Welcome, Guest. Please login or register.

Login with username, password and session length

 
Advanced search

1411490 Posts in 69371 Topics- by 58428 Members - Latest Member: shelton786

April 25, 2024, 12:37:50 AM

Need hosting? Check out Digital Ocean
(more details in this thread)
TIGSource ForumsCommunityTownhallForum IssuesArchived subforums (read only)CreativeWritingprocedurally generated stories
Pages: [1] 2 3 ... 10
Print
Author Topic: procedurally generated stories  (Read 25879 times)
Graham-
Level 10
*****


ftw


View Profile
« on: June 22, 2012, 09:39:14 PM »

EDIT:

This thread has turned mostly into me talking about my game. I'll gladly welcome any ideas that have something to do with what I'm talking about. The focus is still generated stories.


------

I've wanted to generate stories for a long time. It's kind of my calling in a way. I have a lot of ideas about how to do it. I mean, I'm pretty comfortable with the idea. Anyway... does anyone else think about this? I like the subject so I'll just talk about whatever.

Mmm. I spent a lot of time thinking about how the player can express themselves in a meaningful way. I can tell you that my current designs have no text. I mean I'm planning to have no text in the game. The characters will have a wide range of emotion, which they will be able to express clearly, through animation/sound and behaviour (of which I am most proud).

The player will be able to interact with other characters in meaningful ways. There will be no dialogue options (kind-of obvious). All the interaction will happen through the game's core mechanics. And... the freedom/flexibility of the player's and other characters' expressions will be comparable to that of a real language. It won't be as powerful, obviously, but the idea that you're learning to speak a new language won't be a secret. Everyone will see it that way, you just won't at the beginning. There will be a kind of a "holy shit..." moment part-way through when the player understands what's happening.

I've spent some time studying language construction. I also have some experience designing software that will translate one language into another, accounting for context and other subtle things, and tools that will help a person learn to speak a new language. So I'm kind of comfortable there too.

I spend a lot of time deconstructing stories that I like, looking for their technical constituents, their pieces. It's like creating an ingredient list from tasting a finished dish, but I'm doing it with movies and things like that.

I don't have a lot of experience with animation but I know enough about acting/directing, story telling, and stuff that I think I'll create expressive characters without hurting myself. The characters will animate procedurally to match the generation of their opinions/behaviours. I like procedural stuff....

I like AI a lot. It's probably my biggest strength.

I'm kind of just talking about myself....

Anything interesting about these kind of things I'll reply to. You know, just thinking.
« Last Edit: June 24, 2012, 02:50:24 PM by toast_trip » Logged
moi
Level 10
*****


DILF SANTA


View Profile WWW
« Reply #1 on: June 22, 2012, 10:08:49 PM »

I've wanted to generate stories for a long time. It's kind of my calling in a way. I have a lot of ideas about how to do it. I mean, I'm pretty comfortable with the idea. Anyway... does anyone else think about this? I like the subject so I'll just talk about whatever.

Mmm. I spent a lot of time thinking about how the player can express themselves in a meaningful way. I can tell you that my current designs have no text. I mean I'm planning to have no text in the game. The characters will have a wide range of emotion, which they will be able to express clearly, through animation/sound and behaviour (of which I am most proud).

The player will be able to interact with other characters in meaningful ways. There will be no dialogue options (kind-of obvious). All the interaction will happen through the game's core mechanics. And... the freedom/flexibility of the player's and other characters' expressions will be comparable to that of a real language. It won't be as powerful, obviously, but the idea that you're learning to speak a new language won't be a secret. Everyone will see it that way, you just won't at the beginning. There will be a kind of "holy shit..." moment part-way through when the player understands what's happening.

I've spent some time studying language construction. I also have some experiencing designing software that will translate one language into another, accounting for context and other subtle things, and tools that will help a person learn to speak a new language. So I'm kind of comfortable there too.

I spend a lot of time deconstructing stories that I like, looking for their technical constituents, their pieces. It's like creating an ingredient list from tasting a finished dish, but I'm doing it with movies and things like that.

I don't have a lot of experience with animation but I know enough about acting/directing, story telling, and stuff that I think I'll create expressive characters without hurting myself. The characters will animate procedurally to match the generation of their opinions/behaviours. I like procedural stuff....

I like AI a lot. It's probably my biggest strength.

I'm kind of just talking about myself....

Anything interesting about these kind of things I'll reply to. You know, just thinking.

Logged

subsystems   subsystems   subsystems
1982
Level 8
***



View Profile
« Reply #2 on: June 23, 2012, 12:34:04 AM »

Do it.

Nice to see some interesting thinking and clever approach in this sub-forum. I am little bit afraid thou that this would be new Sims, but then you said something about learning a new language. Game focusing on that would be quite interesting. Then again we are not actually talking about stories so much but player-NPC interaction mechanisms.
Logged

Graham-
Level 10
*****


ftw


View Profile
« Reply #3 on: June 23, 2012, 01:11:14 AM »

Nah, not like Sims. It's more like Final Fantasy, Zelda, Prince of Persia, some platformer. You know how every "make a choice" RPG is 1-dimensional, or a series of 1-offs? Or you have Skyrim which gives a lot more flexibility but doesn't have a structured story? I'm thinking about something that gives meaningful expression in your play, ala Mario (personal opinion), but has the range of something like... reality, the structured adventure of an epic like Mass Effect, but plays deeply and exploratively like Deus Ex mixed with Dark Souls, and of course touches your soul like Journey, name drop, name drop. Though it's not really like any of those things, it's its own thing... obviously.

