Transcript for:
Great Level Design is a Studio-Wide Effort

[Music] hello and welcome to my talk great level design is a studio-wide effort my name is dana nightingale i was born in the us and in the united states and very shortly after that began gaming on an atari 800. in the 90s i absolutely fell in love with games like system shock and thief which awakened my passion for game design i contributed to emerson's immersive sim fan sites for years excuse me i've been practicing so many times today my whole mouth is messed up i earned a master's degree in architecture in 2008 two years later i joined arcane as a level designer in austin texas to work on dishonored i moved to france five months later where i remained to this day finally i became the campaign designer of death loop in 2020 but what's my gamedev experience as a modder i worked on a bunch of missions for thief 2 the metal age and a few quest and location mods for skyrim as a level designer i worked on dishonored's flooded district the loyalists and eminent domain i worked on the clockwork mansion of dishonored 2 and death of the outsiders bank job i did some map layouts for wolfenstein youngblood and finally i worked on a part of death loop's prototype alexis dorsey's mansion which is featured in the trailers as campaign designer i am currently working on death loop's infamous merger puzzle that said i've had a very focused career working almost exclusively in level design i've only worked at one studio two if you count arcane austin which you should so two studios this presents a challenge i do not have the broadest experience so i have researched i've interviewed people from other disciplines with more diverse backgrounds before i get started i'd like to explain who this talk is for in the presentation's goals it's for people who love great level design even if they're not level designers and it's for teams making games which have levels in them it's also for my friends i'm going to discuss the role many disciplines have in making great levels i'll explain how arcane leon does it or the lofty goals we try to reach and ways we could be doing it better but i also want to put a spotlight on everyone who's helped me make levels places that players enjoy and talk about all of that while trying to go as in depth as i can in 60 minutes i should define what i mean when i say a level what a level designer at arcane leon does and what i feel makes a level great yes when i say a level i mean one of many segments that the game is divided into or a location in the game world or a moment in the player's journey at arcane leon a level designer is defining the layout and flow of the map crafting encounters and vignettes with all of the scripted logic that powers them but most importantly are part of the creative process level designers have authorship of their maps from my perspective a great level brings the most out of the game's narrative and systems it captures the imaginations and desire to engage with the content it feels like a place or a moment that matters either to the game itself or to the player and in those truly great cases is something that stays with them after they finish the game they remember and talk about it arcane as a whole has a bit of a reputation for our level design which a few people think is pretty good i happen to be a big fan too i'm extremely proud to be a part of this team and in my other work to bets so what's the plan okay it's it's not this the talk is divided into two parts your team and processes in each part we'll look at four challenges you're going to face and look at solutions to each one part one is your team the first challenge is an evolving level design team undishonored we began with four level designers who had been working with us since dark messiah and also on ravenholm elemento and the crossing with eight new level designers mostly joining from arcane austin this is where i started i was here in uh dishonored we brought in six new level designers to arcane leon because arcane austin's ld team began to work on prey death of the outsider saw three new level designers join and death loop added five more over this decade many team members became leads or specialists some changed to other disciplines while others moved to other studios we had it influx constantly of new talent but how do you keep your work consistent if your team keeps changing it comes down to a clearly defined studio culture and that's not not something that you can just conjure up the ideal is to build a level design team with a foundation in certain design values above all we are a studio that makes a certain type of game and we recruit along those values level designers come here to work on that type of game good level design can't just be a value for the ld team it must extend to the entire studio and almost everything in our games is expressed through the level design the whole studio considers level design as one of the most powerful tools to deliver the experience that we are all after and arcane we never shut up about our design values we talk about them constantly we have posters about them i'm sure people are sick of having one brought up as an argument in meetings but everyone knows them and no one forgets them i could talk for an hour about arcane's design values and many people have but that's not the point of this presentation the point is that your studio has its own values and you are the ones defining them our second challenge is leadership bottlenecks what do i mean by bottleneck uh in this case it's when an idea doesn't move forward until it works its way up the bottle a level designer creates an idea which is reviewed by their lead who will give feedback help them iterate and ultimately validate it or not if validated the producer will evaluate the impact on other teams even validated at this stage which again isn't a given when the creative director gets involved who may be reviewing it without the ld present they may not feel it fits with the vision of the game and kill it not only does this weed out a diversity of ideas but it could also take weeks so what's our solution eliminate the validation process no it's probably not a good idea instead we have a team of leadership our level design team does not have a flat hierarchy but we have several leaders who work together in addition to the lead there's