I Spent Every Spare Moment I Had Today Watching the Dev Stream So You Don't Have To

Hello again!

I'm back with another stream summary, for those who didn't want to or were not able to watch the recent technical livestream, and for those who did but don't remember everything that was said.

First of all, this is the latest summary I've ever put out, so I'd like to apologise for that. As implied by the title, I was completely booked all day today, and wasn't able to do a lot of work on this summary until I got home this evening.

Second of all, the usual disclaimers: although I have made every effort possible to ensure that the contents of this summary are as true, complete and accurate to what was said on stream as possible, these are my words, not BHVR's, and so I would strongly advise you not to read too far into the way anything is worded. Think of the below as official information, but with unofficial wording. The only exceptions are words or phrases in quotation marks " ", which are direct quotes from the stream, and anything in square brackets [ ], which is my personal note or conjecture and should therefore not be taken as official information.

If you would like to watch the stream yourself, you can find it at either of the below links:

YouTube: https://youtu.be/A8SwqJJKlaY

Twitch: https://www.twitch.tv/videos/551426466

Featured on this stream were the following members of the development team:

@Peanits (Community Manager)

Patrick Harris (Game Systems Designer)

Remi Veilleux (Technical Director)

And now, onto the stream itself.

Announcements

DC Penalties

Disconnection Penalties have been re-enabled to punish players who quit games early. They will only affect those who quit while the game is in progress – if you disconnect while loading into a game, you will not be penalised.

The developers are consistently monitoring the system to check how it is performing, and keeping an eye out for any issues that arise. So far, it appears to be working as intended, and they seem generally satisfied with its current performance.

Regarding player concerns about slow or unreliable internet: the developers would not recommend playing Dead By Daylight on “McDonald’s Wifi” to begin with. But in general, if a player disconnects during a match for whatever reason, intentional or otherwise, they have to consider it a “proper [intentional] disconnection” because there is no way to detect a difference between someone who turns their machine off or manually kills the game process, or who turned off their internet connection, and someone whose game or machine crashed or whose internet dropped out. They all look the same from the server side.

The early penalties are intentionally fairly gentle, so if this only happens to you once in a while it shouldn’t be a big deal, but the system is designed to really crack down on players who habitually disconnect from match after match.

Rank Reset

Ranks were reset today. Some players are experiencing issues where their rank gets reset back further than it should under the new rank reset system, which is a bug that the developers are looking into.

Others aren’t getting their ranks reset at all – if this is you, relaunching the game should fix the problem, but if it doesn’t, the team request that you file a bug report on the forums so they can look into the issue. [Here is the link to file a bug report: https://forum.deadbydaylight.com/en/categories/bug-reports]

Chapter 15 PTB

The Chapter 15 PTB will open next Tuesday 18 February.

Upcoming Developer Update

There will be another developer update coming early next week, which according to Peanits will contain “a bunch of spicy info”.

Technical information

This livestream will not be a simple Q&A. Rather, the developers will be going over hitboxes, latency, and what is being done to address player issues with these things.

Hitboxes

Hitboxes, despite the name, are actually more appropriately described as “capsules”.

[I have omitted about 30 seconds worth of jokes about the sturdy capsule in Remi’s pants. Check the stream at about the 6-minute mark if you’re interested in those, or, if demand is high enough, I can provide a transcript here.]

A capsule is shaped like a cylinder with rounded ends. Below are some visualisations of what the capsule-shaped hitboxes look like on killers (using The Trapper as an example).

As you can see, the capsule is slightly larger than the character. This is the collision capsule – it comes into play when the killer collides with a wall, an obstacle or another player.

Here is a visualisation of the survivor collision capsule (using Dwight as an example):

The survivor’s collision capsule is roughly the same width as the killer’s, but slightly shorter.

Here is a comparison of the two side-by-side:

There is another capsule hidden inside the collision capsule called the “slashable zone”. It is slightly smaller than the collision capsule.

Here is a visualisation of the slashable zone on Claudette, taken inside the Editor, which is the program the developers use when they are working on the game.

The slashable zone is the area which the killer hits when they attack a survivor. Most players refer to this as the “hitbox”, although the word “hitbox” is more accurately used to describe any of the different capsules. There is also a third type of hitbox [the hit cone], which will be shown later in the stream.

