Cost of Living: A Study in Stateful Media with Narrational Agency (Poll: What is Cost of Living?)

I clicked the like on this post, but it didn’t seem quite enough.
Mike, I salute and celebrate your enthusiasm, your readiness to engage with new forms of IF, and your willingness to contribute all this valuable editorial input.

Thanks for being so receptive to new ideas, wherever they arrive from. :grinning:


I remarked elsewhere, and will repeat here, that in some ways my absolute favorite piece of interactive fiction in this year’s competition is Will there be any disqualifications? - ParserComp 2022 community - – the thread about whether or not Cost of Living should be disqualified.

I mean, it’s got everything: it’s funny, it’s thoughtful, it’s wide-ranging (even perhaps “sprawling”), and it’s clearly interactive in the sense that there’s a text box at the bottom where you can type things in and, perhaps after a short time, get a response that changes the direction of the game. It might be “interactive nonfiction”, but I don’t think I’ve seen a game disqualified from an IF competition just because it’s an exploration of nonfiction rather than fiction.

Of course, it’s not the writing of a single person but rather the varied, freely-available text written by more than one person–but that’s also true of the original Cost of Living, so I don’t see why that should disqualify it as IF, either.

Did I love CoL, the original game? Not entirely. (I ended up choosing not to rate it–the only game I played but didn’t rate, not counting one where a bug prevented me from really getting very far–because I wasn’t even sure how to rate it.) I agree with much of what Mike Russo said in his review, and I could probably just set up a bot that would post “I agree with much of what Mike Russo said in his review” every time he reviews something and save myself a lot of time. But I did find the concept intriguing (and, like Amanda, I found it reminiscent of Blackbar, not in a bad or derivative way). Did I enjoy CoL, the thread, more than the original game? Possibly. But that just means that the original, like the original original by Sheckley, is fertile ground for exploration.


Hey, my entry “Sombras y cipreses (ifdb)” won the 2019 Spanish Ectocomp “Le petite mort” with just that, a link to a forum thread were to play a “Parsely-like” game without computers, just persons XD

I would’t feel too surprise with all that beeing a social psicology study from Dorian √.√

1 Like

Ha! Thank you so much, Tahnan, for this great comment!

When I first decided to participate in ParserComp, I thought I would, at most, walk away with a handful of ratings and a polite mentioning from a collective attaboy. But, as you witnessed, in less than 24 hours after the competition starts, reality proved that those initial thoughts were completely wrong. Regardless, I’m glad I stuck around and acted silly for everyone. It really was all in good fun for me. So, thank you for appreciating everything and thanks again for sharing your feelings here! :smile:

P.S. I kind of feel that you’ve obliged me to write about what I learned from being disqualified from ParerComp, but to write my reflection as a Stateful Narration on top of the competition version of Cost of Living. :wink: (Hmm, or is this actually a good experiment for a memoir? Nah, probably not! …or is it?)

I want to start by saying, again, that I really appreciate your careful attempt to make some sense of these works. Like most complex subjects, there is some truth to everything you say, even if I disagree with some of it. I also don’t have a fully realized reply, so I’ll just be “lazy” and pick out some quotes to reply to.

Ever heard of bike shedding? That’s why everyone takes this stuff too seriously! :wink: Anyways, don’t take this reply too seriously, either.

Ha! That makes sense now! I think this is the cruelest trope in IF. Design so subtle that it goes — Woosh! — right over an audience’s head. Didn’t this happen to Emily Short’s Blood and Laurels? I think there were moments like this with Cost of Living, where the audience thought it was merely “parroting input, like Mad Libs”. Using Google Translate to play Cost of Living would not work, which is another reason that Cost of Living is more than just Mad Libs.

(By the way, I made some progress on a Spanish translation. I used Google Translate on the stateless and stateful writing, but I stopped short of translating the dictionaries. Should I keep working on this Spanish translation? Or just wait until next time?)

I’m not trying to defend this because you’re right that this is shallow design. This time around, this wasn’t a priority. Even so, I did reply to an audience’s frequent input, right here:

Storyteller: But you really like saying true today, huh?
   Audience: Not sorry, because true is how I feel today!

Storyteller: Ha! Okay.

I’m so happy you point this out, because this highlights the critical difference with my approach. I’ve come to see that mechanics (e.g., textbox, CLI) do not define a genre. To me, mechanics are functionally equivalent to a programming language, and since programming languages are Turing complete, then any set of mechanics are functionally equivalent to each other. In practice, what I believe this means is that “choice mechanics” can implement “parser games” and “parser mechanics” can implement “choice games”. Even in classical parser IF, its input and output theoretically reduces down to such a “textbox” (as evidenced by the dash-outlined boxes in the PR-IF Card). In other words, mechanics are irrespective of genre! So what does this have to do with the critical difference with my approach? Well, my approach is more like an architectural approach (just like one of my inspirations, REST architecture). This means my approach is not about a specific set of tools or implementation, but rather a set of guidelines to implement them into any arbitrary medium. As you rightly point out, there is nothing that prevents my approach from working with story-based agency, or from using a command line prompt with agentic verb-noun commands (i.e., FEEL SAD, ACT HAPPY, FEIGN ANGER, etc.). Indeed, I even experimented with full-sentence parsing from a CLI (but I didn’t like the potential of the results).

I would say that most of the parsing for my approach happens during the authoring of a work, when a human storyteller parses the proxy-audience’s input in order to form an appropriate reply. Now, forming an appropriate sentence for an affective input is not easy, and this is where the most difficult parsing begins. Thankfully, I have a LOT of notes on survey research methodologies and conversation analysis. (I didn’t use any of this research with Cost of Living because it was out of scope for that demo.)

