Designing 'Six Ages', a Storytelling Strategy Game

[Music] so my name is david and i made a indie game with a very talented team and i'd like to talk about that probably some of you haven't played it yet so um the easiest way to describe six ages is it's a king of dragon pass like in the true style of games referencing other games but this is a genre of probably only two games so i'd like to talk about what makes six ages unique and by extension what made king of dragon pass unique so some people say that the game is sort of a cross between choose your own adventure and civilization and really neither of those is true on the other hand i don't think touch arcade is completely wrong it does combine narrative and a turn-based economic game but it's trying to do things that

neither genre alone can do the shortest description of the games i've seen is it's a strategy book this is what one of the players said so the game is in fact a combination of strategy and book you're guiding a small clan of the writer people rather than any one individual or even a group of heroes your clan recently migrated because during the gods war your city got buried under a glacier the core loop of the game is to take actions which are sort of the turn-based strategy side um for example a venture there like expanding your fields as these actions take time to pro to to do uh events will occur and you respond to the events so the strategy game is sort of a skeleton on which we can add stories so that's sort of what it is and i'm

going to talk about how it works unlike civilization it's all about cows we actually do track a few more economic variables but uh really it's all about relationships your clans relationships with the neighboring clans is as important as any economic resources you can go to war but unlike civilization you definitely cannot conquer all your neighbors what you do instead is influence your relations with them for example by sending an emissary with a dialogue there and equally importantly with story decisions the game tracks a lot of different relationships between clans one of the key things i think is that favors are one of the currencies of the game so as a book uh there are over 400 interactive scenes there's three of them shown there

scenes usually have four or five responses it's not choose your own adventure though in the interactive fiction lingo it's probably a floating module there's no single line that goes through the entire story but these floating modules do connect largely through the political and economic model and we'll also try to have as much intention to these as we can for example we might choose a particular clan to intensify the stakes of a situation these scenes can also end up with multi-year lines for example you may have a chance to see those rootless riders in the center again depending on how you treat them and we try to draw the entire story to a dramatic and emotionally satisfying

conclusion so the game has about as many words as the first four harry potter novels but not all in a row it's got a lot of repeatability and there's a lot of variation we set the story in the world of glorantha this is a extremely detailed and realistic world and it helps us have an immersive background you can sort of think of it almost like a character in the story like hogwarts is in the harry potter stories glantha is a a complete world with a mythology and cultures it's sort of as if our background was already play tested and the late greg stafford began writing about glorantha over 50 years ago i'm picturing there the 800 page guide to glorantha which is over 6.3 kilograms of information or 14 pounds if you're american but we actually sort of play in the spaces between that

extrapolating from kind of just a few paragraphs in that the setting gives us the um the main conflict in the game which is that the main god of your neighbors killed your main god and that's what set in place your migration it also gives us a pantheon of deities which provide magic this impacts the uh economic model including combat but we try to make magic more than just sort of something mechanical it's not always going to be well you get plus two in combat we try to make it feel like magic and that means it needs to have much fuzzier feel one of the best examples of that is that we have the ability to let the player enter the world of mythology to cross over to the other side and play through a myth they reenact the stories of the god and they get magic and the outcome of this is usually quite fuzzy you're basically making your own

story and if you've actually played king of dragon pass this these these are sort of the myths that you know from the king of dragon pass days well in the future the goal of all of this is to make an immersive experience sort of the the summary of this is i want you to be playing the story not the game and just a quick reminder that just because it's largely text based does not mean it's not immersive our approach to immersion is that we try to make the player know what someone in the world knows and and not have gamey things intrude on that uh one of the things we have to do is introduce you to what you would know if you're in the world and we start off with a clan questionnaire which helps introduce the history of your clan it's sort of it's not a show don't tell but it's a do don't show so we can avoid an expository lump and

this also makes your clan's backstory the players backstory it's sort of like character creation in a lot of games and it does not only initialize the sort of story situations you're putting your clan into but some of the economic model as well the problem is that the players are probably not members of the writer culture and we don't want to penalize them because they don't know as much as someone in the world would know so the solution is to have advisors these are leaders that will basically have three roles in the game they'll provide information they also act as your playing pieces and they have their own personalities and become sympathetic characters which helps enhance immersion because the advisors are so key the player gets to choose them from a pool of leaders so advisors will give in-world information about like the state of the economy and

