The promise of freedom in story games

Well, i'm not a game designer, but since nobody else did, i'll reply…

aigan said:
Imagine a game where you inhabit a living breathing world with the freedom to go anywhere you want. Find your own unique solutions to the problems you encounter. Befriend any NPC on your way and form lasting relationships.

Probably we already can go where we want in many games.

Agree about unique solutions, which should emerge and enable player creativity. I have not seen this yet, but i assume we could have some progress with making world simulation much richer than usual. But this also means we loose control. And we can't entertain by simply claiming ‘entertain yourself’.
So this seems very difficult to achieve. And sadly i do not see much iterative progress towards this goal either so far. Surely worth to discuss…

But why should i want a relationship with a NPC as a player? And as a dev, why should i attempt to drive people away from having real relationships by luring them into a effortless, virtual alternative?
I want neither of that, so disagree on this point but rather try to minimize any need for verbous player - NPC conversation. I think we can have progress on using dynamic body language, showing more facial emotions and focus, and hinting NPC intents this way to the player. Current games really neglect those options i think. But they should work great, since there is no need to interrupt gameplay to go into some clumsy communication mode.

But i guess that's pretty personal. Some people would enjoy to talk to AI NPCs through a microphone, but i will never like this. It just goes too far for me.

aigan said:
There is a sort of common wisdom in game development that player freedom has to be sacrificed in order to create a more emotional meaningful story experience.

Yes, but it is not just about story.

Imagine a game of chess, but every piece can go anywhere.
Do we still have a game?

No. And that's why we can't have total freedom if we want to make a game.
You can make a reality simulation sandbox, simulating society as well, and it might sell well. But it's not a game then. (You could add mini games to the experience, though.)

And the same applies to story.
You can write a great linear story, e.g. Romeo and Juliet.
You can turn it into a branching story, e.g. enabling a happy end.
But then the interesting point - the tragedy - gets lost below introducing choice.
As you increase the number of branches and choices, there is no more story at all. All that's left is a random path through various events, properly described with text.
But there no longer is a story in the classical sense. It may work, but noone will cite your story hundreds years later as happened with Romeo and Juliet. People will forget it as fast as events pass by in an ordinary day of real life.

That's not a problem we could solve. It's not even a problem at all imo. We may think it is a problem, because dynamic story is hard to implement. But this does not turn it into a problem. Great stories are linear and static, and there is nothing we can do about it. All we can do is accepting some compromise if we need dynamic story elements.

aigan said:
The mechanics connect you to the world. The responsiveness of the control will let you embody your avatar. The world should react to everything you do. The ability to touch the grass, move objects and leave your mark will enhance the immersion.

Yes, but this builds on the assumption that a player - in theory - could do anything in a game.
Which sadly is not true. We can move and look in various directions, which works pretty great. We can aim and shoot, which works well. We can mash buttons for a melee fight, which already is pretty abstract and indirect. We can pick up and move objects, but placing them already feels very clumsy.
And that's it. Nobody has ever figured out to do any more with our bad game controllers.
To do more, we need high level abstractions such as multiple choice dialog boxes, inventories, skill trees, etc.
But using this already interrupts and breaks the interaction with the games simulation.
So no. I can't touch grass or anything in a game. If i try to deliver such experience, i will remind the player that he's not really in the game. Which does not happen if i don't give the option to touch grass.

So yeah… another dilemma we can't do much about, i'm afraid.

aigan said:
For example, if a door can’t be opened because of a scripted quest that requires you to do something else first. If there is an object you should be able to push but it doesn't react. Cutscenes are another example of something that takes your volition away.

But there is a big difference: Cutscene interrupts and breaks immersion, a closed door does not.
And a closed door is just needed to define a problem the player may want to solve. It's essential. We can't make games without defining problems plus their intended solutions.

But i agree there should be multiple ways to open the door, at least sometimes.
I always try to break doors using heavy objects or using long objects as a lever to pry them open.
And then i'm disappointed it never works. For some doors, it should work. The others could be just made more massive, so i see there is no hope to force it open with external force i could apply.

aigan said:
The verb should be directly mapped to a specific movement of your avatar.

Which, due to the extreme restrictions of controlling a 20 bones avatar with just 4 buttons, would rule out a lot of options to the player. So i would not rule out verbs completely.
It's subjective. To me, Sierra Gfx Adventures were too heavy on text input. I did not like to type, and it felt like a guessing game to come up with the right sentence the program understands.
Lucasfilm Point and Click Adventures solved both problems by using a GUI over verbs and inventory items. I loved this, and it gave me the impression i could do anything in the game.
Later, verbs were removed and everything was streamlined to avoid the complexity of building sentences.
Not much was lost actually. It worked and was convenient to use.
But nowadays, looking back, first thing i remember is the verbs when i think about Adventures.
I don't want to bring this back, but i want some simple interface to control complex options. For stuff that is too complex a player could execute the action manually himself. But whatever it is, i think the only way to get there is to sacrifice the immersive first person experience, which kinda sucks.

aigan said:
It should be clear what you can do in the world and what effect you can expect from your actions. Ideally, anything seen in the world should interact with everything else according to rules you can learn to understand and use to your advantage. …

Agreed, but to get there, we need to solve all of the above.
It's a dream, a vision, a goal. And it's obvious.
But we don't know how. :D

aigan said:
The freedom to explore anywhere you want. You should not be locked to a specific area because of a specific quest.

Please do restrict my space of potential options!
Otherwise i'll feel overwhelmed from an infinite problem space very very quickly, and i'll put your game into the folder ‘RPGs i should play to figure out what people like about it’, and then i'll never touch it again.

I need your helping hand to guide me.
Just make sure i do not notice your guidance. ; )

aigan said:
Some linear action-adventure games have started using wide-linear levels where you have the freedom to choose a couple of different paths or at least feel like you have the choice of path, but still end up in the same place at the end of the level.

Pretty ideal. Even with multiple options, we must make sure there are some points of convergence. E.g. the end of a level, and even the moments of choice in a branching story. Otherwise we have nothing to implement level design, a story, a puzzle, etc.

Would you be happy if there were two alternative endings for your level? And you want this in any level, maybe together with related story branches?
If so, notice the cost: You need to create much more content, and each player will only see a small amount of it. You get a very wide tree with many branches, but each individual path through the tree remains short. You have a lot of work, and the player gets the feeling he misses out a lot of content. It increases replay value, but it won't compensate the cost on your side. Thus, designing all the options we want to give to the player manually seems no good idea. Also, since all options and branches are predefined if we work them out manually, it's still static, and the player can not express creativity. He can only select from choices we set in stone in advance without his interaction. The freedom is just an illusion, and that's noticeable. To me personally, it's even annoying.