I’m going to provide a critical analysis of Costikyan’s article. But not in support of whether or not Cost of Living is a toy, or a game, or literature. Instead, I seek to highlight the difficulty of answering the questions, “What is a game?”. (Most of my retorts are silly, but they stand, nevertheless.)

Critical Analysis (Caution: Copious snark ahead! Please do not take anything too seriously!)

Oops! I orignally exceeded the character count of a post (“Body is limited to 32000 characters; you entered 33338.”). Please see the next post if you want to read through this analysis first.

This is an oversimplification of literature. This same reasoning can also oversimply a game — isn’t the objective of a game to reach the “end of game”? If the objective of both games and literature is to simply “reach the end”, then we can discard this similarity as redundant in pursuit of a contrastive analysis between games and literature.

I’m going to be hyperbolic here, but for me, writing is the ultimate objective of society. Humanity formed the pillars of society with writing — law, medicine, math, engineering, literature, etc. — and we all laud the authors of these writings (actually, this is the crux of academia). Who invented the first laws? Who invented calculus or algorithms? Who wrote the first novel? Now, who invented chess? Or poker? Or golf? Why don’t we know this? Why aren’t these games important enough that we remember who made them? Indeed, writing itself (or, at least, some form of narrative) is the precise element that transforms a toy into a game. And vice versa — without writing, a game becomes a toy. How does Senet work? We don’t know, so now it’s just a toy. I’ve digressed considerably, but what I’m trying to underscore is that the objective of literature is not to “end the story”. At least for me, my objective with literature (or any book) is to attain some “self-acceptance and societal acceptance”. If this doesn’t happen, either the author or I did something wrong.

Like all interesting debates, I think there is an underlying model that validates seemingly contradictory opinions. In this instance, you are right to claim there is no “struggle” and, at the same time, I am right when I claim there is a “struggle”. Let me explain this with my context-problem-solution model and my story-plot-narration model.

Here’s my first rebuttal, which is also an excerpt from my notes:

Readers engage with stories through the Problem (i.e., here’s what I would do) by either accepting the Solution (yup, I was right!), rejecting the Solution (no way that would happen!), or learning from the Solution (wow, I’m going to try that!).

To me, with Literature, engaging with the Problem is the “struggle” while learning from the Soltuion is the “victory”. You can also see here that Solution has different expectations within both Literature and Game. With Game, a player expects to provide a Solution, but with Literature, an audience expects a storyteller to provide a Solution. (This is also why I believe the broader literary audience rejects story-based agency, because it often propositions an audience to provide a solution.)

Here’s my second rebuttal, which is also an excerpt from my notes:

Ludology and Narratology considers themselves incompatible, and I think the reason is due to a consequence of a hierarchically-defined relationship between between play and game (i.e., play → game), where those concepts take on a diametric opposition to each other. Since storytelling intuitively resembles play more than game, it’s a mere consequence to claim a diametric opposition between game and storytelling. But, in fact, game and play are independent activities, sharing only physical reality. The hierarchy play → game is false, making false the diametric opposition between game and play, and false the opposition between game and storytelling. Also note that by combining game with commentary, it’s possible to create a functional equivalency (≅) between storytelling and game + commentary.

Transition context problem solution story plot narration
Physical Reality context story
Physical Reality → Game or Play
Game context problem solution story
Play context story plot
Play → Storytelling
Storytelling context problem solution story plot narration
Storytelling ≅ Game + Commentary
Commentary plot narration

Let me unpack this by starting with a definition of my story-plot-narration model.

story = ordered sequence of the events
plot = unordered human-curated subsequence of the story
narration = "human element" of the plot

Now I can use this definition to provide a sequence of my reasoning.

  1. Physical Reality is a Context that has an ordered sequence of the events (i.e., Story).
  2. Game applies a set of rules and conditions (i.e., Problem and Solution, a.k.a. “struggle and victory”) over Physical Reality.
  3. Play focuses on the interesting parts (i.e., human-curated subsequence of the story) of Physical Reality.
  4. The common refrain is Play becomes Game after adding “struggle and victory”. But for me, Play and Game are separate concepts that happen to share Physical Reality. In other words, Game does not build atop Play.
  5. Instead, this table makes the claim that Storytelling builds atop Play. By starting with Play and then adding a Problem and Solution (a.k.a. “struggle and victory”) and Narration (i.e., the “human element” of the Plot), this creates Storytelling.
  6. What makes a Game compelling to anyone besides its players? What makes a Game as societally-relevant as Literature? Commentary! (Ever watch a televised sporting event without the commentary? So boring!) Plot and Narration add the “human” touch to a Game, and this is what gives Games their popular appeal. This is especially true for sports commentary, whose commentators often fetch astronomical salaries. For instance, Tom Brady accepted a 10-year $375 million dollar contract to be a sports commentator. Yes, that’s right, $37.5 million US dollars per year to add Plot and Narration to a Game. (For myself, this is more evidence that narration-based agency has parity with story-based agency.)

What does all of this have to do with our contradictory opinions? The fifth item shows that Cost of Living is 1) not a game (because it contains both Plot and Narration), and b) does have “struggle and victory” (because it contains both Problem and Solution). So we are both kind of right! And this is why I don’t consider most “text adventure” as a work of literature (for example, Alchemist’s Gold is an author-proclaimed “text adventure” that lacks both Plot and Narration), while I consider most “interactive fiction” as a work of literature (for example, Of Their Shadows Deep is an author-proclaimed “interactive fiction” that contains both Plot and Narration).

So true! This reminds me of how “cinema” started off as a “toy”, i.e., through Kinetoscope and later the Nickelodeon. My approach is definitely a toy today, but what about tomorrow?

