Plot, scene by scene

When I plan plot-heavy IF, I think of it in terms of a sequence of scenes. This doesn’t mean that the gameplay needs to be rigidly linear: scenes can occur in varying orders, or there can be plot branches, or scenes that can be skipped depending on player action. But I nonetheless do the organization in terms of scenes. A scene has a definite beginning and a definite end. It usually has to take place in a specific area of the game map (which may mean that the player triggers it by entering that area [as in City of Secrets] or that I move the player myself when the scene is scheduled to start). Following some writing advice I got long ago, I try to make most of the scenes end with some kind of clear hook. At the end of the scene, the player should ideally have a new take on what is happening, or a new problem to solve, or a new question about what is going to happen next. Exciting the player’s curiosity about something is especially powerful in getting the player to keep playing.

But the conventional writing advice tends to be insufficient when it comes to the types of scene that IF supports. I find that in interactive fiction my scenes tend to come in several styles, identifiable by the sort of interaction I expect from the player.

In rough order of intensity, they are

Atmosphere. The scene is about setting a mood for the player: it’s usually a short transitional sequence of a few turns, during which there is not actually anything important to do other than to observe the surroundings, examine a few things. I find atmosphere scenes especially useful when I need to show that the player character is in a strange or dangerous situation; they make a good pacing element between two more active scenes, as well. If every scene in a game has lots of dramatic action, then (by a curious reversal) none of them feel that important.

In some games, dream sequences, unconscious moments, visions, or fits of delusion also fulfill this function. (City of Secrets has a number of dream sequences, which are there partly to provide mild hints but also partly to offer the player a different flavor of interaction.)

The challenge about this sort of scene is to make the most of the limited number of turns the player has at his disposal: just about every action he can take during the scene needs to do the scene’s work. Sounds, sights, object interactions all need to underscore the mood of the moment. To prevent combinatorial explosion, it’s often a good idea to focus the player on just a few objects; besides, if you give the player only two or three turns in a space that is obviously very densely implemented with many items, you may make him anxious that he missed something important.

Trap. This scene is about the player being stuck and unable to do anything, and I list it immediately after Atmosphere because the two are quite similar in their effects — except I think of a trap scene as one in which the player is restricted from doing anything interesting. The limousine ride partway through When in Rome 1 is like this; the opening moves of Hitchhiker’s Guide to the Galaxy; the carriage ride in Masquerade… there are doubtless others.

Classic Puzzle/Manipulation. The scene is about the player accomplishing something, usually solving a physical manipulation puzzle. In plot-heavy games, I tend to use these sequences sparingly — because they are not inherently very plotty, because they can take an unpredictable amount of time to solve (and thus throw off the story-telling pace), and because it is not always easy to come up with a good reason to include puzzle challenges in the middle of a largely narrative work.

That said, they’re not useless. Several testers said that they liked the (very few) passages of physical interaction in Floatpoint because these grounded them in the environment in a way that all the talking and sightseeing and computer-chip-reading did not: and as setting is one of the things that IF does naturally and well, it seems a pity not to take advantage of this.

One option, if true puzzles seem out of place in a work, is to offer the player some easy but rewarding physical interaction. Make it obvious what to do, but make the doing of it interesting — that is, write in interesting sensations as a result; have NPCs around comment on the activity (Lost Pig gets enormous mileage out of letting Grunk do obvious things while the gnome and the pig react); have a seemingly straightforward action reveal some rather unexpected outcome. (Act of Misdirection is a great one for this: it leads and prods the player through performing a magic act, getting him largely to perform actions that the game has more or less laid out, but the outcomes are fascinating.) It is remarkable how much richer one can make a game feel just by writing special responses to ordinary actions. How does the bar of platinum feel when you pick it up? How does the player character feel about taking it? Can we come up with something better than “Taken”?

Attention to these details can make even fairly standard puzzle-game kinds of interaction into something with a fair amount of narrative content.

Exploration. The scene is about the player finding out something, or perhaps seeing some series of clues. The scene ends when the player has found out everything the player needs to find out. This can be about exploring a physical space, or solving a research puzzle with books or character conversation or a computer database.

