Chris Crawford’s Unslain Dragon Of Interactive Storytelling

Chris Crawford is a contradiction: mythologized for his bold vision for the future of games; criticised and dismissed for his lifelong failure to accomplish that bold vision. His numerous critics see him as out of touch and over the hill. A washout. An unrepentant failure. Someone who should give it up and walk away. But he can’t bring himself to do that. He has a dream, and he hasn’t finished laying the groundwork for someone else to carry that dream forward. Crawford is driven by a singular vision — by an idea that he’s pursued doggedly since he left the games industry over 20 years ago. And even now, as I speak to him about Siboot [official site], his latest attempt to spruik his dream of character-driven interactive storytelling, he remains tormented by a dragon he knows he will never slay.

The famed designer of Atari 800 wargame Eastern Front 1941 and 1985 geopolitical thriller Balance of Power (a game in which your goal was to prevent war from breaking out) left the games industry in 1992 during a speech at the Game Developers’ Conference — the yearly meeting of minds that had begun in his living room five years earlier. In a talk immortalised as The Dragon Speech (you can see it in five parts on YouTube), he decried the narrowing breadth and growing monotony of commercial games.

He spoke of his dream for games that would express the full breadth of human experience and emotion. And he called for games that combine the interactive immediacy of play with the dramatic punch of a great story — not as separate entities that you swap back and forth but as one and the same. (Then he charged out of the room screaming at a metaphorical dragon that he swore to one day fight for these dreams.)

His latest effort at fighting that old foe attempts to rectify the biggest mistakes of Storytron, which was an over-ambitious, overly-complicated unusable mess that even he could scarcely manipulate to his storytelling whims. This new thing is called Siboot, which may sound familiar to the two of you who remember Crawford’s well-reviewed but commercially-disastrous 1987 relationship sim Trust and Betrayal: The Legacy of Siboot. The new Siboot borrows the world and premise of the old one — figure out whom you can trust and whom to betray by judging what they say against what’s visible (or subtly hidden) in their facial expression. But Crawford has rebuilt the whole thing from scratch with a new design that cherry picks the best bits from Storytron and from his decades of research.

“With Siboot I told myself I’m going to strip [Storytron] down,” Crawford says. “Any complexity that isn’t absolutely necessary gets ripped out. I got the balance much better now, with a much simplified system. It is much easier to work with; it’s easier to make design decisions with it. It is much much easier to write the algorithms that control the character behaviour.”

The key to Siboot’s supposed power seems to be that its story is run through algorithms. Its key moments are not written in any traditional sense, but rather designed. Big whoop, you say? There are lots of games in which the story emerges organically according to your decisions and those of your artificial opponents — take Crusader Kings or Civilization, for instance, which both provide complex systems upon which all sorts of interesting decisions can be made. But Crawford argues that these sorts of games don’t count because they “are not very dramatically interesting stories.”

“It might be fascinating to the person who experiences it,” he says, “but ultimately we have to ask does it really approach the quality of what we normally call a story? And the answer with almost all games is, well, no.”

It’s a bold, somewhat dismissive statement that appears at face value to trivialise the achievements of game developers over the past two decades. And it’s likely to immediately get some people on the defensive. But to do so would be missing the point. Crawford doesn’t say this out of malice, and he no longer says it out of ignorance — despite past comments that he’d stopped playing games entirely, he now pays attention to what’s coming out and reads about or watches videos of many new titles (although it’s still rare that he plays something himself).

Games today are incredibly broad and varied. They tackle themes such as poverty, depression, love, war, duty, regret, longing, and death in myriad ways. “The indies,” Crawford says, “are the best thing that have happened to this industry in a long, long time.” The proliferation of cheaper, better tools and more efficient distribution systems has brought with it an explosion in creativity in terms of both development and criticism of games. It’s brilliant, he agrees, and many fascinating ideas have come out of it. But for Crawford it’s not enough.

“The types of changes that I am observing are really fine improvements on fundamentally ancient architectures,” he says. “I can respect the way they took this classic architecture and rendered it for modern technology and modern capabilities, but deep down inside it’s still the classic design. That doesn’t make it bad. Shakespeare is still good today, even if Hamlet is using a machine gun instead of a sword.”

Crawford feels there is something very specific yet vitally important that remains absent from videogames past and present — something that he believes would bring with it a paradigm shift in game design. To explain, he turns to a concept from psychology: mental modules. It’s a theory that suggests there are innate functional structures in the human brain dedicated to different processes.

