The Pixar Lesson

 Posted by (Visited 23288 times)  Game talk, Watching
Dec 072005

A few folks have noted that the On Point radio show host seemed to be rather interested in drawing analogies to theater, and in discussing games as story. This lays the groundwork for a nice little ludology vs. narratology discussion! 🙂

I suspect the “games as games” versus “games as story” thing was too subtle for the show, though both Heather and I kept trying to steer the host away from his “it’s theater! No, it’s movies!” sort of approach. We may have overemphasized the opposite side because of that, with our constant references to games as models.

It’s hard not to see the game world in terms of the narratological side these days, when all the buzzwords are about improving presentation: the HD era, the need for better facial animation to convey emotion, and so on. Yet the “games as models” viewpoint is truer to the core nature of games; a modern blockbuster game, like say Half-Life 2, would make a terrible movie if yu stripped out all the FPS parts, but would still be a passable game experience if you ripped out every cutscene and character bit.

Of course there is room for both, but IMHO most of our learning to improve games as story is going to come from learning to make movies, frankly. Better writing, better cinematography, better shots, better CGI, etc. It’s outwards-looking learning, and it won’t make the games any better as gameplay, but it will make them better interactive entertainment experiences.

Inwards-looking learning, improving games as games, is where the growth in audience seems to come from over time. New audiences arise from new models and new metaphors: the Sims, god games, RTSes, and so on. This seems to happen much more sporadically and slowly than the other sort does. Frankly, the boardgame industry invents more new forms of gameplay in a year than it seems videogames do in five…!

In the interview, I said that I wasn’t sure we had had our DeMille yet in the games biz. But I think it’s fair to compare the original King Kong to some of our games today — maybe Ico? In terms of how it probably felt to the audience back then, and in terms of the sort of ending it has. But looked at that way, we can also measure just how far we have to go in order to become less blatantly manipulative as gamemakers; there’s a long way from the original King Kong to a modern movie that reaches for similar goals… say, Thelma & Louise. The complexity in the issues and story are vastly greater. And Thelma and Louise feels dated to us already too.

In good movies, shooting a guy through a window isn’t a plot point. Why the guy was shot through the window is what matters, usually. The body count in even a straightforward thriller like The Bourne Identity is quite large, but none of the incidental deaths of guys in stairwells are the issue. The plot points are actually when Bourne chooses not to kill someone.

In games, we tend to say that the shooting bad guys in stairwells is the game, and that the choice not to shoot someone belongs in a cutscene. Well, that would leave the part that has the largest possibility for emotional engagement, for art dare we say, in the bit that can be cut with no impact to gameplay whatsoever. 🙂 This is why I say that many of the peak emotional moments we remember in games are actually “cheating” — they’re not given to us by the game at all, but by cutscenes. The death of Floyd the Robot in Planetfall is an example of a cheat like this; so’s the death of Aeris. Both are cutscene moments, effectively (well, in text in the case of Planetfall…) and not gameplay moments.

Games are indeed doing this OK — not well, by the standards of other media, but OK. There’s tons to get better at here, I think, foremost of which would be just raising the general caliber of the writing throughout (which is why it’s so important that the industry get off its butt about hiring writers). We just shouldn’t kid ourselves that this is in any way “game design” or advancing the state of the art in gameplay. It’s not. It’s us learning to be better entertainers, an advancement in interactive entertainment, and it’s us learning to be slightly less sucky filmmakers.

You wanna deride interactive movies? That’s what we mostly make now.

The structural issue with doing more become apparent when we try to do something other than make a bad movie with some minigames interspersed throughout, which is what something like Final Fantasy, or what something like every WW2 shooter, basically is. That, I think, is why so many people gravitate towards Ico as an example, because the emotional hook was actually conveyed through gameplay: the (literal) hand-holding of princess Yorda was what made it work, not cutscenes.

There’s the distinct possibility that Ico was “too smart for the room,” however. It sold poorly, even though it is greatly admired among designers and game cognoscenti. This might make us think that the bombastic games are the ones that will earn the most; one thinks of the way in which empty-headed summer blockbusters seem to dominate box office takes. But that’s ignoring what I think of as “The Pixar Lesson.”

The movies that in the long run have earned the most are the ones with the most emotional connection, not the biggest explosions. I mean, there’s zillions of examples. What big but empty movie sits atop the lifetime charts? Don’t get me wrong, the biggest movies deliver both the emotion and the spectacle, but c’mon, the memorable moment in Titanic is the scene on the bow where he’s king of the world and the scene of DiCaprio drowning, not the actual sinking; in E.T. it’s the kid silhouetted across the moon and it’s the finger lighting up in the closet, not the bike chase itself; in Star Wars it’s all through the movie, from the moment we find the blackened corpses of Uncle Owen and Aunt Beru (notice the deaths happen offscreen), to the swing across the chasm in the Death Star and “do you trust me?”, to the moment when Ben chooses not to fight.