the campaign designer hi that's me and the lead technical level designer [Music] together we sit down with the level designer and review their ideas with the campaign producer and the creative director if one of us can't be present the process happens anyway we keep things moving we involve the ld so they can defend their ideas which then don't have to be validated over and over sometimes we disagree and that's the point we have different tastes and areas of expertise we talk it out we try to find a consensus or a compromise and if and sometimes we just trust the person who's the most qualified to make that call this means a bigger variety of ideas can make it into the game okay so what's each of our areas of expertise who is the most qualified for what the lead level designer defines the level design objectives and guidelines they work closely with systems design and level art the campaign designer crafts the backbone of the player's experience they work closely with narrative and ui and user experience the lead technical level designer is all about the team's tools and processes they work closely with tools development and tech art as you can see we have a pretty balanced load in terms of what other disciplines we work with leading each of us focus on our particular area and trusting the trusting each other to have our backs for the rest [Music] in a nutshell discuss ideas together as a group not as rungs in a ladder where each rung could break we have a variety of leaders with different areas of expertise and different tastes include the level designer in the validation process let them defend their ideas some people can present their ideas better on paper and some can do it better in a discussion give them a chance and keep the process moving deference and trust are required challenge three is masters of none there is a huge variety of thing a level designer has to be good at you know jack of all trades a level designer creates the map's layout the challenges the logic of events and so much more every choice a level designer makes can impact or define the narrative every vignette involves a ui or user experience element whether it's an interaction prompt or an objective notification many game systems happen inside the level design how does the maps design get the most out of those systems does the gray box layout match the architectural intentions with the right proportions scales and rhythms of spaces is the map following all metrics and constraints and optimization guidelines is the level designer and expert on every tool including scripting logic in various form of markup this is a lot and no one person can be an expert at all of it i know i am not so what can we do to mitigate it the diversification of level design specializations this is divining level designers into two types map owners and specialists specialists ease the burdens on map owners a narrative scripter is part of the narrative team but with a level designer's skills this allows them to autonomously carry out narrative setup in the maps the campaign designer makes sure the ui is in sync with what the level designers are doing and guides the lds and how to best take advantage of it a level systems designer can be one or the other but has the skills in both they provide the lds they provide an ld's insight into the development of some systems and assist lds in getting the most out of those systems level artists work directly with level designers on the geometry so lds don't also have to be masters of the fundamentals of architecture needing to create gray boxes that transition beautifully into realistic buildings that's not the point of our gray boxes tech lds assist in tools development but also work with level designers on highly technical tasks this has helped arcane so much it is not necessary for map owners to be experts at everything there are people on other teams who have level design skills let them contribute but most importantly be sensitive to what each member of your team needs they don't all have to be map owners let one focus on technical tasks in another work unnarrative let them focus on their strengths challenge four is overwhelming dependencies no discipline works in isolation and game development has many moving parts the whole design must work with nearly all of them but what does that look like what are level designers key partnerships first what is the work that happens before level designers begin what allows them to do their work in defining the shape and scope many questions have to be asked the broadest one is what does a level even mean in our game are the levels connected by ellipses or are they contiguous how big are the maps how dense are they do several maps make a level like we had in dishonored or is it something else entirely like we're currently doing in death loop it's critical that technical that the technical director and level designers are in sync with creative direction and art direction if they're not be careful what feels like the final decision can change and it will be painful if it does [Music] engine and tools developers have a hand in laying the foundation for the ways lds and level artists do their work or what work is even possible at all to ensure that these tools fit the needs and processes of both teams tech lds and tech artists are involved early tools are created workflows are mapped out best practices are defined now lds and level artists have exactly what they need tools that are concretely designed for the exact work they're doing now and then we try to do something wacky like dishonored 2's time travel mission crack in the slab or the clockwork mansion our engine and tools teams have the flexibility to make our dreams come true but we don't just use tools to build maps we also need to implement game systems technical lds will work alongside the systems designers in a mix of engine developers and gameplay programmers the work done by these teams is critical to is critical to making immersive sims void studio is built from the ground up for exactly this type of game we've created a visual scripting language called kiscool tailored exactly to our needs and the needs of the games we're making we have a modular component system that ensures behaviors are consistent game wide which is a key to systemic game design we have external tools like a content