Raph Koster's Home Page

Influences

This was a keynote given at Project Horseshoe in 2006. You can comment here on the blog if you like.

Howdy, y�all.

[”Howdy!”]

I reckon I left my hat under the table there, but� I got a stuffed head and I think the voice goes with the hat.

You know, I�m not gonna do slides [”Sweet!”]. There is some stuff I wanna show you later, but we�ll set it up later.

You know, I went back and forth trying to think about what exactly to say here, because I spent the last, oh, maybe three years working on understanding fun, and on thinking about how games work at a really fundamental level, like an atomic, tiny, game system, how things tick and stick together kind of level, and you know, I concluded games were mostly about learning. And if you wanna know more about that, you can go read the book, right? And that was good. It was useful, it helped me a lot.

Around the same time that I was doing that, you know, I gave that initial talk (PDF) back at the Austin Game Conference in 2003� and I started working on game grammar stuff in 2004, and then I wrote the book. So there was a whole bunch of game grammar stuff that was in the book. And this whole game grammar thing was about common elements that are in all of the game systems, and how if some of these little elements are missing, then the game system won�t be fun. It was about how some games are symmetric, and some are asymmetric, and how all games happen within a topology, and you always have assets to maneuver, and you know� a whole bunch of stuff like that. You know. How games are structured around� I have it written down. �Solving statistically varied challenge situations presented by an opponent who may or may not be algorithmic within a framework that is a defined systemic model blah blah blah.�

I�m not the only guy doing this, OK? I mean, when Mark Terrano shows his graphs [of measuring user experience via metrics on input channels and outputs provided by the game at any given moment], I go, �Funny, I saw that same graph on Dan Cook�s website [LostGarden.com] about six months ago.� And for that matter, I saw something very similar on Ben Cousins� website about a year and a half ago. A lot of people are working on basically the same stuff. Right? I mean, I put up a very similar graph to that one arguing that socialization in multiplayer gaming requires downtime because you need attention troughs, which look exactly like the troughs that were in Mark�s graph.

Ben Cousins, Stephane Bura who�s been working on game grammar, there�s all the work that Eric [Zimmerman] and Katie [Salen] did in Rules of Play, you know � we�re all getting results that seem to work. But we�re finding is that there are essentially grammatical rules, or physics rules, or whatever, under this stuff, that can be found and that are helpful.

I could talk to you for hours about that crap.

I learned a lot from it, right? It�s had a huge impact on how I see games, and Ironically, I haven�t hardly gotten to work on any games since I came up with it. [laughter]. But it has actually had a huge impact, and I think it can have a huge impact on the industry as a whole, right? You can see how tools like what Mark proposes, or gosh, if we had something, if you could have a notation � this is what I argued in the Game Grammar talk � where you could sketch out how the pieces of a game system fit together, you could actually see where the balance would go out of whack, or why a system wouldn�t be fun, before you started implementing. It�s just another way to play test.

But honestly, I don�t want to talk about that. For one thing, it�s kind of dry and technical and full of diagrams, and you know, for another, sometimes it seems a little stupid and pointless. Like you�re caught up in the theory and where is the fun? And arguably, it breaks the wall between science and art, which is a wall full of holes already, but�

I really want to go a little further into the future, actually, and just take for granted that all of those efforts aren�t stupid. That they are actually showing us something about games. Because of the conclusion it led me to.

I concluded that just about all games were about math.

They are not about arithmetic, unless you�re playing an RPG [laughter], but about all fields of math. In fact, they tend to be about � I�ll go all geeky on you � NP-hard or NP-complete problems. They tend to be about, I actually went and looked up on the Web a list of NP-hard problems, and they sure as hell sound like our games. Motion planning on a plane with polygons. Yeah. There goes every single platformer ever. [Laughter]. Isomorphism. All matching games. Traveling salesman problem. Packing problems � Tetris is a packing problem. Almost, well, a lot of puzzle games are packing problems. Cover problems, the knapsack problem, blah blah blah. A lot of them are about NP-complete problems. In fact I actually started the exercise: on Wikipedia there�s this wonderful list of Karp�s 21 NP-complete problems, and I started at the top and started designing a game for each one of them. [Laughter.]