The reason why the capsules are the size and shape that they are is to prevent players from being able to stack the characters on top of one another, creating a “human pyramid”. Stacking two capsules is like stacking two balls – almost impossible. Players are welcome to try, however, and the team invites us to send them our best attempts at in-game human pyramids.

Another interesting thing to note about collision capsules is regarding how they interact in an environment like the basement stairs. If the killer stands in the middle of the stairs, it is possible for them to block the entire staircase, since there isn’t enough space on either side of the killer’s collision capsule for the survivors' collision capsules to move past. The Technical Team have therefore made a small change in the code such that if a collision capsule attempts to move past another, the first capsule will become slightly narrower, allowing the player to move past like “a fish in butter – a buttery fish” 🤨

The main reason why capsules are used is because the game was originally designed without projectile attacks [like Huntress’ hatchets] in mind, and the capsule system is optimal for melee play and does what the developers want it to. It’s also very efficient, not to mention standard for games in general. Remi claims the capsule technique has been in use for “centuries”.

How Hits are Calculated

There are two “steps” to registering hits. When a killer attacks a survivor, the slashable zone controls how the hit is registered on the “receiving end”. On the killer side, there is a kind of cone extending outwards from the killer which dictates where the swing lands.

Below is a visualisation of what this looks like.

Here you can see both characters’ collision capsules, and two cones – a red and a yellow – extending forward from the killer. These cones are used to measure whether or not an attack lands. The rules for this are simple: if the red cone overlaps with the survivor’s slashable zone, it counts as a hit.

All killers have the same sized “hit cone”, regardless of weapon length. There was an update to the game last Summer in which Legion’s animation was changed to have a longer visible reach, so that their visual movements aligned better with the size of the hit cone. For killers with larger weapons, the team will try to achieve the reverse effect, by “pulling back” the hit animation so that the weapon remains inside the bounds of the hit cone. This is so that players can have a better idea of how far their attacks will go. Having the killer visually and significantly miss but still hit the survivor because of the hit cone was described as “silly and we don’t like it”, which is why they have been changing the animations in order to make hits appear more consistent.

“Auto-Aim”, or “Aim Dressing”

The reason there is a red cone and a yellow cone is to do with what players typically refer to as “Auto-Aim”, but the team prefer to call “Aim Dressing”. The yellow cone is very slightly longer than the red cone, and its purpose is to detect whether a hit will land, and turn the camera slightly so that the player is able to see the hit happen. So it serves as a kind of visual “dressing” to make the player feel like their hits connected properly.

It will not give you “free hits” that you wouldn’t otherwise have landed. Its purpose is to address edge cases, such as when the survivor is at the very edge of the cone when the hit lands, by doing things like rotation the camera and moving the character closer after the hit has already been determined. The hit is decided on a technical level, and then the camera is moved slightly around if necessary so that the player sees the hit land, and so that they don’t get disoriented by getting hits that they don’t visually register as hits because they don’t see them connect.

Much emphasis was placed on the fact that, again, this does not give the player a free hit. It doesn’t help them mechanically, it’s just about dressing up the animation so that the player sees themselves get hits and has a better experience as a result.

If a survivor is not on your screen when you attack, you will not hit them. The only rule that is used to determine hits is whether or not the hit cone touches the survivor’s hitbox [slashable zone] when the hit is made. If it does, then it counts as a hit; if it doesn’t touch, then it doesn’t count as a hit. And if it does register as a hit, then Aim Dressing may come into play so that the player sees the hit connect.

Latency and Ping

The reason why players sometimes feel like hits have connected from too far away is not to do with hitboxes but rather with latency.

When a player is playing Dead By Daylight, every action is performed by clicking a button – attacking, moving, opening lockers, hooking, etc. It takes time, not just for the machine itself to register and process that input (which typically takes next to no time, less than is perceptible by the casual observer), but for that message to be transmitted to the game server and processed on that end. The amount of time this takes is referred to as latency, and latency is hugely variable, as it is dependent upon a number of factors. These include the quality of the player’s internet connection, and the quality of their connection to the host server.