I always use the example of Die Hard. Yeah, you get the big explosion and jumping off the roof of the building with the crashing helicopter behind you, but the crucial emotional moments are a phone call or two, and picking broken glass out of Bruce Willis’ feet. Why? Because the whole movie is about him walking over metaphorical broken glass to get back to his wife. The whole rest of the movie is just the obstacle. In a less bombastic movie, you might well not have had any explosions at all, and yet the emotional connection would still be there, and then you get Sideways, which is just as much a movie about an emotionally damaged man trying to connect with a woman.

In the more bombastic case, you get movies that nobody remembers. Car chases, explosions… stuff that you put on for a party or to fill an empty evening, but you’re not quoting the rest of your life.

That’s the Pixar Lesson: it’s not that story comes first, it’s that story deserves respect and so does the audience. Particle systems, shader 2.0, and physics are tools, like the terrorist in the building was a tool to tell Bruce Willis’ story. And just as Finding Nemo delivers the story, and then goes on to deliver the spectacle, we need to deliver the gameplay, and then go on to deliver the story.

  24 Responses to “The Pixar Lesson”

  1. Blogroll Joel on SoftwareRaph Koster Sunny Walker Thoughts for Now Sex, Lies and Advertising

  2. we just need to catch up and take it seriously, like the rest of the world does.” If I interpret this literally, it almost seems as if Raph sees stories as something added after mechanics, rather than designing mechanics with story in mind. In The Pixar Lesson, Raph clarifies. “That’s the Pixar Lesson: it’s not that story comes first, it’s that story deserves respect and so does the audience. Particle systems, shader 2.0, and physics are tools […] tell [

  3. Wow, nice post Raph. That’s really clarified a lot of things for me. It seems that often an issue with pushing the emotion in to the game is not going far enough. In KOTOR the best bits for me were when deciding to do the right thing or not, helping the victims or profiting from them. When I sold someone down the river I felt bad, but the game basically didn’t care. It gave me some dark points and then got on with things. When playing Half Life 2 I felt really protective of my squad, but a friend really resented not being able to shoot them or the other characters in the story. Is making emotional elements slightly interactive worse than making them not interactive at all?

  4. Is the purpose of the emotion to…

    1) Keep players wanting to continue to play… as Fable does.

    2) An emotional ride is enjoyable for its own sake, as in a tear-jerker movie.

    3) The emotion leaves a permanent memory, like ET. The permanent memory is a useful meme device.

  5. Jim, ya got me as to whether making them slightly interactive is worse than making them non-interactive. Certainly having interactivity without consequence (I’m reminded of the way the branching story worked in the Wing Commander games) can undermine the effect.

    In Prince of Persia: Sands of Time you had a charming little story with basically no interactivity to the story arc, and it worked really well. Then again, Jordan Mechner pioneered that back in Karateka… In Beyond Good & Evil, you had a story you participated in a lot, and it still worked well. So I suspect it comes down to what storyteller skill is being applied to the game. Frankly, usually the answer is “not much.”

    Mike, I think the answer is “all of the above.” At least, with a book or movie you would never settle for just one of those. You always shoot for all three…

  6. Of course, this ties into your other post: You need to get away from killing 10,000 orcs in a session, perhaps only killing one or two, and adding meaning to the murder.

    You also need to be able to protoype ideas quickly…

    And you need some mechanism for creating the emotion: I suggested Personal NPCs on Mud-Dev as one approach, and at least a few people got what I was saying. I have another idea that I’m toying with too… but I don’t know if either will work. Both approaches are more game-like, and you seem to be interested in world-like approaches.

  7. Emotional impact in games?

    Raph posted an interesting bit on his blog called “The Pixar Lesson”. In it he describes the importance of story and how games need more emotional impact. It’s a great post and I recommend that you read it.
    I was going to comment on his blog, bu…

  8. […] Are games like Final Fantasy "bad movies with minigames interspersed throughout"? SOE’s Raph Koster gives his opinion on what makes an emotionally gripping game.Raph Koster developed Ultima Online, and is currently Sony Online Entertainment’s chief creative officer. Here are the highlights from a recent entry on his blog. … […]

  9. Raph I have been following many of your keynote addresses and other articles you have written for quite some time. My only question is why do you write and herald all of these things yet in games you have been in charge of etc you completely turn your back on everything you have written? Look at Star Wars Galaxies for instance, this is a game that has turned into everything you have ever said a game should not be. Would be nice to hear your thoughts on this and an explanation why you say alot of great things but, when it comes to practical application you turn your back on it.