integration system or performance monitors but we sometimes like to give them funny names like that delivers warnings or tintin our bug reporter porting the golden path is where i come in our games are very narrow to driven but narrative and gameplay are one what the story is and what the player can or must do are intertwined this forms the campaign the campaign designer will collaborate closely with narrative team lead level design and creative direction to form the high level outline of the player's goals they then collaborate with the lds to turn this outline into detailed scenarios challenges and vignettes the game world is fleshed out from a collaboration between level design narrative and art direction a level design problem can lead to a solution that makes the game's world richer one's narrative and art get a hold of it which what was a problem now fleshes out the world and dishonored we wanted urban maps which implies streets but we also didn't want to put up invisible walls so art and narrative invented massive street blockers which fictionally controlled the spread of the plague but also enclosed the gameplay space a narrative designer may propose something uh surprising a concept artist runs with it exciting everyone leading to a level designer creating a vignette or maybe an entire level around it but this can only happen if each team member is attentive to one another and feels free to collaborate [Music] now we're in the thick of it what keep partnership come into play while the level geometry is being made the crafters of the maps may be our most important partnership the level designer and level artist will begin work at the same time the ld creates the layout and flow of the maps in gray box which establishes the size scope paths basically what the level is the level artists creates the visual dna of the map this proof of concept establishes the scale and proportions the style and the mood once both plans are validated these two join forces the level artists are not doing an art pass on the maps they're creating a new map from scratch based on the ld's layout but with the visual dna of their proof of concept pairs of level designers and level artists will work together until the end of the project collaborating to bring their combined vision to reality little artists work closely with environment artists to establish a modular system called a breakout kit that can be snapped together these are intended to be replaced with authored assets late in production during the polish stage level designers also work with environment artists to create unique props this can be a complicated process involving concept art and level art as well and can require a lot of iteration since this piece may be part of a puzzle that players need to understand just by looking at it the moving bank vault and death of the outsider was constructed by someone who could do both concept art and environment arm concept artists will absolutely become involved to visualize the spaces doing paint overs of the gray box or the level artists rough blackout sometimes this can bring out new ideas with the concept artist providing a vision neither map owner had anticipated practically speaking it is useful to define the details that the environment and lighting artists will use other times it's just exciting you get to see what the map could look like in the finished game possibly years before the art and rendering are ready to make it possible it can be an inspiring moment that invigorates the passions of everyone involved we require illumination our level design and level art wouldn't mean much without lighting how would the players see once a map's layout is stable a lighting artist steps in and the three will work together a map's lighting could be subtle and functional there but never thought about or it can be emotional heavily pushing the desired mood or it can be directional giving the player critical information guiding them forward the lighting artist will use all of these methods to make the map shine pun intended environmental storytelling is the multitude of stories that exist in all spaces in the game expressed either explicitly or left to inference it has two forms high level is why does this space exist why is this place like this it is often conceived by the narrative team creative direction and uh level art leadership this forms the core concept of a location low level is what happened here what do these objects communicate to the player low level comes from all directions often proposed by the narrative team but just as often from the level designers or the level artists constructing the level geometry is now well in progress but we're making games what about the game play what about taking a drink because my throat is getting worn out just as we have emergent gameplay we have a merchant level design our level designers have a lot of freedom to invent wild the work done by the systems designer and gay play programmers allows lds to invent novel weird or just plain game breaking stuff that that that that you they create new uses for everything it's not uncommon for an ld to push the envelope with our tools and create a spaghetti pile of logic just to get it working yes i am very guilty of this i think it's great but we still must be careful this is when tech lds and gameplay programmers and systems designers get involved it could be coded into a fully supported new feature this is a best case scenario for everyone and it happens very often it could be cleaned up and optimized by the technicalities even without code support a specialist can do a lot with our tools in either outcome it can't become a nightmare for qa this is especially true if they come across what looks like a new feature but there's no documentation and no metrics and the only person who even knows it exists is the level designer who cobbled it together player affordance is near and dear to our hearts at arcane expressed simply if it looks doable let the player do it but also if we can't let the player do it don't let it look doable this is expressed through a union of environment art level art and systems design level designers will define rules about about what looks interactive or what looks traversable can i fit into that tunnel can i open this door it looks like i can systems design will set the player's expectations for