The speed of the latter is referred to as “ping”. Ping is essentially the latency multiplied by two, since it includes the time taken for information to reach the server and to return to the player’s machine. Latency generally refers to a trip in one direction, from the player to the server, and ping refers to the time it takes for information to travel in both directions, to the server and back again. The two are often confused, but there is an important technical difference between them.

When the player initiates an action by clicking a button, that information is transmitted to the server, which then distributes it to the other players. For the vast majority of the Dead By Daylight playerbase – around 90-95% – the ping (time taken to send information to the server and back again) has been less than 150 milliseconds since dedicated servers were introduced. This is fairly good ping, at just over a tenth of a second.

The developers on stream showed a video demonstrating what a game would look like on an ideal connection, with no lag whatsoever. Here is a link to the stream at the start of said video, if you are interested in watching it for yourself: https://youtu.be/A8SwqJJKlaY?t=1085

Basically, what happens in the video is Dwight runs into the shack, vaults the window, and the killer attempts to hit him through the window after the vault (when Dwight is maybe around 2-3 feet, or a bit under one metre, away from the window), but misses. The same thing is then shown from Dwight’s point of view, with little to no noticeable difference.

The video was created on a connection with extremely low latency, and shows what the game would look like with little to no latency, meaning that the killer’s and survivor’s machines are communicating fast enough with each other via the server that they see essentially the same thing happening on both screens, with no perceptible difference between the two.

They then showed a video of the same scenario with 150ms ping, which is a small, but in some cases quite noticeable, difference, as seen in the video: https://youtu.be/A8SwqJJKlaY?t=1133

In the second video, the killer very clearly hits Dwight as he is going through the window. However, from Dwight’s point of view, he manages to make it through the window and is at about the same point as he was in the first video (2-3 feet away from the window) when the hit connects.

This makes it clear that even a small amount of latency can have a major impact on visible gameplay, especially when dealing with very quick actions like window vaulting. The developers are, obviously, not satisfied with this state of affairs, and are therefore working with various techniques to minimise situations like this going forward.

It is also worth remembering that the killer’s connection takes priority over the survivor’s when it comes to hit detection, so if the killer gets a hit from their point of view, then the server will register it as a hit and the survivor will be affected accordingly. If this were not the case, killers could watch their weapon go right through a survivor and never get a hit due to latency, which would feel great from the survivor’s perspective, but terrible, not to mention potentially game-breaking, for the killer.

How Lag is Felt in a Multiplayer Game

The developers have plans to address such situations, and prepared a presentation to explain said plans, which is, incidentally, the same presentation the Technical Team used to explain their goals to the rest of the development team. So naturally, it's very professional. [I've summarised it for you below, with visuals included]

In a first-person shooter, there might be one player in green, who we’ll call Arnold, who is running to the battlefield, and his opponent in red, who we’ll call Brandock, who is also running to the battlefield at the same time. Both fired up, guns blazing, etc.

As we know, there is typically a small amount of latency (say 100ms ping) between both players and the server, meaning that the server will perceive both players not as they are currently but as they were 100ms in the past.

From Arnold’s point of view, he will see himself as he is currently, but his perception of Brandock will be slightly behind.

Brandock will see the same thing in reverse – himself as he currently is, but Arnold slightly behind.

In all three cases – from the server’s perspective, from Arnold’s perspective, and from Brandock’s perspective – the distance between the two players is the same and only their exact positions differ. The fight that is about to happen between the two players will proceed roughly the same way from all three perspectives, just in a slightly different location.

The difference between a generic shooter like this and Dead By Daylight is that the survivor and killer are not usually running toward one another, but rather, the survivor is fleeing from the killer. Below is a visualisation of how such a scenario might appear from the server’s perspective:

From the survivor’s point of view, because he sees himself in real time and the killer slightly behind in time, he will perceive the killer as being further away than the server does.

The killer, similarly, will see himself in real time and the survivor slightly behind, and thus he will perceive the distance as being shorter than the server does.

If the killer attacks the survivor in this scenario, the system will register it as a hit for both players, because hits are registered locally to the killer.

The hit makes sense from the killer’s perspective, but the survivor, because he perceived the killer as being too far behind him to land a hit, feels cheated.

This is why survivor players tend to complain a lot about hitboxes, which is a normal and understandable complaint to take away from this situation.