“There are a lot of different ways of slicing the pie up,” Crawford says. “But five of the most important modules are motor module for musculature [think reflexes and dexterity], visual-spatial for navigation and recognising 3D worlds, cause and effect — our knowledge of realistic things. If you drop something, it falls down. If you put your finger in a fire it burns your finger. Those types of relationships. Those are the three types of modules that games most commonly challenge.”

Games rarely challenge the two modules that Crawford is aiming for with his interactive storytelling. These are language and social intelligence. He argues that the latter is what other forms of entertainment are based on. “You don’t jump around in your seat in the movies,” he says. “They don’t train visual-spatial — in the movies people don’t spend a lot of time navigating their way through mazes. They just say I’ll go there and the next shot is them there, not traversing it.”

Movies and games are different, of course. They have different strengths. Movies show you people making big, life-altering dramatic choices. Games let you make innocuous, inconsequential choices, interspersed from time to time with meaningful ones. Movies are driven by relationships; games are driven by you, the player.

We as an industry tend to celebrate that about games. We revel in the little things and the freedoms that games give us. But Crawford wants a game where there is neither an open world to live and explore in nor a path laid out before you. He wants something that provides the narrative focus of a good novel or film coupled with the agency of a game, without any pre-authored story branches and with complex characters who interact with one another independently of you but who each have a definite role to play (like an antagonist or mentor).

I suggest that an MMO such as EVE Online might challenge your social intelligence and be filled with interactions that could each potentially be dramatically significant, but he waves it away as too unfocused. “There’s no such thing as a supporting character in a multiplayer game,” he says. “Everyone wants to be the protagonist. In interactive storytelling you create artificial characters who you deliberately establish as supporting characters, primary opponents, antagonists, and so forth.” Crawford wants an artificial agent — an algorithm — to be pulling the strings behind the scenes.

His second-in-command on Siboot, art lead Alvaro Gonzalez, explains that an MMO is akin to a social intelligence sandbox whereas Siboot is meant to be a forerunner to a “dramatic sandbox where the most important thing is to have control of the story, even though the player has total freedom on how that story develops.”

Gonzalez explains further with an example of how this model of interactive storytelling might be applied to role-playing games: “At the moment, a lot of RPGs don’t have meaningful quests,” he says. “Like *booming voice* ‘you have to kill 40 pigs to get 1000 tails to take it to’ — that’s not meaningful because the player doesn’t have to use any module to get that quest. It was easy. But if the player has to make meaningful decisions using this social intelligence while they approach these NPCs, then that quest is going to become very important and very meaningful, and that will breed dramatics.”

Essentially, if I understand him correctly, RPG quests would become meaningful because there’d be fewer of them, their goals would involve an intrinsic motivation on the player’s part, and they’d take a whole lot of talking, conversational probing, and emotional investment to get in the first place.

What it really comes down to is that, in Crawford’s vision of an interactive story, your every choice could change the course of the narrative. Branching paths aren’t written so much as generated, at every turn altering the story to optimise the dramatic punch lying ahead.

Siboot is meant to prove the concept and kickstart a revolution in game design. Seen at face value, it doesn’t seem so different. Its claims of complex characters and symbolic language (with 40 to 50 verbs and several degrees of nuance) aren’t perhaps as unique as Crawford makes out, and more cinematically-minded games are already starting to weave in more subtle facial animations that you have to consider as you make your decisions (although Crawford plans to go further by having the micro-expressions be algorithmically varied according to multiple inputs rather than hardcoded from motion capture).

In fairness, Crawford is quick to admit that Siboot is painfully constrained in terms of this grand vision of character-driven, algorithmically-defined, social intelligence-focused interactive storytelling. “I’m sure many people will look at that and once they understand how it works they’ll say, ‘I can do better,'” he reasons. “That’s the whole point and purpose of this project.”

Siboot is on Kickstarter right now. It doesn’t appear that it’ll reach its goal, but it will be finished and released regardless — to be followed some time later by open source code for both the game and its underlying engine, the StoryWorld Authoring Tool. “I’m no longer interested in the money or the fame,” Crawford says. “I’m not rich but I’ve got enough money to be comfortable, so I don’t see any need to commercialise this project. The Kickstarter is really to raise money for the team members; it’s not for me.”

Siboot, meanwhile, is really to inspire developers to tackle storytelling from a new angle, to give academics something to play with in their interactive storytelling research (he’s actually been asked to keynote the next International Conference on Interactive Digital Storytelling), and to prove that Crawford hasn’t been barking up the wrong tree all this time.