Here’s some figures from Statista that make a claim about recent digital sales in the United States.

So, yes, by comparing these claims of past and future annual sales, I might have a handicap by not being a game. However, story-based (i.e., game-based) enhancement of literature is largely explored territory. In other words, we know that many game conventions are incompatible with interactive fiction (racing interactive fiction? RTS interactive fiction? FPS interactive fiction?). I’m not sure how expansive my narration-based territory is, but’s it’s largely unexplored. If I can figure out how to interlock my stateful narration-based metaleptic conversational frame to an arbitrary stateless frame, then I have an opportunity that’s quite large (stateful narration of any past, present, or future literary work). So, yes, by comparing the opportunity of story-based agency against narration-based agency, I might have an handicap by not being a game. (Note the definition of handicap is “the disadvantage or advantage itself.”)

1 Like

Here’s a critical analysis of Costikyan’s article. This is not in support of whether or not Cost of Living is a toy, or a game, or literature. Instead, I seek to highlight the difficulty of answering the question, “What is a game?”. (Most of my retorts are silly, but they stand, nevertheless.)

Critical Analysis (Caution: Copious snark ahead! Please do not take anything too seriously!)

What makes a thing into a game is the need to make decisions.

Is ordering a meal in a restaurant a game? “The menu is tightly constrained, the ingredients clear, and satisfaction requires you to think about several considerations, like quantity and quality, and price.”

But the game is not intrinsic in the toy; it is a set of player-defined objectives overlaid on the toy

And what exactly are these “player-defined” objectives? And how are they “overlaid on the toy”? I talk about this after this analysis.

But SimCity itself has no victory conditions, no objectives; it is a software toy.

The objective of Cost of Living is to have an audience use their own abductive reasoning to either accept, reject, or learn from the abductive reasoning in the metaleptic conversational interstitials. However, this objective is not disclosed to an audience. So, do we move the goal by claiming that all objectives must be visible?

SimCity is a game – at least when a user plays it as a game

This is basically a tautology, right? Anything is a game if you make it a game. Frankly, most of Costikyan’s arguments feel like tautologies.

“Cooperative games” generally seem to be variants of “let’s all throw a ball around.” Oh, golly. What fun. I’ll stop blowing deathmatch opponents into gibs for that, you betcha.

It seems that being a jerk is a common trait among us designers, huh?

So – why don’t they just get rid of the puzzles? Why not just make it an interactive story? […] But never mind that; without the puzzles, it’s no longer a game. […] The puzzles, and the struggle involved in solving them, is what makes Grim Fandango a game.

Earlier, Costikyan states that a puzzle is not game. But here, if a work contains a puzzle (or sequence of puzzles) then that work is a game? How?

Life is the struggle for survival and growth. There is no end to strife, not this side of the grave. A game without struggle is a game that’s dead.

Is life a game?

Making something difficult makes it more enjoyable? That’s not how we view everyday life […] But it is absolutely true of games. We want games to challenge us.

I want life to challenge me. But I suppose a sample size of 1 (i.e. Costikyan himself) is adequate proof of evidence for his assertion.

Cookie Clicker is a commercially successful incremental game that had no struggle, which also spawned it’s own industry of mobile games referred to as “incremental or idle games”.

They aren’t any fun if they’re too simple, too easy, if we zip through them and get to the endscreen without being challenged.

Costikyan is correct, but only because he is implicitly referring to his target audience. The problem is that this statement reads like it applies to all of humanity across all time. I bet the rise commercially lucrative idle games are still a surprise to Costikyan.

As a result, Ultima Online is, under most circumstances, a Hobbesian war of all against all, the game filled with a palpable fear as people flee from one another, trying to avoid potentially deadly encounters. In EverQuest, by contrast, players frequently stop to help each other out, strike up conversations with random passers-by and in general behave with a degree of social solidarity. Clearly, I prefer the latter.

Didn’t Costikyan just pooh-pooh cooperative games? But here he claims to prefers cooperative games?

The structure of the story, however, creates a single, unchanging narrative that the reader cannot alter. Narrative structure is one dimensional, because you can follow only a single path through a story.

Barthes’ The Death of the Author challenges this statement five decades ago, and reader-response criticism is still around. I bet the rise of our post-truth world is still a surprise to Costikyan. (For me, our “post-truth” world is more evidence that narration-based agency is a viable alternative to story-based agency.)

But a game shapes player behavior; it does not determine it.

Exactly! With a game, the player expects to produce the solution in the context-problem-solution model. I’ll return to this soon in this reply.

But economists do assume that people behave rationally, by and large.

But also, this economic assumption is strictly in the mathematical sense. This assumption of rationality is akin to banning a division by zero in algebra; allowing division by zero invalidates all of algebra (because then anything can equal anything); likewise, accepting irrational actors invalidates many classical economic theories. Economists know that people are (psychologically) irrational. For instance, a person can intend to be rational while still acting irrational. This shows that psychological “rationality” is subjective. This reason, and many other examples of psychological rationality, is why economists just assume this matter away from their mathematical economic models.

It helps, in other words, to think of a game’s structure as akin to an economy, or an ecosystem; a complex, interacting system that does not dictate outcomes but guides behavior through the need to achieve a single goal: energy, in the case of ecosystems; money, in the case of economics; victory, in the case of a game.

There are more goals to a game than victory, right? The author belittles on one MMORPG for allowing PKs, then lauds another MMORPG for banning PKs, which the author claims this promotes “prosocial, moral” behavior. Well, “community” is another goal of playing games (perhaps the ultimate goal?). Also, none of those systems are alike! Costikyan is committing what economist Friedrich Hayek accused other economists of doing, “it involves a mechanical and uncritical application of habits of thought to fields different from those in which they have been formed”. In other words, everything can be anything with some oversimplification.

