Monday, July 24, 2006

More Stories!

A bunch more people have posted about stories in games. I have lots of opinions, but I'll limit myself to one response, because I have other things to do today.

Tadhg Kelly seems to have several people replying to him already, but that's because he's refuting to something that doesn't exist and spraying the same refutation on things that do exist by misguided fiat.

In English, his flaw is the same as if he had said: "Because space ships are too difficult for you to build in your back yard, nothing that flies exists!"

The thing he's saying is that generated stories aren't the future of games because they are complicated and fragile, and therefore cannot be generated "cleanly". That's true, as far as it goes. I have no doubt that computer-manufactured stories will suck.

But since when - WHEN - did computers manufacture stories? Nobody honestly thinks they will, unless they are hopelessly naive. Maybe fifty years from now, but not today.

No, games with adaptive or generative stories aren't about building a story. They're about helping the player build a story. Or an experience - whatever you choose to call it, if you're the sort that overdefines everything. The game doesn't pop in with grand ideas of its own invention - the game simply utilizes the tools and pieces provided by the writer/programmer to give the most fluid and interesting experience available.

And there IS a basic algorithm for getting a computer to figure out what things the player is interested in and twisting his arm with them. Just because nobody's AAA game has used it doesn't mean it's impossible: it's simple token substitution.

Lastly, it is entirely possible to use PLAYERS as story writers.

The idea that stories are fragile and complicated and therefore not suitable to games is bunk. Advanced 3D graphics are fragile and complicated, too.

Tadgh is evidently a big fan of "simple rules, complex results". So am I. That in no way precludes stories. It's like saying that having a sister precludes having a brother. A family may, in fact, contain both without problems.

Yeah, pure tactical games are popular among a specific crowd. But the experience of, say, Final Fantasy Tactics is not one of simply moving chits around. You're moving people around, and that adds a lot... without diminishing the gameplay. If anything, it offers exciting new gameplay opportunities as missions unfold directly from the plot's push and directly cascading from successes and failures in previous missions.

I simply cannot conceive of how he has decided "interactive storytelling" means "computer-generated storytelling". "Interactive" has a specific meaning, and "generated in isolation" is not it.

Maybe he's been listening solely to zany 16-year-old IF guys and Chris Crawford?

Story - and characters, and graphics - are extremely powerful tools. Writing them off seems absurd, even when a few people are obviously going over the top with their expectations.

---

16 comments:

Patrick said...

"Maybe he's been listening solely to zany 16-year-old IF guys and Chris Crawford?"

Actually I think this is fairly accurate, or more specifically Chris Crawford and AAA developers who're fans of tacked on story to the point of fetishism.

I think a big dichotomy that should explored in this debate is one you keyed me into and thats the difference between a generative and emergent paradigm. I suspect Mr. Kelly might hold more hope for the latter, assuming the goal isn't highly structure play so much as play about people, i.e. characters.

Craig Perko said...

No, Crawford isn't about tacking on stories. He IS, however, one of the foremost people egging on the "computer-generated story" myth.

Patrick said...

Yeah, I meant to make the association as a dichotomy, rather than a direct association. I suppose I did get my syntax garbled.

Craig Perko said...

What you are building and what everyone thinks you are building are two different things.

What you are building is an engine to help writers create interactive stories. What we hear from out here is:

"Look at you! You may look healthy, but you could fall over dead at any second!

"Fortunately, all you need is a swig of storyworld juice, and all that will be cured! You'll be bigger and stronger than ever! Also works on hemroids and hair loss."

Patrick said...

I think you're getting a bit off topic, Crawford's public rendition as given in that interview may not have been the most tactful, but I'd rather hear you two discuss the differences between your approaches to story, rather than exchanged barbs.

Come on, other people are probably reading this, lets give them some substance!

Storytron seems to take stories at different granulaties (verb-based events) and by a process of causation (the specifics of which are authored by the storybuilder) which is mitigated by user options (also authored) generates a resulting discourse. The idea is that this discourse is both interactive, and by the virtue of the verbs between them, an interesting story. The downside, which is only a downside if your needs differ, is the feedback is minimal and the interface is homogenous, in fact the langauge used to make inputs also is the data structure defining the story.

Craig's approach, on the other hand, seems to involve scripting of idiosyncratic characters which handle verbs on a context sensitive basis, its dependant on (rather than insulated from) being tied to a game engine and as such allows much more freedom in presentation and game design, considerably less freedom in making the story dynamic. The characters bounce around within these constraints, angled and spun by player influence, and the result is an emergent story that seems compelling by the inducitive powers of the player's social reasoning, rather than by the structure of the "story atoms".

So thats my understanding as a plebian content designer, discuss.

Craig Perko said...

I'd be happy to discuss it, but email is a better alternative.

I don't have any problem with Crawford's approach, it's his hype that I have problems with.

Tadhg said...

In English, his flaw is the same as if he had said: "Because space ships are too difficult for you to build in your back yard, nothing that flies exists!"

Not really, no.

