Bringing a Star Wars Cantina to Life in 'Jedi: Survivor'

[Music] uh thank you all for coming here I know that there's a lot of other fantastic talks that are going on during this time slot but the Cantina was something that was very near and dear to our developers and um all the departments on our team and we wanted to bring some of the insights and some of the learnings from that Journey uh to you all and hopefully you can bring some of your insights as to how you execute some of your bus narrative Hub spaces back to us couple of reminders before we dig in uh if you could please silence your noise makers so your phones and air horns and Kazo just hold that off until I guess our QA session that's going to be extended across the hall and there's also going to be a survey that is sent out after this via email all right so before we dig in to today's talk uh we just wanted to give you a brief introduction to to both of

us so you understand uh our background our experience and the perspectives and different angles that we hit this pretty complex problem at so um I'm cheyen Pani Moren I'm a senior game writer on the Jedi team I've been in the game industry for about nine years I started off in the mobile and Indie space before I moved into triaa where I focused mostly on Dynamic and state-based and nonlinear narrative structure and storytelling that primed me pretty well to move on to the Jedi fall in order team where I implemented our narrative Hub spaces which were much smaller in scale mostly uh mantis Landing Hub pads and all of that um Pipeline and uh Tech and approaches fed into what eventually was going to become our next evolution of our narrative Hub spaces which is the Cantina and I'll hand it off to co now awesome thank you Cheyenne hi I'm Cole sweny I'm a senior narrative designer at bit reactor right now but I worked with Cheyenne on Survivor uh I did um a bit

of sliding around between narrative design and writing but a lot of it was very Hub focused and making sure that the um design and text side of things was very hand in glove with the writing end of things um and in this talk I'm going to be focusing more on the uh design uh implementation pipeline side of uh of the Cantina thank you all right so we are going to be covering uh how we approached and tackled and executed all this Con content for our state-based narrative Hub space uh Pyon Saloon which is our version of The Cantina uh brief road map before we dig in we're going to do an overview as to how we initially approached this larger fantasy of how do we deliver on a Star Wars Cantina uh Cole is then going to take us through how we mapped out and created some event targets to give us some guard rails and some uh some goals for our writing room and for the rest of the team to Rally around around we're going

to take a detour into the writing room process and talk about the types of topics and um the ways in which we tried to get some specificity uh as well as match the broader tone of the game uh we're going to revisit our pipeline for a bit and talk about how we modified sort of inflight based on how development was unfolding then we're going to touch on Dei uh inclusive storytelling was always a fundamental goal and uh tenant that we Ed to realize a really robust and wild and engaging cast for The Cantina uh but there were a lot of hurdles that came along with that ambitious goal so we wanted to take a brief detour and mention and call out some of those so that if you want to infuse your own spaces with inclusive storytelling uh you'll have a little bit of a map to understand some of the pitfalls and some of the challenges and then Cole is going to take us away with a final wrapup where we talk talk about what we

actually delivered and some of the ways in which maybe would have done it differently and some of the targets that we maybe overwrote for underwrote for and how we might do it differently the second time all right so for those of you who are unfamiliar with the Jedi franchise we wanted to show a little bit of a primer in terms of the main gameplay experience and what the player really experiences most of the time so in the Jedi franchise it's a a third person action adventure game it's heavily inspired by Souls born combat as well as Metroid Vania inspired traversal and exploration so the players on the sticks quite a bit they have a lot of agency to explore these worlds and the narrative team had the challenge of also trying to balance out the darker tones of this specific chapter in Galactic history it's it is one of the darkest periods in the Star Wars Universe you play as Cal cestus one of the last surviving Jedi

and the Empire is quickly closing in and he's trying to find home so Star Wars itself as an IP has a light-heartedness that we also wanted to capture but didn't fight against that broader sort of story goal of really bringing the player to these darker spaces my eyes all right I think we got a good idea of uh what you can expect in the game so that brings us to the larger canteen of fantasy uh when you think about Star Wars as a just a broader Universe what are some of the things that come up in your mind okay any well for us the cantino was was a big one um The Cantina that's featured in episode 4 uh has just left such a sort of ort of cultural imprint

in our imagination and when we were thinking about how to evolve our narrative spaces we felt that delivering on this fantasy would be a huge step in the right direction uh this led us to breaking down that fantasy into uh two different buckets and that was our narrative pillars and our design pillars so starting with our narrative pillars we pulled some of the exact language that we used in our original Cantina design dock that data back to April of 2020 uh we wanted to share this language because this is how we rallied a lot of resources and a lot of developers to become excited and willing to put in their energy love resources into making this goal of reality and delivering it to our players in to the best of our ability so number one we wanted this space to house evolving Tales we wanted there to be meaningful narrative progression in IND individual character

arcs and we also wanted to reward the player for repeated engagement number two we wanted a diverse cast uh the original Cantina was filled with all sorts of aliens and humanoids and Rogues and bounty hunters and everyone under the galactic sun and we felt that if we used inclusive storytelling and Tred to align this with some of our Dei goals that we would realize that um to its fullest and finally because the events of the main campaign bring the player to some pretty dark places there a lot of twists and turns and tragic losses and uh we even have a more tragic five act structure wanted to make sure that this was a narrative space that would be a welcome respit for for Cal during his dark Journey as a kind of light inth dark that brings us to our design pillars uh so just to mention uh the