If I took this article’s section at face-value, and I believed that all of these structures’ components are interchangeable, then I can just put all of my energy into designing the “happy path” (i.e., game structure) because I can assume my players are rational (i.e., economic structure). Can you image the state of IF with this design philosophy? But, again, economic “rationality” is an assumption to simplify mathematical economic models (economic structure), and it is not a psychological assumption of players (game structure).

(I have the feeling that the only editing and peer-review of this article happened when they accepted Costikyan’s membership into their club: “Congrats on making it through the gate. Now you’re welcome to do whatever. Run free, you wild child!”)

Indeed, if I had my way, a solid grounding in economics would be required of anyone seeking to learn about game design.

Gee, I wonder why he moved on to be a consultant…

Nonetheless, its real-world value exists only in the context of EverQuest; if […] EverQuest servers shut down, my notional possession of a Bloodforge hammer will immediately cease to have meaning – and no one will be willing to pay money for it.

Costikyan stated a serious interest in economics. Then states that money has “real-world” value? Doesn’t his “endogenous” argument also apply to monetary systems? We already know that a system’s endogenous value does not transfer across systems; this is intuitively obvious to everyone, right? And monetary systems are a terrible example to use, because they are not a “real-world” system. Weather patterns are a “real-world” system; monetary systems are “social reality”. What happens when a country defaults on a loan from the IMF? What should I do with all of my Italian lira? Why can’t I spend Greybacks in the USA?

Games are fantasy. I don’t mean that all games are about orcs and elves and magic spells, although far too many are; I mean that they ain’t real. The fact that they aren’t real is part of the point.

So CTE in contact sports (e.g., boxing, USA football) is fantasy? What about the endogenous value of these players? Or their exogenous value (e.g., sponsorships)? In this section, the author is trying to provide a counter-example to test this “theory”, but all he’s really doing is cherry picking his way to inconsistency.

Games do the same. “Bloodforge hammer” has no meaning, except in the context of EverQuest.

What about the concept of “checkmate” from the game Chess? Isn’t this a counter-example of endogenous meaning from a game that also has exogenous meaning in the “real-world”? The endogenous and exogenous meanings are the same with checkmate — out of moves, game over!

Achieving a kill in a Quake deathmatch will do nothing for you in the real world, but may elicit glee or satisfaction when you’re playing the game.

Dennis Fong “achieving a kill in Quake” literally earned himself John Carmack’s own red Ferrari 328 GTS convertible in 1997 (five years before publishing this article)! I bet the rise of pro gaming leagues are still a surprise to Costikyan.

At last we have a functional definition of “game”: an interactive structure of endogenous meaning that requires players to struggle toward a goal.

Cost of Living = Game?

  1. interactive structure = text I/O
  2. endogenous meaning = affective labels (emotionally-valenced words)
  3. player = audience
  4. struggle = affective concepts (emotionally-valenced empathy/sympathy)
  5. goal = abductive reasoning


A question almost immediately arises: If “the game” is a subset of “interactive entertainment,” what forms of interactive entertainment are excluded by our definition? My answer: None. Or none worth the powder to blow them to hell, anyway.

Cost of Living = Game?

  1. Cost of Living = interactive
  2. Cost of Living = entertainment


A site that provides articles for you to read, or video clips to watch, or music to download is indeed providing entertainment – but it does not allow you to interact with the entertainment in any meaningful sense.

What if that site provides a social community (e.g., comment section)? Is that site now a game because it’s “interactive entertainment”? I bet the rise of YouTube is still a surprise to Costikyan.

It could be unstructured. But I have a hard time to imagine a completely freeform, unstructured form of entertainment, unless it be simple conversation – and certainly people find online chat entertaining.

I bet the rise of social media (e.g., Facebook, Twitter) is still a surprise to Costikyan.

But merely because I find something entertaining does not mean that it is entertainment

This is an argument about audience interpretation versus authorial intention. This argument is about as fruitful as “parser versus choice” on the intfiction forums.

Entertainment is a side effect, not the purpose.

Where does this side effect come from? The subjective opinion of an audience, right? So if a work is objectively “interactive” and subjectively provides “entertainment”, then it’s a game to that audience? Then why isn’t a puzzle a game? If an audience considers a puzzle as entertainment, then that puzzle is “interactive entertainment” and, by Costikyan’s extension (game ⊂ interactive entertainment), that puzzle is also a game? To me, this contradiction unravels this whole article.

Only if the meaning has direct, one to one connection with the real world is it not “endogenous” – as is the case with history, or the stock market. And if the meaning is directly connected to the real world, you have something of practical value; not an entertainment form

So checkmate (a.k.a. out of moves, game over!) only has practical value? The goal of Chess (i.e., checkmate) is not entertainment, it’s just practicality? What’s practical about Chess? Is Chess even a game?

Perhaps our non-game “interactive entertainment” can eschew struggle? This is feasible […] Or perhaps we can have “interactive entertainment” without a goal? Again, in principle you can.

This is a consequence of an article that’s cobbled together from cherry picking arguments. First, Costikyan uses economic “rational actors”, which is a mathematical concept of economics. Second, Costikyan states that “games” are a subset of “interactive entertainment”. Does he mean subset in the mathematical sense? Since he previously uses a mathematical economic concept in the first point, I’m reasonable to assume that in the second point that subset is the mathematical concept, right? Well, subset also means equality, unless it’s a proper subset. But since the author seems to be cherry picking whatever suits his own intuitions (economically rational versus psychologically rational, layman’s subset versus mathematical subset), the article is ambiguous and thus self-contradictory.