More than any other, this scene is about hooking the player with curiosity and then making the curiosity pay off. Anchorhead does this brilliantly — though the structure of the game is not always fully delimited into scenes, there are several puzzles that are about making the player curious about something even though she also dreads the answer. An artful chain of clues is especially important here. Done right, an exploration scene can build suspense and anticipation for later scenes.

Travel. The scene is about the player arriving in a new location, even if he takes some side trips along the way. This is useful to give the player a feel for the setting, including a sense of distance between one place and another; if the game features a very large and open map, giving the player a specifically routed trip through it the first time can provide some orientation.

The opening of Floatpoint is like this: I wanted to begin the game with the player experiencing this alien world, both so that he could appreciate how strange and unfamiliar and also beautiful it was, and because I wanted to dramatize his coming into town as an outsider: alone, unescorted, and uncomfortable.

Pacing is the difficult part of a travel scene: on the one hand it is desperately heavy-handed if one forces the player along, but on the other if one lets the player make the movement commands (and stop to look at scenery) he may get distracted and spin out the travel for a long time — perhaps longer than the author really intends. In Floatpoint my solution was to set an appointment for the player character, and remind him periodically that he was supposed to be keeping it — which isn’t very subtle, and of course it was possible for the player to go off and do other things anyway.

Another approach is to give the player another character as escort. This is sufficiently common that the TADS 3 library specifically provides for it, for instance. City of Secrets has several sequences where the player is taken somewhere by another character, who keeps the player from wandering off track (on an otherwise fairly open map). An Act of Murder also begins this way, with an introduction in which the player is introduced to all the characters and most of the locations of the game before being turned loose to wander on his own. I found there that it was almost too much of a whirlwind tour — there was so much information conveyed at once that I had a hard time keeping track of it all. But it did serve some important purposes by giving me a clear sense of the shape of the map and of the people in it.

Movie. A sequence of events has to play out in the player’s presence, but there isn’t much the player can do to disrupt this. As the author, I have an exact sequence I want the player to be present for — like overhearing conversation between other characters, or watching a setpiece play out.

The trick here is generally to give the player something to do while the movie rolls to its inevitable conclusion turn-by-turn — and this is something IF authors have been doing with variable success for quite a while. One of the more successful demonstrations of this is Christminster, which finds a variety of ways to ground the player in one place while other characters converse near or in front of her. Sometimes she’s hiding, listening to something nearby, afraid of being exposed; sometimes she’s present but preoccupied.

Offering the player a minor puzzle to do while the action continues sometimes helps, though then one has an additional challenge: what to do if the player takes too long to finish the puzzle? or not long enough?

If the player is able to interact with and alter the scripted sequence, then what you have is not a movie but a timed puzzle or an interview.

Timed set-piece puzzle. Again, I tend to be sparing with the puzzle challenges in a plot-heavy game, but a timed sequence can be quite effective at focusing the player during an emotional crisis during the game.

Several mid-school games (Anchorhead, Christminster) make a lot of use of these during the most tense moments of the plot. In more recent games, designers have tended to move a bit away from using this technique because it can be seen as cruel or unfair to the player: these scenes are intense on the first playing, but if the player fails and has to redo them over and over, they lose some of their impact. IF isn’t the only medium that has this problem: I ran into issues with exactly this when I played Portal.

The tedium of replaying a timed sequence is why I made the timed set-piece at the end of Savoir-Faire deliberately much easier than many of the late-game puzzles: I wanted there to be a sense of urgency but for the player to win the first time anyway. On the other hand, some players complained that it was too easy and this undermined said urgency.

I now think my error was in not making the final steps narratively powerful enough. If the actions required to solve the puzzle were easy for the player but difficult for the player character — acts of sacrifice, pain, or emotional resonance — then the scene would have more clout. Heroine’s Mantle — in other ways a very broken game — does I think get this right, though in a kind of silly and over-the-top way. After lots and lots of fiendishly difficult and unfair puzzles, the final stages (at least as I recall them, some years later) are intellectually easy but emotionally colorful.

Interview. The scene is a conversation with an NPC, and certain things need to get said before the scene can end. Sometimes, those things can be said in a variable order. Sometimes the player can introduce tangents. But the shape of the scene is essentially determined by the things the NPC needs to say to the player, and the things the NPC wants to hear back in return. (If it’s undirected conversation aimed at letting the player find something out, I consider it an exploration scene instead.)