narrative did not come necessarily before the design pillars they evolved very much in tandem hand in hand but this is what it led to number one we wanted to reflect the world state so as the events unfolded in the main campaign we wanted to make sure that this space was also evolving in step with it so every time the player comes back we wanted to have updated fresh content that felt like it was evolving along with Cal's Journey number two we wanted the is to be the primary Hub space for various side content systems to be housed so the main planet that you're traversing in Jedi Survivor is Kobo it was one of our first semiopen world planets it is a sprawl of hidden and side and optional content and so we wanted to feed the player back into one space so they didn't feel like they're getting Whiplash just trying to check in on systems on different planets number three we wanted this space to grow over time and we wanted the player to be actively driving that

and number four we wanted to design the space so that the player was being routed through it as often as possible there was a lot of resources that went into all of these Hub States and so we wanted to make sure that the player was visiting it and actually getting exposed to some of that content so that's very ambitious high goals and all well and good but what did we actually ship with so Pyon Saloon which is our Cantina uh did indeed become the primary narrative hub for Jedi Survivor it housed 18 distinct NPCs with their own conversations and story lines and this is outside of our core mantis cast They possess over 5,000 lines of fully recorded uh vo dialogue in addition we also have 17 main um very distinct World States and countless permutations layered on top of that to account for optional player engagement this also became the unified space for

multiple narrative inside gameplay systems to intersect and that included our Bounty Hunter system fishing gardening Hollow beding rumors and to some degree even our legendary uh Beast system so to demonstrate the evolution of this narrative Hub space through the course of the game we have these two gifts one on the left is a very early visit into the Cantina it's kind of a dilapidated dive at this point uh there's not too much narrative content that is available because the player hasn't tapped into any of these NPC relationships they haven't invested too much into either of the side context uh side content systems either so you don't really see the aquarium in a really good State and it's kind of more sparse in the types of interactions that are offered to the player on the right this is our Cantina in a midgame full swing state so a lot of these NPC

relationships have been tapped you know a little bit about their background you may have invested in a lot of these side content systems an entire second floor of the Cantina has opened you've recruited some optional NPCs out in the world and so we have the music system online and the aquarium is looking like a real aquarium now so now that we've talked about how we initially approach this and kind of broadly conceptualized the Cantina I'm going to hand it off to Cole to talk about how we mapped out and tried to plan this massive sprl of content thank you Cheyenne um so yeah I'm going to talk a little bit about how we set targets for all of the dialogue that we' need and a lot of this came down to questions like how much writing do we need what sort of writing do we need uh how do we apportion that content and how do we present it to the player and then how do we prioritize this massive bucket of content that we have uh we had a lot of different kinds of

dialogue events from walking up to characters and initiating straightforward conversations to um EES dropping in hearing uh ambient dialogues playing out in the background as you walk past and we wanted all of these to reflect the progression of the player the progression of the main story and also to reward um repeated engagement with all of the side content here so uh just to uh sort of conceptualize for a moment what this problem looks like when we're looking at it uh this sort of represents the rise and fall Arc of the Cantina which is mapped onto the five act structure that Cheyenne alluded to for the main story of the game so you have this sort of initial rise in the first three acts culminating in sort of a false Victory and climax in the middle and then you have these darker States at the end of The Cantina and what this meant for us was that in the early States the Cantina was filling up there was this sort of ground swell and the feeling of community as the player had more opportunity to rope in other NPCs to the space uh there was this sort of Marty gr moment where we think we've

dingdong the witch just dead the whole thing and then at the end uh it falls off the Jack boot of the Empire descends and the Cantina starts to thin out and we're left in this sort of somber reflective uh I guess have the good times have died type of space um so when we're thinking about how is the player going to experience this content uh one thing we could count on is that right at the beginning of the game when they were first introduced to The Cantina we knew they had to visit the Cantina twice and this was our opportunity to jam as much onboarding as possible into those visits so we were uh sort of surfacing side content of the player showing that narrative would be uh would be evolving alongside the main story and we got to intro some of our key Side characters um but from here there's there's quite a lot of variance in the narrative experience so maybe our like ideal player uh might have uh sort of sampled the Cantina a few times throughout the game they would have gotten some interesting pulse checks and they could sort of chart this uh rise and fall of the Cantina over the course of the game

they might see it at one of the heightened States and then they might kind of see it during one of the emptying out States at the end and it would sort of resonate with what's happening in the main plot that was the goal but just as valid and and certainly just as possible the player could visit as soon as they had to and then as soon as they were free never see it again and if they did maybe they'd come back after they'd finished the main story uh which is a which is certainly a big departure in terms of the experience of that player um and again just as validly we could have uh this piece of work right here who insists on visiting the Cantina constantly throughout the game uh and we want to make sure as well that this person isn't uh turned off and hearing a lot of repeated lines and whatnot but it's rewarded for sort of uh taking this this path throughout so there again there's quite a lot of Divergence in the player experiences that we're designing for here uh if you'll permit me to uh take a a slight philosophical detour on this uh a lot of the problems that come here uh are are sort of uh the ways that

