Literally the best GDC I have ever seen about non-linear narrative design. Only took me like 3 watches to understand what he was talking about but now I'm gonna scrap the current quest design in my prototype.
This is excellent stuff! Systematizing the flow of causality in a simple and meaningful way that allows for full player freedom and responds to the player... world class.
The magic system in Sorcery isn't just cunning it's utterly incredible. Every spell has a 3 letter word, and some of them have a required item, players can only read the spellbook between game sessions, and can only cast spells they have actually, irl, memorised. There are several spells that are just obviously good, ZAP, NOK (opens locked doors) etc. but many were very contextual and specific, but if used correctly could give you massive rewards/cool new content. I can still remember that small pebbles explode with the spell POP!
It is amazing how a game from a small German developer studio in 2002 does a lot of this stuff he talks about and does open world so well that it is still unrivalled to this day. One of your very first task in Gothic 2 is to enter a city to acquire a certain item. Now, because you look like a beaten bandit the city guard won't let you through because the country suffered from a big wave of escaped convicts, so they won't let everyone in freely. Now there are several ways to enter the city: you can bribe the guards, but this requires a good amount of gold coins that is hard to come by at the beginning. You can buy or work for proper farming clothes, so you can pretend you're a farmer; or you can steal the clothes from the landlords chest (you shouldn't be seen in those clothes, though, because he will attack you for thievery). You can also get a permit from a travelling merchant who gives it out for free but will ask for a favour in return later in the city. If you talked to the right guy you get the information that you could disguise yourself as herbalist merchant but to be convincing you'll need at least ten of the same plant or herbs. Or you could just venture off the safer beginner's path, sneak past hostile NPCs that are way too strong for you, jump down to the coast and swim around the city walls into the harbour and get into the city that way. If you do it that way, the game even acknowledges that you didn't enter by any of those other methods, grants you bonus XP and a NPC comments on your "Didn't see you get in by the city gates. Did you swim all the way?". So what happens is that the game feels much more like a simulation, simulating a living world that reacts to your actions. But you don't need to necessarily trigger a certain quest to open up options. Of course, to act like a herbalist, someone needs to give you that idea. But how you get the money or clothes is completely quest-agnostic and up to you. In fact, you could wander somewhere else entirely, Gothic 2 being an open world game. It's "level design", though, confronts you with enemies you won't be able to tackle on until you're stronger, so the game soft-forces you along a "linear" way. However, there's no level gating. If you put every early skill point into combat, you can get to places with a level as low as let's say 4, while someone on level 10, having spent skill points into thievery and crafting, will have a much harder time.
The first developer I've ever seen that actually understands games. So many games are still locked into mechanics that were created 40 years ago- NPCs that give the same canned chat every time you meet them regardless where you've been or what you've done.
Thank you for this brilliant talk! Now I can't stop thinking about how grinding could appear in a choose your own adventure -book: a varying loop of turning pages and throwing dice, which annoyingly repeats every time you're trying to do something? It would be an infuriatingly ridiculous concept for sure. xD
Interestingly, there is grinding of a sort in Sorcery Part 2 that involves going to a gambling hall and grinding (by rolling dice) at the hall until you have enough money to buy certain expensive items from the market outside the hall.
32:45 - You say "these are not quests" but this sounds like a generalized version of The Elder Scrolls quest system, which is essentially a list of linear / causal state machines, some of which are visible to the player and contain plot information.
Except think of how many Elder Scrolls missions are basic fetch quests like he describes. Fairly different in how a player experiences the game from what he’s discussing.
@@sub-jec-tiv Have you ever spun up a TES game editor and looked at how the quest system functions under the hood? (Or even played a TES game?) I'm guessing no. Because what you're talking about is not the quest **system** but the implementation, so you don't have a clue how the system itself works. And the "basic fetch quests" implementation you're referring to accounts for a fractional percentage of the number of quests in any given TES game. All of the procedural bounty quests in Skyrim operate using the exact same quest system as the entire main quest, the entire civil war quest, all of the crafted side quests, and hundreds of hidden quests that drive thousands of different behavior trees, creature spawns, boss fights, weather systems, map changes, and dialogue options throughout the game. (Same is true for Oblivion, though it didn't have the procgen system yet. I won't comment on Morrowind since I haven't cracked open its construction set, but I would imagine it's very much the same since it was built using the same gamebryo/netimmerse engine.) I don't wanna be rude, but your comment oozes ignorance.
Yeah of course it LOOKS like that system, which I remember using in Morrowind to just punch into the console something like "questname_state_65" to unbreak a quest because it had become impossible to fulfill due to some circumstance unforseen by the devs. That's because, as pointed out in the talk, hitting those states is a player target. This is in contrast to the ink system of encounters which are robust enough to happen no matter what, regardless of whether you have hit any state at all. Hence "Ready Myself". (Even if hitting some encounters totally unprepared results in a game over). I will say there was some subtlety in a few of the Morrowind quests that had optional "grace note" states that players wouldn't necessarily reach, and I think there were also quests that could be started in different ways. I believe all quests are complete at 100, but you could get lower values in any order and start the quest with any value. But at this point I think devs like Bethesda have no excuse not to step it up to this far more robust system. Especially if you're going to have gigabytes of voiced dialogue this system allows your writers to work economically with dialogue, make the stuff that doesn't matter "competent but not interesting," so that the interesting stuff can actually be interesting and not just more drivel. There are so many useful lessons for devs to learn from Inkle in designing stories. I will say, the simplicity of being able to use the console to fix quests in Morrowind was really convenient. I made Gentlemen Jim Stacey disappear by completing the Thieves' Guild quests before starting his Bal Molagmer quests, so I simply summoned him using the console in fun and evocative locations like a meetup with a retiree. People do the equivalent of "console cheat" runs of Inkle games using save editing, but that requires basically an all-day research project so it's usually more convenient to just reach whatever desired state via gameplay. I'm still bitter that the No Beacons All Serpents run in Sorcery 3 requires staring at your map screen for an hour waiting for a hole to open up. I'd almost rather just save edit than do that.
