Working on a new branching narrative project for inkle has crystallised in my head a problem I think I’ve often skirted around, but not ever pinned down before; and it’s a problem that sits right the heart of interactive narrative design.
The problem is this: if we’re tracking what the player chooses, and using that to alter how events play out, then how do we decide when to cause, and how do we decide when to affect?
There’s a conversation going on here – the reader says something, and the author says something back. The best interactive writing matches the author’s reply to the reader’s comment so perfectly it feels like there must be a human being inside the machine, typing furiously away.
But how do we decide who gets to hold the talking stick at any given moment?
You Are Being Watched
To be specific – the project we’re working on tracks a few, emotion-based stats. Are you friendly, or surly? Are you selfish, or generous? Are you hasty, or cautious? When you take decisions early in the story, each of these balances is tipped until we start to get a sense of what kind of character you want to be.
Let’s take the last one as our case study. At the start of the game, you can get up and straight on with your task – or you can take a while, exploring and finding out the groundwork first. That adjusts the stats we track. When you approach the first location in the journey, you can scope it out, or go straight in, and that adjusts the stat again.
Then we use that stat to alter things in the story and the game-world. Do you sense a trap coming in advance of it being sprung? Does the stranger you meet feel inclined to be friendly towards you because you’ve got that kind of face? Does your gut feeling about his trustworthiness match the reality of the situation?
Me vs The World
This is the simplest model of player-tracking: a straight conversation between reader and author. The reader picks their option, and the author adjusts the story to match. Want to be surly? Then every character who meets you will be grumpy. Want to be flirty? Then every character who meets you will like you, except the married ones, who’ll find you oddly threatening. Want to be sneaky? Then everyone who meets you will be surprised to see you.
Every interactive story should have a bit of this – it feels responsive when the game nudges, winks and says, I saw what you did there. But if that’s the only way the player’s choices get used, the results can start to feel a little uncomfortable. After all, if everyone in this world hates me, and I hate them, why do I keep getting the option to give them flowers? (Even if, when I do, they throw them back in my face).
In the real world, it’s true that we can change our mode of being on any day we decide to, if we try hard enough. But in a story, characters are not allowed to do that. They have to be true to themselves; whatever that might mean.
So shouldn’t we, at some point, be locking down the options the player gets as well? If we’ve decided the player is a hasty kind of player, shouldn’t we stop offering them cautious options? If they’re a friendly player, shouldn’t we take away the option to be miserable and rude?
If we don’t, we get something a bit like the LA Noire problem, in which the detective flips psychopathically from being polite and gentle, to furiously interrogating, perfectly innocent bereaved housewives — with somewhat comical results, (not to mention making the game rather harder than it should be).
Set up and Punch
There’s a need for a model here. How about, “set up and punch”? We can use the beginning of the game – the first scene, the first act, whatever – as the set up: allow the player to outline their character, pick their sides, lay out the land. Then for the rest of the game, we play those choices out. For the first scene, you get to choose whatever you like, whenever you like. From that point on, we start to edit what choices you get to the kind of character you are.
The Choice of Games games do this explicitly, with RPG-style character creation choices that remove later options.
This is good, because it matches what should be happening in the scope of a story anyway. To begin with, details are important. How does the protagonist talk to people? How much care do they take over what they do? But by the second act of a story, things should be hotting up. It’s no longer about how they do what they do, it’s about what they do, and the repercussions that their actions are having. The main character is still integral, but the reader is now learning about the world, and not the character.
But in the interactive context there’s a downside, as the game will inevitably feel more and more railroaded, as though the writer simply “gave up” on making the story adaptive. (Of course, the far-spectrum alternative is the game feels arbitrary over the long term, with the character contradicting itself from moment to moment, and that’s no good either).
It would seem that set up and punch works well for short stories, but starts to suffer over the long term.
A Series of Ever-Increasing Boxes
One solution might be to think of each act as a separate bit of gameplay. In Act 1, let’s make choices that define if the main character is friendly or moody. In Act 2, let’s use the results of those choices, and now decide if the character is brave or cowardly; and also if they’re physically or mentally inclined. In Act 3, let’s set all of those results in stone, and start deciding if the character is interested in self-promotion, or helping those around him, which will take us to the final turning point – save themselves, or save the city?
That’s how film structures often work (the above is, loosely, a description of The Dark Knight Rises). Looking back on it now, I can see that’s what Dave Morris was doing in Frankenstein – the early parts determine Frankenstein, and his monster’s, empathetic outlook. Later parts develop new models, with the final, most intriguing one being does the narrator trust you, the reader, to make his choices for him?
If Life Is Just A Highway…
But a structure like the one above means there are certain stories you can’t tell – namely, stories with a long scope but a tight focus. If you’re doing an epic about the life of Emperor Claudius, the respect his peers have for him is constantly in the balance, as is his own pride and selfishness. The interactivity should be driving these elements the whole way to the end.
So perhaps there’s room for a more flexible model. Something like driving on a many-laned motorway: at any moment, the player is in one lane or another. Their choices are determined by that lane – they can keep going as they are, they can move in a lane and become more extreme in one direction, or they can move out and move the other way. Swapping lanes isn’t the result of a single choice, but a build of several choices, but once it happens, it’s a definitive switch in the character’s temperament and the options available to them.
At this point, a player reaching Act 3 as a disrespected, prideful despot still has time to save their soul — but, just like getting off the motorway when the slip-road is approaching, they might be running out of time to do it, and it might require some drastic actions to get there before the end arrives.
Of course, from the point of the view of the author, this is a lot of work: by the end of the book you might be supporting five or six variants of a particular action, depending on which “lane” the player is in. But there are ways to handle that kind of thing elegantly, of course: it’s not like every level the character finds themselves on needs to be a different story-thread all of its own. Sometimes, the same material can be reused and just have its language altered.
What’s it all about, anyway?
Underneath all of this is the related problem of telling the reader that any of this is happening. And the key to that is probably not about interactivity, or logic, but rather, about theme. Whatever it is that’s being tracked, followed, and adapted to had better be core to the themes of the book. A reader should either be customising the nail varnish or rewiring the spinal cord of their story: anything in between is going to pass by unnoticed.
(Or you could just simply showing the stats, but that doesn’t necessarily solve anything – if your text doesn’t match what the numbers are doing, then your numbers will start to look fake).
So if your novel is about spies, track loyalty and betrayal. If your novel is about love, track, um, loyalty and betrayal. If your novel is about religion, track faith and despair. If you novel is about an epic journey across a wilderness, track… well, I’ll let you know when we’ve written it.