You grow to interact more fluidly with the people around you as you progress. There's this Pokemon/Monster-Hunter vibe where you encounter, study and capture creatures, but unlike those games you have to form relationships with them. The creatures are more like a half-way point between humans and creatures; they're still creatures but more directly relatable the way a person is.

The growth of your relationships, with the creatures and other people, are necessary for progressing through the game. There will be clear barriers in-front of which you will need to communicate effectively enough with others, or understand others enough, or search for the right partners before being able to proceed. It's kind of like getting "the hookshot to cross the gap," but you're getting "the developed inter-personal skill-set to climb the mountain."

The game will start off a lot more like a really straight-forward action/adventure, but there are beings and plant-life everywhere going about their business in independent, though interrelated ways. The idea that you belong to this community, or the abstract idea of a community, comes to you slowly. You spend the game playing it like a normal game, but also building up these relationships, learning about the world, learning your own abilities, growing, making choices, and learning how to communicate.

The "communication" is more like a platform from which to deliver story and more complex mechanics. Partner characters make an excellent vehicle for giving comprehensive player feedback. Whatever "language" they speak will be tuned to the mechanics of the game, so interacting with them will be like speaking to buddies on a headset in a multiplayer death-match, except there's no bullshit, the story is pre-built (i.e. good, because effort went into it), and everything blends with what's happening on the screen.

A lot of the time the value in having a relationship is just to have the relationship. Like in real life, your friends in the game give you constant feedback on what's going on around them, including how they perceive you and your actions, filtered through their own personal world view and current state of mind. They're always teaching you how to play, driving you in new directions, stuff like that. They're like a game-long tutorial in a way. That way I can make the game as deep as I want, then expect players to learn it, because learning from an attentive person is much more interesting than tutorials, text, and figuring it out yourself. The tutorial is the game and vice-versa.

The diversity of the characters' behaviours are the building blocks for generating a story. The world "behaves" too, the character's act on the world, stuff like that. A lot of "inanimate" objects have subtle personalities, the way trees can "have personalities" if you look at them in the right way, though they express themselves really slowly. In the game the inanimates would be a little more timely.

There's a basic "structure" for the game's plot. Individual players would experience it differently, certainly, and there would be a lot of replay and extended story and stuff like that, but there would also be core elements that every player goes through. The world goes through general transitions, characters have goals, certain things will definitely happen. There will be an invisible "story director," like the "action director" in Left 4 Dead, except you know, not platonic, that ensures basic structure is held. The result is a story-line that fills in a probability space that I have a lot control over and can reason about. I'll build it as carefully as I would a movie, I just have - what would you call it - "functional primitives" (?) instead of dialogue, blocking, and camera angles.

You could say my goal is to make others feel the way I felt when I played FF7 when I was ... 11(?), except they have far more control over the way things develop. It will feel like you're being guided along, but you still get to express yourself, be yourself. It's literally you in an adventure. It's hard to state in simple terms.

Also I want it to be deep like Go. Go is like a conversation if you rock at it. Sports fans see a ton of stuff in their favourite sport. I'm also influenced by Poker, fighting games, shooters, hard-core action.... You know, why not.





« Last Edit: June 23, 2012, 01:55:15 AM by toast_trip » Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #4 on: June 23, 2012, 08:40:13 PM »

Anyone who thinks about procedural stuff, come at me. I'm ready for you.
Logged
mirosurabu
Level 4
****


View Profile
« Reply #5 on: June 24, 2012, 02:17:06 AM »

I don't think about procedural generation because it's merely a way to create content (the other way being by hand). It says very little about games.

I am, however, very interested in believable, open-ended and complex worlds with believable role for players to fill in. Any such world would necessarily result in playthroughs that would be interesting to retell, the so-called "emergent narrative".

I wonder what kinds of actions would players use to communicate with NPC's? And what are they going to be talking about? Like, is it going to be proper communication, or just a puzzle masquerading as communication?
Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #6 on: June 24, 2012, 03:39:16 AM »

Haha. Well, generating content is not a substitution for doing things manually. First you have to do enough stuff by hand so you can teach a computer to generate things just like it in greater variety. Turning to generating content as a way to avoid manual labor will only result in sub-par content. I'm not talking about a short-cut.

I think generation is the future of games. It's hard to say why easily. We certainly have some way to go before we're generating voice acting at movie quality level. We've still got a-ways to go before we're doing it consistently manually....

Yes, the emergent narrative. That's what I'm going for, the mythical holy grail.

Yeah, ok, the actions thing. That's hard to put into words at this point.... hmm.

Think of it this way. When an expert plays Starcraft, or Go, or Halo, or whatever, you can get a very good idea of what's going on in this player's head if you're an expert too. You can get a whole lot if you know the guy. The little ticks in his behaviour will communicate distinct personality traits and feelings. Even Mario communicates a lot. I love Mario. The key issues with these kinds of communications are thus:
  1. You need to know the game really well to get it.
  2. The player generally needs to know the game really well before he expresses himself through it in meaningful ways.
  3. You probably need to know the player personally to link behaviour to emotions in the most interesting ways.
  4. What you do glean from the player's play is not necessarily information that would allow you to "converse" with him in-game, through game-play, in the most relevant ways to what's going on in the game.
  5. Even if the above things aren't an issue, observing the mental state of the player actually takes a lot of careful observation on your part. I've watched a lot of pro Starcraft. I always need commentators or I don't know what's going on. The game was built to be played, not to communicate. Viewing the game as a collaborator/opponent is very different than viewing it as a spectator.

