I recently created a Twitter bot that tweets one randomly generated metaphor every two minutes.

I’ve decided to make the source code available. Get the source on Github here. Read the README.md, as it requires installing some software and getting your own API keys.

The whole thing is pretty simple and hacky. The program is pure JavaScript, running in node.js, which lets you run server-side JavaScript totally distinct from a web browser. I host the bot over at Nodejitsu — you can actually see the app “running” here: http://metaphor.jit.su/ (Nodejitsu requires all its apps to respond to HTTP requests, so I have Express, a very simple web server, running just so the hosting service doesn’t mess things up.)

The program uses the Wordnik API, which is a fantastic service that lets you specify things like “give me 4 random nouns, and now give me words commonly used in phrases with those nouns.” I don’t really use any of the fancy stuff; I just ask for random nouns and adjectives within a certain tolerance of common use (I’ve calibrated it so that the words can get pretty weird, but not so weird that the sentence is incomprehensible). What the bot does is call the Wordnik REST API using Chris Williams’ super-simple restclient, which returns a JSON object with my words that I add to a very simple string. The algorithm is:

(a/an) noun (is/considers/of) (a/an) noun: adjective (and / , not / , yet / but / , / , but not) adjective

That’s it. I play around with frequencies a bit, where “is” is much more likely to show up than “considers” or “of”.

When it chooses articles it can be buggy. I don’t check for phonemes (though you can do it with Wordnik), so sometimes you’ll get “an university” or “a hour,” things like that.

Weirdly, the Wordnik API needs to be coaxed to just give me regular nouns or adjectives. For example, to get singular nouns I had to ask for it to include nouns, but explicitly exclude proper nouns, plural nouns, proper plural nouns, possessive proper nouns, suffixes, family names, idioms, and affixes. Adjectives I didn’t filter so much as it seemed to mostly work, but I still get weird words that don’t seem to fit, like “generall” (archaic adjective), cist (a noun), youngling (archaic noun). But I also get wonderful conversational words like “lovey-dovey,” “sky-high,” and “Smithsonian.” So… I’m pretty happy with the results.

Finally, I have the program tweeting on a setInterval using the twit npm package, which is the simplest JavaScript Twitter API interface I was able to find. It works exactly as advertised.

The only other interesting thing the bot does from a programming perspective is that every five hours, it grabs the last 20 retweets of its own tweets and favorites them. The result is that you can see a best-of list right here.

I wanted to write up an essay in response to the third chapter of Ian Bogost’s Alien Phenomenology, “Metaphorism.” Instead, I decided to do philosophical carpentry. I made this:

Metaphor-a-Minute!

It uses the Wordnik API to grab nouns and adjectives and does some very naive processing to form a basic metaphor. (It’s not polished; for some reason when I ask Wordnik for an adjective it only gives me one 90% of the time.)

Essentially, I’m skeptical that we can ever have good metaphors for anything; or rather, anything can be a good or a bad metaphor. I think metaphorism requires someone to take the stance of a poet, to have confidence in metaphor. And a poet I am not…

I would say more, but I think the work speaks for itself. Take heed, Ian–this object has spoken!

a speculation is a wicket: long-chain and soft-hearted

— Metaphor-a-Minute! (@metaphorminute) May 10, 2012

Here’s the latest in a series of articles chronicling my unhealthy obsession with videogames and ontology. Just in case you thought I was back to blogging about normal stuff.

Most videogames have dynamic objects that you can interact with and static objects that you cannot interactive with. After a few minutes of experimentation, experienced players intuitively understand what is static in a game and what is not. If we imagine a typical first-person shooter, “the environment” is static: trees and ground and sky, but also buildings, rubble, street signs, and so on. This environment provides ambiance and strategic cover, and perhaps motivation for progress, but little of concrete utility to the player. Interactive, ‘real’ objects are usually things with obvious utility or drawbacks: guns, ammunition, vehicles, and people usually top the list. But abstract games have a similar ontological divide: in Tetris, the tetrominoes are real, interactive objects. But the score board and playing field lie inert and mostly uninteresting.

It’s easy to see that the ontology of most games relies on a division between types of objects: static and dynamic, borrowing the lingo that many game engines use. Things the player can directly affect, and things she cannot. No matter how many rockets you fire at that stop sign, it’s not going anywhere, sorry. (Furthermore there are objects that are not modeled in the game world itself. For example, a spawning system or a menu system could be represented as an object. I’m bracketing that discussion for this essay and will get back to it at some point.)