When thinking about how to address the issue, the team are operating to a philosophy of “gameplay first”, meaning that they want the game to feel responsive to player input. So, if a player performs an action, they want to see it register in the game right away.

To this end, they want to design the system such that it can predict player movements and adjust their position accordingly on the server side, instead of just relying on delayed information received through the network. This was tested on the most recent PTB.

To prevent abrupt changes in direction, which cannot be predicted as easily, they also made changes on the PTB to “dampen” the input and smoothen the character movements. However, there were mixed reactions to this, as it rendered certain “advanced” movements and techniques which relied on abrupt changes in direction, and which were core gameplay elements for some players, impossible or extremely difficult.

The challenge the team are currently facing is that of striking a balance between the first point, wanting to keep the game feeling reactive and responsive to player input, and the second point, wanting to reduce the disparity in distance between players as demonstrated above. The aim is to find a way to make the game feel as fair and synchronised as possible with minimal impact to a player’s sense of movement and control, something which they are ongoingly working to achieve.

It is impossible to get rid of latency altogether, but there are small optimisation tweaks that can be made to minimise the impact it has on the player experience.

These changes won’t really change how often killers are able to land hits, because the killer’s connection already has priority as far as hit detection goes. Hits are and always will be registered locally to the killer, so if the killer sees themselves get a hit on their screen, then the server will register them as having landed that hit, regardless of what new measures are put in place to reduce latency.

The plans the team has for optimisation should reduce the discrepancy between what the different players see on their screens, so that everything each individual player sees will appear a lot closer to how the server perceives what’s going on. Currently, under moderate latency conditions, killers may see a hit connect while survivors see themselves being hit from several metres away, which doesn’t feel very fun. With these new fixes, the survivors will see the killer land a fair hit whenever the killer also sees themselves land a hit. Both players will see pretty much the same thing. It won’t be about giving killers more or less hits, but rather about having what happens onscreen make sense and feel fair to all players involved.

Q&A

As a small aside before we get to the questions, the developers made it known on the stream that they really appreciate receiving these questions, since they see them as feedback, and all feedback is good feedback, because it means that the players care enough about the game to take the time to give it. “Constructive feedback is always welcome” - Peanits

Why are Survivor hitboxes a capsule? (from @Hoodied)

Characters are actually made of several thousand triangles. If the hitboxes were made “realistic”, collision would have to be applied to every polygon that makes up each character, which would be very demanding on the players' hardware. It’s both customary and practical, therefore, to use “simplified geometry” like capsules. It is also much easier to program things like friction and collisions on a simple shape than on a complicated humanoid one with a lot of jagged edges. So essentially, capsules are used instead of “realistic” hitboxes for both technical and optimisation reasons. If this were not the case, the minimal system requirements for the game would have to be raised much higher. (It also stops the characters from getting easily caught on obstacles in the in-game environments.)

The goal of the game has never been exact and precise hit detection, so the choice was made at the beginning to go with the simplest and most efficient solution. Dead By Daylight is not a precision-based game, in the sense that a first-person shooter might be, so having hitboxes be the same shape as player models is not strictly necessary in this case. It’s not impossible, it’s just not to the overall benefit of a game like Dead By Daylight, and it is not how the game was intended to be played. If weapons had to connect very precisely with part of a survivor’s body, instead of just being able to swing at close range and hit them, it would feel “weird and not very fun”. Simplifying the hitbox shapes, in the case of Dead By Daylight, actually makes the game more fun overall. It also makes it fairer, so that smaller survivors don’t have an advantage over larger ones when it comes to hit detection.

Additionally, having hitboxes be the same shape as the characters would make things like latency far more difficult to deal with, because every player would need to see the same animations happening, down to the individual frame, at the same time, in order for hits to feel fair or reasonable. It would also require many, many exceptions to be programmed into the level designs, which would make that process much more complicated and time consuming.

To summarise: having hitboxes as capsules instead of complicated shapes makes the game more streamlined from both a development and a gameplay point of view.

Do killers have unique hit detection ranges? (from @edgarpoop)

No, all killers have exactly the same sized hit cone, or hit detection range, on both their basic attacks and lunges. However, some killer powers do change hit detection variables. The Nurse, for example, after blinking, will attack faster than usual. Killer weapons like the Chainsaw and the Oni’s Kanabo also have unique “rules” attached to them. Basic attacks and lunges, however, both have the same range across all killers.