But maybe we can design in a way so those options emerge from the simulation. Like in the real world, where i can do so many things that there are almost always multiple ways to solve a problem. That's possible i think, although idk how precisely.

aigan said:
The design of the level with rubble and highlighted paths are of course meant to serve the pacing of the hand-written narrative. All the resources can be focused on a limited set of locations that are needed for the story.

It's not necessarily because of hand written narrative. It's usually done to achieve any level design at all. And that's not bad, it is good.
Imagine a FPS game, but open world on flat terrain. Freedom to go anywhere, but no fun. You need those walls for partial cover and safety, and to enable some strategic planning. Those corridors are not as bad as people in the 2K's started to think they are. I'm sure they already regret their rant.

What i mean is, imposing constraints and restrictions is more important for action games than options, choices or freedom. And maybe we can arrive at those goals easier if we think about which obstacles, constraints and restrictions actually enable or require multiple options.

aigan said:
The better alternative is to use Systemic Story to adapt quests to anything the player has done.

But this gives quests which are more generic and less exciting and interesting, so you can't claim they're better in general.

But i agree it's cheaper to produce, and thus especially interesting and attractive to indie devs.
There might be a lot of benefits affecting other things, not directly related to actual player freedom itself.

aigan said:
This will allow the virtual game-master to relocate story elements based on which clues the player decides to follow.

Sounds interesting.
I never thought about any from of procedural story generation / adaption. At least not before the ChatGPT moment.
I did not read your blog either, but do you actually think procedural story is doable, using traditional programming?
I know there are some working examples, but i neither played them nor do i know how they work.

aigan said:
Even worse is when the avatar is forced to do things that the player absolutely doesn't want to do, as in making an NPC upset, saying stupid things, or even killing people when you would rather avoid them or make friends. As a pacifist, this is something that frustrates me in almost every chapter of every action-adventure game.

I remember the Prey game, where i had to kill ‘my’ girlfriend because she was somehow merged with an alien / robotic body. But she was alive and could talk. It was far from hopeless imo. But the game did nothing, keeping me stuck until i've shot her. And i really did not like this. The game was great, but they should have removed this one scene and decision, which was just bad.

I also have a related problem. Games became more realistic with time. NPCs look like people actually, no longer like squares, pixels, or images. But i still have to shoot them all. Without purpose and justification. It's embarrassing for us game devs. Some try to compensate by introducing some ethics or political correctness into the story. But that's even more embarrassing and bolt on.

Your proposal could fix the first problem i guess, but not the second.
To address that, my proposal is to defy realism. Make the game a B-Movie alike splatterfest, or an actual Arthouse SciFi story with some intellectual quality, for some examples. Just keep it surreal enough, so nobody takes the bloodbath action too serious.

Besides, avoid any statement on ethics we're not competent to make, and never force the player to make such statement, if it's not a central aspect of the story we try to tell.
Problems solved. But i don't need a system for procedural story telling. I can't be sure every player will like my predefined path, but i can't be sure he would prefer choices either. Even if the player himself tells me he wants more choices, more skins to express himself, more hairstyles and beards…, i would not not be sure about giving him all this would make him happy.

aigan said:
Why bother making the choice if it doesn’t matter in the end?

Exactly.
Assume you succeed, and you achieve dynamic story. Some next gen evolution of the branching story concept maybe.
But then, if every choice has impact, it won't matter in the end regardless, and even less than before.
This is because at this point, a dynamic and unique showdown is the new expected norm.
It would be like procedural terrain. Ok but nothing exciting. Some expected background, but not a reason to play this and not that game.

I don't think this can be avoided. It will never compete Romeo and Juliet.

But i also don't think that's what action games are good for.
I'm afraid most of this desire for good stories comes mainly from the fact that our gameplay became standardized and is no longer that interesting or exciting. This is accepted by both players and devs, so they expect progress on other issues such as combining a dynamic game with a static story.

However, i must admit i have built up my doubt regarding dynamic stories over decades on the belief that it's just impossible technically, so maybe i try to convince myself it is not that important anyway. If my belief would change, my mind would surely follow.

So i wonder, do you talk about general goals and visions, or do you say it can be done? If so, how? What data structures and algorithms could generate dynamic / adaptive story, and how far could this go in practice?

Reopening per OP request.

aigan said:
I believe we can have both the sandbox gameplay together with the top-quality story.

If you have the budget to create the game you described, and you think you can get a profit on it (or are willing to lose the money) then absolutely go for it.

You're likely already familiar with what the biggest games do with a half billion dollars today. Opening it up more is certainly possible if you can afford it.

@khawk Thanks for re-opening the thread.

I do things more long-term than often makes for the best engagement. This is something I started to develop 30 years ago but only recently started to write about.

The post here is about half of my article “The Promise of Adventure”. Its part of my series for creating a systemic story game, i published on my blog.

So welcome to Adventure Alpha Zero 2024.

@joej said:

Well, i'm not a game designer, but since nobody else did, i'll reply…

aigan said:
Imagine a game where you inhabit a living breathing world with the freedom to go anywhere you want. Find your own unique solutions to the problems you encounter. Befriend any NPC on your way and form lasting relationships.

Probably we already can go where we want in many games.

Agree about unique solutions, which should emerge and enable player creativity. I have not seen this yet, but i assume we could have some progress with making world simulation much richer than usual. But this also means we loose control. And we can't entertain by simply claiming ‘entertain yourself’.
So this seems very difficult to achieve. And sadly i do not see much iterative progress towards this goal either so far. Surely worth to discuss…

You are right that the type of narrative that emerges from an open sandbox is not enough to make good stories. Things have to be directed. That's different from a simulation game.

But why should i want a relationship with a NPC as a player? And as a dev, why should i attempt to drive people away from having real relationships by luring them into a effortless, virtual alternative?

Caring about the story often comes with caring about the characters. Character development and relationships are an important part of the type of game I would like to see. Think of main, companion or side characters from popular story-driven adventures and RPGs. It's often characters that will keep your interest in a way that you want to meet them again in upcoming games.

But in addition to that. I often prefer single player games, and also TV-series and “streamers” as a less boring and less frustrating alternative to companionship since I’m introverted and have a harder time making friends.