we interpret the way that we should be writing these stories in very interactive spaces so sort of the uh kind of father of what we understand as narrative Theory and narratology was the old Huer called Aristotle who essentially said uh anything that doesn't uh anything that you could take out of a story uh and the story is still there and still works doesn't belong in the story because a story is a unity and a story should be one thing and anything that it's not shouldn't be in the story that really doesn't hold in the interactive space because we don't really have the luxury of saying the player has to experience these things there are a lot of bits of narrative content that can be taken or left at will so we have sort of this filler content or uh optional content that we need to conceptualize and that means kind of changing the way that we're thinking about writing this content uh well Along Comes Roland bar and says hang on I'm not so sure about this uh actually there are kind of a lot of ways a lot of ways to interpret a story sure we're stacking all the words in order

but there are a lot of entrances and exits to the way that everyone encounters the story so really there's not one way that everyone is experiencing a story but there are a lot of ways that the story is permuting throughout um and this sort of uh this explodes into a bunch of possibilities with more modern narratologists who are focusing on digital media like uh Janet Murray and Espen arth who are saying actually we can have structures like the romatic structure this root system of narrative content where you can loop back on yourself and go off in a and you know go off in a direction hit a dead end come back um and again forgive me for jumping into all of this but this is all sort of to say that we're in a sort of sloshy space when we have a very linear story happening on the main path cow's sort of main Arc and then we have a very nonlinear set of uh stories that are sort of supposed to be a reflection of that so in a way we have to sort of uh balance between those two and try to provision for both ends of the spectrum here so to jump back towards the uh

concrete end of things uh let's take an example uh these are Dana and grock two of the characters in The Cantina who you can meet early on there are a couple of Prospectors and treasure Hunters who've come to Kobo uh the planet on which the Cantina is based in search of treasure and these are truly a couple of overripe old kudin uh they're very unpleasant to be around uh and the player can first encounter them planning a job in dredger Gorge uh near the middle of the game you can encounter them again that job has gone horribly arai uh grock is supremely peeved because Dana has nearly dropped a steel beam on his head but Dana is insisting no no that was just a fluke we can do this uh and you're allowed to interact with them at any of these moments and then at the end uh unfortunately again as as the Empire is on the rise grock ultimately perishes and pays the ultimate price for both of them uh when one of the jobs go arai and an imperial Patrol takes him out and

Dana is left in this sort of reflective space where she's stewing a bit in it and it's not really clear whether she's learned the lesson or not um so again with our little lines we can see it would be really nice if the player experienced this in sequence and they would get a sense of the entire story but it's entirely possible that they will skip the beginning jump in at the middle hit the end and of course possible as well that they could just hear the end of the conversation here which uh the end of this Arc would mean they never meet Gro they don't really understand what's going on and and Dana is sad for no reason um so some of the ways that we try to provision for for these alternate paths would be we wrote uh different on-ramp interactions depending on when in the story you first met Dana and grock so if you met them at the beginning you get sort of the standard introduction to them as a Duo and then you can follow their Arc throughout but if you meet them for the first time in the middle we write a different dialogue where they get introduced which onboards the players to all the facts that have happened in their Arc beforehand and then we can

proceed with what's going to happen there we can also leverage elements outside of just dialogue in the narrative space uh kind of morbidly you can actually find grock's Body out in the world uh after after this job has gone wrong you can use Cal's psychometry to sort of get a vision of Dana mourning his death and that adds depth to the story in a very nonlinear way you don't you don't have to have experienced any of that in sequence so these are some of the ways that we were starting to conceptualize telling sort of linear plot arcs through uh through through a lot of the nonlinear systems and again all of this is kind of part and parcel in what in what it means to tell uh interactive stories so um yeah these were some of the questions we had so back to our standard question how much writing do we need per state per character per dialogue type we found that a lot of these could be sort of divided along two polls on the one end we had heavily state-based dialogue this dialogue was authored it was sort of

deterministic it was going to be the same every time and it focused on Cal's story so uh in in the later parts of the game Cal experiences a great personal tragedy when you go back to The Cantina you want to see that reflected in the interactions Cal has in the space you don't want him to go and crack a joke and leave turgle rolling on the floor that would be a great a great discontinuity in the narrative so we're trying to avoid those kinds of moments on the other end of things though we have more engagement based dialogue and this dial is Evergreen we're not throwing it away once it's no longer relevant um and it it's more focused on the NPC stories getting to learn more about Moran The Smuggler or getting to listen in on someone else's dialogue and learn a bit more about their life um and these sort of played to some of the different fantasies that Cheyenne took us through very early on so on the one end of things we have our light in the darkk story we get to uh Cal we get to see Cal's story reflected through this space on the other end of things though

we have that fantasy of just being a fly on the wall in the mle Cantina and seeing a bunch of wacky aliens doing their thing and getting to uh walk up to them and learn more about them as far as how we prioritized this content when presenting it to the player uh the the ultimate approach we used was we prioritized State specific content so if you've just gone out uh done uh finished a level out in the world and ignited a giant space laser that's now shooting across the sky it feels nice to go back to The Cantina to walk up to someone and they ask you what the heck is that big laser that's now shooting across the sky um so we wanted that level of reflectivity but to reward repeated engagement the Evergreen content was lower priority but if you came back and you contined to listen you would get through this content as well and this proved to be uh sort of the main Paradigm that we used with a few exceptions to sort of stratify the different types of content that we were uh presenting to the player Cheyenne's going to talk a bit more

about how we broke down the game writing needs into these states and how we approached the writing itself but broadly we broke the game down into five macro states which roughly mapped onto the five acts of the main story and then we broke all of those down into substates and these were individual player visits where NPCs were standing in deterministic locations um or or some of them were standing in random locations so uh this is the sort of again the main Paradigm we used when thinking about this content so going back to the uh gifts we presented earlier the one on the left is macro State one the Cantina is delapidated and ramshackle and it doesn't look doesn't look like much uh and then in macro State 3 The Cantina in full swing we we spawn a lot more random NPCs we have a lot more ambient dialogue that we've written to keep it firing and this all is in furtherance of selling that main uh Arc of the Cantina across the game when we were looking at individual substate some of the questions we were answering uh were where is it most

interesting for characters to stand in a given substate if they'd just been in a cut scene where they were interacting with each other it'd be great to position them next to each other in the Cantina so that if the player is interested they can follow up on that relationship so for every substate in the game we placed characters uh very specifically on the floor of the Cantina and this gave us very accurate dialogue targets because we knew if we're going to have these people standing next to each other we we' might as well have them inter intera and we might as well write some ambient dialogue for them to share uh one of the other main elements and probably the final thing I'm going to hit on the dialogue end of things would be how we conceptualized uh these character arcs it was uh ultimately pretty tricky to string a lot of these character arcs in a very linear sequence when the player can be uh encountering the Cantina in a very nonlinear way so one of the ways that we simplified the work on our end was we reduced these into what we called inflection buckets um so probably best illustrated by

example so Moran again is this very frosty Smuggler who you meet early on um and when you first meet him he's very critical he doesn't really want anything to do with you um but as you chisel through that lovable piece of granite you eventually get him to warm up to you and this sort of unlikely friendship forms so instead of having a completely linear set of conversations that take Moran from beginning to end there we had one bucket of conversations where Moran doesn't like you we had a second bucket where Moran does like you and they were connected by an inflection event one sort of high value conversation where there's a very perceptible turn and the State of Affairs between these two characters change so we were able to uh write a lot of different content without tracking a whole lot of the narrative overhead that would come with very specifically trying to plot these out in a nonlinear way so for instance we could also then include State specific dialog in these two buckets so Moran can now

not like you and there could be a giant laser in the sky and we can write a conversation for that that doesn't have to adhere to a strict narrative sequence we used this inflection our planning for a lot of the side characters in The Cantina um for instance uh mosy uh was the owner of the uh legendary beasts system so instead of rewarding repeated engagement with mosy we were rewarding uh in increased engagement with the system the more legendary beasts you hunted the further you got in mosey's inflection Arc and similarly with cage our hardboiled Bounty Hunter the more bounties you hunted the more you learned about cage as you went back and interacted with her so I'm going to hand the mic back to Cheyenne now proverbially uh and then we're going to look at more about the writing process specifically and how we approached all these buckets of content which we've now sort of loosely outlined and how we started to fill them in thank you Cole so a brief overview of what our writing

room looks like on the Jedi team uh for Survivor we had anywhere between three to six writers who were expected to write in any given character voice for the Cantina for any given World State this evolved into more of a TV Writing Room style of review and uh table reads and this gave us the following routine first we would brainstorm in the room and we would break the larger macro State and then move into more specific more granular topics with uh character breaking so how are those characters actually responding to these world events we would then assign out some conversations to each of the writers take a few days to draft them and then we throw them back at the writing room for a table read uh we were all expected to try on every voice even the ones that we were not so comfortable with had to experiment with a lot of accents to really make make it believable and Punchy and then we would get some either

live edits or we would get some feedback from our writers sometimes we would change the topic entirely and we would keep keep going with this process uh rinse and repeat until the conversation was to a ship quality all right we're going to walk you through uh a mini exercise in terms of how we populated these buckets and the types of conversation topics that usually fell into these so the first step was situating the room with macro state brainstorming so we're looking at the broader world events like a laser going off in the sky or maybe you just took down the leader of the bedum Raiders uh this also roughly mapped on to our five acts so if Cal is feeling hopeful that's going to be kind of the tonal backdrop um or if Cal just experienced tragedy then that's also going to dictate some of these conversation topics uh this is more of our Evergreen topic uh bucket of topics because they are substate agnostic they can play out a range of different

substates so some of the usual topics that we would draw on uh would be individual NPC backstories so this would be along the lines of uh when did greez and monk link up what's the story behind them naming it Pyon Saloon uh when did the town matriarch Doma get into uh Kobo and any backstories that she might have of events that have occurred over the last 50 years since she's been on planet we also fit in some relevant Slice of Life topics so this is like a day in the life a day in the ordinary life of one of our wacky characters like what was turgle up to today uh how has his priorities changed based on some world events that was a nice kind of filler flavor kind of topic and then we also used this bucket for calling out environmental or feature improvements so there's a huge difference between macro State 1 and two the second floor of the Cantina has opened up and for some of the resident patrons who've been here for a while that's a huge seismic change