The search for “interactive entertainment” that isn’t games is motivated by repugnance for games – those cheap, gaudy, violent, unpleasant, degraded pop-culture entertainments for ill-read, ill-mannered little boy brats. It’s a search by those who wish to achieve something “higher” and of greater merit and value than can possibly be achieved among such a puerile and repulsive form as “the game.” In short, the search for non-game interactive entertainment is wrongheaded, inspired by a failure to apprehend games and a foolish, reflexive response to what they represent, in our culture, at this point in time.

This article is not a case for a critical vocabulary in games. Instead, this article is a (thinly-veiled) reaction to the vigorous game censorship discussions that were happening at the time. This article was first published in 1994 around the time that the US Senate hearings on video games were occuring and when the ESRB formed.

Any form of “interactive entertainment” that isn’t a game must be noninteractive; or not entertainment; or pointless. […] Interactive entertainment means games.

Again, what about puzzles? A web-based crossword puzzle is interactive, is entertainment, and has a point. But it’s still not a game?

Creating sensory pleasure is important, and when you design, it’s worth thinking about how you will do so. But it’s a supporting factor, not the essence of design.

I bet the rise of Apple products is still a surprise to Costikyan.

I don’t think chess would benefit from a beginning cut scene explaining about how the game is a war between two brothers.

Quick, someone cover @aschultz eyes so he doesn’t read this!

And when designing, you need to identify what it is that players are going to find challenging about your game, and why that challenge will be compelling.

The challenge for the audience in Cost of Living was to find a reason to empathize and sympathize with the awful decisions that Carrin makes in his life. The stateful portions of Cost of Living were meant to explore and support this challenge. And this challenge is compelling because when we ourselves make awful decisions in our lives, don’t we want someone to empathize and sympathize with us?

But actually, [game design] is among the most difficult creative disciplines, precisely because we’re creating structures that people are going to use in every possible way, and use in ways we cannot anticipate.

But the game designer is a rational person, right? So, it’s fair to assume that this rational game designer would address all rational possibilities in a game before publishing that game, right? Then why can’t this game designer address a player’s “use in ways we cannot anticipate”? If a rational game designer cannot anticipate a player’s use of their game, does this mean that that player is irrational? Or is this game designer irrational because that player thought of a rational move that this game designer did not anticipate?

Games are an artform unlike any other, because the product is not passively received, it is not something specified to the last splotch of paint and every comma.

My theory of stateful and stateless media shows that “passively received” is ambiguous, at best. A stateless book changes the state of a stateful audience (this is the storyteller’s intention, at least). If a phenomenon causes a state change, it’s unintuitive for me to label that as “passively received”. If anything, I would use the phrase “passively received” to label the phenomenon where an audience reads a book and that book does not cause a state change; for example, I “passively received” the message that Cost of Living wasn’t welcomed at ParserComp. :wink:

Game design is, therefore, the creative attempt to imagine, a priori, the kinds of experiences players will have with your game, and through that act of imagination, to create a structure to point them toward the kinds of experiences you’d like them to feel.

Doesn’t this describe good design principles in general, à la Norman’s The Design of Everyday Things? Doesn’t this describe my proxy-audience theory? If so, then how am I using game design to produce something that’s not a game?

Rather, a game, as it is played, is a collaboration between the developers and the players, a journey of mutual discovery, a democratic artform in which the shape of the game is created by the artist, but the experience of the game is created by the player.

How is this not the case with everything? How do I love a movie (or book, or building, or game, or clothes, or music, or vehicle, or, or, or…) but someone else hates it?


When emotional values are judged by physical score, the result is miserable failure. Those
Costikyan arguments sound suspiciously like the one Chris Crawford proposed. “Games must have goals!” And Siboot is a failure.

1 Like

I couldn’t agree more!! Ekman’s Basic Emotions? Boo! Russell’s Affective Circumplex? Yay!

Crawford is so tragic, huh?

Please don’t let me stop you!

Apologies for the long-overdue reply here, I’ve been slightly overcommitted on IF stuff lately, but wanted to get back to this conversation, since there were definitely some strands I wanted to revisit.

I think this is right, and to a certain degree is true of choice-based works too – I’ve seen authors complain about reviews that say “this story doesn’t have any branching, it’s completely railroaded” when in fact their games lots of responsiveness, it just felt completely natural and therefore, paradoxically, artificial, to the player. It seems like in these forms of IF, there might be a sweet spot of just enough friction so that you notice the gears turning, but not so much that things seize up.

But regardless, I actually I agree with the drift of where you’re going with that final question – I think a lot of the history of parser-based IF through the aughts and early teens was trying to ramp up accessibility and onramps to new players, through things like Aaron Reed’s various ease-of-use extensions for Inform 7 and introductory games like Andrew Plotkin’s Dreamhold, which were specifically intended to be a gateway to IF. My sense, though, is that while these led to some design improvements that made parser IF more pleasant for people who like parser IF, they didn’t necessarily catalyze a major influx of new players, because the fundamental model of requiring players to internalize limits on their input can’t be gotten around (your AI Dungeon example is on point here, because the magical promise there, that you can type anything you want and it at least sort of understands, did gets tons of non-IF players to try out a text game).

I don’t mean this as a criticism of parser IF by any means – there are lots of forms of media that are somewhat niche and less-accessible, and that doesn’t make them any less rewarding or interesting to the folks who create and consume those works. But just to agree that the possibility-space of parser IF feels relatively bounded to me, and probably isn’t susceptible of too many more major structural breakthroughs that will open up radically new modes of player engagement.