I want neither of that, so disagree on this point but rather try to minimize any need for verbous player - NPC conversation. I think we can have progress on using dynamic body language, showing more facial emotions and focus, and hinting NPC intents this way to the player. Current games really neglect those options i think. But they should work great, since there is no need to interrupt gameplay to go into some clumsy communication mode.

But i guess that's pretty personal. Some people would enjoy to talk to AI NPCs through a microphone, but i will never like this. It just goes too far for me.

One of my inspirations is how lifelike even an icon can appear when it responds to your actions in a way that conveys an internal state. For example your pet in NetHack or the prisoners in Prison Architect or your characters in Rimworld. It doesn’t have to be photo-realistic.

My goal would be something like Cyberpunk 2077, but the first step would be a text-only game with an abstract description of events and without any dialogue.

@joej said:

aigan said:
There is a sort of common wisdom in game development that player freedom has to be sacrificed in order to create a more emotional meaningful story experience.

Yes, but it is not just about story.

Imagine a game of chess, but every piece can go anywhere.
Do we still have a game?

No. And that's why we can't have total freedom if we want to make a game.
You can make a reality simulation sandbox, simulating society as well, and it might sell well. But it's not a game then. (You could add mini games to the experience, though.)

Life has limitations. Nothing has complete freedom. I’m talking about something that can be done. There are plenty of sandbox games. Many of them have some measure of simulated physics. I think they can be a lot better, but it’s at least something. There hasn’t been a game yet that does the same thing with story. That is, doing story through systemic interactions.

It’s not a simulation sandbox. Places, people and events are selected based on story rules.

And the same applies to story.
You can write a great linear story, e.g. Romeo and Juliet.
You can turn it into a branching story, e.g. enabling a happy end.
But then the interesting point - the tragedy - gets lost below introducing choice.
As you increase the number of branches and choices, there is no more story at all. All that's left is a random path through various events, properly described with text.
But there no longer is a story in the classical sense. It may work, but noone will cite your story hundreds years later as happened with Romeo and Juliet. People will forget it as fast as events pass by in an ordinary day of real life.

That's not a problem we could solve. It's not even a problem at all imo. We may think it is a problem, because dynamic story is hard to implement. But this does not turn it into a problem. Great stories are linear and static, and there is nothing we can do about it. All we can do is accepting some compromise if we need dynamic story elements.

“That’s not a problem we could solve” is exactly the belief I wanted to address.

A systemic game doesn’t have branches. It’s not 10 or 20 endings. Like the Sims, the endings are infinite, only limited by the number of interesting combinations made possible by the systems implemented and how they resonate with the player.

The story rules include themes, motifs, tone, character development, and more. The game will use everything the player has engaged with. As hinted by Watch Dogs Legion, you can talk to a person, and that person will be woven into the existing narrative.

The tone can adapt to the player. If the player behaves silly, it could use story rules from action-comedy. If the player is careful, it can take more from a thriller. If the player is flirty, it could use rules from romance stories. If the player is a murder-hobo, it will take rules from outlaw stories.

Giving the player agency doesn’t negate the theme or tone. The story can morph to a different kind based on the player's choice. They could manage to escape, or find another way. But if a theme is established, it will come up again and again in different forms until it’s resolved or combined with something else.

Let's say this is the part of a story where the player hero sacrifices themself. If they are very skillful, they should be able to survive, but then there would be something even more difficult compounding the situation. And they should be able to survive even that trial if they really tried to. For every type of story theme, there are variations that can be mapped to the player stalling, shortcutting, succeeding or failing. The sacrifice part usually is about the hero finally letting go of their faulty ideas, so the important part is the decision to do the sacrifice, regardless of the outcome. And if the player diverts, the theme will be postponed, and combined with more threads.

This will create personal stories. Something that has the potential for deeper impact than any other form of media in existence. A computer can check the constraints against thousands of variations finding a better match than any tabletop game master would.

@joej said:

aigan said:
The mechanics connect you to the world. The responsiveness of the control will let you embody your avatar. The world should react to everything you do. The ability to touch the grass, move objects and leave your mark will enhance the immersion.

Yes, but this builds on the assumption that a player - in theory - could do anything in a game.
Which sadly is not true. We can move and look in various directions, which works pretty great. We can aim and shoot, which works well. We can mash buttons for a melee fight, which already is pretty abstract and indirect. We can pick up and move objects, but placing them already feels very clumsy.
And that's it. Nobody has ever figured out to do any more with our bad game controllers.
To do more, we need high level abstractions such as multiple choice dialog boxes, inventories, skill trees, etc.
But using this already interrupts and breaks the interaction with the games simulation.
So no. I can't touch grass or anything in a game. If i try to deliver such experience, i will remind the player that he's not really in the game. Which does not happen if i don't give the option to touch grass.

So yeah… another dilemma we can't do much about, i'm afraid.

Have you experienced the DualSense controller? You can actually feel the grass under your hands in Horizon Forbidden West.

But what I’m referring to about responsiveness is all the things the game locks you out of doing. For example, when you can’t move because you just triggered a cut scene or a long animation. When you can’t sneak around an enemy because of an invisible wall. When NPCs don't react to you jumping around or trying to kill them. It’s more about consistency. Either make everything climbable or mark all climbables up with the yellow paint. If you need an NPC alive, you can have the player still be able to move the gun but never point it directly at the NPC.

Making a mark on the world can be things like footprints in the sand. In the Last of Us, the player will knock over objects in the environment, even if they can't be picked up.

I wrote an article about a control scheme for embodied manipulation. The difference between real life and a traditional game, when it comes to precision is handled by a combination of reticules, and visualizations of trajectories and targets with a smidge of aim assist and ability to continuously adjust. For example the ability to adjust the landing (slightly) even in mid air.

@joej said:

aigan said:
For example, if a door can’t be opened because of a scripted quest that requires you to do something else first. If there is an object you should be able to push but it doesn't react. Cutscenes are another example of something that takes your volition away.

But there is a big difference: Cutscene interrupts and breaks immersion, a closed door does not.
And a closed door is just needed to define a problem the player may want to solve. It's essential. We can't make games without defining problems plus their intended solutions.

I have seen a lot of quests or arena designs where you can’t exit the scene, not because the door is locked but because you were supposed to talk to a person or take an object or kill an enemy first. Nothing about the door or exit changed. It’s just a type of invisible wall around the arena that stays there until the game has progressed. I probably see more of this since I usually try to avoid fights. Many game designers may not have even considered that the player would try to run back the way they came at the first sign of trouble.

But i agree there should be multiple ways to open the door, at least sometimes.
I always try to break doors using heavy objects or using long objects as a lever to pry them open.
And then i'm disappointed it never works. For some doors, it should work. The others could be just made more massive, so i see there is no hope to force it open with external force i could apply.