Why is it easier to 360 a killer on console than on PC, and are there any plans to make it more fair? (from @Hel24330)

It’s not so much a matter of the platform as it is of the controls being used. It is easier to do the smooth rotation required for a 360 with a joystick on a console controller than it is with a keyboard. Keyboard controls are more comparable to a D-pad on a controller. On the other hand, it’s much easier to rotate fast with a mouse, which means that it is much harder to juke a console killer with a 360 than a PC killer, because the PC killer can respond faster and get the hit anyway (although is something the developers are aware of and are working to fix by improving the sensitivity of controllers). So both of these factors combined make successfully pulling off a 360 on console easier than on PC.

How does the auto-aim work and will you give us an option to disable it? (from @Stitch7833)

As discussed earlier, “auto-aim”, also known as Aim Dressing, is really just a visual aid, not a mechanical adjustment, and does not give the killer hits they would not otherwise have landed. It is not an aim assist. Given that it is not a mechanical feature, there are no plans to make it a toggleable option, as it is part of the way they want to present the game.

How are the dedicated servers performing? (from @cricketscorner)

The dedicated servers run on a Linux machine with no video card, meaning that they can run the game in a stripped-down state and very efficiently, so they can handle running many matches concurrently on the same machine. They are also very fast, because they are hosted on Amazon’s data centres.

The team do track statistics regarding things like latency over time. Currently, the vast majority of players on dedicated servers, around 90-95%, have less than 150ms of ping to the servers, and around 85% are seeing ping of less than 100ms, which is a huge improvement over the peer-to-peer system, and is also much more stable and consistent.

Dedicated servers also give the developers the ability to do things with the game that they weren’t able to do before. Dedicated servers, from a design standpoint, are a “trustworthy authority”, meaning that the team can trust any information dedicated servers provide about what happens in-game. The disconnection penalty system is one example of an advancement which has been made possible by the presence of dedicated servers. They also act as a neutral base on which players from different platforms can unite, such as with the existing crossplay between the Steam and Microsoft stores.

There will be more features like this that the team have plans to implement in the future, which wouldn’t be possible without the dedicated servers.

As a side note, they are still working on bringing the dedicated servers to Xbox. They definitely haven’t forgotten about it and they’re hoping to make it happen in the near future.

Why can survivors sometimes vault immediately after being shocked? (from @Snapshot)

Basically, the answer is latency. It has to do with what was explained earlier in the stream, about how lag manifests in Dead By Daylight. From the Doctor’s perspective, he may have shocked the survivor before they vaulted, but the survivor may not have seen the shock attack until after they completed the vault. In situations like this, the server acts as a kind of impartial mediator who decides which of the two actions happened first, and so if the server decides that the vault happened before the shock attack, then the killer may see the survivor vault immediately after attempting to shock them.

If the two players initiate different actions at exactly the same time, on their own screens, they will both begin performing those actions. However, both pieces of information take time to reach the server due to latency, and so the server will decide, if it receives both pieces of information at around the same time, which one arrived first, and that action will take priority. While it is possible for a shock to cancel a vault that has only just begun, usually when a shock lands after a vault has been completed, latency is what is causing the issue.

As discussed earlier in the stream, the developers are working on ways to reduce the impact that latency has on individual players’ experience, and these new features will be released when they are ready.

It is also worth noting, in this case, that the Doctor does have a built-in delay between when he finishes charging his Shock Therapy attack and when it shocks survivors, so this could also be having an impact as the server does take this into account. So it is not only a matter of who pressed the button first, but of how far through the vault action the survivor is and when exactly the shock effect kicks in.

Why do Status Effects on console cause frame drops? (from @thesuicidefox)

This is something that the Technical Team are aware of. As part of their previous optimisation efforts, they have focused on reducing the time it takes at baseline to process all the different game effects. Some actions are now able to sustain the target of 60FPS, but some effects, status effects included, which have to load an asset when they are activated, may still cause performance dips. This is something that the team are still working on fixing, since they understand the impact that this can have on gameplay, especially at critical moments when players most need their game to run smoothly.