I have plenty of respect for Jon Ingold, but this talk makes me think he should research The Witcher 3 some more. Many of the criticisms he holds are true for most open world games (and some are true of TW3 also), but of all open world games I have played, it also does far more of the things Jon advocates for than it does not - e.g. the analogous wolf scenario. I was personally incredibly impressed by all of the edge cases that led to unique lines of dialogue or outcomes. Watching some of the videos from xLetalis would be a quick way to research just how in-depth some of the branching in TW3 can get. Certainly, the entire game is not this way, but it deserves far more credit than Jon is giving it here. Also, there are plenty of other open world games to use as a bad case for a lot of these criticisms.
I played it for 12-15 hours, and was directly referencing the experience I had playing a particular quest. I should stress I'm definitely not saying its a "badly built" game or any such thing! But the way it was made limited its ability to respond to the micro-decisions that I happened to make as it played out. Also, I used the Witcher as example precisely it doesn't *need* me to give it credit, and I can nitpick at it all I need to without hurting anyone. It's a strong contender for game of the decade, and nothing I say is going to change that!
@@joningold5256 Thank you for the reply, and the talk was really interesting, so was your presentation. I think the OP and I felt the way we did because you are picking on a game that has done rather well.
As someone who played all the Witcher games and the Shamutanti Hills, I'd argue that the comparison in terms of narrative is a tad bit unfair (this is of course my opinion as a consumer, I am not a game-dev), since I find both the games offer me different emotion. I am not playing them for the same reasons. However, playing the devil advocate here, it would've been even more spectacular if Geralt's world responded in a way you described choices and action in your presentation. But was it necessary? Again as a gamer, I didn't think so. Many NPC responses to changes in their world seemed to be quite well done for a game of Witcher's scope. For instance, when you take out a monster plaguing the village, the NPC's walking around does briefly mention it as you pass by. The mission about the Witches in the woods is part of a much bigger plot, and it does make absolute sense that you did not find them earlier and you see them there a bit later in-game. However, your comparison did emphasize your 'key-point' and I guess that was the main goal of the talk.
@@unkindled_dreg Thanks for the response. And indeed, I'm not attempting an analysis of the Witcher; more just using a familiar example to get people thinking about the right kinds of issues. You've got use a real world example, really, so I prefer to "punch up." Are more subtle responses necessary? Who knows?!? For myself, I'm interested in designing experiences that react more fluidly and have fewer things you simply have to accept due to game-logic; but that's a totally subjective measure. Ultimately, if you like a game, you'll forgive it whatever rules it has (and all games will always have some rules.) The one thing I will say is that every time a game does notice a "little thing" the player has done, it's always a great moment, and I think some of the systems we've used can help to generate more great moments.
@@joningold5256 Is there anywhere, a blog article perhaps, where I can see more examples? I'm especially interested in where you start: do you come up with the quest then figure out a state machine, vice versa, or you bounce between the two organically? How do you determine that the peasant needs 2 state machines vs 1? How do you determine that the wolf needs 1 vs 2? PS: I bought Sorcery 4 after watching this and I'm glad I did. Great game!
This whole talk, I found myself thinking a lot about Majora's Mask, and wondering what Jon's comments would be on the narrative structure of that game.
I really like his idea with encounters. It feels like a better idea than Quests in every regard, but it sort of has one problem, which is that it becomes a combinatorial nightmare if you try to tangle too many encounters together. I can sort of see why big games don't try to do this. Since you're writing all the text yourself, the amount of work grows exponentially with each parameter in the web of connections. Now imagine having to voice all the potentially superfluous lines and make the character animations for it. I feel like it might be an unsolvable problem, but I like that Inkle are working on narrative challenges in games, it's interesting and important work. Having played the first 3 sorcery games, I can say they are very good, they make good use of the system and I think it's because they realized that there is a limit to how complex this stuff can get.
Can anyone explain how these linear state-trees/machines are implemented in the inky language he demos please? I see individual variables but no indication of linking
In Ink, you would use 'lists' which can contain one of the options you declare for it. ``` LIST Wolf = HeardOfWolf, SeenWolf, KilledWolf You see a wolf. ~ Wolf = SeenWolf {Wolf < KilledWolf: The wolf isn't dead yet.} ```
I wonder how they dealt with 'reminders'. If the player has started an encounter and run off to do some other encounter, how can a player remember that there is an encounter that is unfinished.
Horizon does it through voice lines. For example, every time you encounter a creature that you haven’t “unlocked”, she says “I wonder if there’s a way to find more data on these creatures” or something. It took me forever to figure out what it meant, though, and then I felt stupid. So I’m not sure if it’s the best solution 😅
Largely, we don't. If you walk away, the plot line finishes itself; or expires; or causes some consequence; whatever makes sense depending on the plot. We treat "wandering off" as an active choice.
@@joningold5256 This is now the 3rd time i’m back to this and your Heavens Vault video. Now i have to go play all your games. Have been secretly making notes for a decade to create my own personal game, and this is really seeming like a process that’d work for my ideas. In fact just “gradually pouring tons of stuff into buckets and creating structure as it emerges” is how i make all my art & music. 😂 i don’t mind it being kind of a pain, as long as the result is magical-ish.
Wow. I'm only half way through and I've already gained so much insight into planning a game like Sorcery. This is exactly what I need--the thinking behind the scenes. The only question I would have is does having an open world and so many adventures cost in terms of the size of the game? Or is it not an issue ina text-based game? Excellent video. Thank you!
Very enjoyable talk, but I'm not sure your solution of basically having the radiant quests from Skyrim but they automatically fail if you return to the characters house too many times is that great.
It's more of, "if you pick up a story beat and ignore it for too long, it is resolved in other ways". Skyrim has no urgency for anything because the story only moves when you do. The base story is the literal end of the world caused by the First of Dragon kind, come to destroy the civilizations of Men and Mer. Completely ignoring that quest chain has absolutely no consequences, and every character outside of that quest chain acts almost exactly the same whether you initiate the next quest step or not. You can literally bugger off to Markarth Solitude, or anywhere else, complete the thieves guild, companions, and college of winter hold threads, and still get treated like an escapee on your first encounter of Riverwood, or an asshole by Irileth when approaching Balgruf for the first time. Dragons don't spawn in the sandbox if you don't fight the first one outside of whiterun. The only ones that do are the story beat Dawnguard bosses for their respective battles. If skyrim had an encounter-state system like the one described in this talk, the world would react to the overarching threat regardless of your progress, with the possibility of an actual fail-state.