Then, he says, “I can retire with the sense that my life was not a waste.” And maybe someone else will finally slay that dragon, stopping for just a moment afterwards to pay respect to Crawford’s failed attempts.


  1. Crimsoneer says:

    Before I read this, I have a question for the hive mind – was MASQ ever finished? Is it available anywhere?

  2. RagingLion says:

    Thanks for this. Having watched his Dragon Speech a few years back and a videoed conversation between him and Jason Rohrer after that I have half an eye and interest in whether anything of substance will come from his toil.

    • Supahewok says:

      I’m rather the same. Thanks for the article, Moss, I like to hear what Crawford’s up to every couple of years. Man’s a bit of an irascible coot, but he’s got more vision than half the industry combined. (including indies AND AAA)

    • Oozo says:

      If you’re interested in knowing more about him, google “Into The Night with Jason Rohrer and Chris Crawford”. It’s an episode of a really great series on Arte, where they pair up two interesting people and just have them talk through, well, the night.

      The whole episode is on Youtube, and it really makes for a fascinating, albeit sometimes almost depressing, portrait of two people with an interesting outlook on games. Highly recommended.

      • Oozo says:

        Ah, that should actually have been a reply to the comment just below. Obviously, you were referring to that very episode yourself. Nevertheless, for all the others who haven’t seen it: it’s worth checking out and readily available.

  3. cpt_freakout says:

    A great read. I had read about Crawford somewhere a few years back, but it was just one of those names that stuck with me; I couldn’t really place him or his ideas until now, and I’m glad for that, so thanks! Hopefully he can realize his visions, ’cause they sound very, very interesting.

  4. cristoffson says:

    “but ultimately we have to ask does it really approach the quality of what we normally call a story? And the answer with almost all games is, well, no.”

    I don’t think this one statement is wrong. I mean, I love stories in games, I have played many that tell great tales or do it in an interesting way, but have I played any game with a story as good as for example, Moby Dick? Fuck no. A portrait of the artist as a young man, The invisible cities, Life a user’s manual, The Aleph?. Holy gods, no, not even close. What I find interesting is that Crawford insists that this can and should be done with games, while I , and many more I think, continue to believe that stories are not that important in games. Again, they can be good, but I think they have nothing (so far) on the novel, or even cinema. Also, I’ve heard many times that games can make the player more involved in a tale than traditional mediums. In my experience this is completely false. Many books changed my life, destroyed my personality and forced me to rebuild myself, while no game, not even through systems, has ever managed that. Not even Papers, Please, which I thought was a work of genius when it comes to narrative in games. Maybe I’m just “bookish” and don’t get it. I still think it’s great that this guy keeps trying, though.

    • Bradamantium says:

      The difference is that the novel (and to a large degree, cinema) weren’t born as limited and commercial as gaming. It was chiefly a novelty, pure entertainment, simple escapism right out the gate, on top of being tremendously constrained – like if the novel had been preceded by a lengthy period of writing limericks because we hadn’t yet solved the problem of putting so many words on so many pages. Games have reached new narrative heights in recent years as they’ve diversified and tools have gotten into the hands of people that aren’t part of a team working under corporate mandate, so we’re edging ever so much closer to games that manage that depth and greatness.

      • cristoffson says:

        Yes, I guess I hadn’t considered a more historical perspective. Also the novel has had centuries to develop, and before that there were many precedents set in other types of literature. I guess it is a much more complex historical context.
        Thank you for your response, I hope we live to see some of those great works being released to the world!

    • malkav11 says:

      I also agree with that statement, but I think you’re eliding a crucial bit of context. He’s saying, as far as I can tell, that game-generated/emergent narrative, which is often touted as “the way for games to do narrative” over hand-authored narrative, is not actually interesting as what we would normally call a story in other media. I’d describe it as “anecdote generation”, myself. And anecdotes are fine and all, but they’re never going to be great stories.

      Maybe that’s not what he means. I can’t speak for him. But if not, it’s certainly something I would contend. I don’t think games are best served by trying to deliver hand-authored narrative the way other media would, either, don’t get me wrong. I think we can and should explore how to deliver that narrative in ways that utilize gaming’s unique tools like player input and the immersive quality of being represented in the game’s fiction, if only in avatar form. Her Story is a great recent example. But narrative in games is absolutely important to me – crucial, even – and if forced to choose I’d far rather have a well-considered hand-authored narrative delivered without exploiting the tools of the medium than settle for anecdotes.