Besides those classical problems, games rely heavily on estimation of probability, which isn�t actually even complicated [laughter] but the human brain sucks at it, and appears to be hardwired to suck at it [laughter], and therefore we rely on it in our games a lot, because we know that we can sucker �em every time. [laughter] Just like Vegas.

So the more I dug into grammar and all of this stuff, the more I kept thinking, crap, it may be that games, by nature, are inherently math.

And I think that sucks. It makes me really unhappy.

A game about flapping

I wanted to actually show you something. I was recently working on a game about flapping. This sounds like a really odd thing to make a game about, but I wanted to try anyhow. I was working on this game about flapping, and it�s actually up on the blog, and you know, it�s basically a little game where one key controls one wing, and the other key controls the other wing, and the longer you hold them down, the longer your downstroke goes. When you let go, the wings slowly returns to where they are supposed to be. But because you have independent control of each wing, you can start doing this [slides sideways], turn this way, and if you give long strokes, you move in a different way than if you fold them slightly, and if you beat on a regular pace you actually rise in the air, and if you hold them tight, then you stop flapping and gravity starts taking you down. A flapping simulation. It was keyboard input because that�s what I have on my laptop.

So the thing about this little flapping thing that was y�know, it wasn�t about flight, it was about flapping, which is an important distinction. I built a little model here, and I even made it so that the keys were on opposite ends of the keyboard so that you had that sense steering like that. I was thinking in terms of, you know what, a lot of games are driven by interface, and that experience, so if I completely change up the interface for how you think about flying something, maybe I�ll head to something that feels fresh.

Just flapping the bird around on the screen actually turned out to be compelling for me. Other people, however, bounced off of it instantly, because they just didn�t get it. And they asked for mouse control, actually [Laughter], which I resisted mightily, because, well, then� the game was about the flapping. If you can�t hack the flapping, then tough shit. [laughter].

But the thing about this was that it ended up being compelling enough that I posted the prototype to the blog. And it actually got on BoingBoing, which meant that it �sold� a few thousand �units,� because a bunch of people downloaded it [laughter] and it got reblogged all over the place. So it was compelling enough that I added islands and scrolled down below, and a blue background with sparklies in the water, and color-shifting clouds and a day/night cycle with beautiful pastel colors, and a motion blur across the screen, and all this other stuff.

And that was cool, and lot of people � those who could hack the flapping � started commenting about how Zen it made them feel, particular once I added a little DADGAD-based soundtrack on acoustic guitar.

So it was a toy, right? It was designed for you to flap. If you successfully made it to the top, you were rewarded � I built a tiny little dynamic music system, and if you made it to the top, basically, I recorded lots of little chopped up riffs, and you had different riffs depending on where you were on the screen, and what angle your wings were at, and stuff. You know, really basic, and people said it was really Zen. But it wasn�t what I would really call a game. You know, it was a toy, an amusement.

The “Zen” version: Win32, Mac (11MB)
The “game” version: Win32, Mac

So I put a game into it. I made a 3d path in the air that went up, went down. Now, keep in mind, this is an overhead view game, so 3d means it got bigger or smaller. And you could see this rollercoaster path wending around. There was wind blowing you from side to side that you had to cope with as you flapped. You were scored, and you had to stay on the path, not too high, not too low, you know. Basically, an overhead version of flying through rings. If you got off the path for too long, you croaked. Game over. Which was really frustrating, because the main attraction was getting further so you could see what noon looked like, because of the color ramps on the clouds.

And pretty much, all the charm that the prototype had went away. [laughter] It ceased being fun because the math came in big time.

So here I am, and I�ve been looking at all these games in terms of the math. I�ve been looking at how you make these games and you work with them on this systemic level, but lately I�ve been dreaming about making games that make you feel what�s it�s like to be a wolf living in the winter scrounging scraps from a nearby mining town. A game about the sensation of a kaleidoscope. A game that exudes �treeness.�

And fuck! Math doesn�t tell us anything about that. It�s really really bad at it.

I think that these aren�t things that reduce down to math. As designers, we�re really trained from really few sources, overall. When I really dig into the sources and influences � we already saw this, even with the talks today � we point at other games, and we point at movies. And that seems to be the total extent of what we� [sigh]

When a designer adapts the sweep of history like what Sid [Meier] did with Civ, and the development of civilizations, he builds a spreadsheet. History is a spreadsheet. It�s represented by pretty icons, but it�s a spreadsheet. When a designer tries to capture the inner soul of a person, what forms the inner core of someone, Will [Wright] makes eight bars, and they can be raised and lowered by doing dishes, watching TV and peeing at the right time. [Laughter] It�s a spreadsheet again. All of human aspiration, and it�s eight variables moving up and down. When we try to represent a mystic moment in life, what Lorca called duende, you guys ever run across this word? It�s a wonderful world, it�s that elusive moment of transgression of the unreal into the real, you could almost call it magic.

It�s a mana bar and a list of spell about glowing missiles.

And I think we�re prisoners of our math.

It�s ironic, because we keep chasing story, right? We chase after story, and we suck at story. I�m sorry, but we are really really really fuckin� bad moviemakers, people [laughter]. We really are. We�re just no good at it. So what we do is we make bad movies, and we stick in some gameplay which we mostly stole from five years ago [laughter] and we say, �in order to get to the next episode, you must do this�� Right? That�s advancing the medium. That�s what we�re doing. It�s like, great, so in order to get to the next page of Gone With the Wind I have to play with a spreadsheet, and then you get the next chapter. [Laghter]

So� just to broaden our sense of influences a little bit. You guys all know that I came out of academia before I was a game designer. I was, to my everlasting dismay, I was mostly studying modernist literature while getting a creative writing degree. I really hated modernism, which was why I took so many courses in it, so I could argue with the professors [Laughter].

When you look at the modernist movement, and I�m actually going to broaden Modernism a little bit� really talking about everything from the Impressionist period on forward, because all of that stuff is kind of influenced by the modernist ethos. What was going on there was� There were artists and writers and musicians who got tired of showing things as they are. The height of craft before that was to say look, this is what it is. So, when you listen to Mozart�s Requiem, he is representing as accurately as he can with his medium, a certain set of emotions. And when you look at a Rembrandt portrait, he is showing you what is. It�s not strict realism, but it�s showing you what is. And the same in the realistic novels � oh God, the Victorian novel, ugh. [Laughter]

Instead, once Impressionism kicked off and as we move forward, we start getting people who wanted to talk about how they showed what they showed, with what they used, if that makes sense. So we got paintings about color theory. Or paintings about how light reflects. Impressionism is about, in pretty much all the media, Impressionism is not about what�s there. It�s about how what�;s there is perturbing everything else. This is why I say that Minesweeper is actually an Impressionist game, because you never actually see it, you see everything around it and how it�s been bumped and jostled. So Impressionism isn�t painting the lilies, the water lilies, it�s painting the light bouncing off the water lilies. It�s about what isn�t there.

Virginia Woolf wrote a book called Jacob’s Room, about this WWI soldier named Jacob who is dead. He�s the main character in the book, and he�s not in the book. You learn about Jacob only through what everybody else says. Gertrude Stein wrote an autobiography of herself from the point of view of her lover who was real, so it�s called The Autobiography of Alice B. Toklas, who is a real person, and she wrote it from her lover�s point of view and she�s a character in the book � it�s actually her autobiography but she�s not IN it.