What would a game with a flat ontology look like? Jason Rohrer’s Inside a Star-filled Sky comes close, as I posited the other day in a roundabout way. The game with the flattest ontology I can think of is none other than 2004’s sleeper hit for PlayStation 2, Katamari Damacy.

The Prince of Objects

If you’re unfamiliar with Katamari Damacy, and since I’m too lazy to write one myself, I’ll quote Wikipedia’s current description of the game:

The game’s plot concerns a diminutive prince on a mission to rebuild the stars, constellations, and Moon, which were accidentally destroyed by his father, the King of All Cosmos. This is achieved by rolling a magical, highly adhesive ball called a katamari around various locations, collecting increasingly larger objects, ranging from thumbtacks to people to mountains, until the ball has grown great enough to become a star. Katamari Damacy’s story, characters, and settings are bizarre and heavily stylized, rarely attempting any resemblance of realism, though the brands and items used are based on those current in Japan during the game’s production.

Katamari Damacy eschews the static/dynamic divide seen in most videogames–at least, to a point. Everything is an object to be picked up. Domino, snail, seat cushion, pencil sharpener, lipstick; police car, picnic table, graffito, telephone pole, shoe store sign; hot air balloon, butcher shop, iceberg, storm cloud, and Jumboman alike. All are objects to wadded up into your magic ball.

For myself, Katamari’s primary delight came from the slow realization that things which appeared to be static when the Prince’s ball is small turn out to be dynamic when the scale changes. You might spend the first five minutes of a level rolling around a house. The house is immovable and noninteractive, comprising of a maze for you to traverse and collect the smaller items that it make their home within its walls. But five minutes later you’re 50 meters in diameter and can pick up the entire house as easily as if it were a ballpoint pen!

And yet the way this scaling works is extremely interesting. I’ll look at another example: a park. At the small scale, you find yourself inside the park, working your way up from popsicles to picnic baskets to punk rockers. But at a certain scale, those objects become abstracted into the notion of a “park” as an object on its own. You can no longer pick up a single popsicle–you’re too big. This is related to a concept in game engines called “level of detail” or “LOD” (basic explanation here) In a game like Skyrim, when you’re far away from a mountain range the game does not render every single goat and tree on that mountain. It shows you the mountain in the abstract in order to save on resources. What’s interesting is that LOD normally applies to things that are far away from the player or from the camera. In the case of Katamari, LOD applies to a case where an object is no longer on a scale you care about anymore. What we’re seeing here is a case of a Latourian black box appearing in a game. Yes, we can and do appreciate the individual elements of a park, but at some level we close that box and begin to refer to the park as an object in and of itself. What was inside the park no longer exists, and is subsumed by the “park” object.

Contrast this to a game like Minecraft, where huge objects are constructed from raw materials, either by human hands or the hands of the world generation algorithm. While Minecraft features objects at many scales, it never abstracts those objects. Every object in Minecraft is either a building block or an aggregate composed of building blocks. It’s a strictly atomist ontology: there’s the stuff the world is made of (wood, stone, coal, iron, etc), and everything else can be reduced to those atoms, literally with the blow of a hammer. (Perhaps it’s more Empedoclean, since there’s not a single substance at the root of things but rather a set of primal elements.)

While large objects in Katamari are indeed composed of smaller objects, they are not mere aggregates. They are qualitatively different from the sum of their parts. Here’s Graham Harman on Latourian black boxes:

“Finally, we have seen that Latour’s relationism allows him to replace the usual twofold rift of philosophy with a plurality of levels. [...]There is no final stratum of brute material from which flimsier, more ostentatious entities would then be molded. [...] There is no substance, only black boxes, and like Pandora’s box they can be opened at will to examine the delicate internal negotiations that made them possible. Traditional realism’s hobgoblin of the ‘mere aggregate’ or ‘mere thing of reason’ is no longer a worry, since in the end everything is an aggregate, assembled carefully or carelessly from numerous components. Gaps multiply to infinity and are constantly crossed by the work of translation, not by the impossible perilous leap so deservedly ridiculed by James. In short, there is no final layer of reality from which all relations will have been cleansed. The stunning metaphysical implication here, which Latour never discusses openly, is an infinite regress of actors. If there are only black boxes and never a final substance, then we will never come to a final stage in any analysis.” (Harman,  Prince of Networks, 106) [emphasis mine]