  10. You’re missing the point about Bruce Willis. It works because he’s a good actor (save it Willis bashers, he did a fine a job in Sin City). Make a game with generic Willis like character doing the same things he does in the movie (with crappy voice overs of course) and would it work? No, it would just be another bad game based on a movie.

    People know they aren’t Willis, but they can want to be like Willis. In a game you have neither, because of the limits of character modeling you know you aren’t the character on screen and you don’t want to be like him because he’s unnatural.

    Look at characters that people do relate to well, Master Chief, whats her name in Metroid, both are blanks that you can project what you want onto so they’re not unnatural. Mario also works well because a character who is so obviously a cartoon doesn’t seem unnatural.

    Until you get past the bad acting and offputting unnatural characters problem you’re not going to get the emotional connection you’re looking for in games. Best bet would be well acted cartoony characters.

  11. The story gives meaning to playing, and playing provides interaction with and exploration of the story.

    Story ties experiences together in a meaningful way.

    In theatre and cinema, there is, traditionally, no interactivity. One is attempting to tell a story with the given medium. The story binds together the scenes and acts.

    The same is true for games as a mode of playing. Game choices (moving, shooting, speaking) provide for atomic experiences of interactive events. Story provides a structure in which to present games and stitches together multiple interactive events in a meaningful way to provide for a consistent contextual experience from one event to the next. Games today actually rely upon very little interactive elements that are played out in multiple ways with the use of story.

    Story, though, clearly serves as an attractive force for interactivity and exploration, as seen with the release of Star Wars Episode IV, the legion of action figures manufactured in its wake, and the existence of Stars Wars: Galaxies. How? Why? Because story initiates fantasy, and we explore our fantasy with story, and interact with fantasy as play.

    Bigger explosions, better special effects, interesting topics, are all things, I believe, that tie into some cognitive need for fresh and familiar experiences. Everything about our experiences require some degree of similarity for familiarity, and freshness for fascination. Aspects of freshness and familiarity seem to act as motivators, as in fashion.

    Emotion, in story, provides a mechanism to identify with the story, immerse us in it, and motivate us to remain until its conclusion.

    “would still be a passable game experience if you ripped out every cutscene and character bit.”

    Cutscenes and characters are not the story, but tools for telling stories. Without character development and cutscenes, Half Life 2 still tells a story in setting, theme, lighting, character, pace, etc.. This story creates context which provides for expectation and expectation creates interest. Cutscenes are indicative of the difficulty in using an interactive environment to tell stories, as cutscenes are specifically used in the control of pace and focus during critical story moments.

    The model, for example, is navigating a space with a mouse and keyboard and positioning a reticle over an object with a mouse and pressing a button to either pick an object up, use an object, or drop (or throw) an object. Successful reticle targetting will result in an object variable increased or decreased depending upon the object used. Complexity is increased by increasing the number of objects and variables and their relationships to one another. The metaphor, in its simplest sense, is spatial corporealization and ranged conflict. This metaphor is repeated over and over, with some variation, in a meaningful way through the use of story.

    Now, what constitutes a good story? That’s a different topic.

  12. […] Next-Gen notes a blog update from SOE’s Raph Koster where the topic is emotional engagement in games. Shhh, don’t tell anyone but the real secret is nerfing the crap out of everything and doing “revamps” every year or so. Just kidding! I don’t play Star Wars Galaxies. Here’s a snip:It’s hard not to see the game world in terms of the narratological side these days, when all the buzzwords are about improving presentation: the HD era, the need for better facial animation to convey emotion, and so on. Yet the “games as models” viewpoint is truer to the core nature of games; a modern blockbuster game, like say Half-Life 2, would make a terrible movie if yu stripped out all the FPS parts, but would still be a passable game experience if you ripped out every cutscene and character bit.Of course there is room for both, but IMHO most of our learning to improve games as story is going to come from learning to make movies, frankly. Better writing, better cinematography, better shots, better CGI, etc. It’s outwards-looking learning, and it won’t make the games any better as gameplay, but it will make them better interactive entertainment experiences. […]

  13. Raph –

    Interesting as always. I think we really need to get away from this obsession with “games as stories” model. It has lead to increasingly crappy games.

    Games are about choices that the player gets to make. The role of the game creator is to present interesting choices for the player. Novels, movies, are about watching choices and plots. The novel or movie creator makes the decisions.