and so they're going to call that out for the player it's also used to point the player up to these newly accessible features so once we had that broader tone we would add some reactivity and specificity by breaking down the substates and doing less of this Evergreen content and more of this state-based content this led to the following topics things like NPC reactions to critical path updates if you just took down Ravis a character like turgle who has been IM mesed and entangled with that faction for a while is going to have a very personal reaction to that update this also doubled as a way to remind the player of their current objectives and some newly opened content whenever a new Cantina NPC would move into the space or a NPC relationship hit in inflection we also use this bucket to call that out and prioritize those changes in

relationships uh cage is a really good example of someone that we use as sort of a spokesperson for calling out new NPCs in the space as a bounty hunter who has her eye on the door constantly she's checking out whoever's moving into the space and determining whether or not they are someone she can manipulate if they're a threat if they could be a potential ally what have you and finally we also used this bucket to uh update the player on any side systems that have evolved and to remind them of any rumors that will direct them back out into the world a good example of this is scua Steve will have a high priority conversation if you just collected enough fish and you've updated the aquarium so this brings us to one of my favorite parts of the talk uh we wanted to demo what our table read process is like and how some of these characters evolved from an earlier State and an earlier understanding that the room had of these characters and how further in the game they have a little more depth

and nuance and maybe a little more personal stake in the events of the game so we're going to do a live table read I will be playing Moran and Cole will be playing turgle very apologetically hey Moran oh please no you wouldn't turn down the opportunity of a lifetime would you stop it because I have a business proposition that will make your eye stems spin let me make this eminently clear I have in my time struck deals with every flavor of unsavory Scamp you can imagine uhuh I have funded Ventures so risky they would make a Hut skip breakfast but I would sooner feed my credits to a monkey lizard than go into business with the likes of you with a small onetime investment no so that's one of our early uh almost like characters of each of these characters very quintessential turgle quintessential Moran this could probably play during a range of early substates and uh macros States and now we're going to move on to

a conversation that takes place more in the middle of the game um and Cole will be reprising turgle and I will be playing cal cal is it true is ravas really gone yeah you relieved you think so wouldn't you but I'm actually sad it almost feels like I killed him he did try to kill you oh yeah a few times but none of this would have happened if I wasn't trying to make a quick credit I'm a murderer hey monk what's it take for an amp pod with blood on his hands to get some service around here so um these two conversations again they demo a a simpler character version of turgle and then later in the game once you've had enough time and you've been exposed to the Dynamics of the Cantina and how some of these NPCs relate to the broader Gamal uh we turn turgle who was initially just an agent of chaos up to irrelevant Antics into someone who has a stake into what Cal's doing and how he's affecting the world

so I'm going to wrap up this segment with a little bit of writing room etiquette that we landed on first of all the writing room requires a lot of mutual trust and willingness uh we do these table reads and we have a lot of fun and Selenas with it but we also need to accept that your fellow writers uh start off with an initial intent and if it's not Ling for you you got to build off of it and you need to use the classic improv rule of yes I accept the premise and let's build on it this kind of leads into the second rule which is you can't be precious with initial drafts because we had a really quick turnaround and we would often throw our first drafts and some really wild swings at the room and sometimes it was a really good initial idea but it wasn't quite Landing in the way that we wanted the player to experience this character so it's more of like a yes butt so accept the initial premise but be willing to Pivot and shift directions based on writing room Fe Fe back so that we deliver the best

representation of this character and lastly got to have fun and be silly and another reason why we did those table reads we did those weekly sometimes twice a week and that was maybe the highlights of that project at least in my experience I'm going to hand it back to Cole to talk about how our pipeline kind of changed along the way awesome yeah so I'm going to take us a bit more into some of the spefic specifics of the ways that our narrative decisions were informed by our pipeline uh and I think that ultimately this was not sort of a one directional process where the narrative was conceived written and implemented and then it was done it was uh being implemented as we were writing it so we got a lot of feedback over the course of the writing process that allowed us to sort of steer the ship very slowly and lay track in front of the train and we could make some corrections as we were going um and so ultimately the pipeline played a lot into the decisions we were making as a narrative team so at a very high level

uh the four main tools that went into the writing process were Meo the uh virtual whiteboard tool that you saw earlier writ or duet our screenwriting software we used a a dialogue Matrix in Google Sheets and then we had our content management system a tool called Heckle which certainly deserves it its own talk but uh for right now essentially takes all of the all of the dialogue lines hooks them up to wise events we tag them up and it sends it into the engine so as far as how the process as we moved through this went um at first we would brainstorm these seating charts in Meo and we would send that information into the dialogue Matrix The Matrix would use that information for a given state to calculate okay we have these character adjacencies uh you know turgle standing next to Moran in this state so we need to write an ambient between them uh we wanted to have one conversation for every character for every substate okay then we need 18 different conversations and it would generate a blank template which we would send back back into writer duet and this ensured that we weren't constantly cross-checking with