Now think about soccer. People fucking love soccer. I'm learning to love it. A lot of shit goes down on the field. What you see is just a bunch of dudes animating a mile away on a green background with a high-res texture. But people are attached to their teams like a religion. The humanity gets communicated.

Think about Guitar Hero. What if it actually taught you to play guitar, had an escalating free-form campaign - in which you expressed yourself increasingly - and had cooperative play that leveraged these facts? A lot is said between professional musicians in a jam. It's all emotional, sure. But in a game it can be a lot more. Guitar playing is a small sub-set of games.

People love Machinima. Red vs Blue is a small edition to an evocative engine.

It will be proper communication. It won't be a puzzle masquerading as communication. That's like Monkey Island, or bits of Earthbound - which does a surprisingly good job of making you believe otherwise.

In the game you will talk about the things that are most relevant to the game. In Fallout 3 for example, conversation is always about nothing. Gameplay is about choosing destinations, preparing for a journey, fighting, customizing, upgrading, buying, exploring, yada-yada. The conversation is about "the story." The game would have been a lot cooler if both the mechanics and communications we're focused on the same things. Really they're just two separate things that intersect in controlled ways, providing contrast and assist.

I'll bring in Team Ico to the explanation. In Ico you drag Yorda around, you call out to her, she comes to you, you hold her hand to make leaps, you help her up ledges. In Shadow of the Collosus you interact with the Colossi, you interact with the horse; both interactions are very subtle. In both games it's more like you're making decisions based on the mechanics, and then the world is reacting to you as if you're communicating with it, re-enforcing the meaning of your actions.  After playing for a while you start to think of your actions in more narrative-like terms. Your actual decisions are controlled by the rules. The story just guides your attention to where it should be. So when playing it's like you're making narrative decisions, and the world is responding with character, even though you're not really. It's a clever illusion. (I love those games btw).

Ok. So. Imagine if the characters had a wider range of emotion. Say Ico could interact with Yorda in 3 ways, instead of just one. Now say the meaning of each interaction is partially influenced by context. For simplicity, say 1 interaction is the call, as it is in the game; the second is an "attack," like an emotional berating, yelling at her for making a mistake, punishing her, being a dick; and the third is a conversational, like chatting, maybe consoling, mostly sharing your current feelings.

Ok. Now say each interaction existed on a scale, just 1-dimensional. The scale represents the intensity of that interaction. Let's say there are only 3 notches. So for example, there are 3 degrees of calling to her, each indicating how urgent or demanding you are being. Maybe you can't choose which intensity you use; it's entirely dependent on your "state". As you play the game the system is measuring what you're doing. It assesses if you're panicking, if you're ruminating on a puzzle, if you're goofing off, if you're in the zone (i.e. succeeding), if you're playing slickly, if you're playing smartly but executing sloppily, stuff like that. Ico's state of mind shifts around depending on how you've been playing, and based on what's going on in the scripted story.

Okay. So let's say you're stuck on a puzzle and frustrated. This is a hard thing to measure, but let's just assume that we do it well. Ico demonstrates his feelings subtly in the way he moves (procedural generation!). The biggest key to his state (of mind) is when he communicates with Yorda. Maybe you do the "conversational" interaction. This queries Yorda for help. She behaves in a way that indicates how she feels about her current situation. Maybe she's making nonsense noises - as she does - as she walks around with you, gesturing at things (procedural!). If she's in a good mood she'll be more helpful and springy, and she'll offer more insight (keeping character, obviously). Note, the insight would be very, very subtle. The puzzles in Ico are very delicate. They would have to re-designed to accommodate this stuff. If Ico is frustrated Yorda will be more helpful; Ico will show his frustration through his tone. However Yorda might slowly lose trust in him as a result (only slightly). Each interaction is a window into both characters' minds.

Now, as you play, Yorda is transitioning constantly. Since you're always interacting with her you're consistently building this idea of how her mental state is developing. Understanding her is like looking for ships on a radar screen. You get a blip every so often then extrapolate movement. From this you can determine how your behaviour affects her - also, you're constantly giving off your own cues in your animation/sound and mechanical execution (the things you can't control).

Yorda could then be leveraged by the devs (me) to smooth out the rough edges in the puzzle design. Yorda is already supposed to be your "key" to escape; she literally opens locked doors. That stuff is all scripted. With procedurallity (??) she'd literally become a necessary component of your journey. You would learn the link between your play, Ico's changing state of mind, and Yorda's reactions. Slowly the barrier between what you'd like to say and what you would say would disappear. Yorda too would be able to communicate increasingly complex things.

I don't know how well that communicates the idea.... I've had to work on this a lot already... I mean near full-time for years. There's a lot of stuff to it. Also note, Ico isn't really a good game to justify the work-load of the procedural stuff and the communication. A game has to be designed around those features from the beginning. That's what I'm doing, sort-of.

So yes, real proper communication.

I'll use one more analogy. I love fighting games, but I suck at them. It takes too long to learn everything. RPGs however I rule at. I don't grind, get treasure, and I'm careful about side-quests. I think some (lots of?) people are like that. Anyway, I love repeating a boss over and over, slowly tweaking my strategy until I own his ass by a slim margin. I get super pissed when the ceiling in the game's mechanics is low, meaning I can't go a huge distance just on my own ingenuity. Most people aren't like that. Most people are confused by stats, or they lack an understanding of basic fundamentals in-order to see the strategy in making good decisions. I didn't know grinding was a thing until I grew up. I mean, if you make poor decisions about how to level your characters, or where to put your resources, you can mess yourself up for way later on. Like you need to know your stuff. Anyway, I suck at fighting games.