    I think this is why movie tie-in games tend do be so bad. The type of creativity that makes a good movie/book is that the author is a dictator. Defining what is shown and hidden to the audience/reader.

    The ultimate rail ride/game.

    Good game designs are about the author giving meaningful choices to players. Emotional impact (if that is the goal) comes from allowing the player to make a difficult choice.

    decide-ative, something… not narrative.

    In a good game, we don’t notice the constraints of the game or its borders – we follow flow of decisions that the game designer opens for us. No one complains about Chess being overly restrictive for being played on an 8×8 board with specific pieces and rules… rather these tightly defined mechanics open up a game of intellectual depth and challenge.

    I have been struggling with this over on my blog:

    Beyond Storytelling in Games


  14. Games are about related choices the player gets to make. That relation is the irreducible nub of narrative that exists in every game, even Tetris. One of the keys to Tetris is that it’s ruleset is compact and tightly related. This gives the game tone and direction, so that every game has perceptible continuity, which gives the player the framework to understand their choices.

    Tetris has as a narrative direction; the struggle of the player against a universe both hostile and without meaning. Every game is doomed to end the same way, and at the end, nothing has changed. The player, like any existential hero, must find their own meaning in what they do. On the other extreme, tabletop roleplaying, which has endless permutations of often fuzzy rules, always has one core rule; the events must ‘make sense’ to the players. All rules but this can be dispensed with.

    Note the commonality in the two extremes above. Niether is a story, but both create story. Thus I do agree that games are not stories, but story is a core element of the game experience, just as rules are also essential to games. Indeed, I would argue that story is a primary product of all games. Ludology and narratology are bound together in games, and I think that binding is a key part of the description and definition of games. Both can be reduced, but both must exist.

    To me, the unresolved synthesis of gaming is the linking of narrative and gameplay on the most fundamental level. Story that is not in the code is meaningless window dressing, and code that does not support story is distracting and disruptive. If the rules and the story, however rudimentary either may be, are not in harmony, the game will suffer for it.

  15. Evangolis, you’re talking about two different sorts of story, I think, and blurring them together a little bit. The narratological school focuses a bit more on the traditional “author’s intent” form of story than the emphasis you’re giving, I think. Have you read Hamlet on the Holodeck?

    I strongly agree with you on the value and importance of “after the fact” storytelling — see Two Models for Narrative Worlds.

  16. I’ve read Hamlet on the Holodeck, Two Models is new to me. I agree I’m blurry, but I don’t think I’m wrong, I’m just not yet able to express this point well. I’ve been struggling with this notion for a while, really since I read your book. Not so much the book in general, it is the Tetris example that sets me off, it feels like something subtle but important is slipping past there.

    I really think that storytelling in games is both integral and different from the written stories common to our culture. Something about listening to my grandfather tell his stories over the years, and the traditions of oral storytelling in our family, with the interaction between teller and audience; that too is leaking in here. I see before the fact setup by the designer and after the fact perception by the player as a single process, if I understand your drift here, which may be why I’m blurring things together.

    With a book, there is a separation between the story and the reader that I don’t see with games. In games, the player is part of the story, and the story therefore should arise from the game, but also extend beyond it. As I said, I think games create story, whether that is what the designer intends or not; if that is so, then what remains is to do it well.

    As I say, I’m blurry on this, and thus was going to withhold comment, but secureplay’s phrasing tripped something new to me.


    The above was written before I realized that Two Models was a presentation I could read fairly quickly, and not another 300-page tome of wisdom. I still hold with what I said above, but here are the further thoughts Two Models leaves me with.

    I talked a bit about code, rules in computer games, code for shorthand. I’d take the Impositional and the Expressive, and I’d say that really what you have is two sets of code running. One is the game code, the rules the game designer makes up, and the other is the person code, the code the player has running in their head. The game is in large part the interaction of these two code sets.

    Now traditional text narrative has these two code sets as well, but the written code set is very fixed, so the impact of the person code is purely internal to the reader. Now this can be really powerful; I can still recall the jolt I got from my brain when I hit the climax of Sometimes a Great Notion, and I couldn’t stay in my seat, couldn’t stay in my dorm room, had to go out and walk, and even the Montana sky seemed too small for what I felt. And that is in part a post-narrative story; I still feel that jolt 20 years later. But it is also an extension of that story out of the past and into my present. In a sense, that story has never stopped telling itself to me. The code crossed over.