I also agree with the strong form of your argument that AI-based approaches to story-based agency – what I think you refer to as generative ones – will inevitably fail, or at least, I think they’ll always fail for me. I know a lot of folks much more steeped in the theory and practice of narrative games find this branch of things incredibly exciting, but to me it feels quite sterile, because I mostly read (and play text-based games) to encounter another human being and their view of the world, whether that’s through the jokes they find funny, the clever tricks they embed in a puzzle, or – ideally, in terms of what I tend to enjoy – how they present characters as they navigate change and conflict. Sure, I’ve seen AI-based approaches superficially present these things, but they always have the flavor of a clever trick or uncannily successful mimicry rather than anything I can seriously engage with (I’m willing to believe that this attitude of mine will seem hopelessly blinkered in fifty year’s time, but so it goes).

So yeah, this all makes sense to me, and I get why you’re pushing in the direction you’re trying to push.

Yeah, definitely! I think in these terms, what I’m describing as “magic” is the moment where the game program bridges the gap between the author imagining a notional audience (here, an audience that’s responding with sarcasm), and the player experiencing the author’s imagination (realizing that the author anticipated this response), to create the spark of engagement. That’s the feeling I get, at least, and it’s the same when I try some obscure command in a traditional parser game and find that the author’s already there and has been waiting for me to try that this whole time (this is an experience that’s significantly harder to capture in choice-based works, I think, though it’s not impossible).

Yeah, I think it’ll be worth talking through a bit of a debrief of this jam vs. competition question – I sketched out some initial thoughts in my review of Anita’s Goodbye. It’s tricky when different communities when different norms brush up against each other! I will say that part of the friction here was due to the specific nature of ParserComp, which is a reaction to the fact that the main stream of the IF community has gotten much, much more accepting of non-parser text games, to the extent that traditional parser-based IF is a minority in the major competitions and festivals – in some ways it’s too bad that Cost of Living wasn’t a Spring Thing entry, where I think it would have fit right in and nobody would have felt the need to even gesture towards gatekeeping.

Anyway I wouldn’t worry about fallout – in my opinion it’s awesome that the experience has wound up with you engaging more deeply with this community, and the downside was really just that you’d get some negative ratings for your game.

This is a fair point! The social media example points to the idea that part of the power of engagement can come from being recognized – I’m not on social media so I could be very wrong about this, but I think part of the draw is being able to project a certain image of oneself and see that received and recognized by one’s followers, which isn’t too distant a reward-state from how it felt when I realized you’d anticipated me being sarcastic. Though I suspect the overhead is higher in the interactive context. Like, it’s easy for a person to post something on social media and imagine how their followers are reading and understanding it, and what a “like” represents; for IF where it’s harder to picture the person behind the words, the responsiveness probably needs to be higher to trigger the same response. But anyway, this is part of why I’m interested to see what the engagement loops look like when your approach is deployed in other contexts beyond this story.

I did know that! But yeah, I’d agree about the unpredictability of some of his solutions, at least as to the story that’s said to have invented the genre, The Murders in the Rue Morgue. It is the jumping-off point for an anecdote that’s made me laugh harder than pretty much anything else this decade, actually – see “We Don’t Talk About the Orangutan” (this spoils the story, but honestly, there’s not much to spoil and the anecdote is worth it).

That makes sense – in some ways, maybe I have a more depressing read of the story after all, because while I agree that that’s what I would do in his circumstances, my take was that the reason he didn’t wasn’t due to external factors, but because his psychology had been so distorted by this messed-up society that the capacity to dream and create art had basically withered away.

Yeah, I think there’s an anecdote about the economist John Maynard Keynes – one of the real luminaries of the early 20th century – thinking that automation would lead to 15-hour work-weeks, and a central problem of the societies of the future would be how to deal with folks having all that leisure time. There is a lot about capitalism and social norms to unpack in explaining why he got that wrong, I think…

I agree here too, I think you should feel good about this! The nut isn’t fully cracked by any means, but the theory that this new approach to IF could work really well is still very much a viable one to my mind.

This makes sense to me – when I wrote an autobiographical work (which was definitely somewhat scary to do, though I have to confess I help a lot of stuff back) it was notable to me how many of my testers and reviewers spontaneously shared a lot of details about their own lives and understanding of themselves. There’s a very human reciprocity that gets prompted, I think.

Anyway, definitely food for thought! I’m sure you’re taking stock post-Comp, but I’m definitely curious about where you’re headed next.


I wonder if counter-intuitively would be a more precise adjective than paradoxically? When I read about Emily Short’s Blood & Laurels, for me the standout feature was computer-authored prose that was as natural and subtle as human-authored prose. I thought this made B&L a big deal — basically, the dawning of Murray’s Holodeck. Complete state-of-the-art in rule-based generative IF! Then I learned that the audience was quite underwhelmed with the generated prose (likely because that prose was story-based content, such as process descriptions or room descriptions). I thought, this is a true paradox, like a contradiction in the sense of Russell’s paradox, and not like a counter-intuitive insight with benefits. In the CoL developer notes, I claimed that “I demolished everything I knew about interactive storytelling”. But really, learning about this paradox made the foundations of story-based agency shake so much that there was nothing else — not even arguments like, “there might be a sweet spot of just enough friction so that you notice the gears turning, but not so much that things seize up.” — that could stop the foundations from finally collapsing. In hindsight, this was an over-reaction (despite now being happy with the consequences of my reaction). Please know that at the time, I thought this paradox (Emily’s Paradox? Versu’s Paradox?) was responsible for stripping away someone’s IP. I’m still terrified to think this. This is also why my approach is now 100% human-generated storytelling (details forthcoming).