This is exactly the type of thing I’m referring to. There are so many times then the avatar could clearly just walk right through the giant hole in the wall, but aren't allowed to. Or climbed over the tiny heap of rubble.

aigan said:
The verb should be directly mapped to a specific movement of your avatar.

Which, due to the extreme restrictions of controlling a 20 bones avatar with just 4 buttons, would rule out a lot of options to the player. So i would not rule out verbs completely.
It's subjective. To me, Sierra Gfx Adventures were too heavy on text input. I did not like to type, and it felt like a guessing game to come up with the right sentence the program understands.
Lucasfilm Point and Click Adventures solved both problems by using a GUI over verbs and inventory items. I loved this, and it gave me the impression i could do anything in the game.
Later, verbs were removed and everything was streamlined to avoid the complexity of building sentences.
Not much was lost actually. It worked and was convenient to use.
But nowadays, looking back, first thing i remember is the verbs when i think about Adventures.
I don't want to bring this back, but i want some simple interface to control complex options. For stuff that is too complex a player could execute the action manually himself. But whatever it is, i think the only way to get there is to sacrifice the immersive first person experience, which kinda sucks.

Can you give me any example that couldn’t be handled by the physical movement of my embodied manipulation control mapping? The thing I want to get away from is the “use” verb, just because it can lead to surprising results when the player has another idea of how to use an item, or when it differs because of the exact cursor position changed directly before the activation. For example when you use X both for talking and jumping.

@joej said:

aigan said:
It should be clear what you can do in the world and what effect you can expect from your actions. Ideally, anything seen in the world should interact with everything else according to rules you can learn to understand and use to your advantage. …

Agreed, but to get there, we need to solve all of the above.
It's a dream, a vision, a goal. And it's obvious.
But we don't know how. :D

That was a description of NetHack. It was done in the 80s. And I guess the same holds true for Minecraft and several other systemic games. Just not for any story games. Yet.

@joej said:

aigan said:
The freedom to explore anywhere you want. You should not be locked to a specific area because of a specific quest.

Please do restrict my space of potential options!
Otherwise i'll feel overwhelmed from an infinite problem space very very quickly, and i'll put your game into the folder ‘RPGs i should play to figure out what people like about it’, and then i'll never touch it again.

I need your helping hand to guide me.
Just make sure i do not notice your guidance. ; )

The combination of freedom and guidance is a big topic. And players are already used to many techniques, from quest markers in one extreme to level geometry that gives the illusion of choice on the other end.

The avatar or a companion can comment on a suggested path. There can be dynamic events that draw attention if the player goes in the wrong direction. NPCs could be smarter and actually be able to give directions if the player asks about “did you see anyone run past here”, or “where is that antique shop” or “I forgot, where was that ancient temple”. It would be a lot of content for a traditional game. But not for a systemic game. This was actually something you could do in Daggerfall.

A game where you can say “anything” can still have contextually three suggested responses for the most obvious alternatives.

But the player getting lost or going in the opposite direction is not nearly as big a problem in a systemic story game, since the game will adapt to the player's choices. The world exists as a probability waveform until the moment the player observes it. That means that the game will create a location in the direction the player is heading, based on the need of the story. It could be that the player catches up with the thief whatever path they take, if it’s a path that would make sense for a thief to take. Or it could use another of the threads or create a new thread from the theme and tone that would match that point in the story.

@joej said:

aigan said:
Some linear action-adventure games have started using wide-linear levels where you have the freedom to choose a couple of different paths or at least feel like you have the choice of path, but still end up in the same place at the end of the level.

Pretty ideal. Even with multiple options, we must make sure there are some points of convergence. E.g. the end of a level, and even the moments of choice in a branching story. Otherwise we have nothing to implement level design, a story, a puzzle, etc.

Would you be happy if there were two alternative endings for your level? And you want this in any level, maybe together with related story branches?
If so, notice the cost: You need to create much more content, and each player will only see a small amount of it. You get a very wide tree with many branches, but each individual path through the tree remains short. You have a lot of work, and the player gets the feeling he misses out a lot of content. It increases replay value, but it won't compensate the cost on your side. Thus, designing all the options we want to give to the player manually seems no good idea. Also, since all options and branches are predefined if we work them out manually, it's still static, and the player can not express creativity. He can only select from choices we set in stone in advance without his interaction. The freedom is just an illusion, and that's noticeable. To me personally, it's even annoying.

But maybe we can design in a way so those options emerge from the simulation. Like in the real world, where i can do so many things that there are almost always multiple ways to solve a problem. That's possible i think, although idk how precisely.

Yes. But through a systemic story. Not world simulation. Branching paths is the wrong path. I wrote about that in “Everything wrong with hand-written narrative”. Simulation games are boring. We need the “director” that creates meaningful coincidences. It’s done by mapping all encountered places, people and events with existing threads and themes. If you choose to interact with something, it will be connected to an existing thread or start a new one that will be connected at a later point.

aigan said:
The design of the level with rubble and highlighted paths are of course meant to serve the pacing of the hand-written narrative. All the resources can be focused on a limited set of locations that are needed for the story.

It's not necessarily because of hand written narrative. It's usually done to achieve any level design at all. And that's not bad, it is good.
Imagine a FPS game, but open world on flat terrain. Freedom to go anywhere, but no fun. You need those walls for partial cover and safety, and to enable some strategic planning. Those corridors are not as bad as people in the 2K's started to think they are. I'm sure they already regret their rant.

What i mean is, imposing constraints and restrictions is more important for action games than options, choices or freedom. And maybe we can arrive at those goals easier if we think about which obstacles, constraints and restrictions actually enable or require multiple options.

You are right. The constraints create the challenge.

But that is in-world constraints. The rubble problem comes from when even the best level design can’t stop some player from misreading where they are supposed to go. Especially in cases where they can’t find the “right” path but see one or more paths their avatar should be able to take, but can’t traverse. Having a game robust enough for letting the player make their own path if they are clever enough is preferable. As long as it doesn’t become a dominant strategy. “Levels” should be varied in a way that the constraints lead to different kinds of solutions. Some games do this by making the enemies adapt to the player's strategy, as in MGS5. A systemic game can even adjust the level design based on the players capabilities, such as in Judas.

JoeJ said:

aigan said:
The better alternative is to use Systemic Story to adapt quests to anything the player has done.

But this gives quests which are more generic and less exciting and interesting, so you can't claim they're better in general.

But i agree it's cheaper to produce, and thus especially interesting and attractive to indie devs.
There might be a lot of benefits affecting other things, not directly related to actual player freedom itself.

I don’t believe they have to be more generic. They have the potential to be much more impactful than anything hand-written because they can adapt in so many ways. There is a large group of stories that isn’t that complicated if you look at their high level plot points. They are still engaging because of how they combine a bunch of story techniques in combination with character relationships and themes. The game can layer several of these systems and get the same type of complexity. Combining topic, theme, mode, conflict, relationships and social and psychological dynamics.

I don’t think it’s as hard as people think. The trick is to do it top down, starting with the abstract. Trying to create dialogue is a trap. The first version would only tell the story in the manner it would be retold from memory. But doing it in this way would allow for the right amount of layering for it to be just as good as standard tv-fair.

And with a lot of work, a AAA studio could expand this to something much more living. Especially now with all the advances in motion-matching. The engine would have all the data needed for systems showing emotional expressions. Same for sentence building and speech pattern with inflections.

aigan said:
This will allow the virtual game-master to relocate story elements based on which clues the player decides to follow.

Sounds interesting.
I never thought about any from of procedural story generation / adaption. At least not before the ChatGPT moment.
I did not read your blog either, but do you actually think procedural story is doable, using traditional programming?
I know there are some working examples, but i neither played them nor do i know how they work.

The existing examples use large example libraries of large sets of symbolic rules. But the type of systemic story generation I’m talking about would be with traditional techniques. It hasn’t been done but is really not that complicated. Just that every single attempt has combined it with traditional hand-written content, and thus gone in the same trap, every single one.

A high level story is built on conflict. And we don’t have to implement every kind of story. I have an idéa of where to begin with something limited but that can be expanded with more types of stories and more systems.

Start out in an isolated location. Do it in medieval times, since that limits types of objects used by people in their daily lives. Should start out with people having secrets, creating a type of murder-mystery. That would be the building-block included in most type of stories. I’m about to write more about it soon. Social, psychological and relationship dynamics could come later, even if it's an important part of dialogue interactions. My plan is sort of implementing the systems top down. Starting with the “season arc” or “main story”.

@joej said:

aigan said:
Even worse is when the avatar is forced to do things that the player absolutely doesn't want to do, as in making an NPC upset, saying stupid things, or even killing people when you would rather avoid them or make friends. As a pacifist, this is something that frustrates me in almost every chapter of every action-adventure game.

I remember the Prey game, where i had to kill ‘my’ girlfriend because she was somehow merged with an alien / robotic body. But she was alive and could talk. It was far from hopeless imo. But the game did nothing, keeping me stuck until i've shot her. And i really did not like this. The game was great, but they should have removed this one scene and decision, which was just bad.

I also have a related problem. Games became more realistic with time. NPCs look like people actually, no longer like squares, pixels, or images. But i still have to shoot them all. Without purpose and justification. It's embarrassing for us game devs. Some try to compensate by introducing some ethics or political correctness into the story. But that's even more embarrassing and bolt on.

Your proposal could fix the first problem i guess, but not the second.

My solution is to make the game richer, more like other types of stories, so that you wouldn’t have to behave like a psychotic mass murderer. My goal is to inspire AAA adventure game developers to do this type of systemic story. But so far as I would create a prototype system, I would not prioritize any type of combat. Just look at what usually happens in stories. It’s more about negotiation, even when it’s life or death.

A lot of the gameplay would be about finding the truth, convincing other people to see your truth, confronting the big bad with your arguments, chasing after him when that fails, messing up and getting caught, trying to escape your captivity, sneaking away, finding the evidence you needed, getting out only to find yourself framed for a worse crime, being confronted by the truth you didn’t wanted to face, finally admitting your mistake, gaining support from the people you previously avoided, getting supporting testimony, finally getting the big bad to confess.

Violence can be added later. I would make the “default” more realistic. But a systemic game can adapt to different preferences. The game could be tuned to murder-hobo gamer logic for those who prefer that style. Everything can be tuned to player preferences, or by default adapt according to the player's style. Configure the size of the world, degree of danger, stress level, story complexity, degree of brutality, guidance, deceit, and so on.

To address that, my proposal is to defy realism. Make the game a B-Movie alike splatterfest, or an actual Arthouse SciFi story with some intellectual quality, for some examples. Just keep it surreal enough, so nobody takes the bloodbath action too serious.

That can work.

Besides, avoid any statement on ethics we're not competent to make, and never force the player to make such statement, if it's not a central aspect of the story we try to tell.
Problems solved. But i don't need a system for procedural story telling. I can't be sure every player will like my predefined path, but i can't be sure he would prefer choices either. Even if the player himself tells me he wants more choices, more skins to express himself, more hairstyles and beards…, i would not not be sure about giving him all this would make him happy.

A systemic story game could be made to be basically indistinguishable from a traditional story game with just as much guidance, but with the difference that it would make the players on the margins less frustrated, because the game has the ability to adapt to their alternative play-styles.

A table-top adventure game like D&D has main quests in the form of “modules''. The game-master has the ability to move around the components of the story to match the player's choices. The player will have the option to do something completely different and ignore the quest. But whatever it may be, the system can take every opportunity to connect threads with the existing module.

aigan said:
Why bother making the choice if it doesn’t matter in the end?

Exactly.
Assume you succeed, and you achieve dynamic story. Some next gen evolution of the branching story concept maybe.
But then, if every choice has impact, it won't matter in the end regardless, and even less than before.
This is because at this point, a dynamic and unique showdown is the new expected norm.
It would be like procedural terrain. Ok but nothing exciting. Some expected background, but not a reason to play this and not that game.

I don't think this can be avoided. It will never compete Romeo and Juliet.

That comes back to the first thing I said about befriending NPCs and forming lasting friendships. It’s about creating stories you care about. And that is usually done by showing someone struggling and letting you help them, and then follow them in their personal growth. X-COM showed how that can be done even without deeper psychology systems. The story is connected to all the people you met along the way and they all have relationships with each other.

I’m interested in deconstructing what it is that creates engagement. Some of it is the themes I talked about. Many of those are things that are part of the human condition that will make people read/watch/play the same type of story hundreds of times because it helps them process how to cope with life. Things like carrot/stick, hope/fear, love/hate.

And then, you can continue on another adventure in the same world. Or hand the save to somebody else and see how they handle the same situation. Situations will naturally diverge. But you could still have scenarios that would initially be the same

But i also don't think that's what action games are good for.
I'm afraid most of this desire for good stories comes mainly from the fact that our gameplay became standardized and is no longer that interesting or exciting. This is accepted by both players and devs, so they expect progress on other issues such as combining a dynamic game with a static story.