combat tactics like how big of a gap can uh can an enemy see through or what counts as cover and what doesn't environment artists interpret these guidelines but they're often the start of the conversation they may want to make a certain type of asset like a box of fireworks well if the player shoots it they'll expect it to explode we achieve a narrative rooted in gameplay through level design systems design and narrative design bouncing ideas off each other letting them evolve here's one example narrative decides that the protagonist should have a personalized weapon that expresses their themes and death of the outsider billy lark had a voltaic gun that represented her unconventional methods and pragmatic ingenuity systems design uh devises a unique ammo type pencils and pens [Music] level design could then have you visiting locations where pens and pencils are laying around say why not a bank it can move in the other direction level design creates a puzzle uh there's a button that you have to shoot systems design invents a rubber bullet that can impact silently and then be recovered narrative design american design brings that tool into the game's world i think it was part of a fish but players can't beat our minds the game has to communicate possibilities in play what is allowed and what is needed to progress the campaign designer will work with level designers to transmute their designs into communication via the user interface this can take a lot of forms like objective texts or markers on the hud a quest journal or failure states the narrative team is involved from the start weaving story through objectives and doing the final writing pass so now the map exists what happens next a static map is not an immersive map what brings the maps to life [Music] sound designers vfx artists and animators work directly in our maps to create vibrant reactive living places using the same tools as the lds these team members are artists and designers and may see a way to do something a bit differently than what the ld asked for who doesn't have the same eye for this kind of work in the clockwork mansion the electroshock sequence was pitched as being quite subtle these three teams punched up the intensity over and over both trying to reward the player with a dramatic outcome and drive home the horror of what you were doing this couldn't happen if these teens just had a list of assets with no context and no say in how they're used there's a deep symbiosis between systems design and animators together they bring the npcs to life but she gets real when l level designers are involved a little the level design could be causing trouble for npcs the player could often be crouched inside a low tunnel leaving enemies baffled even though they're within reach animators could step in and create a new set of animations to solve this level designers could need to convey something about the area maybe a dust storm has started which is hazardous to the player the animators have the nbc's bracing themselves and shielding their eyes making it clear without exposition animators can push hard on certain behaviors taking it beyond what was proposed our games have a feature we call distractions where npcs move off their patrols to do a trivial action a handful of them would do but our animators create dozens and dozens in a vivid and tactile world so much is communicated through sound a button is modeled and animated the sound is what it makes it feel like you pressed it there's that click and then the machine roars to life or maybe it gives a dull empty sound and a crackle maybe there's something about the sound that tells you the machine needs power or is broken or if it does turn on it will only stay on for a short time sound can communicate in intuitive ways instinctive ways that are faster and and more intuitive than a ui prompt or the player character explaining things so much of what you see can also be heard atmospheres equal visual equally visual as auditory tech artists and rendering engineers create visceral experiences with the sound designers the gentle hum in the background perfectly matching the haze surrounding you the crunching under your feet letting you feel the shaders on the ground the player will feel like they're there the map is now a place [Music] details tell a story a light touch of vfx can deepen the player's involvement in what the lds and level artists have built you could add flies over a trash can next to a door to tip the player off that that house isn't abandoned and something fishy is going on in there beams of light through cracks in the ceiling signal that your long journey back to the surface has come to an end maybe this is an exciting moment of relief or a weary somber moment the way the beams of light move can push either feeling a cloud of dust could appear when moving an object deepening the feeling of intruding upon a forgotten place details like these aren't just immersive they carry information the player can learn or understand without realizing that they have and finally how does all this work how how how do you get this work into the players hands how do we get there someone has to keep the ship afloat and eventually you need to ship our testing teams are divided between quality assurance and dev testing typically qa testers will be assigned maps or parts of the campaign and switch every few months to keep ideas fresh a dev tester will attend each weekly level design meeting so their dev team always has the latest info on what the lds are up to big complex features that are tied deeply to level design may get special focus from a dev tester who will attend the kickoff meeting and provide feedback players usually like all of the frames per second optimization is both the first and last thing our technical teams work on in a production our game systems are cpu heavy so from day one everything is designed with optimization in mind this impacts the scope and density of our maps if you run a foul of this you may need to split your map in two or three or four and that's harder than it sounds it will define how many active ai can be added location at once we have to do all sorts of tricks to limit this invisibly without making the maps feel empty sidelines are one of the fundamental tools of level design and rendering metrics determine everything that we can do our systemic