My attitude is not that it's simply a very difficult problem, because a difficult problem is solvable. My attitude is that it's a paradoxical concept. For instance:

They're about helping the player build a story.

This is just a nonsense statement. If you accept that a story is a structure, it follows that that structure is interdependent on its early and late elements for it to work. You must have setup and you must have reward, and they must be timed just right. Therefore for any kind of story, player built or otherwise, to work, it must all go perfectly, and that's a big logic break. Because if the player is active in their own experience, then it cannot go perfectly.

Secondly, for a player to feel that they are the ones playing, the system of the game that they are playing must be robust, which requires simplicity and openness and so on to work. The more guided it gets, the worse a game it becomes, therefore. That is self evident.

This means that games cannot have stories if they are to work and stories cannot be games if they are to work. This is why I said that games are fictions, not stories. From the basic interactive choice gamebook through D+D and all the way to some mythical complicated technology system, these things all function according to the same rules, and those rules are of robust systems and a 'fiction' approach that ditches story in favour of play.

You see, the phrase "helping the player to build a story" is just meaningless. Every game of football helps the player to build stories of their games. Everything we do in life helps us build stories in one shape or form. So "helping the player to build a story" means basically nothing.

We've had some lively further discussion on this over at the original article, if anyone's interested.

Craig Perko said...

Re: Players building their own stories.

Nonsense? It's already been done! KoToR, Black and White, Deus Ex, and dozens of other games allow players to progress through a story, modifying it to their preferences as they go.

The reason you think it's nonsense is because you've defined "story" funny. Perhaps "drama" would be a more appealing term?

I don't see any proof that it is anything other than a difficult problem, but even if it is impossible, that doesn't change the fact that it isn't the problem MOST of us are tackling:

We're not making computer generated stories. We're making computer enhanced stories, or computer arranged stories.

Furthermore, a story can be separate from the game dynamic, save for driving level progression. The player doesn't need to "play" a story, simply "interacting" with it is enough.

There are lots of techniques for enhancing play through story, and enhancing story (or drama, whatever) through interactivity. Many of them are already in wide-spread use: it seems like sticking your fingers in your ears and humming to pretend they aren't possible.

Duncan M said...

Secondly, for a player to feel that they are the ones playing, the system of the game that they are playing must be robust, which requires simplicity and openness and so on to work. The more guided it gets, the worse a game it becomes, therefore.

I disagree. You can write a simple system that feels conversational and natural. You can even write it to feel as if there is more player control than there is. So long as the player feels in control, they might as well be. Even if their control is limited, they have the same experience as if it was not. This can be done through immersion, careful direction, and good writing. Bad systems feel flat, good systems are invisible and you never notice the walls.

Tadhg said...

Nonsense? It's already been done! KoToR, Black and White, Deus Ex, and dozens of other games allow players to progress through a story, modifying it to their preferences as they go.

None of these are stories. They are fictions. The player is not generating self-satisfying stories for themselves in them any more than a tennis player is generating a self satisfying story for themselves in a match. In both cases, they are playing a game.

There are key differences between story and game and those differences are robust systems and a lack of the brittle structure of stories. All these games are paced at the player's pace and ramble all over the place as the player uncovers and plays with the fiction.

The reason you think it's nonsense is because you've defined "story" funny. Perhaps "drama" would be a more appealing term?

It wouldn't. Drama is an inherent part of a story, as it is the climactic result of good story structure. Drama does not exist in isolation. In games there are occasional moments which feel dramatic in the more colloquial sense, meaning that you're emotionally excited. This usage of the word drama is nothing to do with the idea of Drama as it applies to stories. As with much to do with the concepts of interactive stories, it's built on a linguistic conjunction, trying to patch together two entirely different things because they happen to use the same word.

We're not making computer generated stories. We're making computer enhanced stories, or computer arranged stories.


Arranged, generated, sequenced, whatever. What you're cleaving to is the idea that a story can be systematised and broken down into many small chunks, and therefore guided to a player. As I've explained, it can't. It's a brittle self-dependent structure that becomes more difficult to change and alter as it gets improves.

What you're also cleaving to is the idea that this can be made robust at the same time. It can't. Robustness works best the more abstract and elegant it becomes.

The brittleness of a story's structure contradicts the elegance of robustness in every way. They are paradoxical entities, and there is no getting around that, not with all the most complicated systems in the world. Previous attempts have conspicuously failed to do so, and have proved less entertaining than games or stories whose creators have understood the distinction.

"Fiction" approaches, on the other hand do work and have been proven to do so many times. They recognise that play is what matters, that exploration of the fiction is what matters and that guidance of the fiction into some kind of story is counterproductive. It makes the game bad.

As I mentioned my own blog in the comments section, this is plain to see in games like GTA where the robust scenarios are always the better ones, and the ones where the game tries to get more plottish are tedious.

Furthermore, a story can be separate from the game dynamic, save for driving level progression. The player doesn't need to "play" a story, simply "interacting" with it is enough.