This “plurality of levels” is immediately clear in Katamari Damacy. While at one level the house is a container for smaller objects you can pick up, at another level it becomes an object you can pick up in its own right. When I read that gaps “multiply to infinity and are constantly crossed by the work of translation,” I think of the moment in Katamari when you hear a chorus of harps and get a sort of magic transition to a new size category. Moving from small to large is not a smooth, continuous process. This is for technical reasons related to LOD and memory management: the PlayStation 2 simply can’t model every object in the world as dynamic, so when we jump the gap between scales, the engine places small objects in various black boxes and closes them up for good.  This is the work of translation between the multiple levels of reality, represented in Katamari Damacy by your magical katamari ball, or perhaps the King, who appears with a pithy comment whenever translation between scales occurs.

On another level, at the end of a stage the King of All Cosmos shoots your balls of stuff into the sky and they are transformed into stars. It’s an occasionalist world, but instead of Allah, the King is waiting there to provide the magic needed for objects to truly touch one another and become aggregates, for the black boxes to close. The prince brings objects together, but the king allows them to translate.

The ontology of Katamari Damacy implies the “infinite regress of actors” that Harman identifies in his take on Latour’s metaphysics (which Harman finds unacceptable and dedicates a significant portion of Prince of Networks to offering an alternative). Unlike Rohrer’s Inside a Star-filled Sky, where the player opens up black boxes and can travel ever downwards or upwards into a literal infinite regress of objects, Katamari stops at a certain point. In the credits you roll up the nations of the Earth itself. In its sequel, We <3 Katamari, you can roll up the solar system, including the Sun:

It’s worth asking what you can’t roll up. What objects exist outside the realm of things-that-can-be-rolled-up? Well: the Prince himself, and his father the King of All Cosmos, and the katamari ball all seem to have some kind of special status. In some of the sequels, you can roll up the King. But it’s clearly not the true King — rather it’s some kind of projection of the King. The real King is the one who lives in the source code, doing the hard work of translation from objects to closed black box of park or of star.

Of course, there are also the limitations of the game engine itself. The game runs on hardware. The Prince can’t break out into our own world and roll us up. Perhaps a Katamari game with a true flat ontology would allow the player at some point to roll up addresses of memory in the game console’s hardware that would corrupt the game, bringing the world to a screeching, freezing halt in a cacophony of confusing colors, repeating sounds, or perhaps just blackness.

The World is Full of Things

“My, Earth really is full of things!” –The King of All Cosmos

One last point I want to make is that the encyclopedia of items in Katamari is an example of ontography as defined in Ian Bogost’s Alien Phenomenology. Every object the player picks up gets added to an encyclopedia that you can then review. There’s a possibly-complete list of every item in the original game available at Fanboy.net.

Oddly, I remember the quote as “My, the world really is full of things!” I’m not alone in this, apparently. I think this has to do with the universal scale of Katamari Damacy. The Earth just seems too small to contain it. As Anna Anthropy says in a great short essay about the game:

“we are replacing the stars with stuff; with human stuff, all the little minutia that litters human environments in abundance. the game suggests that the whole of the vast, dark cosmos contains as many interesting things as the surface of the one small planet earth.”

I’ll take a break from my extended philosophical ramblings to note that I’m teaching three upcoming one-day workshops on HTML5 game development! Dates and signup info:

• May 5, 2012 at NYU, New York City – 1 day workshop for $250 - Sign up now!
• June 18, 2012 at CRIM, Montreal, Canada – 1 day workshop for $295 - Sign up now!
• July 21, 2012 at the Bocoup Loft, Boston – 1 day workshop for $250 - Sign up now!

Topics covered include Canvas, animation, audio, and physics – at the end of the day, attendees will have a simple 2D physics game built in JavaScript and Canvas. It’ll be similar to Angry Birds, yet 100% guaranteed to be less commercially successful!

Feel free to ask any questions in the comments below. Hope to see you at one of these!

I wrote the following piece on a transatlantic flight yesterday after reading a big chunk of Bruno Latour’s (absolutely brilliant) Aramis, or the Love of Technology. I was inspired by a passage in the book where a personal rapid transit system is talking to its priest about ontology and free will. (Yeah, I know.)

This a dialogue that takes place between two philosophers who live inside Jason Rohrer’s game Inside a Star-Filled Sky. (Yeah, I know.)