Side note from McClean: There was a bug which caused Madness 3 screams to interrupt fast vaults, and a fix for this is on the way.

How do pallet/window hits work? (from @AWildLuke)

Hits at pallets and windows work very similarly to the interrupt system. They are based off the same hit cone (explained above) as ordinary attacks, and the same rule applies – if the killer’s hit cone overlaps with the survivor’s slashable zone when the killer attacks, the survivor will get hit.

With window hits in particular, issues with latency often come into play. As demonstrated in the earlier videos, sometimes the survivor can get hit when they are already on the other side of a window because on the killer’s screen they were still in the middle of the vault. The survivor, obviously, is not supposed to be hit if they have already made it far past a vault, but they will still be hittable if they are close to the window, as long as they are still inside the hit cone.

All the rules around how pallet and window hits work are essentially just a logical extension of everything that was discussed earlier in the stream regarding hitboxes. This is, however, impacted by latency, and so different players may see things differently on their screen from how those actions are actually registering on a technical level.

Why is Dead Hard broken on dedicated servers? (from @Muntcuffinz)

Dead Hard is not broken on a fundamental level. The issues around Dead Hard are not issues with the perk itself but with latency, because Dead Hard, more than almost any other perk, relies on extremely precise timing to work properly. This means that without having special cases programmed into the servers to handle such situations, the perk and its success can become unpredictable. If a player had little to no ping, say 30ms, the visible difference would be very minor, but at a higher ping of 150ms, a perk like Dead Hard which requires split-second timing will be impacted.

Cases like this are being specifically looked at by the team as part of their efforts to minimise the impact of server latency on players' experience. In some cases, Dead Hard being one example, unique systems may need to be implemented to get them to work as intended. But this is certainly a problem that the developers are aware of and are working in improving.

Do different characters have different sized hitboxes? (from @Pythonheir)

This was covered earlier in the stream – the answer is no. The developers decided to make each character’s hitbox the same size for a couple of reasons: firstly, for maneuverability within the levels and ease of level design, and secondly, for fairness, so that certain characters don’t have an advantage over others due to having smaller or larger hitboxes. And the same goes for the collision capsules on killers – visually larger killers, in spite of appearances, do not loses more distance around loops than smaller killers on account of their size.

Not only would having differently sized hitboxes for each different character make things extremely difficult for the level designers, having to design maps and routes with 50 differently sized characters in mind, it also wouldn’t add anything of significant value to the game in terms of player enjoyment. Incidentally, map design is also the reason why killer and survivor collision capsules are very similar widths, so that all characters are able to navigate through almost all of the same spaces on the maps.

How does the game know when a survivor can be stalked? (from @Edys)

The technical aspect of this is referred to as the “Occlusion Query”, which is where the game checks what proportion of the other character’s torso is visible by locating the torso visually and then checking for obstacles and line-of-sight blockers in between you and the other character. It also factors in how close the other character is to the centre of your screen – if they are too close to the edge, you won’t be able to stalk them – and the distance between you – if the survivor is too far away, you also won’t be able to stalk them. The same is also true of survivors trying to break Ghost Face out of stealth mode, where the exact same mechanism is used.

The other aspect of this is related to latency, which involves the connection between your machine and the server, and how the positions of various players are registered at both ends. Issues like this largely have to do with the discrepancy between what your machine thinks is happening and what the server thinks is happening, and latency can cause delays and therefore further discrepancies. This is something that can be optimised further, in order to give players more of a sense of immediate responsiveness, and it is something that has been improved overall with the launch of dedicated servers, particularly for killers like Michael Myers. It is also something that the developers are continuing to work on as part of their efforts to reduce the visible impacts of latency across the board.

Many of these problematic mechanics which players have questions about have their own unique considerations that need to be taken into account, especially with dedicated servers and how latency is handled there, and each one will require specific attention in order to be properly fixed. The team have a big list of all such mechanics that need to be dealt with, and they are working through them as fast as they are able.

Having that dynamic where your machine has to essentially ask the server for permission to perform an action makes you “a slave to latency”, which can happen either when the server won’t let you stalk someone you can clearly see, or when they let you stalk them but slightly delayed, so that by the time you are able to stalk the survivor according to the server, the survivor has moved, which can result in it seeming as though you were able to stalk a survivor through a wall or other line-of-sight blocker. All of these situations can result in feelings of confusion or dissatisfaction for the players involved, but it’s not actually a problem with the stalking mechanic itself, just with latency between the player’s machine and the server.

