ICIDS: The future of interactive storytelling, plus some Versu thoughts

Hartmut Koenitz submitted a talk for ICIDS that was essentially a manifesto about what needs to happen next in interactive digital narrative, and accompanied this with a workshop on the future of interactive storytelling. The points of the manifesto are as follow:

  • We need a new theory of narrative for interactive digital narrative in order to get rid of accumulated preconceptions.
  • Interoperability is key: tools need to be developed in such a way that they can be hooked together and progress on one hand can be used by others.
  • Sustainability is essential. Lack of archiving has already destroyed a lot of valuable research work.
  • Interactive digital narrative needs to be author-focused. There is a challenge in training new authors in procedurality in order to get useful feedback from them.
  • User experience is crucial. We need to focus on how people actually experience and enjoy this work.

I don’t have strong feelings about the narratology issue.

In re. interoperability: I think this is an interesting point, but in my experience you don’t get tools that can work together until after you’ve already reached a pretty strong community consensus about what you’re even trying to do. This pertains in some areas in the IF world (Vorple being designed to plug into multiple back-end engines, for instance), but it’s not so applicable in areas where people are doing really fresh research. It’s not clear what it would mean to be able to plug together even such related projects as Facade and Versu. They’re both doing some agent modeling and tracking social practices, but in many other respects under the hood they’re extremely unlike.

Sustainability: yup, I agree, and this is something that affects everyone working in this field. The IF Archive is terrific, and one of my biggest concerns about the current IF diaspora is that it’s no longer standard for a majority of interesting IF in a given year to end up archived there. A lot of Twine work is hosted individually and could vanish again at any time; likewise a lot of individually created projects. I’m not sure what we should be doing about this, though possibly more advertising the existence of the IF Archive to creators outside the community would be a good step.

Focus on authors and on player experience: these feel to me like issues related to the distance between academia and actual practice (by the IF community, by indie devs, etc). This is something I talked about a good bit while I was at the conference: a number of researchers lament the difficulty of finding authors with an interest in procedural or interactive aspects of storytelling, but have not been in the habit of reaching out to communities where experienced IF authors could be found.

Likewise the issues about thinking about player response. When I gave my talk about Versu and Blood & Laurels, I framed it by explaining that B&L was a commercially released app, that I hadn’t done academic-style player response studies but that I was going to be talking about some feedback that I’d gotten from reviewers and players of the game. I knew that this wasn’t how things are usually handled in this community, but even so I was surprised to get several questions afterwards from questioners who seemed taken aback by the whole concept: basically, that the process of releasing a game to the public, and seeing it reviewed by individual players and by publications, was so unusual that they actually did not understand what I was referring to.

I think this is a problem. There’s much to be said for rigorous data collection about specific aspects of what players think or feel about your game; the work done by Juhana Leinonen and Aaron Reed in analyzing transcript and other IF-related input is really cool and I wish we could do more of those sorts of things. It might be interesting also to do some survey work into particular aspects of people’s IF experiences. But the flip side is that with a survey you get back information of the kind you were looking for, and you can often skew that information pretty hard (intentionally or not) depending on how you ask the question. Open-ended reviews and feedback are harder to control but much more likely to return types of information that you didn’t know you should look for.


Subsequent discussion. Koenitz also publicly discusses his manifesto with Chris Crawford in several sections here: 1 2 3 4 (so far — it may be that they’ll post more of these).

In the discussion, Crawford writes (predictably, for those who have been following his work):

The whole boolean approach imposes an overly simplistic black and white mentality on the infinitely subtle processes of drama. Michael Mateas recognized this with Prom Night and used 4-bit numbers to obtain greater resolution, but I think we need to go all the way to floating point… The fundamental thing that both sides miss, in my physics-centered weltanschaung, is the necessity of modeling character interaction using numeric algorithms.

Here’s the thing I found with Versu: yes, it is useful to have character reactions and mutual evaluations modeled numerically, to have a sense of accumulation and of actions having large and small effects. It’s useful to measure several different axes of affinity between characters, because there are frequently events that need to be gated on more complicated criteria than whether people like one another. A lot of drama comes from internal conflict, and having warring feelings about another character is a good source: the person you find charming but don’t trust, the person you dislike but grudgingly admire, etc.

It’s also useful to be able to scale, for different works, how much characters are affected by a particular action. If you’re writing a fifteen minute comedy piece, there’s not much time for character feelings about you to accumulate, so it’s important to make big changes and drive exaggerated reactions, whereas if you’re writing a two-hour epic, you want to move emotions in smaller increments, to show a relationship evolution over multiple scenes. (This is one of many reasons there’s no simple one-size-fits-all library of actions and interactions that will be able to serve all drama models. And no, I don’t think a simple multiplier would resolve it.)

But the numbers and thresholds by themselves are not enough to produce satisfying drama or a sense of agency. You also need to communicate to the player what the current status is, and when the characters are close to one of those thresholds, and which if any actions are likely to have some effect. “George is .952 angry” is not story content, and “George is very angry” is only somewhat better. You really want, if possible, “George is so angry that one more rude remark will push him over the edge.” That puts stakes on the player’s activities and gives them a context for their own character’s actions. It may feel as though the player ought to just be able to sense how the interaction is going, that it gives too much away to say “George is nearly ready to punch you”; but in a system with so many moving parts, being explicit is preferable to being vague if we want the player to have any hope of acting with intention.

As for a sense of drama, a lot of that comes not only from what we see happen, but from what we apprehend might happen. In the case that George is very angry but winds up calming down and not punching the player, there is dramatic charge only if the player knows that the punching was a possibility, in which the player character deliberately backs off from prodding him or perhaps even attempts to pacify him. If the player doesn’t know about the implicit threat, then it’s just a scene in which seemingly not much happens.

I don’t claim, by the way, that in our released Versu work we have entirely got this down. But these are my conclusions from the work we’ve done so far, and this is something that I talked about in my own talk about what we’ve learned from Versu so far. One reviewer mentioned not understanding why a certain character was present in Blood & Laurels because it wasn’t clear how that character affected the plot — meaning that, without realizing it, he’d avoided triggering that character to betray him. If the player had been conscious of that jeopardy, presumably the relationship would have been more interesting and its plot relevance clearer.


There’s a lot else in this discussion, much more than I can take on here — about whether progress can be made by community action or by individuals working on their own visions, about how and whether we ought to be trying to relate our work back to existing narrative and dramatic forms, and about the education of a procedurality-literate audience.

4 thoughts on “ICIDS: The future of interactive storytelling, plus some Versu thoughts

  1. On your last point, about being explicit, a principle from indie roleplaying that I think is related is the idea that while character secrets are interesting, _player secrets_ that aren’t shared with the rest of the group are disruptive and kind of boring for everyone else. So instead you make them visible and let people react to them as they like.

    Yes, we’d like to be able to replicate the told-to-you media effect of a sudden reveal, but maybe it’s just as good to build up anticipation for the event beforehand. I see it as being related to the fear of spoilers. My personal inclination is to aspire to build stories that are interesting even when you know the outcome. (The most difficult choices in Fallen London, to give an example, are very likely to have the consequences be pretty explicit. It’s the choice itself that’s hard.)

    It occurs to me that it’s also a way to deal with the difficulty of foreshadowing. You can’t determine exactly what will happen in the future, but you _can _ spell out the consequences of the player’s choices. Instead of a narrator composing a plot for cause-and-effect, we substitute the consequences of the system which naturally has cause-and-effect.

  2. Pingback: Chris Discusses Interactive Narrative Issues with Harmut | Objective.Me

  3. Pingback: Tightening the World-Plot Interface: or, Why I Am Obsessed With Conversation Models | Emily Short's Interactive Storytelling

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s