Reading a post on Grand Text Auto I came across this; one slide from an interesting-looking GDC talk by Noah-Wardrip Fruin. The talk discusses approaches to making interactive stories and covers the basic concepts of branching, and stats, and currency and conflict through systems.
While I liked the talk, he said something about CYOAs that I didn’t like, so I thought I’d briefly beat on that.
He briefly makes the point that CYOA games don’t work with the following two slides:
While I can see the point he’s making, I felt both of these were rather too simplistic to dismiss the CYOA-format (which I’ve recently been playing around with, thanks to the rather lovely Undum js framework.)
Primarily, making a game a CYOA is a choice about complexity of interface for the reader: by offering clearly delimited options, a player knows exactly what choice he’s making at any given turn, he has no rules to learn and no conventions or UI to grapple with. This makes the story very accessible, at the expense of rich, finely-grained actions.
So to look at those two slides, in reverse order: the second slide is perfectly possible to implement in a CYOA fashion. You’d only have to break the choices in several stages, rather than presenting them all on one turn: how much on gold? Now how much on gas?
There’s no reason that CYOA games need to be stat-free – they weren’t in the days of gamebooks, they certainly don’t need to be now. There’s no reason that the options offered shouldn’t adapt to the state of internal variables.
Above all, there’s no reason to assume that the old, book-based model of “few options, large text chunks” is appropriate or even desirable if you’re building your CYOA on a computer. Screen-based text is generally most successfully delivered in small units (think Twitter, status updates, or the one-sentence-one-paragraph layout of BBC News online).
Then there’s the first slide, implying uncontrollable explosions of content. Ignoring for the moment that text can vary on the state of variables, and you don’t need a separate paragraph for every eventuality, the slide is still built on the process-orientated assumption that different play experiences should result in computer-differentiable outcomes: that is, in alternative endings, or at least alternative stat values by the time the ending is reached.
But the story being told isn’t being told inside the computer, it’s being told inside the head of the reader, and the reader is tracking everything in a lot more finely-grained detail than the computer is. Our readers, if they’re interested in the story, care about the flow of the conversation, they’re interested in the back-story and detail acquired, they will find a human differentiability between, say, a character decisively taking an action, and one forced into an action by circumstance. We don’t need a stat to track it, we don’t even need another state. (But a stat would be nice.)
Perhaps a good CYOA is light on interface, easy to read, and probably not worth replaying more than a few exploratory times. That makes it a terrible game, maybe, but not a bad story or experience.
(And it’s maybe not so much about choice. Or you. Or adventures. Unravel your own narrative, maybe?)