@@redfiend1670 This is a very late response (apologies for the ping it will generate for you!). Skyrim's problem is the narrative DOES have a sense of urgency, but the game rules don't require you to act upon the urgency. For example, after you kill your first dragon, everyone says you must go talk to the greybeards right away (there's even side dialogue to the effect of "Don't keep the greybeards waiting"). But there is no consequence or change for going off and doing your own thing. The thing is, in an Elder Scrolls game I prefer the ability to go off and do my own thing at any point. This is supposed to be the main calling card of TES - be anyone, go anywhere, do anything, at any time. For a TES game, the narrative should be structured in a way that makes this natural. Morrowind did this well. The main story was clearly important with a sense of dire weight, but it also didn't push you to act urgently. There are even moments near the start of the narrative where the main questgiver says "Good, you're here. Take some money and go establish a cover identity. Join a guild, do some quests, level up, then come back when you're stronger. Or if you think you're ready, we'll get started." The narrative naturally ENCOURAGES you to do your own thing, but leaves a natural option for moving forward if you wish. And then there are some rare moments in Morrowind's main quest that do take on a level of urgency (when the Tribunal Temple becomes more hostile), and there are real consequences (i.e. if you're a Temple member, you can't do any of its quests until you advance the main quest to the point where the Temple is no longer hostile). I love Skyrim but I often avoid the main quest precisely because it fights against the game's message of "be whoever you want, do whatever, go wherever, at any time". If I do decide to pursue the main quest, I often have to build elaborate roleplaying reasons in my head for why my character is ignoring the urgency of the main quest (an approach that worked once or twice but lost power for me after a while). I don't think the problem with Skyrim is that the quest design is bad and an encounter design would be better. Rather, the problem is the narrative pacing doesn't support the game's own open world/freedom quest design.
@@_purplewinter_ I think they do this step by step: first behind a payment wall on their page, then public for all but still on their page (I guess to increase traffic/awareness) and last here on YT. Sometimes it takes years until it ends up here.
Reminds me of Deus Ex where they tried to a large extent to make the game reacts to whatever decisions you make. Atleast in the beginning, it does get more linear.
The Witcher example is astounding, because so much of The Witcher 3 *does* let you do quests out of order. I played one of the quests on the main path before I was prompted to go there just because I was wandering around in the area looking for fun quests to do. It blew my mind when Yenn told me "you need to go help this werewolf (different quest) and that should help lead us to Ciri." Because I had already met that werewolf and helped him, and was surprised that they let me break the sequence that way, Geralt even had a bespoke line to tell her that he already did it. The lead quest designer from CDPR gave a GDC talk where he talked about another main-line quest that lets you complete it before triggering the first step, too. Maybe it's because these are examples of side quests that are interwoven into the main quest and not just side quests that are just dotted around the map? Not sure what exactly the difference is, if the main quest won't progress until you complete it, it's not a side quest. But whatever lol, point is, you can do a lot of quests out of order and it's interesting to me that they neglected it in other places.
The "Serious fatal story bug" is a massive reason why this system is not implemented in more games, while it sounds acceptable in a Sorcery game, I'm not sure Assassin's Creed fans would react kindly to that. Multiply the audience by several factors of magnitude and "obscure bugs" are now a "vocal minority".
AAA studios are putting out trash games now, and people still buy them. A massive AAA publisher/studio like Ubisoft could absolutely execute on this, and do it well, but they won't, because their fans will buy anything they throw at them, in the already uninspired state it comes in. Why do more work when they're already eating out of your hand. Beyond any of that, the bugs are fixable. Developers just have to acknowledge that the more complex their systems are, the more time they have to allot for testing before publishing. If developers are not going to do something because their might be bugs, they might as well not do anything at all.
He says the special sauce of these state machines is the implication that later states imply earlier states. Why would “peasant dead” imply “peasant impatient”? It might imply “peasant was impatient” but not just “peasant impatient”
I think "Peasant dead" is just a bit of an oversimplification, what that state actually means is "Peasant killed by wolf" which can only come about because of him getting impatient. Realistically "Peasant dead" (If relevant) would be a separate state that could come about from multiple possible events (The wolf killing the peasant, you killing the peasant, etc)
I disagree on what he thinks an open world is by definition, they don't need all that he said they did to be considered open world. They just need to be open, as in finish the tutorial quest and then pick a direction and go that way until you hit the map border or follow the main questline if you want.
I bought Sorcery! And I felt ripped off because it wasn't clear that it was part 1. I got too the end and felt extremely unfulfilled, and never bought the other ones.
@Min Lungelow it does, sort of! 'Sorcery' is based on 'Fighting Fantasy', of which there is an 'Advanced Fighting Fantasy' ruleset. The tabletop rpg Troika! is based on that.
What if as a punishment for not being “active” on this wolf encounter ... like the peasant; Other NPC’s in this town also started to “disappear”? Maybe even that important vendor who sells you those rare mushrooms to craft that epic potion? “God those mushrooms are a pain in the ass to loot” 😪
I know this is from 2017, but I hope this dude has reigned in his ego by now. He came across in multiple parts of this video as very holier-than-thou and full of himself. Some of the information was great, but the rest came across as "my game is great, if you don't like it you're wrong."
i dont think its fair to compare a text based game to a game like the Witcher. You already point out how much branching you need to do to achieve your intended level of narrative in a text based game. Now imagine how much more you have to code for a massive game like the Witcher. Thats like picking on a Skyscraper for being not as creative build as your lego tower. I still learned a lot from your Video and it gave nice Ideas so dont get me wrong
If a game is about exploration and finding special things, shouldn't the value of any one thing be an inverse with the number of players expected to find it? To specify a bit, let's rate the value of any discoverable thing by 100 points, then we take that base and deduct the percentage of players who experience that one thing. If all players experience a thing, it has a value of zero, for example. Then once all findable things are calculated, you spend your time working on those things that have the highest score.
which ones? What's funny is that some hentai games have some interesting systems that if improved, could be really cool. Artificial Academy 2 for example had some interesting ones tracking npc relationships and emotion states, like characters being angry reacting in a different manner than calm ones, or certain personality traits like being lazier making the character skip classes, have less inteligence staff and that cascaded into other things, all with I imagine was less than 50 lines of code, becuse most of it were just simple lists.