However, i must admit i have built up my doubt regarding dynamic stories over decades on the belief that it's just impossible technically, so maybe i try to convince myself it is not that important anyway. If my belief would change, my mind would surely follow.

So i wonder, do you talk about general goals and visions, or do you say it can be done? If so, how? What data structures and algorithms could generate dynamic / adaptive story, and how far could this go in practice?

I have a plan for doing a proof of concept, and how to continue in several more steps. And a plan for how a AAA developer could eventually use that in their most cinematic games. Will write about it in more detail in future articles. Will also do step-by step mockups. The story-beat level is fairly easy. But my plan was to start with the season-arc level. The actual fun would not necessarily be part of the first prototypes. For that, there needs to be some layering with multiple conflicting conflicts, personal goals, secrets, and so on. That’s why I also write about these things, for explaining how a boring plot-outline text game is part of a grand plan.

@frob said:

aigan said:
I believe we can have both the sandbox gameplay together with the top-quality story.

If you have the budget to create the game you described, and you think you can get a profit on it (or are willing to lose the money) then absolutely go for it.

You're likely already familiar with what the biggest games do with a half billion dollars today. Opening it up more is certainly possible if you can afford it.

The traditional way is to use hand-written narrative. My suggestion is to make the story systemic. I believe I know how to do it just as good and wit a lot of work eventually better than the best story games produced today. But it wouldn't have to be super-high quality to be much better than the procedural quests used in some games today.

aigan said:
Character development and relationships are an important part of the type of game I would like to see.

I'm putting just some random responses here, for the sake of inspiration. I don't think i understand your vision, at least not yet, having not read your blog. It's not clear to me what ‘systemic’ means, for example. Maybe you can reduce your vision to some one liners, to give an overview up front.

I also lack a lot of recent advances regarding modern AAA games. There seems to be a lot about handling relationships, e.g. thinking of TLoU 2, but i have missed those games. Played CP 2077 at least for a while, which seemingly was about a protagonist with a dream of achieving something remarkable, so he will be remembered. That's certainly more about character development than i've seen from my major inspirations, basically games from the 80's up to 2010.

But i have thought about relationship too. To avoid a half billion dollars bill, i reduce it to just two characters. Protagonist and companion. And i want it to be a love story, or at least raising the expectation on that. It should happen in the players imagination for most, not so much in the game itself. Due to costs, mocap and cutscenes are no options for my budget.
So how should this work?
Mainly with facial expressions. I was drawing comics as a kid, and getting anatomy right is hard, but i was good with facial expressions. And i'm optimistic i can develop a system to do this with game characters, although it's a lot of work.
Another point is to reduce a need for speech - another thing i can't afford. To explain the lack of speech, i need a background story to fulfill the constraint. Planet Of The Apes is an example. Man lands on planet, meets a girl which is somehow uneducated and can't talk. Something like that.
The final point is to keep it shallow. There is too much action in the game, so the love story can't fully develop due to external events hindering the relationship to progress. Maybe some Indiana Jones movie would be an example of that.

That's very limited compared to AAA storytelling, but it can be fully dynamic. Technically, animation is not required. Facial expression and body language, that's all. But i have not really thought about the problem of detecting which emotional state should be caused from emergent situations.

aigan said:
Think of main, companion or side characters from popular story-driven adventures and RPGs. It's often characters that will keep your interest in a way that you want to meet them again in upcoming games.

Does not work for me at all. I saw you relate to HZD on your blog page a lot, skimming over and seeing screenshots.
I have tried it. I was very impressed from the cutscenes for the introduction. I thought ‘wow, that's animation at Disney level’. But the story was kinda low brow, clichees, predictable. So i 've focused mor on how silly the hairstyles are, than on the story. Finally when the game started, it was the classic problem: I feel thrown into a world where i can go anywhere, but i have not build up any motivation to make a choice. I expect more work than fun, and then i do not really come back but try another game.
Although modern games do this much batter than back in 2010, it does not work for me.
For similar reasons, almost all RPGs fail on me right at the start.

A game where it did work was the recent cat game. The introduction is great. You see your cat having relationship with friends, then due to accident the cat falls down and becomes disconnected from it's friends. Pretty simple, but characters are introduced well, and i have motivation to bring the cat back to its friends.
Well done, and more than good enough.

However, the conclusion for me here is: I can not list a single example of a game where character centric ‘live’ storytelling (cutscenes, talking NPCs) actually works for me.
But i can list many working examples of alternatives: Narrative FPS or Walking Simulator, picking up audio logs or notes, showing reprises, mostly about events from the past, avoiding the live events.
I think this is because those latter approaches spur imagination, while the former prevents this, and also NPCs even if recorded never feel alive in games.
Personally i tend to believe that maybe we just can't do storytelling like the movies do. We have our own ways, which are better no matter how big the budget is. But not sure, since my personal impression may be subjective and shared by a majority.

aigan said:
My goal would be something like Cyberpunk 2077, but the first step would be a text-only game with an abstract description of events and without any dialogue.

Why not both? Dynamic characters in an action adventure, but avoiding the static restrictions of static dialogue. I mean, for the last step as well.
Personally i see almost no other way, aside of some key scenes maybe where dialogue is required.

To me, the disliking of dialogue dates back even to early graphics adventures. I loved Zak McCracken, but in Monkey Island they talked too much. I was bored and never finished the game.

I realize just now that it's probably dialogue which breaks story telling in so many games for me.
Probably that's just me. But omitting it makes ideas about dynamics, freedom and emergent options so much easier.

aigan said:
There are plenty of sandbox games. Many of them have some measure of simulated physics. I think they can be a lot better, but it’s at least something. There hasn’t been a game yet that does the same thing with story. That is, doing story through systemic interactions.

A story sandbox - how on earth should this be possible?
I'll read on, but in case you don't, i really would love to hear about some examples…

aigan said:
As hinted by Watch Dogs Legion, you can talk to a person, and that person will be woven into the existing narrative.

Interesting. But i know only the first WD game, which was pretty good imo.

aigan said:
The tone can adapt to the player. If the player behaves silly, it could use story rules from action-comedy. If the player is careful, it can take more from a thriller. If the player is flirty, it could use rules from romance stories. If the player is a murder-hobo, it will take rules from outlaw stories.