The depth of a lot of games are out of reach without the proper mindset because there are too many things you have to learn to get peak engagement. I tend to love RPGs more than the average gamer because of how I think. It took years before I started to love shooters, because I have issues with staying focused on the necessary elements when navigating in 1st person. Timings and combos make fighting games elusive to me. Starcraft, I'm sometimes pretty good at, and other times terrible; it depends on my mental state. Either I focus enough to tend to the fundamentals or I don't. I don't have enough practice to make them automatic.

Ok, so here's the thing. The difference between a skilled, deep-playing, fighting game player and a novice is large. The barriers that one needs to overcome to cross that gap are often not spelled out by the game. This is true of Starcraft, high-level Diablo, pretty much everything. That's why casual is hot; games are hard. Anyway, if a game experience could guarantee that a player makes it to the deep-playing level, say by completing the campaign, then you'd have a game that teaches players to communicate in that game's "language". (Games never do this. Some sort-of do, but not without serious bumps. Completing the campaign usually means completing the story-line.) If that same game had you communicating with other characters in order to proceed, in a way that was mechanically and narrative-ly consistent, internally and with each other, then that game would teach you to communicate in general.

Imagine the Ico example, but by the end of the game you're inputting commands as complex as you would in Tekken - automatically without thinking, intuitively. The result is an expression that could be appreciated by an un-initiated viewer, and would be understood and reciprocated by in-game characters and other players.

The game is in a way a training ground for learning to express yourself. The expression and the gameplay are linked, so the expression and the mechanical meaning are mirrors of one another (like how an experimental jump in Mario looks experimental, even to someone who's never played Mario before).

Mmm. I guess there's a lot of pieces to it.
« Last Edit: June 24, 2012, 06:30:17 AM by toast_trip » Logged
mirosurabu
Level 4
****


View Profile
« Reply #7 on: June 24, 2012, 09:12:18 AM »

So the game tries to mind-read you so that other NPC's can see what you, as a player, not as an avatar, feel?
Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #8 on: June 24, 2012, 09:37:49 AM »

Mmm. It's more like how you feel filtered through the lens of the avatar. Assuming there's a single playable character, you (the player) would be expressing yourself through him. NPCs would react to some blend of your personality mixed with his.

Think about Mario. He's pretty Mario no matter what. If you've never played before, maybe your personality doesn't shine through that much. As you play, Mario becomes a closer reflection of your ideas, what you want. Obviously you hit some boundaries early. It's very hard to be un-Mario.

In Skyrim you slowly embody your ideal person. You don't really animate/look different - you have different armor; you play different.

Ok. So it's like a combination of the two. At the beginning you are role-playing this character. Maybe it feels like the beginning of FF7. The story is baked but you have this control over exploration and pacing. Over time you gain more control over how you develop and what you can do. I think at some point you become exactly you, or something very close to you within the context of the game. You're super distinct like a leveled up Skyrim char, but with the freedom of expression you feel in Minecraft, except of course your decisions are reflected through your person instead of your creations.

In simpler terms you transition from what feels like a linear role-playing experience to something more open ended and expressive. Your character is always bound by his personality, i.e. shaped by his past decisions, but is always growing so can eventually become whatever you want, as long as you stick with it. For example, you can never behave in a way that breaks with your character. It is impossible to create an inconsistency. If you want to be someone new you have to create the long history for that person the be the natural conclusion of your actions. So there's always this filter between what the player feels and the character expresses. You're never not role-playing.

It's kind of like waking up in someone else's mind and slowly gaining control over them. They slowly become you. Except of course you're in Never-Never Land.

And of course the game is built in a way so that you pursue activities that let you be a person that you want to be. Nobody wants to be straight themselves, not usually, unless they find an awesome version of themselves.

In a sense, yes, the NPCs "mind read" you.

You also have a great deal of choice. Like life, some things you can choose to express or not, and some things trump your choice.

Does that make sense?
« Last Edit: June 24, 2012, 10:02:59 AM by toast_trip » Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #9 on: June 24, 2012, 02:38:19 PM »

Hahah, I know that talk. I love Jon Blow. Very, very wise designer.

He, uh, has lower hopes for the role of story in games. He sees the
potential of games in other areas. He gave a talk with the creator of
Miegakure at some event in the last few years. In that he described
things like orthogonality in design. The talk was called, "Truth in
Game Design" or something like that. Or maybe that was just the point
of it.

He contributed the most important phrase that he has given me in that
talk. It was something like, "games exist to deliver a truth." I used
to think, "games deliver an experience." I prefer his perspective.

He likes to talk about games as a vehicle for discovery, for both the
creator and the player. He says you should find something that is
true, represent that through rules in your game, then explore that
truth and see where it takes you. Game development is like a
conversation, not a plan-and-execution.

Some of that stuff isn't original... okay, whatever. No one is
original, but I liked it a lot. It did feel fresh, and I just like Jon
Blow.

Anyway, assuming you love Jon Blow too, I'll leverage that fact for a
counter-attack. There is nothing about a story that can't be
understood and deconstructed into its constituent parts. When a
screenwriter sits down and loses his(/her) mind for 6 months straight
over a script, he(/she) is going through some, definable, mental
process. He forms relationships between ideas in his head. Maybe he
goes out and gets some new experiences. At some point he expresses
them on paper, according to the restrictions of his contract (or his
initial impulse or plan or whatever), his understanding of good movie
form, and his understanding of the perceptions of others. Each
construction is a neurological event transcribed. One idea connects
with another, and another and so on. The writer screens these ideas
using some formula, then ends up with a script.