If you haven’t played the game, you really should do so (Win/Mac/Linux), but this video will give you an idea of what it’s like to play a recursive shoot ‘em up. (Yeah, I know.)

METAPHYSICIANS INSIDE A STAR-FILLED SKY

–What is an object?

–Why, an object is a thing that sustains qualities.

–Very well then: what is not an object?

–Simple! It’s something without qualities, such as a wall.

–Ah yes, of course, silly me. It appears I’ve temporarily forgotten my education at L’Ecole [Star Filled]! But how do we know whether a wall contains intrinsic qualities?

–I say, is something wrong? Have you been drinking? You’re asking the most peculiar questions. Knowing whether a wall contains intrinsic qualities is a simple empirical matter: when we try to dive into the depths of the wall to see what is inside of it, we find that we are unable to do so!

–Ah, so the fact that the wall does not allow us to plumb its depths is what makes it not an object. But how can a wall exist without qualities?

–The nature of the wall, as a not-object, is contained within objects themselves. You might call it a virtual object, or an intentional object? The wall is a kind of negative space. It appears to have qualities but in fact does not have any: you and I have the quality of Do-Not-Pass-Through-Wall. This in turn makes it appear to have the quality of Denizens-Do-Not-Pass-Through-Me. But do not be fooled. The wall has no qualities. Similarly, bullets know to disappear when they encounter this wall-space.

–Then let me ask: is “Wall-ness” a quality?

–It is not a true quality, not in the same sense as Do-Not-Pass-Through-Wall is a quality. It’s… let’s call it a surface quality. This is why we cannot plumb the depths of the wall: it is nothing but surface, no depth whatsoever.

–What of bullets, then? A bullet moves, but we are not able to plumb its depths. Yet it sustains velocity, attraction, deflection, replication, and other behaviors! Surely those are primary qualities in the same sense as Do-Not-Pass-Through-Wall?

–Hmmm. You make an interesting point. Very well, I revise my metaphysics: the primary divide in the world is not between objects and not-objects, but rather it’s between objects whose depths can be plumbed and those whose depths cannot be plumbed: the plumbable and the unplumbable!

–That sounds all well and good, but I believe you’re short-sighted. For let me ask: what do we see when we plumb an object?

–By the stars, I know you know the answer to this question but I’ll play along anyway. When we plumb an object, we see mazes of walls, exits, powerups, a sometimes dense spray of bullets, and of course other denizens.

–Yes, and further, what relation does the interior of an object have to its exterior?

–Well, the maze makes up the general geometry of the object’s exterior. The powerups contained within represent the object’s potential for change: collecting these powerups provides the transformative moment when the exterior object changes its properties and behaviors. And the density of bullets and denizens is somehow related to the overall power level of the object itself.

–Yes, very good! So allow me to ask: what do we see when we plumb the depths of ourselves?

–Why… why we see the very same things!

–Is it not possible, then, that only objects that are similar to one another—that is, that sustain the same qualities—are able to plumb one another?

–But that is absurd! Powerups sustain the same qualities as we do—we can see as much when we enter them—but a powerup never plumbs the depths of another object!

–Yet perhaps it could, were it able or willing to do so? It is perfectly normal to enter a powerup, then enter a series of denizens, then exit those denizens, then find yourself back inside the powerup. It’s normal because, by convention, we know that we must have entered the powerup at some point. And while you think you know that you are yourself, that you are a Primary-Path denizen, how do we know we are not 100 levels deep inside some powerup, deluded this whole time into thinking we are not a recursive branch—secondary, tertiary, or worse?

–You mean to say that on my exit from the current maze, I might not find myself in a new body, but rather a body inhabiting the interior of some powerup? Or that I am currently inhabiting a powerup, and on my next exit I will find myself face to face with that powerup for the consumption?

–Yes, I mean to say that. What I mean to say is: like denizens, powerups contain both powerups, mazes, and denizens. Perhaps a powerup is just a denizen who chooses not to shoot or move for whatever reason.

–That’s absurd. Next you’re going to tell me denizens can be consumed.

–I wouldn’t go that far. What I’m saying is that ontologically speaking, denizens are a hair’s width from being powerups, and vice versa.

–How does this talk have anything to do with my theory of the plumbable/unplumbable divide as the two primary ontological states of being?

–Has it occurred to you that perhaps self-similarity is the hallmark of plumbability?

–It has occurred to me, yes.