the spreadsheet but we knew we were going to hit our Baseline targets if we just filled in all the blank events in the Writ or duet script uh on the Writ or duet end of things we would take uh these conversation ideas that we've been brainstorming in Miro and start to drill down into the actual language do voice Explorations and then ultimately arrive at the dialogue events that we would use in the game and we would send them into the database where again we would tag them up and they'd be ready to go uh some of the information though uh would flow back the other way so for instance every time we made a push into the database and got all the content tagged up it would update the dialogue Matrix which had a big um metrics readout showing what we'd written versus what we said we were going to write so there would be Big Red Blocks all over the place if we hadn't written something that we had established and set out to do in our dialogue targets another reason this was valuable is because ultimately these dialogue targets weren't meant to be high water marks what we want the product to be they were just meant to be baselines and then

ultimately the goal was to empower the writers to if they had a great idea for a certain conversation in a certain substate uh the system would be capable of taking those facts and putting them into the game um but since the writing process was a little more diffuse like that it was very useful to have a metrics tool where we could see oh we're kind of overriding a lot in this in this in this substate or shoot this character doesn't have uh any conversation dialogue uh we we we need to fix that something might broken we overlooked it um so that was just a very useful tool as we were steering the ship and then finally uh this this allowed us to change our planning on the Miro end of things and kind of that completes the cycle some of the like decisions that uh arose from this process would be we realized that one of our characters had no ambient dialogue in the entire game and that made us realize oh shoot two locked our soothsayer isn't actually standing next to anyone she isn't standing close enough to anyone over the course of the game so we were able to move move some characters over to her in given substates so that they could have

fun little interactions and we' get to Showcase some more of the character um so that's just an example of how the the pipeline here really was a critical feedback tool and of course it must be said there is no feedback quite like jumping into the engine and playing it on the sticks and seeing how it feels and that was also a great tool that we used to shape the writing process I'm going to hand it back to Cheyenne now and she's going to take us through some of our inclusive storytelling approaches in the Cantina all right so inclusive storytelling in Dei was always a fundamental premise and goal that we had for the Cantina we felt that that was going to deliver the most robust Cantina experience and we wanted to make sure that that was supported not only by our character Concepts and our voice Explorations but also very intentional uh player interaction models so you'll see a lineup of many of our uh

underrepresented identities who display a range of different voices um visual representations and each of them also have pretty distinct and specific player interaction models uh just to caveat this is talk isn't going to cover why Dei is so important or how to do it perfectly but it was a big problem that we approached and there's a lot of production hurdles and and challenges in the pipeline in terms of scoping and getting the appropriate resources to authentically represent these identities and so we wanted to flag some of those learnings for you so that you can Infuse your own spaces with some inclusive storytelling and hopefully bring those learnings back to us in a future GDC talk all right we're going to First Step through the standard character creation pipeline for those of you who've been engaged in it for a while this is probably pretty familiar uh First Art and writing would collaborate and pitch a concept lots of back and forth really

just trying to explore voices and how that would best be represented visually after we landed on a concept that we were really satisfied with the writing room would then hone in on the voice and create some casting sides these are some quintessential character scenes where we get to see who they really are and how they might perform once we get the auditions back we'd have key stakeholders weigh in and choose their favorite performers uh this usually involved more of directors and leads and people who are going to be very Hands-On with that character and then we'd hand it off for the rest of the in-game translation pipeline now to contrast this the Dei casting pipeline had a few more steps that we had to build into the process the first one being we did a lot more preliminary Dei research so this involves uh Consulting internal ergs uh employee resource groups this involved looking at internal Dei primers it also

involved polling our internal Dev team to see if any individuals were willing to share any experience that was relevant or wanted to weigh in on this kind of representation this is also the step where if we felt that we were really treading in unknown Waters we could bring in a cultural consultant then we'd go into the art and writing uh Concepts phase there's a few more rounds of back and forth and we may have the key stakeholders start to weigh in even at this early stage then we would produce a enhanced casting side so the writing room wanted to be a lot more clear on what this character's backstory was about what their voice was like and give at least two scenarios or two casting sides to make sure that we were getting a couple of different performances from each of the potential candidates who might eventually play this role this led into uh potentially additional cast rounds I think for most of our uh Dei characters we had at least

two rounds of casting and that's more of an industrywide problem there's a pretty limited pool of underrepresented talent and so make sure that you may have to build in some extra time to make sure that you're getting uh casting a broad net and getting as many performers as possible then we would do an enhanced stakeholder weighin so anybody who was called out during that preliminary research round uh any internal devs who wanted to share their experience and weighin gets to be a part of this step and we have a slightly more diverse Council who is evaluating our potential performers then we would do in-game translation as usual and we built in one extra step where we would do additional review where we'd set up internal slack channels and get some more team feedback all right so that is how we landed on the concept and the initial Explorations for these characters but we also wanted to support their representation with very intentional uh

design interactions so we are going to walk through a couple of different characters who represent a fully inclusive story uh Winnie and zigg on one side and Ash who is more of a light inclusive story and what I mean by these terms is we had a fantastic uh race and sexuality expert come in and talk to respond internally and brought us these terms to describe how much of that character's identity is playing like a primary role and really being exposed to the player and how much that's actually driving their story so we'll start with Winnie and Zig Winnie and Z represent um a lovely adorable queer romance and the way that we approach this is we were faced with the problem of wanting to expose the player to their content but not making the player feel like they were playing Matchmaker and that they were controlling this romance and how it unfolded so we employed a state based inflection Arc so what that means is the