How? I mean, we could raise the budget to 5 billions, making content of 10 AAA games, then tailoring to the players style simply by selecting content.
But this really goes too far?
I'd rather hope on procedural content generation in this case, which now becomes possible using generative AI. But i'm not optimistic regarding quality. I expect boredom worse than ever, artificial and strange experiences, and i even hope such attempts will fail.

aigan said:
For every type of story theme, there are variations that can be mapped to the player stalling, shortcutting, succeeding or failing.

Still sounds like exponential growth of production efforts.

But it depends on how you tell the story eventually.

Personally i have a vague idea which maybe is similar. It's about generating cool action movie situations around the player. Some background task could analyze the gameplay, preparing things like an ambush, an helicopter coming buy so the player can jump on it, or adjusting traffic during a car chase so pursuers crash into oncoming traffic, etc.
That's difficult, but surely doable. And if action movies tell a story, then this would count as story too.

So what do you mean? Physical events and spectacle like in my example, or character centric events affecting relationships, dialogue, drama, …true story with a deeper meaning?

aigan said:
A computer can check the constraints against thousands of variations finding a better match than any tabletop game master would.

Yes, but that's not the problem. The problem is how to generate the content?
If it's generated offline so static, you simply need too much content. It might be better to make only a fraction of content, targeting a specific player type. Lower costs, but a happy niche of players served well.

Contrary, a general game, which can adapt to any player type, has costs (and also storage requirements) multiple orders of magnitudes higher.
And assuming it works, each player will only select 1% of the content. He might not even notice his freedom of choice, so to the player there might be little difference between both those games.

That's surely a disillusioned way to look at it, but you can't ignore the 100 : 1 ratio of the risk involved.

aigan said:
Have you experienced the DualSense controller? You can actually feel the grass under your hands in Horizon Forbidden West.

Ha, no thanks. No trackball or proper touchpad for mouselook, no immersion.
Feeling rumbles is not immersive, but rather just annoying imo.
Controlling your view with precision and without lag is simply required for a immersive experience.

I really like the box of those consoles. HW and SW made for games. I'm envy.

But until they realize that mouselook is the biggest innovation for games since Pong, i'll pass. ; )

Console gamer: I can hear the gunshots out of my controller! So immersive!
PC player: hahaha, lol, rofl.

I'm just from the latter camp. But i don't believe in a 'glorious PC master race', and i rather think our HW totally sucks.

aigan said:
For example, when you can’t move because you just triggered a cut scene or a long animation.

Yeah, that's a no go. Never take camera cotnrolls or movement away from the player, and never interrupt gameplay.
Same page here. But ofc. ther are always good reasons to break such rules.

aigan said:
If you need an NPC alive, you can have the player still be able to move the gun but never point it directly at the NPC.

That's a really good idea. : )

aigan said:
I wrote an article about a control scheme for embodied manipulation.

ugh - ‘console player trying hard to self assure his controller does not suck’ :D
Just joking, did not read. Too many button combos maybe. I always fail to remember more than one button. Atari was good. : )

Personally i'm working on robotic ragdolls. It will work, look realistic, and they can do many actions.
Intuitive action controls seem impossible. And i hate to use many buttons on PC even more than gamepads.
For now i believe the solution might be something like mouse pointer and pop up menus. :O
So you see i don't have much clue about controls, and lack the competence for feedback. :D

aigan said:
For example the ability to adjust the landing (slightly) even in mid air.

I might have an idea to unite the laws of physics with the requirements of game controls.
But my ragdolls are not yet good enough to try it out…

aigan said:
I have seen a lot of quests or arena designs where you can’t exit the scene, not because the door is locked but because you were supposed to talk to a person or take an object or kill an enemy first.

Yeah, that's bad. But i'm sure every game designer is fully aware about those issues, and has the same goal of addressing them. Likely they can't fix every little imperfection, simply due to reasonable budget limits and deadlines. It's forgivable.

But it becomes a real problem in case the player simply does not notice what he must do, and he becomes stuck because of that. Happens pretty often to me. Then i feel dumb. And the game also feels dumb - sitting there, doing nothing, not realizing i need help.

Some games do realize, and react by some NPC repeatedly telling me what i should do.
Which fells dumb as well. What a dilemma. :D

aigan said:
I probably see more of this since I usually try to avoid fights. Many game designers may not have even considered that the player would try to run back the way they came at the first sign of trouble.

I can't do that. I never know from where i came from. Every direction looks the same. I'm constantly lost.

I often feel like game designers also forget to consider that some people have problems with orientation.

aigan said:
This is exactly the type of thing I’m referring to. There are so many times then the avatar could clearly just walk right through the giant hole in the wall, but aren't allowed to. Or climbed over the tiny heap of rubble.

Sure, but that's also forgivable imo.

But i don't forgive a total lack of options.
For example, i was playing CP 2077. Those guys have no idea about shooters. It's no fun. The enemies don't even move. They're just bullet sponges. Initially i could not figure out that i have to go to the damn inventory to equip a weapon, so i did just run which worked pretty well.
However, at some point in the game, i was at some shopping mall. Many enemies.
It was obvious what to do. Climb up, and sneak on top over the enemies. The climb would be easy, and they could not see me. I could prevent the boring gunfight.

But in this game you can not climb at all. And although i knew this, i could not longer accept the limitation. Rage quit, and never coming back.

Back then the game was still very buggy. So after a lot of frustration, i just had enough. It felt like the end of the AAA industry to me, and since then, i have not really played a AAA game.

Initially i was excited. Exploring the city without any gameplay was awesome and mesmerizing. They did it. A big scifi city in a game, and i was sure this is the game of the decade. But then the game itself was such a disappointment.

Now i play simple indie retro shooters, like Amis Evil, Prodeus, or Dusk, etc. Those games work really well. I'm surprised how much fun they are. They feel perfect.
Initially i as dreaming about big games, where you can do anything. Total freedom. Infinity.
But times teaches me a lesson about how important restrictions are, constraints, and a lack of choice.
It's like my expectations and ambitions are shrinking with age, contradicting that after decades of work, i feel close to being able to realize many of my former dreams and convictions about richer options.

That's somewhat confusing. But maybe it's just temporary…

aigan said:
Can you give me any example that couldn’t be handled by the physical movement of my embodied manipulation control mapping? The thing I want to get away from is the “use” verb, just because it can lead to surprising results when the player has another idea of how to use an item, or when it differs because of the exact cursor position changed directly before the activation. For example when you use X both for talking and jumping.

My head already hurts form reading so much small text.
But assuming you have some perfect control scheme so the player could do anything,
the problem is not to find some exception that would still not work. It's not about finding bugs.