Wow! Anyways, I didn’t mean to make this into an origin story. Haha! I just wanted to say thanks for the reply, and also please take as much time as you want or need to reply. Like AZ said, this post is

And also, this is just a placeholder reply. I’ll follow up in a few days with a better one. :grin:

I’d like to share some details on a new control flow that I created. Why did I do this? I’d like to say that Conversation Analysis influenced me to seek new modes of expression. But really, I was sick of working with the very elaborate state machine diagrams that structured all of my previous works. So, this is my attempt not rely on diagrams while writing IF.

I also wanted my stateful writing process to follow the flow of conversation, which has a main through-line that provides forward momentum, and also has riffs that can explore or recontextualize a topic of conversation. In other words, conversation is simultaneously free-form and rigid. So, how do I translate this into writing stateful narration?

My search for inspiration stopped at the switch statement. This control statement has a top-to-bottom flow with fall-through behavior. For me, the switch statement was like “a main through-line that provides forward momentum, and also has riffs”.

So I developed a control flow structure that is similar to a switch statement. And there are no if-else statements or for loops to control the flow of a conversation. Here’s an excerpt from the introduction, where the Storyteller is checking if the Audience is the target audience for the work.

Storyteller: I pay attention if people are good or bad. $naThree
             I enjoy finding many solutions to a problem. $saTwo
             I form opinions about everyone. $naOne
             I enjoy puzzles that I must solve. $saThree
             I prefer complex instead of simple things. $saOne
             It bothers me to remain neutral about people. $naTwo

Storyteller: Thanks for answering my questions. Now, let me have a look…

^^^convo quiz.narration.isHigh
Storyteller: So far, this is looking good.

^^^convo quiz.narration.isHigh quiz.story.isLow
Storyteller: Okay, these results look great to me! I think you're really
             going to appreciate reading this experiment in literature.

^^^convo quiz.narration.isHigh quiz.story.isHigh
Storyteller: All done, and not bad. You'll appreciate this more if you try
             not to expect too much "game" stuff.

^^^convo quiz.narration.isHigh quiz.story.isNone
Storyteller: Actually, never mind. My apologies.
             I think you'd be happier reading a book instead.

^^^convo quiz.narration.isLow
Storyteller: Hmm, I think this might still work somehow.

^^^convo quiz.narration.isLow quiz.story.isLow
Storyteller: I just hope you don't mind all the "literary" stuff here.

^^^convo quiz.narration.isLow quiz.story.isHigh
Storyteller: Just think more "literature" and less "game" and I'm sure
             you'll be fine.

^^^convo quiz.narration.isLow quiz.story.isNone
Storyteller: Oops, never mind. Sorry about that.
             I think you'd be happier reading something else instead.

^^^convo quiz.narration.isNone cool.hasPE
Storyteller: Wow, that's a strong change from how you felt earlier.
             You really had me thinking you were interested.

    Audience: Guess I changed my mind.

^^^convo quiz.narration.isNone cool.hasNE
Storyteller: Looks like I was right earlier.

   Audience: Yeah, I tried to tell you.

^^^convo quiz.narration.isNone cool.isN0
Storyteller: Well, you are definitely not on the fence anymore. So…

^^^convo quiz.narration.isNone quiz.story.isHigh||quiz.story.isLow
Storyteller: I think you'd be happier playing a game instead.

   Audience: You got that right.

^^^convo quiz.narration.isNone quiz.story.isNone
Storyteller: Sorry, but I don't think you're going to like this.

   Audience: No. Probably not.

The first thing you need to understand is that the interpreter evaluates all of these ^^^convo expressions from top to bottom. If a ^^^convo expression is empty or evaluates to true, then the interpreter displays that ^^^convo to the audience. For me, this gives me some of the properties of a switch statement, such as the top-to-bottom flow with fall-through behavior.

Let’s take a closer look at the conversation around the quiz answers with high narrational preference.

^^^convo quiz.narration.isHigh
Storyteller: So far, this is looking good.

^^^convo quiz.narration.isHigh quiz.story.isLow
Storyteller: Okay, these results look great to me! I think you're really
             going to appreciate reading this experiment in literature.

^^^convo quiz.narration.isHigh quiz.story.isHigh
Storyteller: All done, and not bad. You'll appreciate this more if you try
             not to expect too much "game" stuff.

^^^convo quiz.narration.isHigh quiz.story.isNone
Storyteller: Actually, never mind. My apologies.
             I think you'd be happier reading a book instead.

Using a switch-like control flow accomplishes two tasks. First, it helps to group a conversation by topics. Second, it provides for naturally occurring (yet structured) moments where a storyteller can correct a prior statement. For instance, refer to the third topic riff, ^^^convo quiz.narration.isHigh quiz.story.isNone. Diegetically, this third riff is a contradiction, so now the storyteller has to self-correct in regards to a prior statement, which is a common feature of real-life conversation.

Well, this is all I want to share for now. Thanks!

1 Like

Looks like table based flag system. Neat!

1 Like

Thanks again, Mike, for the reply. I’m very relieved that my vision might be starting to coalesce for you and others.

How did this sentiment become a blanket expectation over all stateful media? It’s a reasonable statement for comparing a work to Colossal Cave Adventure, but who normalized this for everything else? Experience is unidirectional, from unknown to known. In other words, “branching” is also a mental phenomenon (i.e., noematic interaction) and not just strictly story-based agency. Cardona-Rivera’s Achieving the Illusion of Agency has some interesting data on this.

Exactly! And not only must new readers learn this extra skill, they must also convince themselves that such limits are aesthetic to the medium. In my experience, this is not an easy task for new readers. The only available head start for a new reader is a history of playing video games or role playing games. Otherwise, a new reader starts with only reading skills (and, of course, “internalizing limits on their input” is not a reading skill). I was conscious of this aspect when I designed my stateful narration approach. Assuming a future when my affective label dictionary is complete, new readers do not need anything beyond reading skills to engage with my approach.