Why does healing have such a big hitbox? (from @TraitorousLeopard)

Contrary to popular belief, entering the healing animation doesn’t actually change the size or shape of the players’ capsules. They remain the same height and width as when the character is walking or standing still. When a survivor is standing up, their head will align roughly with the top of the capsule. The same applies when a survivor is injured – even though their character is visually hunched over, the capsule is still in the same position as when they are healthy. When crouching, the capsule will lower in order to match the player’s movements, to allow them to do things like dodge hatchets, but when healing, the hitboxes do not adjust to fit the animation.

This is a known issue that the developers have recently been discussing and have decided they want to work on fixing – not only for the healing interaction, but for various other interactions which suffer from the same problem. This was not an issue when the game was first released, but it has since become one after the introduction of projectile-based killer powers, such as those of Huntress, Plague and Clown. Now that it is becoming an increasingly major issue, it has become more of a priority for fixing. And the team want to fix it, because they are always aiming to deliver the best game they can to their players.

Why do grabs often get cancelled without a hit? (from @kamisen)

This is part of the “infamous” interruption system. When the game was first released, it operated exclusively on a peer-to-peer system, where the server was the killer’s machine. So whenever the killer tried to grab a survivor, it would work, because the killer had little to no latency. Since dedicated servers were introduced, this mechanic has become increasingly problematic, since not only does the killer have to register a grab on their end, but the survivor’s machine also needs to register and acknowledge the grab for it to work. There can be a significant enough delay between these two inputs that by the time the information from the killer reaches the survivor, the survivor has usually already stopped performing the grabbable action. Essentially, there has to be agreement between the two machines, and the killer can’t grab a survivor who is no longer grabbable. This feels unusually bad for the killer because they not only miss their grab, but they also don’t get a hit either and just end up “waving” at the survivor.

This is another important latency-based issue that the developers are not satisfied with currently and are working on fixing, with the aim of creating an experience that makes sense and feels fair for all players. Interrupts and grabs, in particular, can definitely be improved upon because the developers have a very clear idea of what needs to be done to fix it, so in this case it is “just a matter of time in an ongoing process”.

Why do you guys think it’s better to prioritise hits on the killer side over the survivor side? (from the live chat)

The main reason is because the killer is the one performing the action, and so they need to make sure the action goes through in order to maintain a sense of responsiveness. But even though the killer’s connection is prioritised when it comes to hit detection, they also want to make sure it looks and feels fair from the survivor’s side as well. This is not happening as much as they would like currently. One solution they are looking into is having the server do the acknowledgement of the hit, rather than the killer's machine, so that it can mediate between both parties and reach more of a middle ground instead of prioritising the killer’s connection above all else.

If there is no auto-aim, then why are there times when the game will prioritise hitting an object over the survivor? (from the live chat)

This could be an unintended side-effect of the Aim Dressing feature. So the system may register your hit, rotate your character towards the survivor you hit so that you can see it connect, but there happens to be a wall or object in between you so you end up hitting that instead. This is definitely not intentional and would be considered by the designers to be a bug, or flaw in the system, which they would like to fix. [This is just Remi’s theory off the top of his head as to what could be happening in that situation, not necessarily the exact cause as proven by rigorous testing.]

Essentially, the suggestion is that this could be an “order of operations” bug with the Aim Dressing, where the system checks for nearby objects after rotating the camera, instead of ignoring environmental collision at that point as it should. This happens particularly often around places like sharp corners. There is definitely something to be done to fix this issue, and it is requested that you report it as a bug if you encounter it in-game. [Here is the link to the Bug Reports subforum again: https://forum.deadbydaylight.com/en/categories/bug-reports]

End-Of-Stream Surprise

A short video clip was played at the end of the stream [here is the link to where it begins if you want to watch it yourself: https://youtu.be/A8SwqJJKlaY?t=3408] which revealed that the Trapper is now able to reset his traps without having to pick them up and replace them. [Presumably, this feature will be introduced to the game alongside Chapter 15.]