      • ChrisGWaine says:

        His aim is game generated narrative that has more of the qualities of hand authored narrative.

        • Baffle Mint says:

          I’m curious how this works when so many of the techniques of modern storytelling rely on the fact that things aren’t random. Things like foreshadowing or dramatic irony rely on an author who knows where the story is going ahead of time.

          The other problem I have is that the stories in games with pre-written narratives are almost uniformly terrible. Portal is the only game I’ve ever played that I think rivals really good novels and movies in terms of storytelling craft. If video game creators haven’t really solved the problem of telling good stories WITH the tools of modern storytelling, I don’t think I trust them to tell great stories without those tools.

          • Arglebargle says:

            A lot of commercial writing is not that great. It may be popular, may make a lot of money, but a lot of it is going to end up on history’s Bulwer Lytton scrapheap. Other forms of entertainment media (films, games, etc) have a lot of even less competent writers calling the shots.

          • malkav11 says:

            I can name a whole lot of games with better pre-authored narrative than, say, The Da Vinci Code. Or Left Behind. And plenty that hold their own with decent movies and books. I’m not sure games as a medium have mustered something on par with the timeless classics of other media, but they’re also a lot younger and we’ve got a bunch of people involved with the medium actively fighting against the idea that they should involve narrative at all. I figure it’s coming.

          • Premium User Badge

            Ninja Dodo says:

            >so many of the techniques of modern storytelling rely on the fact that things aren’t random. Things like foreshadowing or dramatic irony rely on an author who knows where the story is going ahead of time.

            What you want there is likely something akin to the Left 4 Dead AI director, but for story arcs. Easier said than done though.

        • malkav11 says:

          Yeah, I got that. I was mainly agreeing that the results to date haven’t been impressive. Where he and I differ is that I think there’s a reason for that and that getting the game to generate narrative is a basically quixotic quest. But hey, if he manages to pull it off, more power to him.

    • Ed Burst says:

      If you compare games to novels in terms of what novels are good at, games aren’t going to win, any more than the Mona Lisa is going to beat Mario at doing what Mario is trying to do. It’s not a flaw in the Mona Lisa that it lacks a good difficulty curve and satisfying power-ups. You can’t expect a game to be more moving than a great novel, or a novel to be more interactive than a great game.

  5. kwyjibo says:

    It’s an interesting problem that Crawford has tried to solve, but given how he’s completely failed to do anything in the last 20 years, it’s very difficult to have any faith in him. I played with Storytron when it was released, an updated version of Balance of Power. It was fucking awful.

    There was nothing new or interesting in the game. It was essentially the diplomacy menu of a Civ game, each country would have their own goals, an internal state, and would horse trade with other countries to achieve that goal. And now Siboot looks like the same thing, only with pictograms. “Play with a limited number of cartoon AI actors” is not a strong pitch. I do not know how this remotely contributes to felling Crawford’s dragon.

    • kwyjibo says:

      As a measure of how little faith and excitement people have in this project, this is the only article on Siboot I’ve read. It has not been covered anywhere else.

  6. DrTalos says:

    It’s been a long time since I’ve heard the name Chris Crawford:

    “All this leads to a suggestion for what might work for women in games: social reasoning. The ideal game for women, according to this simplified model, would be some sort of interactive soap opera or bodice ripper, presenting the player with complex social problems as she seeks the ideal mate.”
    link to

    Yes, by all means, give this man your money.

    • Supahewok says:

      Yes, by all means, let’s cherry pick from an article that’s very nearly a decade old in order to set up a derisive one liner to condemn the subject on political grounds.

      Hrm. When I put it like that, it doesn’t sound very nice, does it.

      Why don’t we be honest and quote the 3 paragraphs before that?

      “At this point, I need to cover my butt against the picky-picky natterers who will point out that there are some men who are better at social reasoning than some women. Yes, of course that’s true. But social reasoning is the skill that most differentiates women from men. In spelling, arithmetic, cooking or any of a thousand other skills, women and men are closely matched, but when it comes to social reasoning, the advantage women enjoy is greater than any other advantage they possess. Thus, women are highly motivated to exercise and develop their social reasoning skills.

      We should therefore expect that modern women might well want to exploit this talent in their entertainment. And in fact that turns out to be the case. The classic female mass entertainments are the soap opera and the bodice-ripper. In each of these, women face intense and intricate social conflicts requiring elevated social reasoning. In every case, it’s the protagonist’s special insight into people that carries the day. It’s not the size of her breasts, how many antelopes she can kill, how many berries she can collect or how much money she can acquire. It’s her social skills that make her the heroine, the champion, the role model.

