June 2, 2006
Friday Night Links
- Microsoft is holding a chatterbot contest with $40K in prizes. Bots due Oct 15.
- E3 panel of game designers on interactive narrative
- Brandon Rickman critiques Glassner’s Interactive Storytelling
- Ernest Adams on “Perlin’s Law” (oh please)
- An ongoing series of game structure analyses at Only A Game (1 2 3 4) — identifying key player moves (aka verbs, or discourse acts) in several games
- Excerpts from Katherine Isbister’s Better Game Characters By Design
- Second a-Life
- Very cool recursive interactive images
- DARPA life recorders (previously blogged 3 years ago as “LifeLog“)
June 2nd, 2006 at 10:39 pm
“Perlin’s Law” isn’t so bogus in itself, but I think a better way to see it is in terms of conserving agency in a system rather than “credibility” which you obviously can’t measure robustly. I wrote a letter to Gamasutra regarding this, but they didn’t publish it.
The E3 Panel is complete groping in the dark. I love it how David Jaffe’s resignation on the issue is implicitedly a worthwhile opinion.
Rickman’s critique is right on.
Isbister’s book is really sharp stuff, looking foward to getting my hands on a copy.
June 4th, 2006 at 11:22 am
Yeah, the Lawgiver has spoken, eh –
June 9th, 2006 at 2:21 pm
On the notion of Ken’s Law: “The cost of an event in an interactive story should be directly proportional to its improbability.”
This is always sort of an interesting discussion to me…
First, I should say, while I think Ken is awesome, and a friend, I would hesitate to call this a law, or to call it Ken’s Law. And, I think Ken would agree.
What it is getting at, though is the very basic and well established existence of what most people call setting, but what I prefer to think of as narrative physics. By which I mean that when an author tells a story, they decide by what rules their characters and environments operate: e.g. in 100 Years of Solitude, ghosts walk around, in Murakami, animals speak, in John Cheever, men have gone to war, and come home to live in Connecticut and commute to NYC. These physics that the author establishes are the rules by which all the characters and events must operate.
Where this gets interesting is when these rules are broken.
The rules are always broken intentionally (and I think this is important), and yield only two results:
1) A total break in (for lack of a better term) the willful suspension of disbelief. When the physics of a narrative are broken, the participant/reader/player is forced to ask themselves, “In this context, is this believable?” Now, it’s the first half of that sentence that’s important. The author establishes the context/physics by which all other events are measured for their believability. When this break happens (thanks a lot, Brecht) I, personally, almost always find it highly annoying. If you’re a programmer, think of it as an unexpected state change. This is also why I tend to dislike narrative/cinematic games in which I must constantly switch between playing and watching without clear delineation. As an example to disastrous end, watch Spike Lee’s Summer of Sam. I don’t know anyone who didn’t laugh when the dog spoke. Which, I’m pretty sure is not what Mr. Lee intended. In the very least, it’s not clear that that was the intended response.
2) The transcendent moment. Sometimes, and this is particularly hard to do, the break in narrative physics is that transcendent, cathartic moment. There are many examples, and I won’t go into them here. What’s important is that the user experience from this sort of event is much different in linear narrative than it is in game play. In linear narratives, naturally, everything has been leading – in a direct line – to this moment, and then it’s done, and nothing is ever the same again. In game play, I think the best analogy is the Easter Egg. And, likewise, once you’ve found an Easter Egg, it’s really hard not to spend the entire rest of your playtime searching for them. – This is what the Mario Bros. franchise understood so well and took such great advantage of.
What’s interesting about the idea of a transcendent break in narrative physics is that it seems at odds with the notion of free-form exploratory play. GTA does have direct consequences for behaving strangely; if, say, you decide to go on a car-jacking/crashing/shooting spree (as I think everyone does in the first 5 minutes of play) you are quickly taken down by the police. So, how would you build the diner scene in Pulp Fiction into GTA? I don’t think you could, at least not to the same emotional, Aristotelian effect. I also don’t think that’s a problem, it’s just not what GTA was built for. Façade, on the other hand, I think, is built in that direction. The other problem with trying to engineer this sort of moment into a world-based game is that in linear narratives, the catharsis happens, and then the story ends. The story ending is no small part of this system. It leaves the participant with the momentum of that cathartic moment. Because it’s really not very impactful to have a cathartic moment, then read about the character brushing his teeth and taking out the trash for another 250 pages. This is because all good narratives are about editing, just as good game design is as much about what the player/world can’t do as what it can. World-based games tend to (there are, or course, exceptions) offer the same sort of cathartic opportunities as life, which is to say, small ones, which stay with one briefly. But even when we have the huge cathartic moment in a game, if it’s powerful enough, we tend to lose interest afterwards. Or try to recapture that moment endlessly, slightly diminishing its power on each subsequent encounter.
The beauty of games which I think game designers/writers tend to forget is that we’re giving the participants the opportunity to have or not have those moments at their discretion. I think one type of good interactive narrative design creates a space in which these moments may happen. Just as the participant crafts the story of their time in a world, so to do they decide where and when that story ends.
In summary, the notion of Ken’s Law already exists and has been well documented and discussed, albeit with slightly different terminology. I also think this (lack of) agency and consequences are already built into good games. And I think it’s more interesting for us to consider how and to what end narrative physics can be broken.
June 9th, 2006 at 8:37 pm
Adam, great to hear your voice here! Great insights.
Just a few follow-up thoughts, building upon yours…
I’m reading Murakami’s latest right now, Kafka on the Shore, with speaking cats, ghosts of living people, and a Japanese pimp named Colonel Sanders. Do I feel a credibility budget being drained as I read?
