Jan 06


Tensions and Difficulties Created by the Vignette Structure

Posted by: Adam Strong-Morse | Comments (7)

While we believe that a vignette-based structure with relatively little branching lets us build strong text-based games, it produces some inherent tensions.

On the one hand, we want choices to be meaningful, but on the other hand, we want to reuse later code.  The urge to reuse code can create pressure to have choices be purely cosmetic.  Sometimes we explicitly make choices not affect the rest of the game.  Those choices are about the player’s imagining of a scene or triggering thoughts about motivations, not about affecting results.  But taken to an extreme, this could destroy the meaningfulness of the choices.  In “Choice of the Dragon,” there’s a vignette where the player’s dragon seeks a mate.  In the original version of that vignette, it was impossible to actually succeed in getting a mate, because the vignette’s author wanted to avoid any downstream effects to having a mate.  We changed that because it seemed too frustrating for players and because it made the vignette less meaningful.  Now, your dragon can get a mate, although it remains challenging.

Likewise, the more we reuse code and avoid branching, the less replay value the games have.  Imagine a game that begins with a simple binary choice and then uses entirely different text and subsequent choices for each option.  The bad thing about that is that players who choose choice A don’t see any of the work that went into the results of choice B.  But the good side is that a player can play the game a second time, choose B instead of A, and find a whole new game.

The flipside is that some of our individual vignettes have a lot of branches, and some then set variables that affect later vignettes.  Taken to an extreme, branching within a vignette and then using a variable to condition later vignettes can cause the same problem of proliferating options that makes a simple branching structure undesirable.  It can even result in writing essentially dead text—text that nobody ever sees in play but that’s there to cover the outside chance of a series of strange choices.

Ultimately, we think we struck a pretty good balance in “Choice of the Dragon.”  In part, that’s because Dan always pushes for simpler models, with more cosmetic choices, whereas I always push for more complexity to make the choices more meaningful.  He tends to rein in my efforts to make the game ever more complicated, while I make sure that choices in the mating vignette actually affect outcomes.  But at the end of the day, it’s a balance that we wrestle with and discuss over and over again, and we’d love to hear your thoughts about it.



  1. Ian says:

    How do you get a mate! It always says that she turns me down!

    • Dan says:

      @Ian That’s a bit of a spoiler, but we’ll probably submit a blog post soon with Dragon Mating tips. Stay tuned, and thanks for playing!

  2. Julius says:

    Avoid cosmetic choices like the plague, if you can. I definitively don’t want to play text games to give myself the illusion of choice, otherwise I’d play any kind of normal computer rpg out there. Choices should be meaningful, otherwise they’re not really choices, and text games are all about choices, so making fake choices is self-defeating. I know this means a more complicated game, but this shouldn’t be much of a concern since these are text-only games. Bottom of the line : quality before quantity, thanks.

    • Dan says:

      @Julius I also prefer quality over quantity, but in this case I think quantity is also quality. On the app store, many players criticized “Choice of the Dragon” for being too short; the game is 30K words right now (and took quite a bit of work). The trade-off between cosmetic choices and “real” choices is the trade-off between game length and replay value. Fewer cosmetic choices could easily double or quadruple our effort without making the game even slightly longer. You might not even notice!

  3. Julius says:

    It might be too short, but if it’s replayable, then who cares? That’s the point of making different, meaningful choices – replayability. Wouldn’t you say? The parts that I don’t notice are what is fundamentally interesting about all of this – knowing that my choices opened certain paths and closed others. If the majority of the game is opened to me in one playthrough, then not only there’s very little reason to replay it, but also there’s less incentive to play it at all – why not read a book instead?

    Basically, I think this concept can be taken to very interesting heights, and I wish you wouldn’t shun away for the sake of complexity or to appeal to a larger public. I know I’m asking a lot, but hey, I had to say it… let’s be honest, your ideal game would be with zero cosmetic choices – so I would keep that ideal in mind!

  4. Strangelander says:

    I have taken a whack at branching plot fiction, and realize the problems with both approaches. However, now that we can easily track variables, it makes the problem somewhat more manageable, and makes for more interesting possibilities. I can’t see under the hood of Choice of the Dragon, but it seems that some choices and outcomes are indeed based on the stats you are tracking, and that’s a great first step.

    One possibility (that perhaps lends itself better to a 3D game rather than text) is that certain confrontations with characters can take place at different times and settings. So, you limit the set of confrontations your character can encounter, write endings for the different ways each could go, and then, depending on their choices, may run into those confrontations in weaker/stronger states, with different knowledge or inventory items, etc. This also allows you to randomize things if wanted, since things are more discreet. May also allow for multiple solutions if you have variables attached to objects and characters, e.g. the hammer, pipe, and rock are all heavy enough to break the window, it doesn’t matter which you have at the time, or you and a friendly character’s combined strength is enough to overcome an enemy.

  5. […] too. “How difficult will that be to code” is also a constraint. Earlier blog posts (here, and here) discuss the difficulty of balancing realistic decisions with a manageable number of branches. […]

Leave a Comment

Your email is never published nor shared. Required fields are marked *

Subscribe by E-mail