His definition of open world seems like it would only include completely rng worlds, or just really terrible open world games. And it's funny how his alternative to 'open world design' is just open world design
Most of the time I think you're right, but grind can also be thought of as a way to lower the game's difficulty. The problem arises when the natural difficulty is so hard that even the best players are forced to grind, which unfortunately does happen in way too many games.
@@Venomine95 I might agree, but not every game addresses the problem at all, and giving the player the ability to grind is better than letting them get stuck.
I actually like grind as long as its not too long... i enjoy observing my fight improve and getting the reward of overpowering that monster that previously overpowered me. The key to grind is balance, and most games overdo it.
so both your flow chart of the witcher story and your story is missing one really important possibility, the player never steps foot in the swamp or that peasant house, in a real open world that is not just a possibility but should be high likely, when you make a big story with lots of branches that cost money to make, you put time effort and pride in to it, are you going to be tempted to try to push the player in to exploring it? these encounters do you force them on the player? these encounters are random order narrative plot? why do you think your text game is a open world if you can not have a dead space where the player finds nothing and can go off and do what ever they want. I think this is a straw man speech, you setting up the witcher a actual open world game and comparing it to a text game that has limited story nodes and locations. a finite state machine is FINITE limited NOT OPEN this is why it's a straw man. you could make a open world text game sure but It would require dealing with the dead space, you enter a empty house and find nothing, I burn it down, the fire catches to the rest of the village. all those plots we worked so hard on are now gone, what happens in New Vegas stays in New Vegas but your reputation takes a hit. oh no I save scum, I clean the house, you start cleaning, it you find a tin can maybe you can think of a use for it, if you clean this house 4 more times you can think of this house as your base but beware the tax man because you have modified the world. latter a troll demands some object for crossing his bridge it could be any object. the objects you have available are your sword worth $1500, your armor worth $5000, a important plot item value $0, a tin can value unknown, you are able to give the tin can, but it could have been a twig or wolf head or something else like a chicken. dumby the mad king was going give you $1,000,000 for a "priceless" item never trade unique items, luckily their's save scum. games are not books, yes some games are books just not "open world" ones. please don't try to kill open world games just because they are not what you do well.
"the player finds nothing and can go off and do what ever they want" in this sentence you have made yourself believe a variety of things are both interesting and possible without it being true You will find for most people 'nothing' isn't interesting - it would be similar to having an option appear in the game that says: "Nothing here" countless empty cabinets are not fun, nor empty cabins etc etc Now 'do whatever they want' - wow, I don't even understand how you don't see this - this is exactly the important part of the game and the one the talk is focused on: doing stuff You call his speech a straw-man and call his definition of the open world wrong, while you lie to yourself about what the open world games actually are, even though you describe them perfectly in that sentence: mostly empty space, which you have to traverse to find the interesting bits Most of them delete the traversing altogether and just add fast travel, the only reason you are fooled into thinking this is great is the graphics, but in the end you are auto-travelling to the next plot-point - it works, barely, but it ain't strong design
This is basically just what Emily Short calls quality-based narrative. I think most of this has been understood by the interactive fiction community for decades. It's not remotely innovative.
Awesome presentation, but when you presenting something please do not make repeating "mitch, mitch mitch" sound with your mouth. It is highly irritating after hearing it ten times!
And the Game-I-didn't-know-I-need-to-play Award goes to...
Literally the best GDC I have ever seen about non-linear narrative design. Only took me like 3 watches to understand what he was talking about but now I'm gonna scrap the current quest design in my prototype.
Watching this talk again after playing bg3 is great
Throughout the talk I was constantly like, BG3 does that, and that, and that!
This is excellent stuff! Systematizing the flow of causality in a simple and meaningful way that allows for full player freedom and responds to the player... world class.
This is my favorite GDC talk. I keep coming back to it.
This video blew my mind. I never think about this...
It blew my mind that I _had._
The magic system in Sorcery isn't just cunning it's utterly incredible. Every spell has a 3 letter word, and some of them have a required item, players can only read the spellbook between game sessions, and can only cast spells they have actually, irl, memorised. There are several spells that are just obviously good, ZAP, NOK (opens locked doors) etc. but many were very contextual and specific, but if used correctly could give you massive rewards/cool new content.
I can still remember that small pebbles explode with the spell POP!
The real EUREKA moment was figuring out that there were secret spells that you can cast just by experimenting with the letters. Blew my mind. 🤯
It is amazing how a game from a small German developer studio in 2002 does a lot of this stuff he talks about and does open world so well that it is still unrivalled to this day. One of your very first task in Gothic 2 is to enter a city to acquire a certain item. Now, because you look like a beaten bandit the city guard won't let you through because the country suffered from a big wave of escaped convicts, so they won't let everyone in freely. Now there are several ways to enter the city: you can bribe the guards, but this requires a good amount of gold coins that is hard to come by at the beginning. You can buy or work for proper farming clothes, so you can pretend you're a farmer; or you can steal the clothes from the landlords chest (you shouldn't be seen in those clothes, though, because he will attack you for thievery). You can also get a permit from a travelling merchant who gives it out for free but will ask for a favour in return later in the city. If you talked to the right guy you get the information that you could disguise yourself as herbalist merchant but to be convincing you'll need at least ten of the same plant or herbs. Or you could just venture off the safer beginner's path, sneak past hostile NPCs that are way too strong for you, jump down to the coast and swim around the city walls into the harbour and get into the city that way. If you do it that way, the game even acknowledges that you didn't enter by any of those other methods, grants you bonus XP and a NPC comments on your "Didn't see you get in by the city gates. Did you swim all the way?".