And the next step after that, and Picasso did most of this in painting, it they start saying, let�s do paintings about perspective. Let�s do paintings about facial recognition. Let�s do paintings about color weight. Mondrian? All of Mondrian�s body of work is what they make you do in the first three months of the visual design class in an art major. It�s about �how heavy is red,� which isn�t actually a stupid question if you study art theory. How heavy is red, and how big is empty space between two very narrow lines, and stuff like that.

It�s all about itself, and in music we got music about harmony, or about how a melody works. Which is why eventually, once you get to the postmodern, you start getting stuff like John Cage, who will actually do a recording of four minutes and 33 seconds of silence in order to make you think about whether there actually is any such thing. And when yo go to perform 4�33� you walk to a stage with a piano, you close the lid, and you sit there. And after about ten seconds what you are listening to is the concert of fidgeting, coughing, sneezing, wondering what the hell is going on.

There�s so many examples of this� you end up getting writing about writing, painting about painting, music about music, and lately I�ve been wondering if all game design isn�t modernist. By its nature.

Whether at its core, it may even be postmodernist, or deconstructionist, by its nature. Because it�s all about the models. It�s all about the math. It�s all about looking at this stuff as a bunch of little moving parts, as a system. It�s always about reducing something� you could even say trivializing something. It�s about saying, �here�s this wash of complexity, let�s talk about how it works and break it down into this little spreadsheet.�

What that means is, is that you are always simplifying these complicated things, like the inner core of a person�s soul and turning it into eight bars, and reducing it down to something that is always inaccurate, and always oversimplified, and often frankly kind of Manichaean, good versus evil. Black and white. I think that ends up getting reflected in our stories, which are almost always simplistic Manichaean good versus evil stories too, because that�s all the game can support. Even what we call analog games, tend to be binary when you look into them.

I�m finding this incredibly frustrating and ironic, because people like those in this room, game designers, have some of the most varied sets of interests of any people on the planet, certainly that I�ve ever encountered. I mean, you go to any university, you talk to the specialists in each field, and frankly, they know nothing about anything except whatever it is that they study � which they know in such excruciating detail that you don�t want to hang out with them. [laughter]

Will Wright getting programmed

You get a group of us together� in the car on the way here we discussed the scientific underpinnings of astrology. At a design workshop I was at recently we were trading book recommendations� what�s the latest good hardboiled detective novel versus what�s a better history of America pre 1492� you look around at designers, even the ones who occasionally, well, I�ve got this wonderful photo of Will Wright� they have a version of Fa�ade there which is a full version room. They actually make you put on your little VR headset and goggles, they actually mount a keyboard on your back, so an operator can come and fiddle with what�s going on while you are in there. I snapped this wonderful photo of Will Wright being programmed [much laughter] But even people like Will, who sometimes seem cybernetic to some degree, you talk to him and he�s got interests all over the map. I mean, he liked to build robots, but he likes to build robots in order to leave them piteously mewling on the street in order to see what people will do when they come up against a dead robot saying �Help me!!!!� [laughter] And he videotapes them, so�

So as a group we�re really well-read, we�re interested in all kinds of different stuff� but when I look at the industry, our influences, they�re� I guess the word I�d use is �paltry.� Pathetically small.

Who was playing Cooking Mama yesterday? Yeah, the industry�s reaction is �A game about COOKING?!@!?� And then you play it, and it�s actually WarioWare re-skinned. And it�s another little spreadsheet about reaction times. And� yeah. Grrrr! So frustrating. I mean, we reduce it down to the same little set of reductionist almost deconstructionist mini-games to become a parody of cooking. But at least it exists. And there are those that� we played Werewolf last night, and there are things there that aren�t just statistics, although you play Werewolf, and think �You know there�s am algorithmic approach to this��

[�And that�s why you were killed first!�]

I know�! [Laughter]

And I guess the question is, I look at the game about cooking and I go, why stop there? I mean, there�s an easy step to the game about winemaking. And I can see you right now, let�s make a game about winemaking and in your heads you immediately start thinking, dry season versus wet season, acidity of the soil, there�s all kinds of things like that. SO let�s go a step further, how about the game about wine tasting.