      To offer a prime example, consider Pride & Prejudice. Elizabeth faces the most delicate and difficult social obstacles in developing and weighing her marital options against her own feelings. Other women around her make mistakes and marry the wrong men, but Elizabeth navigates her way though the jungle of English society as well as the uncertainties of her own emotions to a happy conclusion: marriage to the ideal partner.”

      Gee, its almost as if he has REASONING for believing the things he does.

      I’m not taking his side on this subject; the man has studied more psychology than I probably ever will. I’m not qualified to weigh in. But I am very much against taking a quote out of context and sneering at a person for it. That’s scummy.

      • DrTalos says:

        I think I gave plenty of context to it, and I helpfully linked the article so you could see as much context as you would like. It is still deplorable.

        Yes, it’s been ten years, and if he wants to apologize for the article or recant his views, he can. To my knowledge, he never has.

        • Phasma Felis says:

          If you actually read through all that and your takeaway was “Chris Crawford thinks women are only interested in finding husbands,” I’m really not sure what to say to you.

          • Phasma Felis says:

            To elaborate on that: Yes, his conclusion paragraph is somewhat cringingly phrased, especially when taken out of context as you did. I think his actual intention is clearly more benign, and I dislike the practice of aggressively taking anything said in the worst possible light instead of giving the benefit of the doubt.

        • Spluff says:

          Taken out of context, the statement that the ideal game for women would be to ‘find their ideal mate’ does look quite alarming. But, I think it’s reasonable enough, if you read through everything he had to say. Following his argument, the only reason that this was included is because he is drawing on what he states are the most popular genres for women – soap operas and ‘bodice rippers’. I don’t know whether these are factually the most popular genres for women, but it is fair to say that the key plot elements in these genres are the romances. And so, to create a game ‘for women’ in one of these genres (because it follows that genres that are popular with women in other media should also appeal to women in games), the main storyline for the game would probably also be the romance – hence the ‘find the ideal mate’.

        • Hobbes says:

          Yet you seem to be quite content to use it as troll fodder and are unapologetic for that fact as well :)

          I’m just sayin’

      • death_au says:

        Thank you. I knew little of this and probably would have taken that seemingly sexist out-of-context comment at face value, but the reasoning behind it does make the issue of sexism less clear-cut and a lot less terrible.

        I mean you can still argue it’s sexist. No matter how much reasoning, he’s still dividing the sexes, but it’s the internet. Anyone can argue any side.

    • heartnotes says:

      Isn’t that what visual novels already are for guys and otome games for girls?

  7. phlebas says:

    That conversation-via-symbols mechanic reminds me a little of Captain Blood. Which is fairly promising.

  8. housellama says:

    I’m a student studying psychology and game design, as well as a life-long gamer. board/card, Tabletop and Live action RPG as well as digital gaming. I’m interested in developing games that do more than just entertain. I’ve read some of Chris’s books and I’ve dones some game design myself, albeit nothing in the digital realm that got beyond the tinkering phases. It’s a big challenge he’s taken on Siboot looks to be a decent shot at something new. I contributed to the Kickstarter, out of curiosity as much as anything else. I’m not expecting the game to be great, but it will be something different.

    Finding a way to engage the player more deeply into the narrative and activating those social intelligence portions is a big deal for the serious gaming community. If Chris pulls this off, he may not succeed in the gaming community as he wants, but his work may have a big impact elsewhere.

  9. jgf1123 says:

    Ambitious goal, something I would very much like to succeed. But wow was that Kickstarter video wooden.

  10. Sly-Lupin says:

    Great article, and lots of very useful content here for anyone wanting to write about gamic narrative… but I pretty much lost interest in the project he’s working on simply because he’s trying to create a narrative-based game without actually using language.

    And while it IS possible to tell a story without using language, it’s impossible to tell a very detailed, or complex story without language. All of the criticisms Crawford levies against contemporary gamic narratives (from the emergent tales generated by games like Civilization and Minecraft, to the sprawling epics delivered by our best RPGs) seem even MORE valid when applied to his game.

    If you want to create a truly interactive narrative, the best way is to “trick” your audience. The only “real” way to do it is to either use AI to do so (which we probably don’t have the tech for now, unless you’re not going to use language at all) or use something like Twine to create a narrative with dozens of choices for each “point” in the story, with every single branch being independent of all the others.