world and complex ai don't leave much cpu power left for our scripted logic we can only use two percent of the cpu's power per frame as production draws to a close level designers and artists often must hand their maps over to the optimization gods and pray to recap a clearly defined studio culture can give level designers a consistent quality and style and style no matter how your team evolves a team of leadership with unique roles can open wide the creative potential of the level designers to keep their ideas flowing the diversification of ld specializations mean this means not every map owner has to be great at everything specialists can strengthen the entire team by being very good at one thing key partnerships are essential to this deeply collaborative process every team has a part to play in making the level design great and the more everyone understands everyone else's part the clearer these interactions become okay we're we're exactly at the halfway point we're at 30 30 minutes 31 minutes um at this moment i have to stop recording the video and and uh start the next section so i'll see you on the other side of that okay we're back for part two processes challenge five is determinism in design this challenge is about metrics when i say metric i am talking about the dozens of design constraints that allow every part of the game to work with every other part game development is often the march of metrics from their inception to the implementation with strict adherence to the inevitability of rigid clear-cut gameplay experiences there's a deterministic quality to it why can't metrics lead to this kind of design determinism first sorry i get my hair in my mouth it happens there it is again what does the life cycle of a metric look like gameplay metrics are born from the work between systems designers and gameplay programmers these are things like how fast can the npc move or how far can they see uh what's their attack range the gameplay metrics mature into character metrics as they're handed over to the animation team character artists and riggers what is the size of an npc uh how wide is their collision how far off the ground is their vision cone next bringing the metrics into adulthood level artists work with the environment artist to give the world a scale that was defined by the character metrics this asks and answers questions like how wide is the doorway how tall is a handrail what's the floor-to-floor height of a building what's the smallest room uh what's the smallest room can be so that x number of npcs can still move around now it's ready for the metrics which are all going up to arrive in the hands of level designers they follow the metrics to the letter and job's done right job's done right i'm not going to do their voices you could wind up with a team of level designers whose job is heavily restricted by the metrics and they wind up just robotically constructing maps that are defined by the metrics alone we don't want that and we don't do that we can't work that way with the type of games we make we make messy games with systems that are crashing into each other and a passionate dedication to immersion in world building there's a balance that has to be reached between what feels best for the player and strict adherence to the metrics that are still needed to make any of our game systems even work we need flexible metrics you can't know how any of the metrics will work in game play until the level designers are using them and testing their limits the metrics for enemy vision range may not work with the sight lines the level designers need the level geometry may require many iterations with gameplay programmers to find the correct metrics for enemy pathfinding your tools could be displaying warnings when the metrics are broken which is good but this can also stifle experimentation designers like to go off the rails when trying a new idea the engine will often put strict limits on how many npcs can be near the player but level and narrative design have the player crashing a party oh no the engine says you can only have five npcs at the party the interaction range on pickups maybe too large for the object density of the maps making it difficult for the player to target one item over another metrics on audio propagation say that sound can't travel more than 20 meters but level and audio designers are placing diegetic music in the maps which now cuts off awkwardly everything the level artists environment artists planned could fly to bits when it crashes into what the lds have done these metrics are starting point but for the collaboration to work both teams need to be flexible the complexity of the objectives may be too great for the space on the ui three words to explain the next goal may work in ninety percent of the cases but sometimes you need more room to be clear be as flexible as possible but some things simply can't change character limits on ui elements have to take into account localization some languages will need twice as much space if level designers experiment forever expecting systems and gameplay programmers to keep up they won't be able to stabilize the game into a shippable state as annoying as many of the warnings a level designer will butt up against r they exist for a reason and build management will pay the price and hours and sanity for every morning a level designer ignores a level designer may want to go wild with massive vfx and sounds that propagate across the entire map but eventually they'll hit a limit on what the engine can do if the level designer requests something as innocent as changing the height of a table this can cause a domino effect with hours of rework for animation in character art these changes must be requested with extreme caution you probably won't get a second chance if you're wrong it's a difficult process but there's some things the team can do to make it go more smoothly environment artists can get a hold of the metrics created by systems design and in animation early and make a modular gray box kit that will match the breakout kit they'll later make for the level artists if the metrics change this kit can easily be tweaked updating the entire game tools can be created to help level designers visualize the more abstract metrics like audio propagation limits you can try to find a balance between the performance cost of some of the rendering engines fancier features with what the level designers are doing if there's a poison gas feature and the lds use it everywhere it could tank your frame rate since you can't see the future give lds a cheap temp version so they can place it in their maps and then you can see how much that temp version costs and now you have a target to aim for oh and though it should go without saying level designers yes please do follow the metrics but when they aren't working say something this may sound all over the place so what point am i trying to make if your level design and level art teams are forced to strictly follow the metrics and have no say in what those metrics are you'll wind up with rote predictable samey maps [Music] metrics must be iterated on like anything else and you need wiggle room if cover has to be exactly one meter high your map will lose believability give them range it's a balanced give and take sometimes a metric must be strictly followed know when to give and know when to hold firm if you set up processes around the idea that metrics will be flexible everything will be easier and ready to change but most of all if the metrics aren't working or too restrictive you have to speak up and when ld speak up please listen challenge six pivots happen things change during development for whatever reason and the reason is user research i'll get to that later and you may need to react quickly the problem is every team works at a different pace some may be finished in our polishing while others are just getting started when your project must pivot how do you minimize uh wasted time and work the level design and level art teams work together until they stabilize their work and hit a milestone at this point after you get user research the level design team gets some feedback digests it pivots and resumes work and the other on the level art team however they didn't get the feedback it was about the level design after all they polish up their work stabilize it and call it finished but inevitably the changes made to the level design are going to need changes to the level art too but it was communicated too late the level art team has polished up their work and if they change things now not only is it going to waste the polish time but polished work is much harder to change the level design changes are now blocked and the game may ship without some teams can't begin their work until another team's work has stabilized and every time the first team pivots this gets pushed farther and farther ahead there's no way for sound in vfx to realistically work on a map well it's still deep in production the layout and contents need to be stable every time the level design pivots this pushes that moment further down the timeline when the ld work is finally stable enough these teams have barely enough time to do the work and almost no time for their own polish and stabilization the structure of some team's work can be radically different making syncing up iteration impossible features are often developed one at a time and can be delivered to level designers very late game systems can be the biggest impacted by pivots from user research which means new and revised features can come long after the level design work has shifted to stabilization it can be a complicated and risky process to integrate them this late throwing the stability of the maps into question and features may not be implemented correctly in the end to be clear many of these features are requested by level designers themselves and they but the problem is they can impact the entire game and not every map may be in a good state to receive them these issues have to be solved on a case-by-case basis and i don't have the solution to everything but there's two guiding principles behind reducing the negative impacts and that's to de-silo and empower the type of siloing i'm talking about is one when is when one team has information but doesn't share it with other teams give the level artists the same feedback the level designers were given let the two teams iterate as one little artists can be empowered to react to the feedback as well rather than being given tasks weeks later after the level design is already well in motion it is not easy it takes tremendous amount of coordination to keep teams this size and sync ask any producer miscommunication wasted work will still happen but iterating together is always worth the effort the level design team can be empowered to populate their maps with sound and vfx early through a library of placeholder assets this can help level designers get their ideas validated by presenting them more more concretely if needed more temp assets can be created as the ld team reacts to feedback but of course remember a placeholder must be unshippable or it may ship when the time comes for vfx and sound designers to work directly in the maps the maps are already populated by temp assets which dramatically decreases the workloads on all of these teams and gives them a much clearer idea of the scope and variety of work ahead kickoff meetings at the start of a futures development keep everyone in the loop tech lds and stakeholders on the ld team can be involved early de-siloing this for this information they can start working with the map owners to prepare for each features arrival the level design team is then empowered to have their maps ready as soon as the feature arrives ideally of course you can't anticipate everything and many features come very late you do what you can on top of all of that a powerful tool to help the whole team my slice foils de-silo and empower our strike teams the critical disciplines are all present in my example i use a recent onboarding strike team you have to keep it light and agile for it to work if the entire studio is on the strike team it's not a strike team the team members have autonomy leads and directors are in the loop but team members have the authority to make decisions if every plan must be validated then you need 18 leaders to validate everything and it cripples the team's authority uh agility sorry and it must involve a producer empowerment of autonomy means increased accountability and everything you decide will impact other teams there's no single solution it's just communication no teams partnerships are identical the challenges to each will be unique you need custom solutions if you de-silo your work you can empower other teams to be ready to receive it as soon as it's available to get their work in a much better state to begin your pass on it and of course trust in the power of strike teams but remember it has to be lean and autonomous challenge seven markup and sunk cost features that require markup that is to be implemented by hand in every instance are very common and it's common across the industry it's something many development teams will face arcane works very hard to require as little markup as possible but it's extremely complicated to achieve this is because in some cases markup is design work like if you're defining a zone a certain enemy cannot exit so they won't chase the player across the entire map but in others it is simply telling the game what exists so it can interpret the data in a certain way like marking spaces as uh under a table so that enemies can shout she's hiding under the table what is markup or lack of markup do to level design in the standard one sorry i'm getting tired in 201 we had a blink power a short range teleport it required basically no markup we occasionally needed to tweak the collisions a bit by hand to get it to work perfectly on a critical path but for the most part it just worked by blink requiring no markup that's time the level designers can design death of the outsider had the displaceability which allowed our protagonist to place potential versions of herself for presenting alternate realities and then switch between them at will yeah it was complicated because an npc was involved we had to put nav mesh everywhere the player could potentially go this represented a small markup cost but it wasn't design work it was just dropping data in where it fit what's something that requires a massive amount of markup but straddles the line between interpreting existing data and design work enemies don't take cover during combat unless there's data telling them where they can go and what they can do when they're there you could build tools to have the ld hand place each cover point which gives them absolute control over every combat encounter but this is incredibly time consuming and it makes it very difficult to change the maps layout later [Music] systemic cover that operates either by ai rules or much or procedurally generated cover points has pros and cons the cons are a much higher reliance on strict metrics but if we consider this adherence to extra metrics as markup time it's still way less than if it all had to be done by hand that's the pro iteration is feasible again here's a perfect example a personal example in dishonored 2 i was level designer of the clockwork mansion which was this house that uh could transform in all sorts of ways if the player pulled levers this is how i spent my time working on it okay not really but i imagine this is how it might look from the outside this what i'm about to show you is closer to reality the clockwork mechanism excuse me the clockwork mechanisms and basic level design took maybe half of the development time of about three years total [Music] jindash the antagonist had voiceovers which reacted to your behavior setting up the unique logic for each line was time consuming the nav mesh deformations for all the transforming rooms was a monumental task even though the static nav mesh was procedurally generated making the audio propagation work at all took maybe as much time as setting up the transformations themselves it was 100 markup and the tools i needed to use were tricky to handle could it have gone differently sure but at that moment this is what i had to work with so i worked with it i'd like to clarify what i mean by sunk cost usually it means time or money that you can't get back if the team must pivot all the time spent doing the markup may be lost and it can be really lost because it's possible that nothing at all was gained by having done it it's not like cutting features or throwing away design ideas those can always come back later and you could have learned something but the time you lost on the markup is gone [Music] but in more human terms markup can go on for months and months it's thankless tds work if you combine it with long hours this can have a serious consequence for the mental health and morale of your team i have experienced this myself at different times in my life and it can take years to recover from if the tools a level designer is using to do the markup are particularly unergonomic it can cause repetitive strain injury like tendonitis i have tendinitis in my left hand as a result of work similar to what i've described it is kind of a pain this is really important to consider because rsi can have long-term effects and honestly so can wearing down the mental health and morale of your team for this i have two solutions and a process for choosing which one to use the first is to prioritize tool ergonomy by prioritize i mean they are designed that way from the conception of the future not at some time later the second is procedural features that is features that require little to no markup at all here's the process first you have to start with the question does the future potentially require ld work we're going to go with yes because that's the whole point of this topic is it conceptually possible to design the feature so no markup is needed let's play it safe right now and say no it's not this feature requires a designed implementation is there an existing and ergonomic tool to implement the future best case scenario yes you already have one use the tool improve it if you can but what if you don't you should develop the tool in parallel with the feature with input from the level designers but what if markup isn't needed there's two things you have to determine with the time it takes to code and maintain the future be less than the level design cost i feel that in most cases it is here's the tricky one do you have someone on your team who is capable of writing this code this can be a much more complicated task than creating a tool to do the markup you could pour weeks into the work and still fail it's a risk if not we're back at the tools question but if yes create a system to procedurally implement the feature across the game and you're done sort of you must still keep the lds in the loop from early on because now there's going to be metrics involved no matter the outcome the goal is to get the feature into the game as designed or give the level designers a way to do it you must do it in a way that allows for iteration and flexibility you need to avoid wearing down your level design team on repetitive tasks that take months to complete and blocks their ability to pivot and you need ergonomic tools that protect the mental health and the tendons of your level designers last challenge number eight and my clock has stopped so i have no idea how much time i have left oh well let's keep going game development is too much information the solution is documentation too bad the solution is also a challenge we can do better than that but some details first a lot of information flows through level design from the start you have the game's concept coming from creative direction narrative or both the lead level designer and campaign designer will serve up a list of uh the draft of the campaign overview which they've worked out with level art leadership and the narrative team tech lds working with tools developers will be drafting up tools and process specs the systems designers will begin to produce vast quantities of writing about features details and metrics the level and environment art teams will also begin to deliver metrics about the sizes and shapes of everything they're going to be using from there level designers must document their brainstorms and prototypes and then as the lds build their maps they need to begin to write down all the map details and then they should also write up the question vignette details but there's more different stages of a project's life require different types of documentation during the brainstorming phase docs can and should be loose and messy it's the designer's notebook it's also an archive nothing here should be thrown away browsing through your old ideas is a great way to get new ones or find one to resurrect documentation for production is a massive library of detailed specs for every topic in the game i've heard it described as the single source of truth it has to be rigorously guarded against redundancy and obsolete information cut content must be clearly labeled or deleted out of date specs must be updated or deleted documentation suitable for shipping is entirely different beast and often not written by the designers but requires their help to do so it must be lean and holistic with information pertinent to qa to producers to external organizations so in addition to all of the work discussed in the last seven sections level designers also need to take time out of their week on a regular basis to create and maintain their docs so what do we do we're going to treat documentation as game development writing and maintaining docs is a task like any other task in the team's to-do list i am talking about the entire development team not just the level designers level designers need up-to-date documentation on metrics getting enough out of date information is a hazard working on docs should not be treated as an additional thing to maybe get to if you have time it's a task with a workflow and reviews like everything else if you include update the docs as a footnote in a bigger task it can be tempting to mark the task as complete trusting yourself that you'll get to the docs later on the other hand if you complete the work but haven't yet updated the docs the task can still be marked in progress and this can give leads and producers an inaccurate view of the team's progress towards the milestone if documentation is its own task we ease these problems time is allotted to it people can see more clearly what is and isn't done a team member who is blocked until your work in the game's data is finished will be happy to see that that task is now complete and your lead producer in qa will be happy to see when the docs task is complete a few miscellaneous points are rather my personal advice as someone who spends a lot of time of her day writing docs you can tell i'm very passionate about this topic doc should always know who the owner of what is no doc can cover everything people with questions need to know who to ask even if someone is maintaining the docs you need tasks to review them and point out errors i do a lot of doc work for the campaign i have a pretty clear overview but it's extremely helpful helpful to have level designers come in look at the details and know where i've made a mistake [Music] yes redundancy is a surefire way to get misleading docs you need things like data boxes that exist on multiple pages and only need to be updated once old stuff is often good to keep but it should be moved to an archive it is a pain when you search and half the results are from cut features know when to collate certain types of information into a single doc it's mind-numbing to have to search through 50 features pages to find the metrics you need and the most critical point do you want good bugs bugs that are accurate relevant and thorough testing teams cannot do this without good documentation okay we're almost done flexible metrics can save you from determinism and game design freeing outcomes through an attentive give and take between teams key partnerships must be empowering one another to pivot rather than siloing work and racing to the finish line tools creation can focus on procedural solutions when possible and develop ergonomic tools alongside the feature when not documentation should be treated as a game development task not something that's tacked on to existing tasks it needs as much production guidance as anything that will go into the game itself so that's the plan nine solutions to eight challenges as i think you've noticed it's a lot of hard work and involves nearly everyone on the team and i couldn't even get into topics like recruitment the impact of marketing user research all the ways that we depend on producers and build management and i barely scratch the service concerning qa but i think this is a good start and can lead to discussions your team can have on how your entire studio can work together to create great level design but gdc talks are a studio and friendship's wide effort so thank all of you for taking the time to help me with this especially while we're trying to get death loop out the door i couldn't have done it without you it's over thank you [Music] you