Yeah, I don�t enough about wine tasting to figure out even how � yeah, you wanna play that game don�t ya. Live action. [laughter� �The ARG�]

And I go, wait a minute, what about a game� I just learned that all of the peach orchards between Parmer and 35 and Round Rock are gone. Which sucks. [�It�s all WalMarts now..� �It�s the pits� Laughter] Yeah� and why is there no game about the taste of a freshly picked peach, straight from the tree, with the smells and dust of the working orchard? How do you make a game about that?

[�Because my monitor tastes like crap.� �It�s a human interface problem.�]

It isn�t just an interface problem! It isn�t just an interface problem! There are probably poems about the taste of a fresh peach.

The game about the dynamics of a coral reef, and destroying the coral reef, you know, when you think about that, you immediately go �oh yeah, a serious game, and we�ll teach people about global warming,� and it�s not the same thing as really a game about coral reef-ness.

And why isn�t there a game about silk-screening t-shirts? What is it that makes us go, yeah, that just doesn�t come up, let�s go for the orc slaying instead? AT this very moment, there are probably more teenagers out there interested in silk-screening crap onto t shirts than there are in slaying orcs, and yet it doesn�t come up.

The game about the difference between the warp and the weft in the art of tapestry making… Hey, can we make a tapestry making game? Shit, that one�s even math, right? [Laughter] But it doesn�t come up, just doesn�t come up, because our influence set is so small.

Even those areas where we do have control, where we could go in and say, hey let�s broaden our set of influences, let�s draw from more sources: just look at art styles. The styles that we have are Japanese anime, grey-brown photorealism [laughter] and the cartoons that we loved as kids and still love because we�re geeks who won�t grow up. And that pretty much sums up the entire videogame art style.

Why isn�t there a fucking pointillist videogame, dammit? Why isn�t there an Impressionism renderer? And it�s just not there� why has no game gone, visually, where Picasso went? It�s not because it doesn�t work. Once upon a time, Picasso was shocking. Nowadays, you walk into generic hotel on the road, and you�re just as likely to see something Cubist as you are to see a watercolor landscape. We take it for granted., Everyone knows how to read that now. It�s no big deal. But we don�t even scratch the surface on these other art styles and influences and sources of inspiration.

We do rhythm games, and we can�t fuckin� do melody games. We just don�t do them. There are no melody games! We kind of have Guitar Hero, but it�s really a rhythm game with five drums [chuckles]. We�re so caught in this little rut�

Flapping as math

The challenge is, I don�t have any answers. When I wanted to know what it felt like to flap, you know, it�s a little wireframe bird, and up in the corner you get to see all of these floating point variables scooting around, with your angle of the wind and angle of each feather, all swirling by in this haze of numbers. And you know, flying and flapping is not rotation and angle of wing bones and force of air propelled. And yet when we put something into a game, that�s what it turns into.

It�s not that there aren�t mechanical activities that do just fine in this. Flying a plane is flow of air over the wing and blah blah blah, and we�ve managed to mimic that pretty thoroughly and well and at this point, it�s damn close. We have the sensation of flying in a plane pretty good. But I wonder how we take in all of the influences that are not systemic, taking in the games that reflect non-systemic reality without quantizing it.

You guys know the word �quantize�? The musicians, particularly the ones who work in electronic music in any form will know it. This is where you take the music that you play, and don�t quite nail the timing on, and you tell the computer please noodge the notes over by fractions of a second so that everything is lined up with perfect rhythm. That�s called quantizing. Fortunately, there is an opposite process which is to take something perfectly played, and noodge the notes slightly out of place so that it doesn�t sound like it�s being played by a robot� which is called �humanizing.� [laughter] That�s the pull-down�

One of the questions I end up asking is �Have I been trained?� Games are about learning to my mind, and you learn about models from playing games. You learn to understand that model of what the game is presenting you. That�s the learning exercise. What is the shape of the spreadsheet, what are the rows, what are the columns, what are the math operations I can perform on it. That is what games teach you. So I wonder, because I have been playing videogames since I first got that Pong console from Sears. Whenever that came out in the mid 70s, one of the knockoff ones. I am of the generation that has been playing videogames, well, since, basically.

