expert but this is how I see it.
The basis for any research is that it should be able to be replicated to actually be considered reliable. So, as True stated, for this data to truly show how effective killers are on a consistent basis, they would need to show the month-to-month changes. If nothing changes and the rankings are constant, then it means something (that being the strength of killers on average). If it changed heavily from month to month, it's meaningless. But even if it was constant, you STILL won't learn much because there are so many variables to these numbers. Here are some:
1) What is the population of players for each killer? Trapper will not have the same amount of players as Nurse, for example. This in turn will skew the average and paint a false picture. This is why Nurse is so low at all MMRs, as fewer people will play a more mechanically advanced killer even though she is stronger than most killers. But because she demands lots of skill, some people will try her, do poorly, and move on to someone else. Plus, is this PC population or all platforms because Nurse and Huntress (and killers like them) are much harder to play on a console which will have a higher population than PC (more than likely).
2) Are the kills 1 hook? 2? 3? We just don't know.
3) How many were due to disconnects? How do we know some players didn't just kill themselves on purpose? Once again, the smaller the player population for a killer, the worse this issue becomes.
4) Were the kills at the start of the game? Middle? End? Were they all at once or separated? Knowing this would also show us if killers typically snowball or if it's a longer game. For example, does Nurse get all her kills at the start or middle and then nothing if it goes long? That would be vital information to know in terms of statistics, or at least, it would help paint the bigger picture.
5) Lastly, is this with add-ons? If so, the add-ons used for certain killers can make a HUGE difference in their performance. Think Wraith when he was better and used the silent bell add-on as opposed to using the meme-worthy blood point addons that Freddy and others have.