When odd or unusual events occur in stories, if they’re well crafted they don’t withdraw funds from a credibility budget; they (re-)define the setting or “narrative physics”, as you’re calling it, as the events occur. If they’re poorly written and not integrated into the whole of the story — if the story’s particular reality doesn’t hold together — then sure, you are testing the patience of the reader/player, i.e., draining what one could call a credibility budget, but by then I usually put the book down or turn it off before it’s depleted — I’m not interested anymore.
To talk in terms of a credibilty budget is framing an experience in terms of violations that are presumably going to occur; this reminds of me of how some think of player interactions as “interruptions” or perturbations in a story, that need to be “dealt with” or “accomodated”.
I find that a dysfunctional way to think about the role of the player in an interactive story; better to design assuming players are going to do lots of crazy things, experiment, etc., and that the story is going to need to be very flexible and robust — or even better, that the experience is implemented to in effect generate the story on the fly, so there was no canonical story the writer had in mind that needed to be flexible in the first place.
I’d also agree that the ideas behind the law Ernest is putting forth is more or less already incorporated into existing design techniques; for example, Clint Hocking sees that principle in his recent presentation on Simulation Boundaries, and Michael remarked to me that he feels his material-and-formal-affordance-balance theories contains that principle as well.
I don’t really have a problem with Ernest proposing new design laws and naming them after people; he’s known to have generous fits of hyperbole, and who I am to complain! However I’m not sure the design community is going to take him up on this proposal.
June 9th, 2006 at 9:11 pm
Another thought. Let’s get back to Ernest’s original question: how do we balance the player’s desire for freedom with the designer’s desire to tell a consistent, coherent story? What do we do when the player wants to do something that doesn’t work with the plot that we’ve laid out? Refuse him permission to do it, and take away his freedom? Or allow him to do it, and destroy our story?
Yes, allow her to do it, and potentially “destroy” the story — except nothing’s being destroyed if the player truly has agency! The player, by acting the way she is, wants the story to go in that direction, or at least experiment and see what would happen.
Gandhi should be free to pick up a gun. (Michael’s argument, that the player-as-Gandhi should somehow be prevented from picking up a gun in the first place, doesn’t hold together; Gandhi’s world has guns.)
The idea that there is a pre-ordained story, to be followed or achieved and potentially destroyed, is the flaw in Ernest’s question. The word “tell” should be banished from the lexicon of game designers. To give players high agency means they have enough influence to push events in all kinds of directions. As a result, the story may not have a singular overall coherency, but that’s okay; that’s what the player wanted to do!
That said, to help maintain coherency and consistency, the NPCs in the world can recognize that the player is now acting inconsistently, incoherently or crazily. They can resist the player, try to combat her actions. (“Mahatma, calm down, just give me the gun…”) Brief diversions from a strong overall narrative direction (“coherence”) can potentially be recovered from. Totally crazy or off-the-wall player actions can believably be ignored, to a certain extent. But at some point the NPCs decide the player is insane, and perhaps trust is lost in the player, that would need to be rebuilt.
This gets to the idea of a skillful player in interactive drama. There is some skill to performing in an interactive drama such that the overall story has maximum coherency. It’s a goal players can choose to aim for, or not.
June 10th, 2006 at 12:05 am
Lots of threads to touch on, so I’ll go point by point.
Credebility: I’m considering experimenting with Storytron to create a surrealist storyworld where credebility budget is either completely ignored, or perhaps somehow inflates and extends, so that the user is challeneged to adapt the increasing insanity and surf it out. It’d be interesting as a test of Adam’s theory, at least, though I suspect his idea may be useful for realist storyworlds, or something like Facade.
Agency: I think I agree with Michael more regarding agency, its possible to provide an inconsistency or perhaps a player character’s “subconscious” which limits particular verbs by context or in general, respectively. A Ghandi storyworld could have the player pick up a gun, but be stymied in trying to use it by the well-trained “spirit” of the Ghandi AI underlying their navigation of the storyworld. This could actually be benificial, creating terse moments of reflection. Should these measures be absolute? Probably not, though the “subconscious” technique could prove more effective than external NPCs in pushing the player to “stay in character”.
The real question of believability, IMO, stems from the agency the user experiences. If the player is a mage, then they should be able to turn a sword into a flower, for example, but not be able to spit on a tossed coin at it falls into a well and trigger armageddon. The real issue is that a player’s intent is tranlated from the immediate, moment by moment choices into a satisfying macro-scale implementation. I suspect that good agency will automatically provide “coherence” to an interactive story.
A table-top RPG, My Life with Master, provides a good example of how constraining the user’s role and progression along a template of a story arc leads to very strong agency and catharsis, but not nessecarily strong “interactivity”. I recently posted on it, but you should also check out what some others have written on it, in case you’re unfamiliar with the game.
“telling” – A big problem I have with Ernest’s article is that he’s approaching interactive story as if there is some sort of profound conceptual gap between that and the rest of interactive experience. I think replacing “credibility” with “agency” as the crucial variable, which is essentially replacing a noun with a concept intrinsically related to verbs, is a much better approach, since you’re framing interactive story as just another region of design space. Doing that allows you to focus on the main meat of the matter, the material and formal constrains, which for our purposes are characters and the social dynamics betweent them. If you balance your rhizome, that is the porportionality of all the characters’ relationships, you’ll get the desired conservation “for free”.
June 19th, 2006 at 8:12 pm
Talk about unbelievable — I just finished that Murakami novel referred to in my earlier comment; in the final chapter of the book, a large black (talking) cat named Toro is one of the final characters introduced in the story.
I have a large black cat named Toro! And he talks quite a bit.
June 27th, 2006 at 6:52 pm
David Cage mentions the credibility issue in his Indigo Prophecy postmortem:
although he simplifies down to: “you can make people believe anything in a scenario, but only once per story.”