–Well then: walls, even in their complete alienness as a kind of negative space without recognizable qualities, are presumably similar to one another. Perhaps a wall could plumb another wall, should it choose to do so.

–You’re entering moonbat territory here, my friend…

I was interviewed recently for a news article about videogames and the U.S. military. The article was published, and I wasn’t quoted at all — this is fine with me, as the author twisted the words of a friend of mine who was the main interview subject for the piece. (Deliberately not linking the piece here.)

Anyway, I like to play a game with mainstream media journalists where I answer their interview questions in such an eggheaded way that they’ll be unlikely to publish what I say. And if they do publish what I say, it’ll be a small victory.

Here are my brief, unpublished responses to the author’s questions.

Q: What’s your view on why video war games close the gap between reality and fantasy?

A: I’m not sure they do close the gap between fantasy and reality. I think there is a mediating factor: there is the reality of war, there is the fantasy of the videogame, and there is the fantasy of war that is manufactured by both the military and the media. The videogame simulates the larger cultural fantasy of war, not war itself. Take for example the well-known fact that soldiers in the U.S. military greatly enjoy playing Call of Duty and similar games. I would posit that if Call of Duty were truly realistic, playing those games would be the absolute last thing a soldier would do. However, Call of Duty simulates a war that soldiers do not get to experience. Perhaps it’s the war that they signed up to fight, rather than the war they are fighting.

Q: There is war and there is peace. When it comes to video games, why is peace just a non-starter? And how startlingly emblematic is it of the human state of mind?

A: I don’t think peace is a non-starter — however, most games require some form of conflict. And conflict will exist even during peace time. For example, if you make a videogame about growing plants in a garden, you are still experiencing a human vs. nature conflict, and a human vs. self conflict: you’re managing your own resources, and dealing with things like the passage of time, figuring out when to plant which seeds and where. There are plenty of games about trade during peacetime: economic conflict can be a peaceful activity (although it often is not, and often leads to war!).

It’s worth considering that the first large computers were developed specifically to calculate the trajectories of bombs. With about 20 lines of code, I can write a simple game where you fire missiles at a target. In order to build a simple gardening game, I have to write hundreds of lines of code. This is because answering the question “did the bomb hit its target” means solving one elementary algebra problem and displaying the result. Simulating a garden requires math that is analogous to a complex system of partial differential equations and somehow communicating that to the player.

I suppose what I’m saying is: it’s much easier to simulate war than to simulate peace.

About an hour ago, I finished Alien Phenomenology, Ian Bogost’s new book. What follows are my unfiltered (AKA barely fleshed out) responses. The second section is not going to make sense to you if you haven’t read the book.

Style

When it comes to the style of written philosophy, I am in complete agreement with Graham Harman:

Against the program for philosophy written in ‘good plain English’, I hold that it should be written in good vivid English. Plain speech contains clear statements that are forgotten as soon as their spokesman closes his mouth, since they have already said all that they are capable of saying. But vivid speech forges new concepts that take on a life their own, like good fictional characters.
(Harman, Prince of Networks, p. 140)

Alien Phenomenology is a joy to read. The language never veers into that intentionally obscure academic style, yet retains intellectual value (shocking, I know). But beyond mere accessibility, the prose is beautiful. Opening the book at random and skimming a page, I’m treated to a passage about philosophical speculation as a concrete, pragmatic activity, concluding: “The result is something particular whose branches bristle into the canopy of the conceptual.” (30)

Carpentry

Of the concepts covered in the book, the most personally interesting to me what Bogost terms “carpentry.” Carpentry is the act of building objects that do philosophical work. It’s a brilliant concept. The gist of the argument is that while writing is certainly a good way to express some philosophical ideas, it’s limiting and potentially far too anthropocentric to allow us to plumb the depths of the truly alien.

Bogost’s argument for carpentry spends a lot of time setting it up in relation to academic publishing, where writing often happens for the sake of publishing rather than being read and understood by one’s peers. This portion of his argument is a powerful one, but is (happily) pretty much irrelevant to me: I’m not an academic, but I do practice philosophy, and carpentry just makes good sense to me. I don’t have centuries-old institutions dictating how I should practice philosophy, because I don’t rely on those institutions to put food on the table.