Have I been trained them? To only see the world as something that can be quantized? Have I been trained by them to come to Werewolf and immediately say, �so the algorithm is�?�

That�s a little scary, because there is a big difference between shades of gray and a byte�s worth of grayscale. [thoughtful silence] Right? Big big big fucking difference.

When my wife and I were first dating, we were at one of the Apple campuses, when they just got the color Macintoshes. And if you remember that time period, all of a sudden the games had these beautiful color gradients in the background. We would go outside and look at the sunsets over corn fields on the Eastern Shore of Maryland, and we�d say �hey look! Mac colors.� [Laughter]

Long ago, the whole Deist philosophy� the metaphor for God was �the watchmaker.� And I�ve started to wonder whether games make the world into clockwork. And whether this is honest.

It might be that games, and us as designers of them, we�re just filling a role. I mean, there is a huge amount of value in seeing the world as clockwork. A huge huge amount of value. Arguably, it�s a very progressive way to look at the world, because it takes out a lot of frankly, stupid shit that has killed millions of people throughout history. Being able to approach the world in this very scientific way and deal with it in terms of algorithms and so on, it can help, because it makes you see the stuff that is just blatant superstition or false pattern matching, or the real reason why two groups hate each will actually be about calories intake and not about religion�

So seeing the world as clockwork and doing those spreadsheets, it�s an incredibly valuable exercise. This clockwork world has to be in the picture, along with all the other ways of seeing the world. Our worldview has to become, over time, and it inevitably will, right, like lots of layers in Photoshop making up one picture. We will see the world in many many different ways. And the clockwork way is a very good one, but I sometimes wonder whether we�re stuck in our layer, and whether we can participate in the broader side of things.

And when we go to this broader side of things, the world where there is a game about treeness, or the experience of being bark, or whatever artsy-fartsy thing I can come up with � when we bring it back with us, will we always be pixelating it, paletting it, building low-poly versions of reality? Will we always be turning it into something mechanical.

Do a thought experiment right now. Shut your eyes. [Silence]. Good, they�re all doing it.

Imagine something that you dream about. A moment of experience. It could be, I dunno, God, laying on your back watching the clouds shake around, it could be the first time that you and your boyfriend or girlfriend, whichever it was or both [laughter] snuck out in the middle of the night and went for making love under the stars. It could be the first time you ever saw the ocean. It could be that moment� you know, I keep pulling nature examples, how Romantic poet of me. But it could be almost anything. The first time you realized that the computer would jump through hoops and do your bidding. Those moments of wonder that Nicole Lazzaro references.

And now here�s the thought experiment. Build me the game system right now, in your head. And tell me if that dream didn�t just blow to bits.

[silence]

And that�s the challenge that I guess I want to leave you guys with. Because it�s not that this way of looking at the world is bad. It�s not. The challenge that I leave you with is whether or not games are irredeemably spreadsheets in this sense. And I don�t mean the stories we tack on top of them, the pretty art. Because the art and the stories, we know they can do this. The challenge is whether games can do it on their own, without being propped up by all this other stuff.

It might be the answer is no, and if it�s no, then the question I leave you with is, OK, then, what does that mean? For those of us who make them, are we all watchmakers? OK, watchmaking is a noble profession�

And think about what does that mean for all of the kids whose brains we currently control. That what we�re doing is teaching them to see the world as clockwork orreries. And what that means.

That�s the challenge I leave you with.

You can comment here on the blog if you like.