So what happens is that the game feels much more like a simulation, simulating a living world that reacts to your actions. But you don't need to necessarily trigger a certain quest to open up options. Of course, to act like a herbalist, someone needs to give you that idea. But how you get the money or clothes is completely quest-agnostic and up to you. In fact, you could wander somewhere else entirely, Gothic 2 being an open world game. It's "level design", though, confronts you with enemies you won't be able to tackle on until you're stronger, so the game soft-forces you along a "linear" way. However, there's no level gating. If you put every early skill point into combat, you can get to places with a level as low as let's say 4, while someone on level 10, having spent skill points into thievery and crafting, will have a much harder time.
I was thinking about this all the way through this video. And it helped me understand why so many games since then don't feel quite right to me
This is actually a masterpiece. Thank you very much for this talk!!
The first developer I've ever seen that actually understands games. So many games are still locked into mechanics that were created 40 years ago- NPCs that give the same canned chat every time you meet them regardless where you've been or what you've done.
Sorcery! Is the best game I have ever played or ever will play. It changed my life and I don't think that experience will ever be topped for me
This incredibly, exactly what I've been looking for. Thank you for breaking it all down like this.
This is an utterly brilliant talk and answers my biggest question about game structure and narrative design. Thank you so much!
*frantically taking notes*
I am so going to make something amazing off of this talk alone, thanks.
This video deserves a second watch. Maybe even a third.
Very good talk... I already have ideas on how to use that in my current game. Much cleaner and simpler than what I did before.
Thank you for this brilliant talk! Now I can't stop thinking about how grinding could appear in a choose your own adventure -book: a varying loop of turning pages and throwing dice, which annoyingly repeats every time you're trying to do something? It would be an infuriatingly ridiculous concept for sure. xD
Interestingly, there is grinding of a sort in Sorcery Part 2 that involves going to a gambling hall and grinding (by rolling dice) at the hall until you have enough money to buy certain expensive items from the market outside the hall.
Amazing talk. Truly incredible and extremely well explained. Thanks!
32:45 - You say "these are not quests" but this sounds like a generalized version of The Elder Scrolls quest system, which is essentially a list of linear / causal state machines, some of which are visible to the player and contain plot information.
Except think of how many Elder Scrolls missions are basic fetch quests like he describes. Fairly different in how a player experiences the game from what he’s discussing.
@@sub-jec-tiv Have you ever spun up a TES game editor and looked at how the quest system functions under the hood? (Or even played a TES game?)
I'm guessing no. Because what you're talking about is not the quest **system** but the implementation, so you don't have a clue how the system itself works. And the "basic fetch quests" implementation you're referring to accounts for a fractional percentage of the number of quests in any given TES game. All of the procedural bounty quests in Skyrim operate using the exact same quest system as the entire main quest, the entire civil war quest, all of the crafted side quests, and hundreds of hidden quests that drive thousands of different behavior trees, creature spawns, boss fights, weather systems, map changes, and dialogue options throughout the game. (Same is true for Oblivion, though it didn't have the procgen system yet. I won't comment on Morrowind since I haven't cracked open its construction set, but I would imagine it's very much the same since it was built using the same gamebryo/netimmerse engine.)
I don't wanna be rude, but your comment oozes ignorance.
Yeah of course it LOOKS like that system, which I remember using in Morrowind to just punch into the console something like "questname_state_65" to unbreak a quest because it had become impossible to fulfill due to some circumstance unforseen by the devs. That's because, as pointed out in the talk, hitting those states is a player target. This is in contrast to the ink system of encounters which are robust enough to happen no matter what, regardless of whether you have hit any state at all. Hence "Ready Myself". (Even if hitting some encounters totally unprepared results in a game over). I will say there was some subtlety in a few of the Morrowind quests that had optional "grace note" states that players wouldn't necessarily reach, and I think there were also quests that could be started in different ways. I believe all quests are complete at 100, but you could get lower values in any order and start the quest with any value.
But at this point I think devs like Bethesda have no excuse not to step it up to this far more robust system. Especially if you're going to have gigabytes of voiced dialogue this system allows your writers to work economically with dialogue, make the stuff that doesn't matter "competent but not interesting," so that the interesting stuff can actually be interesting and not just more drivel. There are so many useful lessons for devs to learn from Inkle in designing stories.
I will say, the simplicity of being able to use the console to fix quests in Morrowind was really convenient. I made Gentlemen Jim Stacey disappear by completing the Thieves' Guild quests before starting his Bal Molagmer quests, so I simply summoned him using the console in fun and evocative locations like a meetup with a retiree.
People do the equivalent of "console cheat" runs of Inkle games using save editing, but that requires basically an all-day research project so it's usually more convenient to just reach whatever desired state via gameplay. I'm still bitter that the No Beacons All Serpents run in Sorcery 3 requires staring at your map screen for an hour waiting for a hole to open up. I'd almost rather just save edit than do that.
Absolutely fantastic talk. My brain is sizzling!!
I have plenty of respect for Jon Ingold, but this talk makes me think he should research The Witcher 3 some more. Many of the criticisms he holds are true for most open world games (and some are true of TW3 also), but of all open world games I have played, it also does far more of the things Jon advocates for than it does not - e.g. the analogous wolf scenario.
I was personally incredibly impressed by all of the edge cases that led to unique lines of dialogue or outcomes. Watching some of the videos from xLetalis would be a quick way to research just how in-depth some of the branching in TW3 can get. Certainly, the entire game is not this way, but it deserves far more credit than Jon is giving it here.
Also, there are plenty of other open world games to use as a bad case for a lot of these criticisms.
Yup, definitely felt the same way.
I played it for 12-15 hours, and was directly referencing the experience I had playing a particular quest. I should stress I'm definitely not saying its a "badly built" game or any such thing! But the way it was made limited its ability to respond to the micro-decisions that I happened to make as it played out.
Also, I used the Witcher as example precisely it doesn't *need* me to give it credit, and I can nitpick at it all I need to without hurting anyone. It's a strong contender for game of the decade, and nothing I say is going to change that!
