The Value of Story

Posted by on Nov 9, 2009 in Conventions, Story | 9 Comments

This past weekend, I was a guest at Nanocon, a convention put on by the gaming club at Dakota State University. As part of the gig, I gave a presentation and did Q&A, along with fellow guests Richard Dansky and Chris Sims, on Friday afternoon. Many of the people in the audience were DSU game design majors, but from what I heard afterwards, there were plenty of con attendees and non-majors in the audience as well.

My relatively short presentation was about the value of story, speaking not in waffly vagueness, but in precisely measured dollars. When you’re a young designer, especially, it’s easy to assume that the factors that make a game successful and good are its flashy elements: powers, polygons, pixels. But focusing on those things is not only expensive, it doesn’t necessarily lead to a good game.

The “hard data” I used to try to impress the importance of story came from the Significant Objects Project, which completely fascinates me. The project buys essentially worthless bric-a-brac from thrift stores, presents each bit to a pro writer, the pro writer writes a short piece related to the object, and then the object is sold on ebay.

In this way, a $1.00 ashtray is sold for $17.79, a $1.00 creamer-pouring porcelain cow is sold for $26.00, and a $0.75 jar and brush for basting meat with barbecue sauce is sold for $54.00. And this return on investment comes even though the fiction is completely exposed: The ebay auctions bear a disclaimer that “[t]he significance of this object has been invented by the author” (emphasis mine).

As of last week, about 80 bits of written context had turned $112.02 into $2,857.22.

I say “context” rather than “story” because the Significant Objects writings aren’t necessarily even narratives. Some are, but one of the things I wanted to impress on the students at DSU is that it’s possible to make a game better in this low-investment way even without even going so far as to impose a narrative on the experience, but instead, just by creating an in-game context for the game’s elements.

Jonathan Blow, the creator of Braid, who’s gone on record saying that story is at odds with gameplay, turns out to be a master at this weaving of context. When I played Braid, I was blown away (already knowing Blow’s take on the utility of story) at the emotional punch he had packed inside his platformer.

During the Q&A at DSU, Richard Dansky told a story about some dialog he wrote for some guards in Far Cry; one guard tells another guard about growing up in New Jersey and going to the beach. Some Far Cry players approached Dansky at a convention and told him that once they overheard that guard’s story, they went out of their way not to kill him when they advanced on his position, because they were from New Jersey, and they had been down to the beach. That guard was one of them, and no way were they going to kill that guy.

I don’t care who you are, you just can’t accomplish that with polygons or pixels, no matter how much you spend on programmers and painters.

Hi, convention organizers! Keep Gameplaywright in mind when you’re planning the GoH list for your upcoming event! We love to talk gaming and meet gamers, and you know how to get in touch with us…


  1. Tyson J. Hayes
    November 9, 2009

    I have been amused by the Significant Objects Project when I heard about it a while back. It doesn’t really strike me as odd that we buy into stories and are willing to spend more money for them. If my wife would let me I would probably buy more kitchy objects just to tell stories about how I got them (which would likely be made up at the time). She has this thing called taste that always ruin these plans.

    The Far Cry bit intrigues me, why do we as people get attached to characters like that in a video game? “I can’t kill that guy cause he’s been to Jersey.” What attachments were made in that instance? My hat is off to the designers though that level of immersion is something that should always strive to be obtained.

  2. Adam Drew
    November 10, 2009

    There’s a background character in Splinter Cell 3 that tells another character about how his father (presumably a communist rebel in some South American country) was killed by an American raid in the night, with the US troops syncing their gunfire with the lightning, so he always worries that he will be killed in a thunderstorm. It is, of course, storming, and as Sam Fisher, you need to get past him, yet he has a clear view of the walkway you need to go down.

    I’ve played the game through 2 or 3 times, and I always make sure to spare him, knocking him unconscious and then dragging his body out of the rain, both out of compassion and some kind of wistful irony. I like to think that he wakes up, goes home, finds someone of the appropriate sex, and settles down, living happily ever after.

  3. Will Hindmarch
    November 27, 2009

    Adam, if you dig that bit in Splinter Cell: Chaos Theory, you might like this post from my home blog, including an article I wrote for The Escapist.

  4. Trilly Chatterjee
    November 29, 2009

    Jonathon Blow’s comments on the story/game divide seem like a fairly typical developer perspective on the issue (not that I’m blaming him, it is a toughie).