I remember trying to describe carpentry to my friend Darren Torpey a few months ago. His response was, “It sounds like art.” It’s a pretty good response: certainly a lot of art manages to do what carpentry attempts, speculating on the way the world hangs together, and providing metaphors that allow humans to understand that speculation. So how is carpentry different from art? Bogost touches briefly on the question: “unlike tools and art, philosophical carpentry is built with philosophy in mind. [...] Carpentry is philosophical lab equipment.” (100)

That’s an unsatisfying distinction. I find intent-based arguments wearisome, and it’s somewhat ironic to hear an intent-based argument come from an object-oriented philosopher. If I create an object with intent in mind, even if I could somehow imbue that intent into the core of the object, there is no way for others to plumb the depths of that object and somehow retrieve its intent. Further, there is no way for the object to plumb its own depths!

Bogost’s metaphor of “philosophical lab equipment” doesn’t help, either. I’ve worked in labs, and we used plenty of tools that were not built with lab work in mind. Ready-to-hand objects like Scotch tape, paperclips, and scissors are lab equipment, each contributing to the overall work as much as oscilloscopes, spectrum analyzers, centrifuges, and Kimtech Kimwipes.

In fact, at least one example of carpentry in the book was not built with philosophy in mind. While Ben Fry’s brilliant Deconstructulator certainly does philosophical work, it declares its intent outright: to help people develop “insight for how software and hardware work”. I suppose you could unpack “insight” to mean “deep, platform-studies-level analysis of the software and hardware,” but I’m reasonably sure Fry was interested in teaching basic computer science concepts in this case.

My confusion may come down to the fact that the word “carpentry” can mean both “the act of building” and “objects built by a carpenter” (for example: “look at all the carpentry on display in this furniture warehouse”). Philosophical carpentry is the act of building objects that do philosophical work. Yet philosophical carpentry also refers to a body of objects that do philosophy. Carpentry both is lab equipment and the act of building that lab equipment. Does this minor grammatical nitpick matter? I think it does — having a single word refer to both the act of creation and the object created makes it difficult to resolve questions like “is art carpentry?” in a satisfying way.

Practicing ontology

If a physician is someone who practices medicine, perhaps a metaphysician ought be someone who practices ontology. Just as one would likely not trust a doctor who had only read and written journal articles about medicine to explain the particular curiosities of one’s body, so one ought not trust a metaphysician who had only read and written books about the nature of the universe. (Bogost, Alien Phenomenology, 91)

The idea of practicing ontology ties into something I’ve been thinking about at length recently. If ontology is the study of the nature of being, how do you practice ontology? How do we perform applied ontology without being a god? One answer is to settle for being a demiurge. The practice of building a videogame engine is one way to perform applied ontology.

When we architect a game engine, we ask questions like: what is an object? Are objects described by essential classes, or are they simply how they present themselves at this moment? What is an event? Do objects interact in the context of events? When objects interact with one another, what do they have access to? What is time, and how do events and objects resolve themselves within the span of an instant? What is an instant? Etc., etc. Granted, we usually pose these questions in far more technical language, talking about things like class-based inheritance, public and private methods, deltas and whether we have access to the state of variables from previous instants. But they’re the same questions being asked by game engine architect and object-oriented ontologist alike!

While architecting a game engine is obviously not the same as architecting the universe (not only is the former is bound by certain limits, but we know what these limits are), I believe they are similar practices. I claim that architecting a game engine can be good training for doing metaphysics. As I’ve put to several of my game developer friends: just as we talk about building a physics engine for our game, we could just as easily speak of fundamental game engine architecture as building a metaphysics engine for our game.

There was a post on some news site last week with an article by Louis Bedigian quoting an analyst (Arvind Bhatia) claiming that “Zynga loses $150 on every new paying customer.” I read it, thought to myself, “That’s absurd linkbait,” and then assumed that nobody would take the bait. I was wrong: it got picked up everywhere. Sigh.

This morning, Andrew VandenBossche alerted me to an article by Dylan Collins, quoting industry CEO Torsten Reil, that responds, no, you idiots! Your methodology is wrong! “Zynga is probably MAKING $30 on every paying user!” So… here’s what I think: nobody knows what the fuck is going on. (For those of you wondering why I’m writing about this, before I did HTML5 stuff full time, I spent 6 years as a data analyst for game studios, both MMO and Facebook games.)

Surface analysis

Collins/Reil are absolutely right to call the original analysis oversimplified. It was based on a model that completely failed to account for attrition — they’re correct when they state that Zynga certainly gained far more than 400k paying users for their marketing money.