@@joningold5256 Thank you for the reply, and the talk was really interesting, so was your presentation. I think the OP and I felt the way we did because you are picking on a game that has done rather well.
As someone who played all the Witcher games and the Shamutanti Hills, I'd argue that the comparison in terms of narrative is a tad bit unfair (this is of course my opinion as a consumer, I am not a game-dev), since I find both the games offer me different emotion. I am not playing them for the same reasons.
However, playing the devil advocate here, it would've been even more spectacular if Geralt's world responded in a way you described choices and action in your presentation. But was it necessary? Again as a gamer, I didn't think so.
Many NPC responses to changes in their world seemed to be quite well done for a game of Witcher's scope. For instance, when you take out a monster plaguing the village, the NPC's walking around does briefly mention it as you pass by. The mission about the Witches in the woods is part of a much bigger plot, and it does make absolute sense that you did not find them earlier and you see them there a bit later in-game.
However, your comparison did emphasize your 'key-point' and I guess that was the main goal of the talk.
@@unkindled_dreg Thanks for the response. And indeed, I'm not attempting an analysis of the Witcher; more just using a familiar example to get people thinking about the right kinds of issues. You've got use a real world example, really, so I prefer to "punch up."
Are more subtle responses necessary? Who knows?!? For myself, I'm interested in designing experiences that react more fluidly and have fewer things you simply have to accept due to game-logic; but that's a totally subjective measure. Ultimately, if you like a game, you'll forgive it whatever rules it has (and all games will always have some rules.)
The one thing I will say is that every time a game does notice a "little thing" the player has done, it's always a great moment, and I think some of the systems we've used can help to generate more great moments.
@@joningold5256 Is there anywhere, a blog article perhaps, where I can see more examples? I'm especially interested in where you start: do you come up with the quest then figure out a state machine, vice versa, or you bounce between the two organically? How do you determine that the peasant needs 2 state machines vs 1? How do you determine that the wolf needs 1 vs 2?
PS: I bought Sorcery 4 after watching this and I'm glad I did. Great game!
11:31 traditional quest structure
49:43 redundancy
This whole talk, I found myself thinking a lot about Majora's Mask, and wondering what Jon's comments would be on the narrative structure of that game.
Now whatcha do is combine this with Tarn Adams' Villains update talk. You put hooks and state machines in every direction and you're off to the races.
Oh, could tou share the video about villains you are talking about? It seems very interesting!
"Tarn Adams - Villains in Dwarf Fortress", from the channel Roguelike Celebration, is probably the one.
Such a wonderful presentation!
excited to be early to a narrative post!
Inkle’s and Failbetter Games’ presentations are my favorite
But this "Sorcerer!" model *is* how quests work in some open world rpgs like Fallout (at least 1, 2 and NV).
I really like his idea with encounters. It feels like a better idea than Quests in every regard, but it sort of has one problem, which is that it becomes a combinatorial nightmare if you try to tangle too many encounters together. I can sort of see why big games don't try to do this. Since you're writing all the text yourself, the amount of work grows exponentially with each parameter in the web of connections. Now imagine having to voice all the potentially superfluous lines and make the character animations for it.
I feel like it might be an unsolvable problem, but I like that Inkle are working on narrative challenges in games, it's interesting and important work. Having played the first 3 sorcery games, I can say they are very good, they make good use of the system and I think it's because they realized that there is a limit to how complex this stuff can get.
7:30 is there some reason we skip the slide with Sorcery 4? I would have liked to have seen that
Can anyone explain how these linear state-trees/machines are implemented in the inky language he demos please? I see individual variables but no indication of linking
In Ink, you would use 'lists' which can contain one of the options you declare for it.
```
LIST Wolf = HeardOfWolf, SeenWolf, KilledWolf
You see a wolf.
~ Wolf = SeenWolf
{Wolf < KilledWolf: The wolf isn't dead yet.}
```
I wonder how they dealt with 'reminders'. If the player has started an encounter and run off to do some other encounter, how can a player remember that there is an encounter that is unfinished.
Horizon does it through voice lines. For example, every time you encounter a creature that you haven’t “unlocked”, she says “I wonder if there’s a way to find more data on these creatures” or something.
It took me forever to figure out what it meant, though, and then I felt stupid. So I’m not sure if it’s the best solution 😅
Largely, we don't. If you walk away, the plot line finishes itself; or expires; or causes some consequence; whatever makes sense depending on the plot. We treat "wandering off" as an active choice.
@@joningold5256 This is now the 3rd time i’m back to this and your Heavens Vault video. Now i have to go play all your games. Have been secretly making notes for a decade to create my own personal game, and this is really seeming like a process that’d work for my ideas. In fact just “gradually pouring tons of stuff into buckets and creating structure as it emerges” is how i make all my art & music. 😂 i don’t mind it being kind of a pain, as long as the result is magical-ish.
@@joningold5256 What a great solution!
To anybody interested in a strategy game employing these ideas: look at King of Dragon Pass and the recent sequel.
Wow. I'm only half way through and I've already gained so much insight into planning a game like Sorcery. This is exactly what I need--the thinking behind the scenes. The only question I would have is does having an open world and so many adventures cost in terms of the size of the game? Or is it not an issue ina text-based game? Excellent video. Thank you!
Event Sourcing applied to quest state, pretty nifty.
Fellow live audio gang: the reflections in this lecture hall, my god lmao
2:11 “it was incredible, it was magic, it was… Sorcery!” Missed opportunity there ;)
Excellent talk!
"you cannot have grind in a branching narrative game world", tell that to twine or renpy game devs!
SHUT UP AND DISTRIBUTE 239 BUSHELS OF WHEAT.
REPEAT UNTIL DAY 30.
I'm so backlogged right now, but I totally wish I could give this game series a try. It's even on sale right now wtf
Such a great talk
Groundbreaking work! 💥
הי יובל
Outstanding! Cheers
Very enjoyable talk, but I'm not sure your solution of basically having the radiant quests from Skyrim but they automatically fail if you return to the characters house too many times is that great.