    I can’t help but think, however, that its not so much the intrinsic needs of game and story that are in conflict as the time/resource pressure and commercial risk/reward ratio issues that developers repeatedly face when making decisions about trying to integrate narrative more deeply with games.

    I think the majority of developers still see narrative as this adorable yet incredibly demanding puppy that games have – nice if you have the time and resources to look after it, but that will, left to its own devices, scratch the shit out of your furniture and poop all over your game design process.

    As an example, I recently attended a panel on game design at GameCity (Nottingham, UK – October just gone) with some noteworthy developers, indie & mainstream (though none, if I’m honest, with any track record in creating story-based experiences) and when I put forward a question on how the evolution of narrative and games will go in years to come, there was basically stunned silence, then mumbles to the effect of ‘too difficult, can’t really be done without great effort’.

    Okay, I hadn’t picked the best panel for a more comprehensive answer than that, but it did make me think that attitudes towards how games and narrative could be more meaningfully integrated (or even support each others more glaring faults and weaknesses) are still very much in what Noah Wardrip-Fruin termed the ‘static media’ age.

    I understand and to a degree sympathise with this outlook, but think (at least computationally speaking) that its more about working smart than working hard.

    As someone who will always be a role-player at heart, I still believe there are better, more meaningful ways for story and games to get along – particularly ones that take advantage of computers’ unique procedural and information processing capacities to create experiences that account more readily for spontaneous player actions within game narratives – serving to emphasise the fact that their actions have meaning, rather than imposing unnecessary constraints/demands on players and developers.

    But in order to reach that goal I think the following things must first occur:

    1) Game designers’ definitions of what does and does not comprise narrative in non-static digital media have to be both relaxed and refined – so that their definition does not assume beforehand that narrative is of necessity a ‘fixed’ entity, or consists solely of plot-related elements.

    2) Serious effort has to be put into developing situation-based context modelling systems. Such systems should be able to identify, track and accommodate (a meaningful proportion of) the manifold choices players make during the game – their impact on the story world or current narrative thread etc. – with a view to making many subtle (as opposed to few significant) alterations to the game experience on-the-fly as play proceeds.

    Here I’m talking about the ability to compute with story-related semantic elements at a fundamental level – not just how things work, but what they *mean* (this is not as absurd as it sounds).

    3)With a context/situation based approach in place, designers/developers have to work more closely with authors (and ideally those at all levels of procedural literacy, not just the techie ones) as technical facilitators, allowing them to develop more interesting and versatile narrative structures that take advantage of the uniquely procedural facets of game design technologies.

    At this point, I should admit that this issue is fairly close to my heart. I tried my best within the scope of a Masters dissertation to find some way to deal with these issues. Though the dissertation itself was a failure, I found a strong computational basis for the kind of systems I describe above. It needs work (and someone of much greater technical skill & understanding than me to implement it), but the academic foundations for it are there (and have been there for about 15 years).

    I really don’t think this is impossible. I certainly don’t feel its as difficult as most designers/developers seem to think it is. They just have to let go of a few preconceptions, and adopt a more situation-oriented approach to narrative.

  5. Jeff Tidball
    November 30, 2009

    I’d love to read more about your dissertation, Trilly, and the 15 years worth of basis for that kind of computational story-modelling. Given a random year, I vacillate to both extremes on the question of whether it’s possible to rival in interactive fiction even a modestly well-written novel, or well-filmed movie.

  6. Trilly Chatterjee
    December 2, 2009

    I don’t think trying to replicate the experience of either of those things is what developers should be aiming for. I think that trying too hard to do so is actually a large part of the problem.

    We’re so used to taking the tropes of our old media with us as we begin to explore newer media, that we’re blinkered to the more novel and innovative storytelling possibilities provided by those new media.

    I mean, seriously, most games still rely on something not much more sophisticated in structure than the ‘Choose-Your-Own-Adventure’ book format – if they even bother with non-linear storytelling at all (and that isn’t to say non-linearity is the be all and end all of storytelling, just that its one way in which computer games can clearly separate themselves from other media).

    Also, with regards to the ‘academic foundations’ I mentioned before, I’m referring to the ability to incorporate deeper, richer and more author-driven contextual information at the level of game & AI processes.

    See it as the potential for a richer palette of ‘game states’ and ‘inference rules’ about a game world and its characters for AI within the game to work with.

    How much richer? My gut says ‘as rich as you like’. The writing process would be much like ‘coding’ potential scenarios in terms of their individual components (people, actions, times, places) and various outcomes. This information would be integrated into the game world via its AI, graphical filters, camera placement, dialogue, animation etc.