player has the opportunity to recruit Winnie early earlier on in the game in an earlier macro State and then she'll come back to The Cantina and their romance will start to unfold but if the player doesn't recruit Winnie early at a certain State Winnie will find her way into the Cantina and their romance is still going to play out so for players who are really invested in that kind of storyline they get exposure to some earlier content but regardless these characters are still going to have this romance play out on the other hand Ash we felt as a room was a character uh even though they were non-binary the main feature we wanted to focus on is their profession in their craft they are the uh Galactic producer music person with their little dj Droid and that's what we wanted to focus on and so we employed a system based Arc so if the player isn't interacting and not investing in this additional side content they might not get a lot of their

story all right so we're just going to recap on some of the takeways for building in inclusive storytelling into your pipeline you got to plan a lot of time um there's a lot of pre-work that goes into addressing potential biases and there are going to be some hiccups along the way that you may not have expected like having a limited casting pool and requesting extra rounds of of casting and getting casting that broader net try to build in as many sources of feedback as possible so ergs focus groups in-house devs and build those in early and often throughout the character creation Pipeline and then finally explore a range of light to fully inclusive character storylines if a character is only obsessed with this one aspect of their identity even though it's quite critical it could come off as sort of one not and uh just try to experiment with a whole range of characters who have different conflicts on their mind and different involvements and want to

show different facets ofs themselves to the player and now I'm gonna hand it back to Cole to wrap us up with some key Cantina takeaways awesome thank you Cheyenne U so yeah I'm just going to speak a little bit to some of the main takeaways especially on the design end of things from how this sort of the way that we approached and apportioned all of this content uh how it all turned out from from our perspective um so ultimately uh we we we had talked about this these sort of two polls between the engagement based dialogue and the state-based dialogue and at the end of the day we ended up moving more towards the state-based end of things so we filled in our macro States and substates with dialogue that reflected what Cal was doing throughout the story and that mapped on to Cal's tonal Journey uh and then later in the process we went in and filled in a lot of the a lot of the additional characters and work to reward repeated engagement um so uh the majority of our content was State specific and because it's state specific if you'll recall that that poor red line player

who ignores the Cantina for the entire game that means we have hundreds of dialogue events that are very missable including some critical ones for for Side characters um and Telemetry indicated that uh 70% of our players continued completing side content after the story and the majority of those for multiple hours so we had a lot of players who did leave the ca for the end and that meant that a lot of the writing that we put into the state-based end of things uh what wasn't as visible for them um so ultimately the the takeaway here is not just to write less generally uh the takeaway would just be to keep keep your eye on the ball as far as how the player is going to experience the content and try to through through play testing and as much data as you can get figure out where players are spending the most time and then um put put a proportional amount of effort into building those areas out another uh another aspect of the game was we wrote we wrote a ton of characters 18 was a lot of characters to be just completely optional but to have their own stories

in the Cantina um ultimately uh this this uh landed well with a lot of people and we've definitely we we've had instances of pretty much every one of those 18 players has someone who's uh online saying that's their that's their favorite character who's that character's number one fan so that reflected the success of our um um diversity uh goals and a lot of the inclusive storytelling approaches that we took because that meant that a lot of people saw themselves reflect in the content and people were able to find things to engage with across the spectrum of characters um that said we ultimately kind of had to sacrifice some depth for breadth at this point because with 18 characters you definitely have to cover a lot of ground there and if we had been a little bit sharper with our unique selling points for each of those characters there were opportunities to maybe combine characters or cut characters that ultimately weren't adding something new to the roster and then spend that extra scope to deepen the characters who are already there there um so ultimately I think while a lot of these characters landed really

well um the the goal in the future would be to continue refining The Balancing Act here and I'll hand it back to Cheyenne uh yeah so we just wanted to wrap up with some credits uh the Cantina was one of the features that we started digging into as a development team and collaborating from the earliest stages of of the production on Jedi Survivor and so we wanted to do some special call outs for all of the folks and departments who poured their heart and soul into making this an interdisciplinary reality so big thanks to all of the writing room narrative and design teams were in step with us all along the way concept art drove so many of these like wacky initial ideas that the writing room just went wild with um animation brought it all to life environment art audio music tech design code everyone uiux lighting QA and of course our partners at lucasfilm uh really helped us evaluate our characters and how we were adequately representing

some of these um Dei goals thank you it it looks like we've got about nine minutes for for questions if anyone has it hi that was a fantastic talk thank you and thank you for such an incredible game uh I'm very curious since your work involved so many different teams like you know Mission design with what side content you introduce narrative uh the you know ambient Tech design do you have any advice for collaborating with basically every single team on the game yeah so the question is uh given that this is such an interdisciplinary effort do we have any advice for um cross Department

collaboration you want to start sure absolutely um and I think one of the things that uh Cheyenne mentioned was that uh so many of the great ideas for for this like idiosyncratic Rogues gallery of characters came from different you know places on the team and there were there were you know several you know pieces of concept art that everyone on the team just latched onto originally it's like okay we have to we have to write that and and and figure out who that is um I I think uh probably the the planning and dialogue targets and of the spectrum was where we tried to build in a lot of those um supports for other departments so trying to interface with design to figure out you know what what are rumors exactly going to look like this character needs to have four rumors to correspond with four pieces of side content they need to send you to so we need to ensure that we've got room for all of that in in their dialogue targets and and that sort of thing um so this is probably a circular answer to the question but but just like the communication uh and kind of trying to frontload as much of that