It's more of, "if you pick up a story beat and ignore it for too long, it is resolved in other ways". Skyrim has no urgency for anything because the story only moves when you do. The base story is the literal end of the world caused by the First of Dragon kind, come to destroy the civilizations of Men and Mer. Completely ignoring that quest chain has absolutely no consequences, and every character outside of that quest chain acts almost exactly the same whether you initiate the next quest step or not.
You can literally bugger off to Markarth Solitude, or anywhere else, complete the thieves guild, companions, and college of winter hold threads, and still get treated like an escapee on your first encounter of Riverwood, or an asshole by Irileth when approaching Balgruf for the first time.
Dragons don't spawn in the sandbox if you don't fight the first one outside of whiterun. The only ones that do are the story beat Dawnguard bosses for their respective battles.
If skyrim had an encounter-state system like the one described in this talk, the world would react to the overarching threat regardless of your progress, with the possibility of an actual fail-state.
@@redfiend1670 This is a very late response (apologies for the ping it will generate for you!). Skyrim's problem is the narrative DOES have a sense of urgency, but the game rules don't require you to act upon the urgency. For example, after you kill your first dragon, everyone says you must go talk to the greybeards right away (there's even side dialogue to the effect of "Don't keep the greybeards waiting"). But there is no consequence or change for going off and doing your own thing.
The thing is, in an Elder Scrolls game I prefer the ability to go off and do my own thing at any point. This is supposed to be the main calling card of TES - be anyone, go anywhere, do anything, at any time. For a TES game, the narrative should be structured in a way that makes this natural.
Morrowind did this well. The main story was clearly important with a sense of dire weight, but it also didn't push you to act urgently. There are even moments near the start of the narrative where the main questgiver says "Good, you're here. Take some money and go establish a cover identity. Join a guild, do some quests, level up, then come back when you're stronger. Or if you think you're ready, we'll get started." The narrative naturally ENCOURAGES you to do your own thing, but leaves a natural option for moving forward if you wish. And then there are some rare moments in Morrowind's main quest that do take on a level of urgency (when the Tribunal Temple becomes more hostile), and there are real consequences (i.e. if you're a Temple member, you can't do any of its quests until you advance the main quest to the point where the Temple is no longer hostile).
I love Skyrim but I often avoid the main quest precisely because it fights against the game's message of "be whoever you want, do whatever, go wherever, at any time". If I do decide to pursue the main quest, I often have to build elaborate roleplaying reasons in my head for why my character is ignoring the urgency of the main quest (an approach that worked once or twice but lost power for me after a while).
I don't think the problem with Skyrim is that the quest design is bad and an encounter design would be better. Rather, the problem is the narrative pacing doesn't support the game's own open world/freedom quest design.
10:41 SHOTS FIRED
Awesome talk!
Why is this only posted now?
Are you complaining getting something for free now, while others had to pay previously?
@@danielleitner9369 it was already for free on the website, I’m just curious why it wasn’t on youtube
@@_purplewinter_ I think they do this step by step: first behind a payment wall on their page, then public for all but still on their page (I guess to increase traffic/awareness) and last here on YT. Sometimes it takes years until it ends up here.
Reminds me of Deus Ex where they tried to a large extent to make the game reacts to whatever decisions you make. Atleast in the beginning, it does get more linear.
The Witcher example is astounding, because so much of The Witcher 3 *does* let you do quests out of order. I played one of the quests on the main path before I was prompted to go there just because I was wandering around in the area looking for fun quests to do. It blew my mind when Yenn told me "you need to go help this werewolf (different quest) and that should help lead us to Ciri." Because I had already met that werewolf and helped him, and was surprised that they let me break the sequence that way, Geralt even had a bespoke line to tell her that he already did it. The lead quest designer from CDPR gave a GDC talk where he talked about another main-line quest that lets you complete it before triggering the first step, too. Maybe it's because these are examples of side quests that are interwoven into the main quest and not just side quests that are just dotted around the map? Not sure what exactly the difference is, if the main quest won't progress until you complete it, it's not a side quest. But whatever lol, point is, you can do a lot of quests out of order and it's interesting to me that they neglected it in other places.
The "Serious fatal story bug" is a massive reason why this system is not implemented in more games, while it sounds acceptable in a Sorcery game, I'm not sure Assassin's Creed fans would react kindly to that. Multiply the audience by several factors of magnitude and "obscure bugs" are now a "vocal minority".
I think that's true.
AAA studios are putting out trash games now, and people still buy them. A massive AAA publisher/studio like Ubisoft could absolutely execute on this, and do it well, but they won't, because their fans will buy anything they throw at them, in the already uninspired state it comes in. Why do more work when they're already eating out of your hand.
Beyond any of that, the bugs are fixable. Developers just have to acknowledge that the more complex their systems are, the more time they have to allot for testing before publishing. If developers are not going to do something because their might be bugs, they might as well not do anything at all.
Such a great game
He says the special sauce of these state machines is the implication that later states imply earlier states. Why would “peasant dead” imply “peasant impatient”? It might imply “peasant was impatient” but not just “peasant impatient”
I think "Peasant dead" is just a bit of an oversimplification, what that state actually means is "Peasant killed by wolf" which can only come about because of him getting impatient. Realistically "Peasant dead" (If relevant) would be a separate state that could come about from multiple possible events (The wolf killing the peasant, you killing the peasant, etc)
As a consumer, I love to see that kind of stuff
so did BG3 address all this?
Very interesting
Oh god, i tought it was David Cage for a second.
I disagree on what he thinks an open world is by definition, they don't need all that he said they did to be considered open world. They just need to be open, as in finish the tutorial quest and then pick a direction and go that way until you hit the map border or follow the main questline if you want.
I bought Sorcery! And I felt ripped off because it wasn't clear that it was part 1. I got too the end and felt extremely unfulfilled, and never bought the other ones.
This talk is a really interesting way of approaching world design. Might translate well to a tabletop rpg space.
@Min Lungelow it does, sort of! 'Sorcery' is based on 'Fighting Fantasy', of which there is an 'Advanced Fighting Fantasy' ruleset. The tabletop rpg Troika! is based on that.