A story is a real thing. Ideas are kind-of non-tangibles; they are
very tough to define; they are context-sensitive. It's hard to explain
them to a computer. I mean the Microsoft Office paper clip avatar is a
testament to the inability of computers to understand what I want. Or
maybe that's just Microsoft.... But they do exist. You kind of have to
define an idea using other ideas, then those ideas with other ideas,
and around and around you go. But that's the way humans work, and
humans aren't that complicated... sort-of. We're 2% genetically
different than a chimpanzee? How complicated is a chimpanzee... oh
boy....

Games are systems. You create a system, explore it, then arrive at
experiences you could not have originally projected. Ok. Let's assume
that I can take any idea and transform it into its fundamental
components. That just takes some effort and experimentation. Then I
can project those parts into the game. If the rules from the
experience transfer, then the experience I had in mind is created.

The amount of work to bring an idea over (into the game) is
substantial. Creating an experience in a computer is an order of
magnitude more complex than creating it manually; that's game
development. But, once you do bring it over, the value begins to
multiply. Mr. Blow demands that all game development should revolve
around the uncovering of truth that extends from the properties of the
game you're developing. I agree with the intention of that
demand. Stories certainly contain some kind of truth. Ideas can be
measured. We have to generate stories. We don't even have a choice.

The idea of a non generatable story is just one of those things that
we say to ourselves because generating stories is so hard it puts us
face-to-face with the infancy of our collective understanding of
computers and design theory. Though that's definitely not a knock on
anything you said. We're just really bad at expressing our ideas, even
to each other, let alone to a computer.  Creating great stories
manually is a high-power job. Generating them is like a free ticket to
insanity. Who wants that?

I'm not claiming I'll generate things that will rival great books, and
replicate regularily, but I think I can top the average man-made video
game story. Throwing in the high degree of player input that comes
with any great game, I'll get an even better story. Then I'll steal
the credit for all the heavy lifting done by the players until I
overdose on cocaine and stupidity.

Story and plot are different, yes. You say player-generated plot is
the way to go. Ok, maybe this is me misunderstanding
terms.... Consider FF7. I love that game. Any FF game before 13 other
than 11, and maybe not 1 will do. I played FF7 at literally the
perfect time in my life, like a lot of kids, so just assume that it
was incredible. In FF7 I felt a strong degree of control over what was
happening. The plot was largely out of my hands. There were a couple
of points where I could swing events, like controlling who I dated at
the Golden Saucer, whether I got picked by the gangster boss to be his
sex service, when/if I got Yuffie etc. Those things mattered, a lot,
but they weren't the core of the experience. The key plot points were
out of my hands. All the major things happen in a pre-scripted order,
with cutscenes and awesome things. In Skyrim for example, I control
the plot massively, sort-of.... Well, at least I get to control the
order of things. That game doesn't really have a plot to me. It was
more like a dungeon crawl in an art gallery. In FF7 I felt I had more
control over the story than I did in Skyrim.

Earthbound, Secret of Mana, Skies of Arcadia, FF7, they're all the
same; the plot happened, I watched. In-between major events I did two
things: explore and/or prepare-for/engage-in battle. The exploration
was my opportunity to control the pace of the game. In FF7 I chose how
I navigated the environment. As I walked to the edge of one screen, or
the invisible boundary of an area, the camera angle shifted, to
another, often well chosen, shot of me. I was directing the pace of
the cinematic. Maybe I was in a hurry, or taking-in the scenery, or
scouring for treasure and enterable rooms. In town I could choose who
I talked to, how long I wandered around for. I could risk triggering
the next event by hitting the action button in high risk
areas. Sometimes I would just run back and forth between two
screens. Sometimes I alternated between the decision-making required
for shopping and wandering. Sometimes I was compelled to move forward
because that's what felt right. By the time the next major event had
happened I'd invested a good deal more in the character since the
last major event.

Cloud performed the same animations over and over. I looked at static
backgrounds. I read some text that didn't change, some of it
twice. Maybe I missed some. It felt like it was me playing. It wasn't
my friend playing, it was me. I remember there was a guy who loved
FF. He was the only one I knew who did. He was older than me. I
remember him sitting down at my PC and handling the controls for a
bit, obviously more naturalized to the environment - I had just
started my game recently - and I thought, "wow, he's Cloud too. I
didn't consider that." Of course it was obvious, I wasn't
brain-damaged, but it still felt novel watching him run around as my
character. It wasn't an invasion of my privacy. He wasn't controlling
my story. He was temporarily telling his own in a game world that
looked just like mine, that I was moving around in 30 seconds earlier.

What does this mean? I have no idea. The power for a player to tell a
story is very strong, and he/she doesn't need much to do it. FF7
offered very little control. Really you could decide where you looked,
how long you were there for, and when you moved the story along. That
was it, that and fighting, a relatively simple mechanic. There weren't
any complicated algorithms. The computer didn't understand anything
about me. The game was designed just so that I could exercise some
control but never break from what had already been decided would
happen.

Earthbound made me feel like I was controlling a story. It was a lot
like Half-Life 2 in that way (I haven't played the original).  Both
these games are good examples. In Earthbound I felt like I was
discovering the plot myself for most of the game. Half-Life 2 was a
little more obvious, but the flexibility in the approach to combat was
enough to make up for it. Also I was very young when I played
Earthbound. The Mother team just laid things out in such a way that
you were guided instead of fed. They populated the world with little
vignettes that massaged you into the correct mindset. It was like
undergoing suggestive hypnosis. You know it's happening when it is but
you don't care; you play your role.