The primary question is simply: WHAT can a player do in a future game, which he could not do in former games already?
The control scheme does not answer this? Currently we have just combat, puzzles, and finding the way to the exit. What more actions could we want to do in a game? Getting a different hairstyle or more strength points does not count. That's no actions.

But you would be first with providing me some answer to this question, so i'll try to answer yours.

Imagine there is a stack of boxes on a table, and a big empty box on the ground.
Your task is to move all small boxes into the big box, but you have to rotate them so they fit, like Tetris.
Yow would you do this?
There is no way to control 3D rotation intuitively at all. Impossible. And precise placement of objects also is pretty cumbersome already, even if orientation is already right.

In 3D modeling software we use gizmos for that, which requires some experience, but nowadays most people are used to it already.
So maybe my thinking towards a mouse cursor is not all that bad.
I have implemented this to interact with physics objects. I can lift hem and drag them around, rotate them, push

aigan said:
That was a description of NetHack. It was done in the 80s. And I guess the same holds true for Minecraft and several other systemic games. Just not for any story games. Yet.

against other objects, all with a usual 3D gizmo. I did this for debugging, but it's actually fun, i tell you.
But i'm not really serious about it. ; )

Often heard about NetHack, but i never played it. Might try.

Personally i've made a mobile puzzle game about interactive blocks and predictable NPCs. Movement like Sokoban, but i 've added any puzzle mechanics you have ever seen in such games, some new ones, multiple floors so it became 3D, etc. It even had crafting, and could do everything Minecraft has done much later. There were many potential interactions of multiple objects, it was rich in options.
It did not make me rich, but everybody really liked it. They felt creative and smart while playing it. It made them smile.
There always was just one way to solve a puzzle, but it did not feel like that. It felt much more like creative exploration.

This sticks with me. Rich, but not overwhelming. You did not need to memorize and master complex mechanics to build up skill incrementally, since every level was just a entirely different problem.
Surely something i'd like to bring to a 3D action game.

NetHack looks somehow similar, although i guess it's turn based.

I think more than freedom, players want a sense of purpose. In real life, many people feel like they have no meaningful purpose whatsoever. There are some who find their purpose by reading, listening to a lecture or sermon, experiencing the arts, producing creative works to share with others, etc. But games are unique in that they can provide instant purpose even if it is purely fantasy.

In meta-gaming, speedrunners find purpose in outsmarting the designers and developers by playing the game in a way that was not intended.

JoeJ said:

I don't think i understand your vision, at least not yet, having not read your blog. It's not clear to me what ‘systemic’ means, for example. Maybe you can reduce your vision to some one liners, to give an overview up front.

Most games have systemic combat. Many games have systemic environment. Take Zelda BotW or Baldurs Gate 3 for example. Many base building games have systemic character behavior, such as Rimworld, and Dwarf Fortress. Many RPGs have a few systems connected to the story such as faction reputation or relationship status.

No game has extended to use systems for the main story comparable to hand-written main quests. And that is because you can't do that as long as you use hand-written dialogue, branching story or custom events. There has to be developed systems for everything included.

On the top level you would pick a conflict such as envy, greed, health, poverty, and so on. Using that constraint, pick more things on lower levels, until you get down to the inciting incident. And keeping everything not observed open for change in order to adapt to the player's actions.

@Hypnotr0n
Yes. A sense of purpose is important.

Something to care about. I'm interested in what it is that makes us care about stories, and use that to generate them dynamically.

They would, by definition, be formulaic. 🙂 But that wouldn't differ from many other stories we enjoy. And there are so many “tricks” we can use. (insert sad puppy)

JoeJ said:

I also lack a lot of recent advances regarding modern AAA games. There seems to be a lot about handling relationships, e.g. thinking of TLoU 2, but i have missed those games. Played CP 2077 at least for a while, which seemingly was about a protagonist with a dream of achieving something remarkable, so he will be remembered. That's certainly more about character development than i've seen from my major inspirations, basically games from the 80's up to 2010.

What are your favorite games? What are your gamer motivation profile? People play for partly different reasons. As you may tell, I care about immersion as well as mastery.

But i have thought about relationship too. To avoid a half billion dollars bill, i reduce it to just two characters. Protagonist and companion. And i want it to be a love story, or at least raising the expectation on that. It should happen in the players imagination for most, not so much in the game itself. Due to costs, mocap and cutscenes are no options for my budget.
So how should this work?
Mainly with facial expressions. I was drawing comics as a kid, and getting anatomy right is hard, but i was good with facial expressions. And i'm optimistic i can develop a system to do this with game characters, although it's a lot of work.
Another point is to reduce a need for speech - another thing i can't afford. To explain the lack of speech, i need a background story to fulfill the constraint. Planet Of The Apes is an example. Man lands on planet, meets a girl which is somehow uneducated and can't talk. Something like that.
The final point is to keep it shallow. There is too much action in the game, so the love story can't fully develop due to external events hindering the relationship to progress. Maybe some Indiana Jones movie would be an example of that.

That's very limited compared to AAA storytelling, but it can be fully dynamic. Technically, animation is not required. Facial expression and body language, that's all. But i have not really thought about the problem of detecting which emotional state should be caused from emergent situations.

I started out mostly with C64 games. It's amazing how much life can be expressed by characters just a few pixels high, such as in Lode runner or Choplifter. More recently, I'm inspired by the expressiveness of the systemic behavior in Prison Architect. You can see when they are happy, calm, angry or frustrated, mostly from their movement patterns.

There doesn't have to be an excuse for being silent. The relationship in Ico is powerful. But if you want a reason, make them speak different languages.

Reactivity is so powerful in games. Simple reactions to the environment and the players actions. That's why NPCs look back at you.

For my vision of a systemic story game, I would start simple and use icons or other abstractions for things not implemented. I considered starting out with graphics similar to Dwarf Fortress. But then I realized that I can do even better and not use graphics at all. But the same type of system can be used in any type of story-based game. But the higher graphical fidelity, the more systems would have to be implemented, such as facial expression, voice generation and so on.

As mentioned, the first versions would abstract dialogue with a description of the exchanges. An alternative to abstracted descriptions is the things many RPGs do when they sort of repeat the same line every time you do the same thing. Then that is done, the actual words lose meaning and the whole line becomes an icon for what it represents. I would prefer that to be replaced with a system to the point that the character wouldn't say the same thing every time. The 100th time you walk up to the same character to initiate a conversation, it would be more like “yea?” rather than that long line you wrote for the first time.