I agree with this, but I also hope we’re both wrong one day. It goes without saying that the next innovations will come from tomorrow’s readers. That’s why I’m glad this community is still thinking about how to make today better for new players.

I especially agree with your ideal reason for reading. You’ve seen me clinically summarize this as “self-acceptance and societal acceptance”.

I totally understand. Machine learning is a very clever application of statistics. But to me, when it comes to forming an idea, “AI” (i.e., machine learning) will only ever produce a pastiche of an idea.

This is what I love about stateful writing! And yes, this “magic” can also happen in stateless writing: a storyteller writes to engage with a future audience, while an audiences reads to engage with a past storyteller. But with stateful writing, this “magic” is up close and personal. Unfortunately, with story-based stateful media (e.g., agentic parser-based works), a storyteller engages with semantic structure (got disambiguation problems?), which has an unbounded engagement space. My hope with my approach to narration-based stateful media (i.e., stateful narration) is for a storyteller to engage with affective labels, which has a bounded (but not small) engagement space. If my calculations are correct, then my approach might provide that “magic” to an audience while requiring less overhead from a storyteller.

I couldn’t be happier with the outcome, and I begrudge nobody for any sort of gesturing. Honestly! However, I’m still not convinced that ParserComp was the wrong showcase for Cost of Living because it has more similarities to classical parser games than other forms of stateful media; there’s textual input, “frustration” mechanics, “hidden” mechanics, and “magic” moments, which are expectations for classical parser games.

Recognition is a major component of my approach to narration-based agency. In particular, affective recognition is especially engaging because people are naturally inclined to interpret communication as either affiliative or disaffiliative.

There is more overhead. Soliciting feedback from an audience is difficult, regardless of the genre of stateful media (e.g., choice, parser, hypertext). I underscored this for myself after reading about survey research methodologies (which, to me, is basically an academic disciple of choice mechanics). Fowler’s Survey Research Methods is an great overview of the overhead of an interactive context. (I’ve mentioned this elsewhere, but my scope for Cost of Living didn’t allow me to incorporate any lessons from survey research methodologies.)

To get a perfect picture of the person behind the words? Sure, high accuracy could be necessary. To get a good enough picture of the person behind the words? Well, we do this so easily that the Code of Conduct of this site cautions against it: “Please avoid: Responding to a post’s tone instead of its actual content.” Hopefully, I’ll discover some approaches that give me maximum impact for minimal overhead.

What a hilarious story!! :joy: Thank you for sharing! Now I’m trying to figure out how to cross We Don’t Talk About the Orangutan with the Poe’s The Murders in the Rue Morgue using a stateful narration. I’m definitely going to come back to this idea after I get several more experiments out of the way.

I can see this psychological withering, too. I definitely spotted the first signs of blight in Billy. But I also know that retreating into imagination is a common tactic in the face of constriction, which gives me enough reasonable doubt to seek blame in something greater than the willpower of a single person.

And this new approach is as large (as mysterious) as a coco de mer nut. So a great big emphatic Thank You! to you (and to @parsercommander and @ramstrong ) for engaging early and often. (And also, thanks to all you silent engagers. Yeah, I see you out there. :index_pointing_at_the_viewer: Thank you, too!)

I’m toying with the idea of writing an experimental autobiographical work that is insightful but not necessarily vulnerable. Maybe something like a slice of life story crossed with an adventure story. I’m just not confident in my current abilities to pull off anything more than a “lighter” autobiographical work.

Definitely been taking stock post-Comp. Soon, I’ll be collating everyone’s feedback and will incorporate all that I can of it into my next work.

Thank you so much for being curious — curiosity is one of the greatest gifts that anyone can give!!

And where am I headed next? Where else? A disqualification from IFComp! (I’m implementing another radical input scheme. But don’t worry, I’ll release more works with the input scheme from Cost of Living, too.)

1 Like

Thanks! I’m sure tables would be handy here. This time it’s just a plain object. Here you go:

  score: audienceScore,
  hasNoAffect: !hasAffect,
  isN0: audienceScore === 0,
  isPE: _.inRange(audienceScore, 0, 2),
  isSPE: _.inRange(audienceScore, 2, 5),
  isNE: _.inRange(audienceScore, 0, -2),
  isSNE: _.inRange(audienceScore, -2, -5),
  hasPE: audienceScore > 0,
  hasNE: audienceScore < 0,
  isExactlyHowIFeel: storyteller === audience,
  startsWithUN: audience.startsWith("un"),
  endsWithING: audience.endsWith("ing"),
  endsWithED: audience.endsWith("ed"),
  endsWithEST: audience.endsWith("est"),
  endsWithLESS: audience.endsWith("less"),

(You can see from the last five properties that I implemented a few prefixes and suffixes for ^^^convo expressions. Unfortunately, I didn’t get around to trying them out.)

But I’m so glad that you’re trying to picture the implementation of the stateful writing’s manuscript. I tried to make it unopinionated, so it’s however the artistic engineer decides. Of course, a stateful storyteller and artistic engineer should work together to spec out the ^^^convo expressions, but the implementation of that expression is left to the artistic engineer.

1 Like

Poll: What is Cost of Living?

I’d like to enter more competitions, but I’d rather not grief organizers and participants (or at least, not more than usual). So now that you’ve read Cost of Living, and maybe even looked through this post, I’m curious to how you would describe this approach.

Do any (or all) of these describe Cost of Living?
  • parser
  • choice
  • hyperlink
  • chatbot
  • madlib
  • puzzle
  • other

0 voters