Unfortunately, Collins/Reil pick a number out of thin air (20% attrition rate) which results in a rough estimate where Zynga spends $120 per paying user, and makes $150 per paying user, resulting in a net profit of $30 per paying user. I say unfortunately because if the number is 10%, then by Reil’s metric they’re losing $21 on every paying user. If it’s 30% they’re earning $57 per acquired paying user. It all hinges on their attrition rate, which we don’t know! Some games see 10%. Some games see 90%. 20% seems like a roughly correct ballpark for a mix of successful and unsuccessful games, but honestly we have no idea what it is because we’re on the outside looking in. But the truly weird thing to consider is: Zynga doesn’t know what their attrition number is either.

Models and black boxes

All numbers like this are built on models that analysts put together, and models are built on assumptions. Simple example: when we talk about attrition, what phenomenon do we refer to? Typically we mean “the moment when someone is no longer a player of the game.” Yet in the context of a social game, how do you define that? Facebook users don’t typically uninstall an app — they usually just stop using it. So you have to pick an arbitrary cutoff point. Does someone fall into an “attrition” bucket after 1 week of inactivity? 2 weeks? A month? Remember, this number is arbitrary, so you can adjust that number all you like (within reason, you’re not going to pick 100 years) until you come up with an attrition percentage that meets your criteria. Whether those criteria are “seems more realistic” or “would appease our shareholders” is another question!

But regardless, this attrition percentage then affects all of your other calculations. Now, ideally you want to remain internally consistent once you pick this number, but a dirty secret is that even if you maintain perfect internal consistency in always using “2 weeks of inactivity” as your cutoff for attrition, there will always be dozens of other fiddly and less directly consequential definitions that you can tweak. And the thing is, on some level you have to tweak these numbers! Otherwise you might find yourself stuck with a model that doesn’t reflect what looks like the reality of your game.

To put it another way: the internal game studio analyst’s job is to assemble a black box known as the concept of “attrition” — to the CEOs and CFOs and shareholders and external analysts and pundits at home, this concept seems pretty straightforward: it’s the people who leave your game. End of story. The black box behaves and does its job, reporting a number between 0% and 100%, and presumably you panic if the number is closer to 100%.

But the internal studio analyst needs to assemble this concept from a variety of sources. They might ask the game designers what they see as a “normal” or “natural” amount of time away from the game — if a game is designed to be played during the work week, then you shouldn’t sweat it when someone isn’t playing over the weekends or on Christmas. They might look at historical data for the game and notice that 80% of players who are inactive for 12 days never come back. And 90% of players inactive for 15 days never come back. So maybe we pick 90% and say 15 days is our cutoff. But of course we’re looking at historical data for the current game, which is different today than it was back then, so it’s not a perfect analogy! So maybe we want to rely on data from the last 30 days, when the game was most similar — but now our definition of “never come back” really means “people who were inactive for 15 of the last 30 days and haven’t been back.” But of course, those people “haven’t been back” for a maximum of 15 days since we’re looking at a 30 day window. So now that our historical data is more representative of the current state of the game, our very definition of “never” comes into question!

An infinite regress of assumptions

“ An internal game studio data analyst does in fact work in a vacuum, and will get fired for sharing with outside analysts. This means that the chances that our assumptions are off-base are pretty good.

In summary: games are very complex systems, and the numbers that get thrown around in the media are built on black-box-style assumptions. These black boxes can always be broken down into components, and those components into subcomponents, forever and ever into an infinite regress. If this seems mind-bogglingly weird, well: it is. On some level you need to stop digging into the infinite and come up with assumptions about the way the game works that become the foundation for your models. There’s nothing wrong about that in principle: science does this all the time, and manages to come up with some great models to describe the world. But there’s a huge difference between science and analyzing the metrics for social games. Scientists do not work in a vacuum within their universities or corporations. Scientists do not work with “proprietary data” and they do not run the risk of getting fired for sharing their results and even their methodologies with other scientists. An internal game studio data analyst does in fact work in a vacuum, and will get fired for sharing with outside analysts. This means that the chances that our assumptions are off-base are pretty good. And it means that the numbers that different companies throw around can’t even be compared. “Average revenue per user,” which sounds straightforward, can be based on entirely different foundational assumptions at different companies and on different games.