There is a big gap between that experience and one that is totally
under the control of the player. One of the assumptions that the
Open-World RPG makes, typically, is that a player needs to be able to
do anything to feel like they're expressing themselves. But this isn't
true. A child who has parents that create no rules is often unhappy. A
child with too many rules is unhappy in a different way. The best home
has the right rules. Children need the structure that guide them to
the areas in life where they can express themselves and have a good
experience. The modern Western RPG is like the lax home, and the JRPG
is like the strict home. Every other game with a strong story
component pulls from one of these two poles. Minecraft has very little
structure, Dark Souls has a ton.

Smiley. Yes it's a complicated problem. I don't think it will take genius,
just uh... persistence, something like that, maybe a little more.

So here comes the kicker. One of the most compelling experiences you
have in gaming is when the narrative, whether loosely structured like
in Minecraft, or heavenly spooned to you like in FF13, lines up with
your personal experience in the mechanics. In FF6 for example, or "3"
on my cartridge, the game felt so real partially because I had
control (in the same way I described having it in 7), but also because
the battles were so intense. I had never had to think and focus and
refine my strategies so consistently in a game before. I'd just sit
and sit and sit, and master. I'd grow, right in front of my own eyes,
with sweat slipping the controller in my plams, my eyes glued to the
screen, and my thumbs bouncing buttons like piano keys. I would
overcome what had seemed like an insurmountable challenge, through
ingenuity, persistence, talent, and growth. Through battles I related
to the characters in one critical way: struggle. I struggled when they
did. When the rest of the story played out I was so much more in-tune
with what was happening. When mechanics and plot combine, the
experience explodes. Players are story generating machines. They pour
creative content into their own imaginations and the machines in front
of them non-stop. There is a potential for story creation in games
that far exceed any other medium, because games are collaborative.

Aside, think about the power of mod communities. Woah, what if you
could harness and channel that into a single product?  Right now it's
a wild-west with mods, or a Mr. Rogers with user-gen content
distribution.

Counter-Strike was a community creation. Hell, indies are a community
creation.

Okay. Where most games go wrong, when mixing story and gameplay, is
not lining up the feelings produced by one with the feelings produced
by the other. They just don't do it. It's not like it can't be
done. It just doesn't happen. Let me name a few games off the top of
my head who do do this... hmm, Zelda, Metroid, Mario. Uh
oh. Coincidence?

You know, players give excellent feedback about how engaged they are
at any given moment. There's a lot a game can do to determine how its
generated patterns are impacting the player. There's a lot of
potential there.

Quote
This is actually the kind of storytelling device that I think
you're talking about: Rules that involve interacting with characters
and the world, and a plot that evolves from that. But I think you'll
have a very difficult time letting a computer decide those
rules. Without some kind of authorial input on the rules of my story,
my plot becomes a lot less meaningful to me as a player.

I wonder what the best way to describe it is. What you've said isn't
inaccurate. The plot isn't something that generates out of
nothing. Minecraft is totally free-form. That's not what I'm thinking
about. The game I have in mind has a general structure. There are
transformations. Imagine you are a rock star putting on a concert. You
show up with a vague idea of what you are going to do, and you
improvise. You react to the crowd, you let it out from the hip. That's
where it's at. You still have songs, your songs have structure; you
just understand them well enough to bend them to your will only using
your intuition. There's no other way to be on stage, unless you're a
big giant loser.

The kinds of intuitive relationship you have with each piece of your
music can be coded into a game. The story will have structure and
flex. I'll shove in themes and ideas that I love, and the system will
find it's way from one point to the next, maintaining the
relationships that I built, and taking into account what the player
has to say. I wouldn't provide any more flexibility than what could be
handled by the AI.

Also, players will learn to compose music. ... oh shit.

The computer doesn't decide rules. I decide the rules. When you write
a script for a movie the first thing you do is spend a shit-load of
time coming up with ideas. I've read that 5-10% of the time in writing
a screenplay is actually writing the screenplay. Nearly all of the
time is pre-work, and the rest is revision. There's this mountain in
the writer's head that doesn't hit the page. The dialogue is just the
tip of the iceberg. Then the director, then the actors, then everyone
else interprets the finished thing. They take those iceberg tips and
build their own associations down into their own minds. At various
points things need to be brought together. The team communicates with
each other and the writer, looking for clarification on how everyone
perceives what he's written. They activiate all his unwritten
knowledge at one point or another. Ideas ripple out from his mind into
the rest of the team, until they're eventually hardened into a final
thing.

The writer has the ability to do rewrites. He can express his ideas to
suit different directors, or the various needs of a single one, and a
million other constraints that stream in. So much never makes it into
the movie. The writer in a way is a collaborative story teller. He
starts with a piece of a thing, a basic structure, then bends it to
blend it with input from the much more powerful voice that is the rest
of the team. What if his mind was in a computer? All the rules he
spent developing in his mind could be put into a machine. Then we'd
get the same result, except reproducible, digitally. Oh yeah....

Note. Writers have a hard time collaborating. You really have to
communicate with someone else to share script duties. The process is
too personal. Look at the diversity of film, the lack of
diversity. Script writing is hard. You don't see companies of writers
for a reason. They only work in teams for some tv shows, and those
come with a packaged structure. If you could illustrate your ideas in
a technical form, it would be possible for two minds to build a story,
or 3, or 4.... go, goo.

There will be a strong sense of authorial input. It will be possible
to generate predictable player behaviour and have the game play itself
if I wanted it to (say for testing). In other words, there will be a
definable median playthrough, from which every other play through will
vary from in some way. You can think of Doom as an analogy if you'd
like. In Doom, if you were to draw out a player progression graph, the
player's play space would expand and expand, climax, then contracts to
a single point, representing the next locked door, or necessary
entryway, then repeat. Imagine the same graph in n-dimensions with a
little more variety in progression. There are abstract "gates" in a
sense, that are defined in rough ways. Some things need to happen
before other things, and so on.

Mmm. Also, since everything is procedurally generated, the quality of
the graphics can be scaled dynamically and context-senstivitiely, if
that has any meaning to you, or anybody. Goodbye loading times.

This is turning into my diary.

Ohh, I'm watching that Jon Blow vid. I had forgotten, in this one he
fights for the potential of story-based games. He did a talk a year
ago - one of his most recent - where he seemed a lot less positive
about it, though a lot more confident in himself. It was also during
that talk that he gave the single best answer that I have ever heard
to how to make good educational games. The particular question was
framed as calculus I think.

Note, he talks a lot about mechanical experience and narrative
experience blending. He calls it story meaning and dynamical
meaning. Those are good terms too. His biggest knock is that it's too
difficult to predict the development of a creative product to stabily
develop a logical one beside it (i.e. software). A small change in the
mechanics can have sweeping interpretations by the player, totally
destroying the narrative meaning; and a small change in the story can
have enormous costs in re-developing all the code to match
it. Software implementations of creative ideas are exponentially less
malleable than ideas on a page...  but they're exponentially more
productive once they stabilize. Oh the ironies....

The solution is to develop the story-line and mechanics in tandem,
retaining flexibility in both. Plan: find the best summary of the
desired experience, implement the minimal narrative and mechanical
structure to deliver this experience, repeat. Grow the experience out.

When I say procedural, I mean everything: the sound effects, the
music, the story, the animations, the characters, everything. 'Tis
necessary... well, sort-of. The alternative is just a lot more
complicated to talk about.

God, I love Daft Punk.

You know, I've never played a Metal Gear game. My god. Who am I?





















« Last Edit: June 24, 2012, 04:46:33 PM by toast_trip » Logged
s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #10 on: June 24, 2012, 02:48:17 PM »

Quote
You can also trick people into thinking that their experience is user-generated, which is something that interests me a little more as a developer. Half Life 2 did this by letting you move around at all times and avoiding cutscenes (or masking them). You didn't have to listen to people, you didn't have to leave City 17, you didn't have to notice the G-man every time he appeared, you weren't instructed to do anything, and yet there was a very clear plot that arose from your actions.
yeah we've become pretty good at giving the player illusion of agency. but how much cooler would it be to get rid of the smoke & mirrors and do this shit FOR REAL.
Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #11 on: June 24, 2012, 02:51:10 PM »

Quote
You can also trick people into thinking that their experience is user-generated, which is something that interests me a little more as a developer. Half Life 2 did this by letting you move around at all times and avoiding cutscenes (or masking them). You didn't have to listen to people, you didn't have to leave City 17, you didn't have to notice the G-man every time he appeared, you weren't instructed to do anything, and yet there was a very clear plot that arose from your actions.
yeah we've become pretty good at giving the player illusion of agency. but how much cooler would it be to get rid of the smoke & mirrors and do this shit FOR REAL.

A lot.
« Last Edit: June 24, 2012, 03:40:48 PM by toast_trip » Logged
Rube
Level 0
***



View Profile
« Reply #12 on: June 24, 2012, 06:11:16 PM »

I was thinking about doing a roguelike with procedurally generated stories at one point. There was a core principle I realised - stories are not simulations.

For example, when writing Lord of the Rings, Tolkien didn't roll a bunch of dice to decide if the good guys won their battles. He just wrote that they did. He didn't plot the position of every individual in middle earth then work out who the protagonists met on their journey, he just wrote that they bumped into whoever they needed for the plot to advance. All stories are written like this, full of contrivance and coincidence. A good story hides the contrivances, but it's still there.

I was going to go about generating a story the same way. Let's say you generate a town full of random people. Normally in a story the protagonist never meets anyone who doesn't further the plot. So I'd have the game remember which generic NPCs the player talks to and when it needs a character to fill in a story position it could pull from the pool of NPCs that character talked to (where appropriate of course).

Perhaps you discover that the bartender you talked to earlier is the leader of the local cultists when you burst into their hideout. Perhaps that weapon store owner you bought a knife from earlier is a lone vigilante who turns up to heroically save you after you almost get killed fighting monsters in the sewers (not because the game scripted you to lose, but because you attacked monsters that were above your level). It adds the sense of narrative that you expect from a story.
Logged
Graham-
Level 10
*****


ftw


View Profile
« Reply #13 on: June 25, 2012, 12:28:11 AM »

Stories are not simulations, yeah. Skyrim is kind of like a story simulator, except it only simulate 4 things. Good stories have to be written.

Reading what you wrote, I'm thinking about it this way. Imagine all of your characters as a resource. Maybe you wrote some interesting characters, and some interesting things that could happen between them - details about their relationships. You have a box of dialogue, feelings, behaviours and so on. The contents of the box are divided nicely into categories. You have exciting stuff on the left, Clara's stuff on the bottom, anything that comforts the player in the middle, and anything that goes well with the 3rd and 5th dungeons marked in red.

When the player is in a situation, you say, "what would I like this character/player to experience next" (this is from the perspective of the machine, or you the developer). Then you search through your box for a piece that fits, given what's already been used and what constraints are on the situation: who is where, what the character's current state is etc.

Some items in the box have generic labels. Some dialogue is like, "when Clara is pissed and frightened and underground," or, "any child under 3 when the hero has just been foolish." You just fit stuff into slots.

You want the player to have control, but you don't want to allow a mediocre story to play out.



« Last Edit: June 25, 2012, 01:44:34 AM by toast_trip » Logged
gimymblert
Level 10
*****


The archivest master, leader of all documents


View Profile
« Reply #14 on: July 12, 2012, 05:16:26 AM »

Have you looked into façade theory and implementation papers?
Logged

Graham-
Level 10
*****


ftw


View Profile
« Reply #15 on: July 22, 2012, 07:57:04 AM »

Have you looked into façade theory and implementation papers?

Are you talking about the software design principle?
Logged
TobiasW
Level 8
***


This can only end brilliantly!


View Profile WWW
« Reply #16 on: July 22, 2012, 07:58:43 AM »

He's talking about this game: http://www.interactivestory.net
Logged

Graham-
Level 10
*****


ftw


View Profile
« Reply #17 on: July 22, 2012, 11:41:23 AM »

I checked it out. I think I've heard mention of it before but forgot about it. It's interesting.

I like the voice work. The game feels very smooth. It's artificially "low budget" - probably from being low budget - so the behaviours are meaningful. I feel a lot of emotion just from the characters' positioning, body language, and facial expressions.

There's a lot of work to make voice blend with procedurally generated content; voice, hand-drawn art, mo-cap animations too - anything man-made. Facade seems to make some assumptions about how the player will fulfill their role, then gently guides him back into place by only re-affirming constructive choices. That's how you train animals, children, unsuspecting people; that's how you do it.

The biggest issue with the game is its lack of guide rails. It is very easy for the player to get into a position where he doesn't know what options are acceptable, so he is left to choose between doing something that probably wastes time or doing something that progresses the situation.

It reminds me of those dog collars that shock the wearer when it steps outside of an invisible, underground, fence. Dogs put in that position tend to roam extremely small areas (far smaller than necessary) out of fear of repercussion for failure. The repercussion of exploring the game is doing something with uninteresting or non-existent results.

Actually, I'm kind of reminded of the freedom I felt playing adventure games - point-and-click - when I was a kid. Back then I felt like I could do anything. My choices were pulled from my vast databank of life experience, so my whole history was engaged. Those games became trivial when I started to see the tiny path you have to walk down; I still like them.

I suppose half the fun of Facade is figuring out what you can do and can't do. There isn't a ton of consequence for being wrong so the value of doing that is purely explorative; it isn't connnected to the game's primary experience. There are two games in Facade: one of playing through the story; the other of mapping out what's allowed, within the "realm" of that story. You need to be in the mood for both to get the value of out either, and one doesn't necessarily rev your engines for the other. A lot of "exploration first" games fit this description. Skyrim, for example, relies on a desire to explore and a desire to role-play, but doesn't use one to inject the other.

But Skyrim is successful, and rightfully so, so whatever.

A lot of story-like, proc-gen, game attempts focus too much on the tech. They often miss the more important point of what decisions the player is allowed to make in the first place, and how they are presented, and how they contribute to the game's vision. Facade doesn't make this mistake, so it's interesting. Too bad it's only half a game though. Mapping its experience even on to something that suits it naturally, like a riddling adventure game (Monkey Island), would increase its value immensely. However, the creative force that builds up the mechanics would have to be in-synch with the one that did everything else.

There's this lack of unity, in vision I guess, that usually prevents proc-gen, and malleable story lines, from being interesting games and real narratives. We're not really short on a new AI paradigm (well, maybe a little), or tech explosion, or whatever. We just don't have enough design experience for blending all of our ideas to feel natural. We're just budding designers. AI is like fruit from the highest tree.

Thanks for the link.

Logged
gimymblert
Level 10
*****


The archivest master, leader of all documents


View Profile
« Reply #18 on: July 23, 2012, 04:35:58 PM »

things you should read

http://www.edge-online.com/features/future-videogame-ai
http://emshort.wordpress.com/2012/07/08/the-act/#more-4857
http://www.gamasutra.com/view/feature/3761/beyond_aiml_chatbots_102.php?print=1Suzette
http://aigamedev.com/open/reviews/facade-ai/
http://www.interactivestory.net/papers/MateasSternDAC05.pdf

EmShort is the industry leading figure for video game dynamic  character and story design. Here is a bunch of categorized group of article she made and is not only insightful but practicable:

http://emshort.wordpress.com/category/characterization-in-games/
http://emshort.wordpress.com/category/conversation-modeling/

Other thing you should look for:
Behavior tree:
http://aigamedev.com/open/article/bt-overview/
http://aigamedev.com/open/article/behavior-trees-part1/
http://www.altdevblogaday.com/2011/02/24/introduction-to-behavior-trees/
Logged

Graham-
Level 10
*****


ftw


View Profile
« Reply #19 on: July 23, 2012, 04:52:50 PM »

Thank you very much. I will read all of it.

Smiley. The behaviour trees. AI is my forte. But thank you anyway. I read all AI literature anyway, always. For perspective, and to be a jerk: I'm currently build a poker AI, which will play the game very well, and will be able construct training schedules for me to play as well as it, optimally. That's where I'm going to get my game funding from.

You talked about, in another post, predictive complexity, like me predicting an opponent's moves in a fighting game. Poker is all that; mechanics are really an excuse to predict. Everything else is about being one move head. That's the connection between the Poker AI and the game one (i.e. "conversational") - also, both are very "soft" logic i.e. intuitive.

I'll post my thoughts on the articles once I've consumed them.
Logged
Pages: [1] 2 3 ... 10
Print
Jump to:  

Theme orange-lt created by panic