Drama Management with Anchorhead-based Test Case

Manu Sharma, Santi Ontañón, Manish Mehta, and Ashwin Ram are publishing some research into drama management using a simplified version of Anchorhead (with a choose-your-own-adventure interface rather than a text parser).

The full paper is worth a read (though fairly technical), but the gist is that they proceeded by taking a series of cases from players, determining which players liked which subplots. They then designed a drama manager that would compare the current player’s behavior against its cases, determine which subplots this player was likely to be most interested in, and hint the player in the direction of those subplots. The result appears to be a better experience especially for non-gamers, though some players (especially the more experienced ones) disliked being over-hinted in the direction of things they would enjoy.

When no confident predictions can be made from player predictions, the drama management model falls back on author-defined rules about what to present when.

I have a bunch of minor quibbles with particulars of the study, but found the conclusions intriguing.

Hall and Baird on Polti again

Have now had a chance to read more thoroughly the article I mentioned last post, Improving Computer Game Narrative Using Polti Ratios, by Richard Hall, Kirsty Baird.

The idea of the article is that the amount of drama in a game can be arithmetically calculated by counting the number of different Polti situations one can recognize in the game, then calibrating against the total number of events and the number of major and minor characters required to bring these events about. They further use these calculations to argue that games with no characters (“unembodied” games) may not reasonably be interpretable as stories. They propose that games in production should be improved by restructuring the story to include more Polti situations and to decrease the number of characters.

I’m in general pretty skeptical of anything that claims to be a recipe for narrative construction, but some are still more plausible than others. This one strikes me as especially doubtful, since the natural limit of the proposed equations is a story in which all 36 Polti situations are included in a tale with just two characters and a minimum of distinct events. A sense of fitness in the narrative development, thematic coherence, etc., do not enter into the equations.

The authors offer a sample revision of a game which does demonstrate concern for thematic coherence in that they eliminate a certain number of Polti situations at the outset, and they do arrange the game to convey a specific message; which looks like an acknowledgement that good narrative is characterized by something other than an optimization of their proposed equations, but they do not account for this theoretically. Nor does the article attempt to distinguish between narrative, story, and plot, though these terms are often treated as distinct in other game/narrative literature.

The most interesting part of the article to my mind is the part that suggests that story depends on how many situations the player recognizes:

We’re going to… initially propose the theory that if people can explain what is going on in terms of at least one of Polti’s units then the object that they are engaged with can be labelled a story.

Even this is a little vague, but let’s assume that by “people” we mean “the player”. This would suggest that a game’s emergent behavior begins to appear as emergent narrative iff the player can recognize the action as corresponding to some fundamental narrative element. But the article raises an important question at the end:

Do people in general (with an understanding of Polti’s units) detect exactly the same units in the same story?

This becomes even more pointed with something like The Sims in which the interaction of characters might be construed to be “about” any of a number of things: one can observe the apparent emotional content, but there’s no verbal dialogue. (This may become more complex and richer in Sims 3; I’ll come back to this point when I’ve had a chance to play the PC version rather than just the cut-down iPhone edition.)

Anyway: the point is that emergent behavior might be construed by the player as corresponding to different narrative elements depending on the kind of arc the player already thinks he’s building. (See: Alice and Kev.) On the other hand, the game has no way of “knowing” which situations the player is recognizing as narratively significant and thus no way of trying to produce episodes that follow (thematically or causally) on those the player has already recognized.

I wonder whether the process of narrative building would tend toward more coherent structures if the player’s interpretation were actually polled and then used to refine the character behavior model.

Whether Polti’s situations are remotely useful for this purpose, I’m not sure. But that’s a separate point.

More on the emergent narrative issue

Thinking about this more, and I ran across an article at gamestudies.org from September of last year, which seeks to quantify the narrative density of a game by how many Polti premises one might recognize during play.

I’m not sure I find this convincing, because it suggests that narrative comes in discrete chunks. What about pacing, development, arc?

But I’ve only had a chance to scan through the article so far, and need to come back to it later.

Sims 3 (Mobile version)

Lately I’ve been thinking more about emergent narrative — in particular, the idea that a sandbox-style game can produce elements that the player then weaves together into a story that he finds satisfying. The story isn’t really a product of the game, and it’s not necessarily true that anyone else playing the game would perceive the same story. The onus is on the player to determine which of the many otherwise insignificant events contribute to the narrative.

I’m pretty skeptical about this idea. Or rather: I can see that some such thing does happen, in that lots of (say) Sims users construct elaborate stories with their characters, and share movies and narratives. But in general this is not what I would call interactive storytelling; it seems more like handing the user a really complicated dollhouse that happens to have built-in tools for recording and editing the best scenes.

Still, I thought I ought to put a bit more research into this topic before I dismiss the possibilities. It’s been a while since I had (and got tired of the grinding aspects of) the original Sims, so I tried downloading Sims 3 for my iPhone.

Alas, I find it a dead bore.

Two From the List

I’ve been playing with a few things on the enormous to-be-played list from a few days ago, though of course there are so many things on there that it will take quite some time to get through.

Braid: I have now finished. It’s of course a masterpiece in the game-play area, and doesn’t need me to say so. I usually have a really hard time getting through platformers, since I don’t have the right combination of patience and skill. As advertised, Braid minimizes the amount of frustration involved in playing a platformer while offering a diverse range of genuinely interesting puzzles; and while there were still a handful of these that were too finicky and that I would have been happy to skip, nonetheless it was the best time I’ve ever had with the format.