    Now with games, the code is more flexible. At it’s most flexible, with tabletop RPGs, the code is little more than suggestion. But at it’s most restrictive, say with Tetris, you still have considerable expressive input, although, as with a book, it tends to be focused inward on the player; it is likely to express in the game only as a change in the player’s performance, as they succumb to despair or transcend into flow. This is not a bad thing; haiku is a restrictive form that still can produce beauty and wonder; but if you can maintain consistency, a more open code set can allow more expression of the player code, and that can be very powerful, as your UO examples in Two Models suggest. Now, I continue the phrase player code in the last sentence to highlight the fact that, while incredibly complex, player behavior is not completely unpredictable (at least in hindsight). Thus I think that game designers need to approach game design as a collaboration with the player. This collaboration is made much more complex by the fact that the designer executes their portion of the work in the past, the player executes theirs in multiple presents, and the whole thing will spin forward into the future as both a recurring event and tales of that event within and between players, but what is life without a challenge?

    Sorry to ramble on like this, I suppose if I understood my thinking better I’d be more succinct, but I think I still don’t fully understand myself yet.

  17. […] Evangolis on The Pixar Lesson […]

  18. Games like Poker, Chess, Checkers, and Football can create post-narration. Storytelling, relating topics of challenge, struggle, loss, victory, and comedy, can all come out of playing these games. These stories occur external to the game, rather than internal to the game. Humans are storytellers.

    If I create a sandbox with multiple open-ended choices, a human can tell a story about it. If I create a virtual world with many metaphors which relate, open-ended, to their own reality, there will be narration as people make their own stories. This is what happens in playgrounds.

    Expressing a story to the player through choices and allowing the player to express a story through choices are, perhaps, two seperate points on the same continuum. Where the two come together may be when the game tells a story in a way in which the player accepts it as their own expression, or when the player expresses a story that the game shapes or completes based on the natural effects or relationships of the players choices. It is this collaboration between the game and player in story expression that may be of interest rather than the simple continuum extremes of film or playground.

    One field I find conducive to this aspect of expressing stories through game theory is the field of literary adaptation and translation. Much of this involves breaking stories down into minimal sets of memes, affects, and archetypes through exploration of the theme, character traits and their relationships, and then adapting that experience to construct a new story, perhaps even to a new medium, be it film, theatre, novel, or culture. This minimal structure can be used to explore the theme further, or use it as an encapsulating form to present a new idea.

    Good games making bad movies is not the result of games being poor originators of stories. Good books make bad movies too. Even good movies make bad remakes. It is more a result of poor adaptation. As I’ve suggested before, many of these poor stories in games seem to come from putting a story together simply to stitch together the same repetitive gameplay, or to repackage gameplay already made popular in another game.

    Just some further thoughts…

  19. […] Guess what? When you change the fundamentals of your online game and jettison characters — and their associated relationships — that took your audience years to build, players don’t like it very much. At all.A Feast for Crows, George R. R. Martin’s new book, is, duh, doing well. Even though The Boyfriend says it suffers from incipient Robert Jordan Disease, which involves taking 300 pages to tell 40 pages of story, I’m looking forward to it. Martin, a former story editor and producer for “The Twilight Zone,” knows how to give his characters layers and complex motivations. While ostensibly fantasy, the Song of Ice and Fire books (particularly the first couple), tend to the Shakespearean in scope and theme. They feature messy, violent politics, Tudor vs. Plantagenet-style, rather than twee fairytaling around the countryside. Oh, and there’s a lot of sex. How is DIE HARD like SIDEWAYS, and how can that info help us make better games? MMOGuru Raph Koster thinks gamemakers need to improve as filmmakers in order to help the medium realize its potential, including getting smart about hiring actual writers to work on games (naturally, I agree).At the heart of this is what Raph calls the Pixar lesson, namely that story deserves respect, as does the audience. Amen. […]

  20. […] Raph’s Website &raquo; The Pixar Lesson posted by [majcher] at 2005-12-14 18:18# <[majcher]> raph make good talk. […]

  21. […] Are games like Final Fantasy "bad movies with minigames interspersed throughout"? SOE’s Raph Koster gives his opinion on what makes an emotionally gripping game. […]

  22. […] In poetry, the genre was the form, and the content was carried by that form. In games, the mechanics are the form and the content; the “content” is something else layered on top for those times we wish to have more of an interactive narrative experience. The mechanics are left to change very little, in the name of accessibility, whilst we forget that the modern game isanything but accessible to those who are not gamers in the first place. The burgeoning evidence thereof rests in the wild popularity of so-called “casual games” on the web, which are anything but casual to the hardcore aficionados. […]

  23. That’s great, looking forward to reading the rest of your entries.
    Take care,

Sorry, the comment form is closed at this time.