the relationships they'll recommend actions and they'll also remind you to do things that you would do if you're running a clan as the game progresses your advisors will get older and eventually die of old age they have skills which are used to overcome obstacles and also relate to the kind of advice you give for example here av2 is heroic and magic so she would probably be someone to listen to if she's talking about magic but if she's talking about a diplomatic situation where she has only a fair skill you might want to listen to one of your other advisors advisors all have personalities we and their advice as you can see can sometimes reflect that personality we do try to prioritize skill-based advice on the assumption that that's

what the player needs in order to succeed so to give advisors personalities uh there's 33 personality traits in the game they are randomly generated every advisor has only two personality traits which is kind of uh pretty limiting for humans but we want to make sure that over the course of the game they can be noticed by a player some of the traits are contradictory so we make sure that when we randomly generate them up character gets only something that makes sense you can't be both optimistic and pessimistic at least not within the scope of this simulation and well a character's religion is not actually a personality trait that also fits into the valid combinations if you worship the war god you're not going to have a personality trait of peaceful

so why do we have personality in the game well as i mentioned before it drives advice but more importantly it'll drive action your advisors are people with their own personalities and their own agendas and they'll go off and do things on their own without you asking them or even wanting them to so we try to be fair to the player and only have something like this happen after you have a chance to witness their personality and if it's visible at the bottom you can see a little bit of script that a character has to be on the clan council for at least two years before this scene will show up where the goat fan starts inviting people in and you the player have to deal with the situation personality can also be a positive thing it can be a a modifier in many situations or you wouldn't send a merciful character or peaceful character off to intimidate someone else you might send your harsh

character so advisors being people sometimes have conflicts with each other the situation here one of your advisors has told the buzzard people they should try harder to fit in and another one is reminding him that the correct term is not buzzard people but adoptees i think in the game some of the best writing is an advice and that's probably because it allows us to have a viewpoint instead of just presenting kind of facts like we have 12 cows less so let's look behind the scene at how strategy and story work together we have our own domain specific language i called it the opal scripting language after an old code name for king of dragon pass it's a domain specific language and it was kind of designed around something that would be easy for our writer to use

at the top there is a conditional response this condition depends on several functions we generate a list of clans and then we filter it to see if we have any clans that are not feuding because this response wouldn't make any sense if we don't have any anyone we could trade it to as the script executes the first thing that happens is a choose clan function and that just puts up user interface where the player gets to choose a clan once they do that we start doing things like looking at the relationship with that clan that the player picked and calculating a modifier there's also modifiers based on your own allocation of clan magic and the just as a shorthand where it says dot ritual magic that's our clan dot ritual magic this makes it a little easier for us to read and whoa that's too fast um

once they have chosen and we calculate the bonuses there's the test which i was talking about earlier this is going to take whoever on your clan council has the best bargaining skill and challenge them to make a bargaining role basically and this has two possible outcomes they could win or lose and we also set a degree of success it's by convention the variable q which lets us have like really terrible failures or marginal successes and i think this one here is a marginal success they get a handsome price but it's the smallest possible handsome price so the scenes are our major a building block of procedural content there's also some other types of scripts for example news can also be interactive but it's not illustrated and it doesn't have advice

scenes in general can have conditions on them this one occurs only if a variable is set to true a lot of the time a scene is a consequence of something else oops i skipped ahead this says um things can happen when conditions occur this one's checking if a variable has changed a value and can do appropriate things it seems there are often consequences here we're triggering another scene three seasons after right now and this is also one of the ways we get um those longer story arcs will trigger scenes that'll happen in a couple of years it also lets us have bonuses that will expire after a certain time limit one difference between six ages and a lot of games is it will only show relevant items and that's because you're someone in the world and you don't really know what might have been if only you had more clan magic you could have drawn on it but

it's not something you would even consider there's no might what might have been the scenes are managed in a scene queue which is a priority queue the granularity is a turn because we're a turn-based strategy game and this is basically how we can add time to the narrative and handle those long story lines uh if there's because something may change once you've you've triggered something in the future the character that relates to you may die so you will have the ability to pull scenes out of the queue generally we look at what's in the queue and run it but if there's nothing right now in the queue when it's time to run a scene we will pick a scene basically at random these unscripted scenes are how the emergent narrative

is comes into play most scenes in the game are unscripted but um which doesn't mean they're they're truly random there's always conditions on them we don't pick a scene that's already scheduled to run in the future because that would start making unnecessary repeats on the other hand we do have the capability of repeating scenes not every scene makes sense to repeat and even if they do we try to keep a scene from happening i think it's nine years nine game years between a repeat of a scene otherwise it would start feeling well repetitive scenes can have conditions because they don't make sense either maybe in a certain time of the year it doesn't make sense to have something that's not related to crops unless it is crop season doesn't make sense to have someone offer to sell you something if you couldn't afford to buy it and some things just don't make sense until you've played for a while

so one of the goals of the game is to have a very high replay ability and procedural content is a great answer for that you really don't know what to expect between games because scenes repeat only rarely you don't get the same set every game the text itself is also procedural and we have what i would categorize into two forms of placeholders the alternatives are just a way to keep the text different each time for example here you could be getting a gift of cedar wood or yellow dye and contextual placeholders are filled in with the value of a variable in this case other clan is whoever it is that came to you offering this gift the value of it is g and the tilde there just means this is something that's worth spelling out as the word like two two

and either cow or cows to make it grammatically correct depending on the number that's being offered it's not just these contextual placeholders pretty much everything in the game is contextual if a clan is coming to give you gifts it matters an awful lot whether that's a clan that you might have an alliance with or if it's clan that's feuding with you so we really are leveraging humans ability to extract patterns from things that may not officially be related and find the relationship between them which is apophenia so i may have skimmed by the fact that there's a lot of tags in the game these are a way that we can help categorize the scenes and that helps integrate the story and the economic model i told our author to make up tags as

needed and we ended up with 135 different tags these are all attached to scripts the ad sign just makes it stand out and there's sort of a way to solve the problem what scenes should be affected by a bonus so if you have done something where the spirits are going to give you good luck with animals we need to know what scenes relate in some way to animals so we tag them with the animals tag at the same time a tag may not apply to an entire scene it may apply only to a specific response you may have the choice to either rely on the gods or rely on spirits and if you chose the spirit branch then we'll add the spirits tag and that makes sure that the appropriate bonuses will apply and some tags ended up just being informational just to look at a scene and say oh this the scene relates to act

three we have an act structure loosely but it's not enforced by anything there's no code that checks that particular one so one of the challenges in making an economic model and a story that relate to each other is if there's a consequence did you notice it and we want to make sure that on the flip side if the economy has a problem is this something that we can have matter in the story explicitly and the solution that we ended up with is something i'm calling concerns their purpose is to make problems in the economy very visible to the player and give us something that's an interesting story decision when you have to deal with that they can actually either be something that's bad which is a stress or it can

be something that's good which we call a bounty and they're always something that should be meaningful it's not just going to be a little plus one that actually does make a big difference if they keep playing the game for 30 game years a plus one will accrue and be very significant but that kind of thing is very hard for humans to perceive so they're going to be big enough an example here is when you have the chance to have your crafters make story tent fabric that will end up distracting them from their other chores and you'll get a craft stress uh concern this means you'll have a penalty on your craft productivity for in this case five seasons these concerns are cumulative you can actually cancel this out if you got a bounty in in some way the concern idea actually originally came from the

the fact we wanted to show the economy visibly to the player and we ended up leveraging that into a larger thing of concerns and they still do get shown concerns are all shown in the dashboard next to the menu buttons and they're also something that your advisors will typically comment on concerns are usually kind of an ongoing thing like that five season uh craft stress but it turns out we have sort of shorter things that we leverage the concern system for like a warning if you're running short of warriors that's not something that's uh an ongoing problem it's just right now we're short of warriors that's a warning and we'll also show that in the dashboard and of course advisors will remind you that you're kind of getting low unlike a lot of games you don't actually have to worry about telling your people what to do we assume that an agrarian clan knows how to feed itself

at the same time there still can be problems that arise with that but you don't need to worry about counting the beans because it's actually really hard again it's something that's hard to wrap your brain around at least for us modern people because food comes in mostly in harvest time and if even if you knew the exact amount of food that you had you'd still have to sort of do mental calculations and compare it against the number of people and but wait this is the this is the winter time and we don't we're going to be not getting much right so it turns out having less information that's sort of less transparent is way more useful and we show the number of seasons of food you have again if this is a problematic we're going to let you know but most of the time you have a few seasons it's it's far more useful than knowing that you have like a thousand bushels of food or something every game has to test a few things that are different in a game like this is that we have a new form of logic error which is problems in the story

apophenia is a two-edged sword the players will expect patterns to be true when they may not be in a way and the fact that everything is context-sensitive makes that more of an issue we may have a scene that runs just fine unless you happen to have an ally and you attack them and if the game doesn't notice that you attacked an ally the player is suddenly no longer immersed they think oh this is just code this isn't a real story anymore so our testing has to focus on on these story problems and as well the fact that everything all the context of a scene could be very different every time it's run and one way i try to make that easier for us to test is to do a lot of automated testing i can run through every response of every scene and make sure that there's no gross errors like crashing out or um

ending up with a zero as a value when it shouldn't be and there's a bunch of sort of stress tests on that well if you have no goods at all does the scene make sense exhaustive qa just means we go through every response of every scene and every branch and possibly with different variables set and make sure that it makes sense as a story another thing qa did was to make aware like what's not in the game this is all testing the sort of the trees and we'd still need to know if the forest makes sense some of the challenges of my approach to showing you only in-world information is that it may not be as transparent as a lot of games difficulty is something that we try to mention with advice players of king of drawing past often complained that they didn't know why something failed so we've added the typically an advisor will explain either how you could have done better next time or an explanation for why something didn't work

we also try to show as many consequences as possible in the dashboard or here in the the magic screen lists all the weird little bonuses you might have accrued and you can also lose immersion by stopping playing and the game doesn't have arbitrary save points that you have you know recognizable places that you were it saves automatically so you're picking up and where the hell was i advisors again will try to remind you of that and there's also a saga which is automatic note taking which you can refer to to see what you had been doing and finally the goal of the game has been kind of a problem unlike king of dragon pass where the title is the goal the goal is kind of behind a plot twist in this game um it would kind of be like telling you what to start hey little farmer you're going to blow up the death star and we can't really do that but

so i'm going to wrap up and six ages is a game that really tries to balance both story and strategy as equal partners we try to make a game where you're playing the story and not the game one of the key ways that we get there is by having advisors who are empathetic characters the overall story is multi-generational our procedural narrative exploits apophenia to get even more meaning out of our assemblage of scenes we try to do everything possible in our scripting language and we have the the concept of concerns which are sort of a quantized unit that helps integrate the story and the strategy game so what questions do you have i think we have a few minutes left about two or three please get up to the mics so people can hear you on the audio

sure hi um thanks for sharing all this so you put a lot of effort into the replayability and the procedural narrative and the logic behind what could repeat and not repeat um but the game is kind of a long game it sounds like you're literally playing through years and years how much was the return on investment for that do people really enjoy replaying it um and and kind of was it worth the effort or yeah just so our the k of dragon pass answer to that because that came out a lot longer is that people have it's insanely replayable and people have been playing it the same person has been playing it for years in some cases um a lot of it was more just that it you so that you could lose the game and try again and see a somewhat different story so i think it's been worth it just just for that it's it's not a single story that you go through people probably don't need to play it as many times as i did with king of dragon pass but

i think it was worth it well if i can follow up the um oh is it lost if somebody lost the game and they tried to recreate the same path you know except for the very end is there enough random event in there that that it would be different anyway i think every game every game is going to be different this the overall goal of the game ends up being the same every time although you can get there even that's got some variation so winning is not always identical either thank you hi thanks for your talk um regarding qa it i imagine must have been a tremendous task is that something that the uh the engine or the scripting language that you guys developed helped you do or was it just you know um manpower that got through it was pretty much other we did have the the unit test that automated a lot of that the scripting language the script compiler can find a small category of

errors itself and i would add more to that um one of the it wasn't quite the language itself but if i a lot of errors ended up being something that would happen multiple places so i would we'd find an error and then i would sweep for similar scripts and correct it before qa found it okay thanks you mentioned that despite the randomization there's still a clear story arc to each playthrough um do you have time to elaborate at all on that on how you accomplished that i don't think there is well there's a there's no single story arc that you'd ever get there are some sort of tentpole scenes that we think you need to get to get the full experience but those are never going to be exactly the same um the context is almost always going to be different for those and you can certainly go through even some of the they're not tentpole scenes but there's some significant

smaller storylines that may never occur in any one play so it's mostly that um you know as in king of dragon pass you're trying to end up being king of dragon pass and your story is always going to be always going to be at the end becoming king of dragon pass but the story along the way is very different and these things would random see even the same scenes will come up in different orders and that the context makes it very very different thank you okay i guess we went over a little bit thanks for coming [Music]