    To go all Tolkienesque on it – “One unified information structure to rule them all…”

    So, for example, what if a game could keep track not just of what you did, but *how* you did it – appropriate to the context in which the action was carried out – and have references to the nature of this action pop up at appropriate points later in the game?

    That’s something we’d like to see more of in games, isn’t it? A way for them to see past our in-game avatar (the ‘who’) to *us* (the ‘how’) – to feel a genuine sense of participation, of consequence, of meaning to our in-game actions…

    And to clarify, the above won’t guarantee “good” storytelling, the same way having a good thesaurus/dictionary handy doesn’t guarantee good writing. In the end, it all depends on what is done with it.

    Perhaps the most daunting part of the whole exercise is that truly exploring the possibilities of computer games as a storytelling medium requires writers to become procedurally literate. Much in the same way we programmers for animation and digital artists, we need writers who can think in terms of ‘expressive processes’ (another Wardrip-Fruin coined term, he recently published a book on this one, though. Check it out).

    I guess my initial point was that the notion of ‘story’ as it relates to computer games won’t gain the sophistication we appreciate in other media until they are genuinely explored in terms of the unique possibilities of the ‘computational’ medium.

    Until then, *any* computer game story (even the ‘best’, by the standards of other media) will just seem like a pale imitation of a movie or novel.

    And sorry for not making my posts shorter. You blog on things for which I have far too much to say. Your bad.

  7. Jeff Tidball
    December 3, 2009

    In talking about rivaling novel and film story experiences in games, the thing that concerns me most is whether the emotional experience of reading or watching a well-crafted story is possible to replicate when the very structure of the story is open to shifting.

    I think you’re more than right to point out that there’s only so much value in assuming that game stories will, must, or should accomplish the same things that stories in other media do.

    I’ve got a copy of Expressive Processing on my to-read shelf; the publisher sent over review copies to Will and I when it came out, but neither of us has managed to get into it yet. The possible advantages of writer literacy in these areas seem massive, but the challenge to working writers—learning the tools and experimenting with the form while trying to earn a living—is also monumental. Especially given the massive unknown of commercial applicability. I guess that’s why we have universities. (At least, for a little while, yet.)

  8. Trilly Chatterjee
    December 5, 2009

    I understand your doubts about whether truly compelling stories can ever emerge in an ‘ad-hoc’ manner, or with a truly dynamic structure. However, you’ve role-played, so you know that these things can be achieved by human beings.

    Of course, I understand that with role-playing the act is primarily imaginative, and less subject to practical constraints. My point in making the comparison is to say that if we can achieve it, then the question is not ‘if’ but ‘how’.

    I think that the shift in emphasis from the former to the latter (for writers, researchers and developers) is crucial, and will determine whether what emerges in years to come will truly be comparable with great works in other narrative media.

    As the saying goes, we could argue until we’re blue in the face about whether or not it’s *possible* to craft compelling stories that, for example, accommodate a high level of user agency. The bottom line is we’ll never really find out until some time and effort is invested in thinking about how this might be achieved, and experimenting with what does and doesn’t work.

    I think part of the solution is to have people who understand the technical potential of these emerging systems acting as guides for writers who wish to explore what can be achieved with them. As you rightly say, this is the reason we have universities.

    Realistically, there’s no way these things could be achieved single-handedly anyway. There has to be some kind of dialogue, and experimentation, and testing, followed by tweaking and more testing. It has to be an iterative process, where the impact of what results can be frankly qualitatively assessed.

    In the first instance, however, efforts have to be geared toward a writer’s understanding and vision. Good writers know what makes a good story. When the story parameters shift, so to speak, a good writer can tell you why a story is more or less compelling as a result. This kind of knowledge is what needs to be encoded and interpreted by computers (to the extent that it can be).

    The all-purpose information framework I mentioned could serve as a language for encoding such information. It will allow writers/developers to define and relate sophisticated concepts in a simple and comparatively elegant way to the methods currently used. These structures could then be used to determine inference rules about what is or isn’t possible/desirable in a given context – applicable in any situation which presents that context.

    The writer’s vision/understanding of a story and its components are part of what this ‘language’ would convey to the systems responsible for presenting and altering the story experience.

    Please tell me at least some of that made sense. 😛

  9. Jeff Tidball
    December 7, 2009

    Yes, indeed, plenty of sense!


Leave a Reply