Combat and romantic scenes are also a kind of interview scene, in the sense that they are about pursuing an interaction with a main character to its natural conclusion (or perhaps one of several natural conclusions). Pytho’s Mask features conversation, a dance, and a sword fight, all of which I would consider interview scenes, though technically the player is doing different things in each.

Interview scenes lend themselves well to branching and choice-making: it’s fairly natural for conversation (or combat) to be able to go one of several ways, offering the player an emotionally or even morally charged set of options. I find that the majority of mine are about a question (the other character(s) want to find out something from the player) or an argument (the other character(s) are struggling with the player, and the struggle comes to a positive or negative conclusion).

Interviews tend to be extremely plot-rich. Every move introduces new information and offers the player new choices. That makes them feel intense — and it can mean that you don’t want to stack too many interview scenes one after the other, lest you create fatigue in the player or simply erode his ability to care. Quiet, mood-building scenes offer a respite and also give the player a chance to dig into the story and setting; without them, the story will be high in drama but not have the impact it should.


When I go to structure a plot, I make a list of the scenes that need to occur: what are the major turning points in the story? Leading up to those, what does the player need to know in order to get the most out of them? What does he need to be emotionally invested in? What questions should he want answered?

Is my sequence of scenes too densely made up of all the same sort of thing? Too many intense events in a row, or too many relaxed ones?

Am I giving the player enough to wonder about? Are there, in fact, hooks at the ends of the major scenes which give the player a reason to want to come back and find out more? (Many of the most successful IF stories are hook-rich: Anchorhead, Christminster, Delusions, Act of Misdirection all offer the player some major questions early on and subordinate mysteries along the way.)

Is the story best served by a single perspective? Or do I need the player to hear multiple versions of the same event? Can I introduce the alternate versions through conversation or exploration? Should I instead consider shifting to a different viewpoint character for some parts?

At any given point, who gets to make the difficult choices? Maybe that person should be the viewpoint character for that scene.

At any given point, who is moved by very strong feelings, especially grief or passion? Maybe that person should not be the viewpoint character for that scene — exactly because it is so hard to guarantee that the player will identify. Television shows often use a technique I mentally label “distance from grief”: when a character learns about a death or something similarly awful, the camera pulls away, letting us see a long shot, the body language of devastation, rather than a close-up of the face and the sounds of sobbing. That distance makes the grief universal and sympathetic rather than specific and voyeuristic. It lets us fill in our own experience of loss, without the accompanying embarrassment that we sometimes feel when we’re around crying people. Similarly in IF, I suspect (though I’m short on examples at the moment) that it’s more effective to (a) let the player know in advance how much a character will be devastated by the bad event to come but then (b) pull back a little and let the player be outside that character’s head when the storm hits.


This all describes a process for writing a fairly linear game. Of course, that is not the only way a plot-heavy piece can work. A number of games consist of open exploration with more directed scenes embedded in the whole — Plundered Hearts triggers interactions with your romantic interest, with the villain, etc., when you get to key points or solve important puzzles, and quite a few later games also follow the same basic structure. City of Secrets starts and ends with sequenced scenes, but the midgame allows the player to wander fairly freely, stumbling into events but then returning to the betwixt-and-between state.

Still. Lately I find a lot of my WIP planning turns on this kind of thinking.

19 thoughts on “Plot, scene by scene”

  1. Fantastic article. I suggest to make a second part of the topic explaining implementation techniques for every kind of scene, and programming resources.

  2. That would have to be almost entirely I7-oriented — I’m sure there are ways to do this in other languages, but I don’t know the TADS 3 library well enough not to make a fool of myself at it.

  3. Emily, this was extremely helpful. I’m heavily into plot planning for my WIP, and was pleasantly surprised to find this article waiting in my reader this morning. I’m already seeing ways to improve.

    I’ve recently learned about the “IF Theory” book project that you are (were?) involved in. Is that project dead? I can’t get to either the original site or the backup that you made of it.