As far as storytelling goes, it’s a bit more confused: there is some really intriguing material in the final level, in which the meaning of events is revised and reinterpreted in a way that naturally connects to the gameplay itself; and even before that point, many elements of the game are framed so that the play is metaphorically significant.

But what I get out of all that is not really a story (good luck finding two people who even agree on what happens in Braid) so much as a series of meditations on some of the common problems in relationships and self-definition. Some of it’s thought-provoking, some a little on the obvious side. Admittedly I usually find this kind of content under ask.metafilter’s human relations tag rather than in a game, and I’m generally encouraged when a game branches out to incorporate new material. So hooray for that.

Nettestadt Troll was recommended to me as an example of good Ren’Py work, and I’m afraid I didn’t get nearly as far with that. The premise is uncomfortable to start with: girl gets abducted and raped but discovers she kind of likes it and/or falls in love with her captor. This is a fantasy to be found in many forms of literature from Menander to a certain genre of 1970s romance novel, but it’s something that would need to be handled with a fair amount of psychological sensitivity in order to be a story I want to read. Otherwise, what you have is basically porn for a specific audience.

I wasn’t crazy about the art or the prose quality, either, and the pacing left me kind of bored during the first few chapters; as for the world-building, it’s extremely vague and careless, featuring both alchemists and telegrams, feudal hierarchies and shops with “receptionists”.

I did stick with it for a while, though, in case this was a case of poor writing craft combined with a strong storytelling sensibility. Unfortunately, once there started to be choices to make, they were often on the level of random and incidental choice: e.g., what dish do I make for my supper while waiting for the troll to come home and rape me again? It’s a bit less inane than Dream Day Wedding, but the choose-your-own whatever aspects display the same lack of significant agency that I complained about there.

Conversation Methodologies

My latest Homer in Silicon column is a bit of a departure from the norm: instead of offering a critique of a game or set of games, I discuss conversation modeling methods, in an attempt to share some interactive fiction theory with a wider audience and to encourage more discussion about conversation modeling in general.

ETA: there is some further discussion of the ideas at TIGSource.

Peeve

People who say “X is impossible in games/IF” when what they mean is “I do not know how to accomplish X in a game/IF”.

Extra demerits if X is in fact something that is already done in various places but the writer of the assertion has failed to do basic research.

Types of Action and Types of Agency

I’ve been thinking again about actions and how they’re expressed, and how the communication of an action relates to player agency.

Let’s say, for now, that agency is the player’s ability to affect the world and story, and it depends, in turn, on whether the player can form a reasonable guess about the results of an action before taking that action. If the player cannot guess or does not care where the action will lead, there is no agency; the player is providing the energy for forward motion but is not meaningfully steering the work.

The thing is, different kinds of actions are themselves susceptible to different degrees of agency. We’ve been moving slowly (and with varying degrees of success) away from having all IF commands be of the sort appropriate to physical action.

Hypothesis

If we consider the types of agency involved in different kinds of action in the real world, we will be able to come up with better ways for the player to command non-physical actions.

Information Flow and Gradual Characterization

Ran across an interesting post from Krystian Majewski on design problems with multiple choice dialogue, which, among other things, draws on some play-testing reports on Emerald City Confidential:

[Playfirst] did a user test of the indie point-and-click adventure Emerald City Confidential and they described how casual gamers reacted when they first encountered a multiple choice dialogue. You might think that point-and-click adventures are a good match for “casual players”. Well, when faced with their first multiple choice dialogue, most players simply froze in panic. They assumed that one of the answer as “correct” while others would lead to failure. From the kind if information they received, they couldn’t really anticipate what would happen. Even worse, after they decided, they didn’t receive a clear feedback on what effect their choice had. They were used to the transparent feedback schemes of most casual games and weren’t able to cope with the uncertainty.

I found this really interesting, because I would have assumed, in general, that a simple multiple-choice presentation would be more accessible to casual players than some other mode of interaction.

Majewski goes on to argue in favor of consistent verbs representing standard strategies usable at every dialogue point. At any given time (for instance) the player might have the option to respond intellectually, sensually, or in a religious way — a strategy that reminded me a little of the dialogue system in Forever Always.

I’m not sure how much application this has from an IF perspective. Certainly the core IF audience tends to have different expectations about how dialogue will work, so is probably not completely frozen by a menu. Moreover, it’s possible to write menu dialogue in such a way that it gives more of a hint about the likely effects; the writing and cluing of the player determines, to a large degree, how much agency it feels like one has. At the same time, I generally agree that menu dialogue feels more distancing (to me) than parser-based dialogue, even if the latter is heavily clued with hints about what the player can say.

One of Majewski’s commenters adds:

Real dialogue flow should be such that no specific choice matters too much, but the sum of the choices does. But as you’ve mentioned this is just a nightmare from a design perspective.

I’m not sure it has to be so bad a nightmare as all that, and it’s been done. To pick the obvious example, in Blue Lacuna… (very mild spoilers, and discussion of a WIP of my own, follow the cut)

Expressive Processing

Over on Grand Text Auto, Noah Wardrip-Fruin is running an interesting experiment in peer-reviewing: a blog-based peer review. Visit the site to read portions of his text and comment on them.