This whole mess is one of the main reasons I stopped being a data analyst for games. I did not feel comfortable coming up with assumptions that weren’t, on some level, complete bullshit. Now, the level on which these assumptions operated was often very low-level, fiddly stuff. But it was an art, not a science. Which, again, nothing wrong with that — except that the black boxes that I generated were being treated as science rather than as art.

In the end, for the purposes of arguments about how much money a company is making, the only numbers that matter are: how much money is coming into the company each month? How much money is leaving the company each month? Everything else should be viewed with utmost suspicion.

About a year ago on a thread on Google Buzz (RIP) I wrote a quick definition of  ”positive feedback” and “negative feedback” in terms of their technical, systems design meaning. I figured I would repost it here, since every now and then someone asks me what it is and I search my email to copy/paste it. It’s pretty slapdash but it gets the point across, and people have found it useful. So… here you go, internet.

In short: negative feedback is a system that regulates itself. Think of a thermostat. You set it to a certain temperature T. If the house goes above T, it instructs the systems to cool the house. If the house goes below T, it instructs systems to heat the house.

Positive feedback is where things are reinforced and amplified. Imagine a thermostat that cranks the house hotter when the temperature is above T, and makes it cooler when it’s below T.

The main difference is that in the first example, the temperature will tend to remain at T. In the second example, it is highly unlikely the temperature will be T, and in fact it will trend towards either very hot or very cold.

The simplest example of negative versus positive feedback in game design that I can think of is possession rules in 1-on-1 pickup basketball. A negative feedback rule is something like, “If I score a basket, you get to start with the ball.” By giving an advantage to the loser, you’re evening the playing field a bit and bringing the temperature back down to T, so to speak. A positive feedback rule would be, “Winner keeps the ball.” This means winners tend to keep winning, making the house hotter and hotter.

Randy Smith likes to use Mario Kart as an example of a game with strong negative feedback: items like the Blue Shell cause people in the winning position to be at a disadvantage so those behind can catch up. This is also colloquially known as a “rubber banding” mechanic.

Mainly this is in stark contrast to the other way the terms “positive feedback” and “negative feedback” are used in game design: to signify something positive given to the player (reward) or something negative given to the player (punishment). “Feedback” in general can refer to either feedback loops like I describe, or to the general concept of “giving information to the player about the results of their actions.” I would vastly prefer if we started disambiguating the two usages. I’d like “feedback” to retain the meaning I sketch above, which is probably due to my bias as someone trained as an electrical engineer!

Any suggestions out there for a word other than “feedback” that would describe “giving information to the player about the results of their actions”? Or, hell, a different word for “systems whose outputs determine their future inputs” (which is what I describe in my examples above)?

(Edit: “reinforcement” has been variously suggested here and on Twitter for EITHER kind of feedback. See my comment below on why I think it’s only suitable for the “giving information” kind.)

I was planning to do this anyway, but then Zack Hiwiller beat me to it and I got all jealous. Here’s a list of some of my favorite games that came out in 2011 (or at least the ones I thought of on my lunch break).

• Murder Dog IV: Trial of the Murder Dog. I wrote at Paste Magazine about why I like it.
• Crime Zone. Another game by the creator of Murder Dog, this is primarily an exploration of the psychogeography of a member of a police squad that enforces the will of a totalitarian state. For real.
• Deus Ex: Human Revolution. It doesn’t innovate over the original Deus Ex. It’s pretty much the same game except not quite as good. I am completely okay with that because I’m a shameless DX fan.
• Skyrim. This took me by surprise. I found Oblivion really boring, but I very much liked Fallout 3. I typically hate fantasy genre stuff. But I’ve logged about 40-50 hours in Skyrim, and keep coming back to it.
• Vertex Dispenser. I wrote about this game as well. (It’s 50% off right now at Steam!)
• Frozen Synapse. It’s the squad based battle planning stuff I love from Jagged Alliance 2. Runs like a dream under Wine on Linux, too. I had fun with this.
• Soul Brother. Hands-down my favorite browser-based game of the year. Incredible soundtrack.

Unfortunately, most of the other games I enjoyed this year technically have not been released yet so I can’t talk about them here. Oh the trials of being an IGF judge and Indiecade-attending-person.

The unreleased game that stands out the most in my mind is Johann Sebastian Joust, which I played with my coworkers at our holiday party last weekend. Here’s a video of a test run:

Playing Joust at Bocoup from maryrosecook on Vimeo.

I spent $200 on four PS Move controllers to play this (I don’t own a PS3) and IT WAS TOTALLY WORTH IT.