I get the feeling that FromSoft may have watched this before developing Elden Ring.
Shout out all the Jerma fans!
Hm. Sounds pretty much like Stalker's initial design for the narration system...
What if as a punishment for not being “active” on this wolf encounter ... like the peasant; Other NPC’s in this town also started to “disappear”? Maybe even that important vendor who sells you those rare mushrooms to craft that epic potion? “God those mushrooms are a pain in the ass to loot” 😪
I know this is from 2017, but I hope this dude has reigned in his ego by now. He came across in multiple parts of this video as very holier-than-thou and full of himself. Some of the information was great, but the rest came across as "my game is great, if you don't like it you're wrong."
i dont think its fair to compare a text based game to a game like the Witcher. You already point out how much branching you need to do to achieve your intended level of narrative in a text based game. Now imagine how much more you have to code for a massive game like the Witcher. Thats like picking on a Skyscraper for being not as creative build as your lego tower. I still learned a lot from your Video and it gave nice Ideas so dont get me wrong
Well bg3 pretty much follows this system
Thank you.
If a game is about exploration and finding special things, shouldn't the value of any one thing be an inverse with the number of players expected to find it? To specify a bit, let's rate the value of any discoverable thing by 100 points, then we take that base and deduct the percentage of players who experience that one thing. If all players experience a thing, it has a value of zero, for example. Then once all findable things are calculated, you spend your time working on those things that have the highest score.
I've seen this idea implemented in hentai games
Name a few
which ones? What's funny is that some hentai games have some interesting systems that if improved, could be really cool. Artificial Academy 2 for example had some interesting ones tracking npc relationships and emotion states, like characters being angry reacting in a different manner than calm ones, or certain personality traits like being lazier making the character skip classes, have less inteligence staff and that cascaded into other things, all with I imagine was less than 50 lines of code, becuse most of it were just simple lists.
Lol the more bitter the desinger; the better the talk.
JERMA
+
This guy never played Paper mario on WII, i’m sorry but that game make you stop and grind lmao
Designing games to allow players to break them. Kinda.
Damn son! Raw stuff here!
Can this lad be any more smug?
His definition of open world seems like it would only include completely rng worlds, or just really terrible open world games.
And it's funny how his alternative to 'open world design' is just open world design
Grind is cancerous mechanics in any kind of a game. Just artificially prolongs the time spent in the game.
Most of the time I think you're right, but grind can also be thought of as a way to lower the game's difficulty. The problem arises when the natural difficulty is so hard that even the best players are forced to grind, which unfortunately does happen in way too many games.
@@scrub_jay Using grinding as a solution to a player's lack of skill is the laziest way to address the problem.
@@Venomine95 I might agree, but not every game addresses the problem at all, and giving the player the ability to grind is better than letting them get stuck.
@@Venomine95 the laziest way to address the problem would be a difficulty setting
I actually like grind as long as its not too long... i enjoy observing my fight improve and getting the reward of overpowering that monster that previously overpowered me. The key to grind is balance, and most games overdo it.
I wonder if this guy's ever heard of fallout new vegas
New Vegas *looks* similar from a player standpoint but the groundbreaking thing here is how so much easier it is to do what New Vegas does
@@339blaster true
so both your flow chart of the witcher story and your story is missing one really important possibility, the player never steps foot in the swamp or that peasant house, in a real open world that is not just a possibility but should be high likely, when you make a big story with lots of branches that cost money to make, you put time effort and pride in to it, are you going to be tempted to try to push the player in to exploring it? these encounters do you force them on the player? these encounters are random order narrative plot? why do you think your text game is a open world if you can not have a dead space where the player finds nothing and can go off and do what ever they want. I think this is a straw man speech, you setting up the witcher a actual open world game and comparing it to a text game that has limited story nodes and locations. a finite state machine is FINITE limited NOT OPEN this is why it's a straw man. you could make a open world text game sure but It would require dealing with the dead space, you enter a empty house and find nothing, I burn it down, the fire catches to the rest of the village. all those plots we worked so hard on are now gone, what happens in New Vegas stays in New Vegas but your reputation takes a hit. oh no I save scum, I clean the house, you start cleaning, it you find a tin can maybe you can think of a use for it, if you clean this house 4 more times you can think of this house as your base but beware the tax man because you have modified the world. latter a troll demands some object for crossing his bridge it could be any object. the objects you have available are your sword worth $1500, your armor worth $5000, a important plot item value $0, a tin can value unknown, you are able to give the tin can, but it could have been a twig or wolf head or something else like a chicken. dumby the mad king was going give you $1,000,000 for a "priceless" item never trade unique items, luckily their's save scum. games are not books, yes some games are books just not "open world" ones. please don't try to kill open world games just because they are not what you do well.
"the player finds nothing and can go off and do what ever they want"
in this sentence you have made yourself believe a variety of things are both interesting and possible without it being true
You will find for most people 'nothing' isn't interesting - it would be similar to having an option appear in the game that says: "Nothing here"
countless empty cabinets are not fun, nor empty cabins etc etc
Now 'do whatever they want' - wow, I don't even understand how you don't see this - this is exactly the important part of the game and the one the talk is focused on: doing stuff
You call his speech a straw-man and call his definition of the open world wrong, while you lie to yourself about what the open world games actually are, even though you describe them perfectly in that sentence: mostly empty space, which you have to traverse to find the interesting bits
Most of them delete the traversing altogether and just add fast travel, the only reason you are fooled into thinking this is great is the graphics, but in the end you are auto-travelling to the next plot-point - it works, barely, but it ain't strong design
Why does this video exist and yet infinite more shitty games keep coming out?
This is basically just what Emily Short calls quality-based narrative. I think most of this has been understood by the interactive fiction community for decades. It's not remotely innovative.
Can you point me in the direction of some innovative gdc talks?
Awesome presentation, but when you presenting something please do not make repeating "mitch, mitch mitch" sound with your mouth. It is highly irritating after hearing it ten times!
The worst part of games are the story. They've never been good, and are no good examples.
bait