as possible so that when it comes time to turn the crank on the content Mill you know you're going in the right direction and you're not leaving a bunch of uh a bunch of you know grass unwatered and just to add to that uh even though we dove into how the writing room would kind of pitch some of these Concepts and start with some voice expirations a lot of these character Concepts originated with concept or someone else on the team who just had a wild pitch for a character and so it just took a lot of willingness um and a lot of conversation and communication between the writing team and whoever is pitching that concept to really realize that uh I remember that our first batch of characters came from concept like this one really wacky amazing concept piece that included most of our alien characters actually and then we went and kind of looked at what some of our design goals were and narrative goals were and then tried to match up some of these personalities and hone in on how they're going to be translated and experienced by the player awesome thank

you time for one more yeah totally um I guess this is a two-parter sorry uh do you have a sense of like prioritized conversations like could multiple conversations um match a current substate and and if so was there like a priority you'd put them in and then um my it's maybe related is do would one conversation be governed by one like set of state or did it Branch inside the conversation that makes sense yeah so so the the question was um what did our concept of dialogue prioritization look like and then within a conversation how did the branching work if that reflects it okay awesome uh on the on the um prioritization end of things um every conversation the game had had a a weight on it and and it was always pick the highest weighted conversation and that was proportional to the number of facts that were attached to it so if something was super super super specific only going to be valid for one state we we'd better make sure that's top of the

stally yeah we did the same thing on firewatch as oh okay awesome perfect uh and there we did have some some light branching inside of uh given dialogues so when you walked up to a character it would pick for you the most uh the highest priority um dialogue uh but uh once it once you were inside the dialogue there were some little choices that you could make to sort of just flavor your experience and then there were a few cases where we would set a conditional tag on one of those and then you'd get a conversation that called back to a choice you made so you compliment someone on something here and then later they they say thank you actually that that meant a lot to me uh those were all pretty uh um surface layer and mostly just wanted to like get the get the player doing more than just like setting their controller down on the coffee table when a dialogue starts to play like they can you know uh engage with it um but we tried to keep the branching pretty pretty light there to reflect just the linearity of Cal's story cool thanks thank you awesome hi time for another one or yeah okay great um yeah thanks this awesome talk I really appreciate it I

was wondering uh your sort of like writer Tex stack was uh really interesting uh how much of that was like emergent or how much of that was sort of predetermined or like you know how do you get writers on board for like what is ultimately a pretty technical process uh yeah I would just love to hear a little bit more about that I didn't quite that if you so the I'm sorry could you repeat the question oh yeah for sure like your your flowchart regarding uh writing yeah just how did that sort of process um emerge or or was it like established prior just was just an interesting yeah concept love to hear more about it yeah for for sure so so I think the the question had to do with the the the flowchart of the different tools and how how information flowed through and then kind of how we onboarded the writers on on onto that process um I think uh it was definitely um designed with and this is kind of a thing people say a lot but like designed with our specific writing team in mind so like we we have our writers um a lot of them have narrative back narrative design

backgrounds and whatnot so so so it was definitely they they were ready to engage with some of the nonlinearity of the system um but there were definitely like um steps built in to try and really streamline things for the writer that they could really just focus on on on homing in on the characters um so for instance like the autogeneration of the of the scripts so that they're not having to sort of tab over and say okay I got to you know go through a spreadsheet and count the number of things they have to write it's just you know throw them in with some blank lines you know z z blank line Cal blank line Z blank line cool that's the scene and then they go in and add all of the all of the detail to it but they're not having to think as much about some of the sort of the broad planning um at the same time I don't want to make it seem like they were siloed to that because a lot of uh I think they especially as the process went on um did a really great job of uh of looking for opportunities to enhance interactivity and um and to to take advantage of some of those some of those systems there um hopefully that answered your question no absolutely

thank you hello hello I'm curious uh what your favorite conversation or interaction is that required hypers specific criteria that most players might have missed what are some Easter eggy type of conversations that you you want to highlight there there was one where question oh sorry repeat the question oh oh yes sorry the question was did we have a favorite uh dialogue event that involved like super hypers specific tagging that that was sort of rare for a given instance uh I do remember one where there's a state in which turgle is infusing about this seashell that he's found somewhere uh and the order in which uh you you encounter uh dialogues in the hub about that determine the the the things that people say so for instance if you go to monk first and monk says listen this guy turgle is is just railing on about this this shell if you go to turgle turle will start to

tell you about the shell and Cal will say okay I already heard about this please shut up and and so there are little kind of like funny gags built in uh I think that was that was a highlight for me if you have one uh yeah I think some of Moran's later conversations once you've hit their inflection event shows a very different side of them uh so it requires a lot of player engagement and it's a very subtle uh inclusive storyline if you if you will and if you're not keed into to that storyline and you haven't been investing visits to him frequently you you might miss that and so when you actually get to see the other side of Moran it's a lot more soft and vulnerable and a really good contrast to that Frosty Smuggler we see at the beginning all right I think we're right about at time thank you guys so much um we're going to continue some Q&A across the hall [Applause] [Music]