No, not at all. Any time that the player is not actively playing is boredom time. That's the time of waiting for the game to stop emotionalising and trying to tell the player how they should feel and going through the motions (Half Life 2 does this a lot) to get back to playing. Discovering background as part puzzle solving and part fiction illustration (As Deus Ex does through emails etc) is play. Watching as the game guides you down pre-served story sequences giving you the odd choice here and there is interacting, but it's not much fun.

it seems like sticking your fingers in your ears and humming to pretend they aren't possible.

Not at all. The complete failure of interactive stories as a concept is there for all to see, and the complete failure of the implementations of it likewise. It remains an attractive idea to some, but so do a lot of notions that have no basis in reality.

People like stories. People like games. People like imaginative games with interesting fictions. These are well determined. Some people think that this leads to a therefore of bridging the gap between the two, but in reality the two forms only bear a passing resemblance to each other and, unlike say the music video, the melding of the two saps the strength of either rather than enhancing it.

Fiction and robustness is what makes games imaginative, absorbing and interesting. Story and complicated attempts to hide brittleness makes them boring.

Tadhg said...

I disagree. You can write a simple system that feels conversational and natural. You can even write it to feel as if there is more player control than there is. So long as the player feels in control, they might as well be. Even if their control is limited, they have the same experience as if it was not. This can be done through immersion, careful direction, and good writing. Bad systems feel flat, good systems are invisible and you never notice the walls.

All systems, by definition, proscribe limits on what the player can do. There is no action Chess outside the board, pieces, turns and the clock (if you're playing speed Chess). All games have rules.

Rules are good. As rules proscribe certain actions, they also provide player control by enabling context. It does not matter if the system is open or not, it is that it provides context which is important. Because then players understand the actions that the emergent actions that they can take inside the context of the game. Emergent actions are the soul of good play.

For this to work, systems MUST be kept simple and relatively abstract. The players have to be able to establish the context of the game quickly, otherwise they go away feeling that the game is arbitrary and confusing (which it basically is). The more complicated and special-case the rules become, the more the context vanishes and so it loses what makes the game a game.

So, to keep it simple, the best rules are abstract rules because they are the ones that are most eaisly contextualised and enforced. Elegance is important here. I can understand what a health bar is immediately, contextualise the actions that I am going to take from that piece of information and the other rules and mechanics that I have also assimilated. I cannot do that from a game whose rules always change in relation to a difficult set of injury parameters and different parts of the environment trying to discreetly alter themselves in response to my "interactive drama" needs. I don't know where to begin.

On the other hand, in Halo I know that if I'm all shot up, I can hide and wait for my shield to regenerate and that gives me a chance to shoot some Covenants. The simplicity of the system conveys the information and context that I need to play.

It just doesn't work any other way.

Craig Perko said...

Thanks for trying to explain, Tadhg, but I just can't see what you're trying to say. You seem to think that there is a division between the kind of story one experiences due to gameplay rules and the kind of story one experiences due to scripted branching trees (or similar).

But scripted plots are a kind of gameplay rule, just as much as moving left in Pac-Man or hitting the up arrow at just the right moment in DDR.

To me, the division you're trying to base your theory on is so faint that I can't even see it.

I think this is just going to have to be one of those "agree to disagree" situations.

Tadhg said...

But scripted plots are a kind of gameplay rule, just as much as moving left in Pac-Man or hitting the up arrow at just the right moment in DDR.

They really aren't.

I'm sorry, I do have a tendency to talk long when I would have better success if I were concise.

The basic difference is that the rules of Pac-man are robust and elegant, where the rules of branched story are special case and arbitrary.

Craig Perko said...

Many games use arbitrary game play, too. Why do you have to be level ten to upgrade to a knight? Why can't you break down doors when you can throw cars?

These are still gameplay rules.

Now, stories are, at present, not "robust". But they are hardly arbitrary - they have a reason, usually an apparent one, and progress cleanly from one scenario to the next.

I won't claim that's the best way to do it, but I don't think it's the only way to do it, either.

Tadhg said...

Not the same thing.

All games involve constraint. Why can't I fly in GTA, why can't I move all my pieces at once in Chess, etc. The thing is not whether those rules exist, it's whether they work together in a consistent robust fashion. It is therefore obvious that the less rules which apply to more pieces that you have, the more robust the game is.

Arbitrariness, on the other hand, is when a game is introducing new rules or breaking its own ones or guiding you down pre-determined "just so" sorts of directions. That's a completely different thing from a constraining rule.

Good stories are not arbitrary at all either (bad ones are), as you point out. The problem is that games that try to be stories are, and that is bad.

Craig Perko said...

That we can agree on: at the moment, at least, adaptive stories are pretty bad.

The rest - I agree that robust rules and emergent play are good. I say so in many of my posts. But that doesn't mean I believe arbitrary, bad rules to be "not rules".

I can see where you're coming from, I just disagree. Not enough to matter: I agree they're bad, I just don't agree that they aren't gameplay rules.