  4. Yes, it’s only an idea. That kind of article could be constructed using formal definitions, algorithms in pseudocode, etc; or as you said, only I7 oriented, or putting two kind of examples, I7 code, and I6 or TADS3 code. The same article but more like a formal study. But again, it’s only an idea.

    And, you have put some insights for each case of what kind of implementation should build each kind scene. So I join the praise, it’s a extremely helpful article.

    But remember, algorithms and programming techniques could be applied to every programming paradigm without fear of left behind any system.

    Let me express an experience I have recently. I’ve tested a Spanish version of the home-brew game “Flight From the Dark” for Nintendo DS:

    It’s the classic game-book (CYOA books) of the high acclaimed saga Lone Wolf. For testing purposes I need to travel through all the map of nodes of the book in quite deep, testing fightings, controlling bonuses, and such; so finally I have learned the structure of that book, how it was designed, and a though strike to me, that it was constructed by scenes, paths that travels through scene by scene branching the book until they join in the final “chapter”. The book has two great main branches with scenes that runs from one to another leading the player to an unique victory ending. So this has more potential about storytelling than traditional Adventures (for example, Adventure itself has no branches or scenes or storytelling, it is just a cave crawling suspended in time, because there are no progressing of plot, or maybe is only a little of it. Sorry my bad English, eh?). So we could learn a lot of CYOA games, although Interactive Fiction has being doing quite well implementing scenes and deep stories about branching, CYOA games by nature must have a lot of them!, so in comparison any CYOA book could have quite more scenes than any Interactive Fiction piece.

    I’m planning to do an article about the topic: “Bisection of a wolf: analysing Flight From Darkness (what we can learn of CYOA books”, or something like that.

    Again, sorry my bad English, I’ve tried to do my best explaining myself, and I don’t know if finally I’m sounding too vague or too wrong.

  5. This kind of post makes me wish for an updated Craft of Adventure — I guess the IF Theory book project was meant to tackle this kind of thing? Either way, it’s interesting food for thought. A lot of writing has been on puzzle design, at least that I’ve seen, but not that much on plotting.

    Just curious… I’d say almost a lot of the storytelling that occurs in Floatpoint is through reading someone else’s journal entries. (I hope this is a fair characterization.) Would you consider these retelling of events scenes at all? And if not, where do they lie? I’m not sure myself, though they’re surely important tools to storytelling.

  6. To the extent that they qualify as distinct scenes, I’d say they’re exploration. The player gets to control the amount, rate, and content of the exposition he encounters, and is free to look at as much as he wants before declaring himself finished.

  7. Great as usual!

    When you say:

    At any given point, who gets to make the difficult choices?
    Maybe that person should be the viewpoint character for that scene.

    – are you talking about changing the PC to that character or just focusing on that character in the writing?

  8. – are you talking about changing the PC to that character or just focusing on that character in the writing?

    I’m talking about changing the PC. (Obviously, not all stories are suitable for this, but it’s an option to keep open, especially for more narrative IF.)

  9. Changing the PC for a single scene might make an interesting kind of game … “The Cyrstal Gryphon” (book) has two main characters, and alternate chapters are from each character’s point of view. I wonder how well that would work in IF (and how you would keep the player from feeling disoriented from too many shifts)?

  10. this is probably one of the best articles on the writing process that I’ve read in a while. Copying and saving to my computer for future reference (NOT to be published elsewhere, btw), if you don’t mind. I’m working on a game at the moment, but will probably be more character driven than plot driven. I haven’t really decided. But, I’ll take your advice to heart, most definitely.

  11. A story can be about a character and still rely on a pretty structured presentation. In fact, I’d say the majority of IF that is heavily character-oriented tends to fall into one of two categories: linear, with a sequence of distinct scenes (Photopia, Rameses, Photograph, etc.); or short and very open, conversation style (Galatea, Shadows on the Mirror, etc.).

    Arguably the conversation games are a subset of the sequential-scene games, and just happen to be only a single scene long.

  12. To David Fisher’s comment: There are IF games that shift viewpoint characters a significant amount — Common Ground and Being Andrew Plotkin come to mind. The trick, I think, is (a) to make it very clear at the start of a new section which character is now the protagonist and also (b) write the sections with sufficient attitude that they don’t feel like they blend together.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: