Narrative Structure in Operator

via XKCD

Warning: info dump of interactive narrative structure ahead.

Operator is a bunch of tiny choose your own adventure stories strung together by having the same protagonist — the player. There are things here and there — often called “qualities” in interactive fiction (more on that later) — which will determine what stories a player has access to, and in what order; at its core, though, it’s just a series of short stories with a shared history and a memory as the player progresses.

Yarr, There Be Flowcharts Ahead

It’s been helpful for me to think of each one of these small stories — we’ll call them “Jobs” because that’s what I call them in my text files of junk — to think of each one of these Jobs as an an example of what Rob Gilbert called “Puzzle Dependency Charts”.

Ron is the designer behind many of the most famous graphic adventure games of all time: Maniac Mansion, the Secret of Monkey Island series, and Day of the Tentacle — a game that is in the upper echelons of influence on my everyday creating. He breaks down the puzzles in his adventure games by looking at them as basic flowcharts of player action:

An example of a puzzle with two independent activities which must be accomplished in order to proceed.
Ron’s example of a puzzle with two independent activities which must be accomplished in order to proceed.

Each one of these atomic pieces can be iterated on and connected to other pieces to extend and flesh out beats of a given narrative. Stack enough of them together, and you start to get a map of the interaction and of the narrative itself:

Ron's narrative begins to take shape.
Ron’s narrative begins to take shape.

With this approach, it’s easy to digest the “flow” of a player through your narrative and make macro level adjustments.

Story Shapez

Another way of looking at this sort of narrative structure is through a “Lock and Key” template, as illustrated by Jay Taylor-Laird in this excellent GDC presentation “The Shapes in Your Story”:

Taken from Jay Tailer-Laid's excellent GDC talk

No matter how you decide to chart out your narrative, it’s clear that having a systematized solution to visualizing narrative progression is key to both plotting the… plot 😒 and evaluating overall structure. Ron and Jay both subscribe to the idea that “linear = better” for both player and developer, as wildly branching narratives quickly spiral out of control. However, using small diversions throughout can lead to an overall feeling of complexity that belies the underlying linear structure. Using Heavy Rain as an example (Slide 21), Jay constructed a narrative map of the game and showed that 86% of the game is linear, and is limited to the first 76% of the game.

The Story Map for Heavy Rain

Every Job in Operator is essentially one of these flowcharts — a series of choices and actions taken by the player to progress from the left side of the flowchart to the right side. Sounds like such a fun game, right? The most fun?

I can see it at the top of iTunes now: Flowchart Crush Saga.

Linear Structure vs. The World

Fortunately for game developers everywhere, reducing a game to a flowchart fails to capture the inherent curiosity that comes about when you’re standing in the middle of forest without knowing which direction to take. Your car could be parked thirty yards away, but you’ve got no idea, and so you continue on walking in the other direction, hoping to stumble into something familiar to mark your progression. Climb a high enough tree, and your perspective reduces the complexity of the task at hand.

What about those times when you can really feel the flowchart in a game? When the nature of its construction starts being so predictable that the flowchart of its construction becomes apparent to the player, even from deep within the forest? When you know that you have to keep moving forward, Point A to Point B to Point C, without divergence or unpredictability?

Interest curve for The Legend of Zelda: the Wind Waker; Bradley C. Buchanan
Interest curve for The Legend of Zelda: the Wind Waker; Bradley C. Buchanan

Sometimes this sort of easily identifiable structure is preferred. It can feel really, really good to complete a linear, authored experience that is paced to perfection and really owns the player experience (see: interest curves).

If you’re not Hideo Kojima, or you just aren’t that into linear progression, there are a few options deftly laid out by Sam Kabo Ashwell in the seminal blogpost (can blogposts be seminal?) Standard Patterns in Choice-Based Games.

Standard Patterns

Briefly (seriously, just go read the whole thing): there has been at least fifty years worth of choice-based games (think “Choose Your Own Adventure” books for the pre-video game example) created, and in those fifty years, some patterns have emerged in the way you can construct a game.

The structure identified by Ashwell that most closely resembles Ron Gilbert’s examples above is the Gauntlet:

from Sam Kabo Ashwell's Standard Structures...

from Sam Kabo Ashwell’s Standard Structures…

Looks pretty familiar. However, there’s a lot more options out there identified by Sam Kabo Ashwell–structures that you would have stumbled upon after a lot of trial and error–but why go to all that work? At some level, you have to be able to visualize an abstraction of the gameplay flow that you’re trying to create. All we’re doing here is working backwards. The easy way. Start from a known structure that works, and build from it.

“But!–” you’re about to say, “–my shit is too next level for a flowchart!”

Shit I didn’t realize I had someone reading this that was next-level. My bad. I am not next level, and for the most part neither is Operator, but Emily Short has you covered with her deep-dive into non-branching narratives titled, appropriately enough, Beyond Branching: Quality-Based, Salience-Based, and Waypoint Narrative Structures.

In the article, Short boils down the whole idea of exploring a game’s structure to one hot-damn of a question: My story is made of pieces of content. How do I choose which piece to show the player next? She goes on to list many wonderful examples of games which tried to tackle this question in assuredly “next level” ways. Quality-based narratives in particular are relevant to Operator, in that it can help provide guide-rails for giving content to the player just in time. Too early, and it’s impossible due to a player’s skill or the abilities awarded to them, too late, and it’s simplistic and un-fun.

How does this apply to Operator?

Operator will use this idea to provide a sense of progression and long-term narrative in the game, despite the day-to-day experience being small, self-contained Puzzle Dependency Charts. By categorizing these jobs with certain levels and qualities, Operator can present them to the player in way that provides variety (“not the same story every time you play”-ness) and appropriate difficulty (“just in time”-ness). I guess you could say Operator dabbles in next level shit, but the meat on the bone is standard gameplay structures that have been in use 50 years.

This mix allows me to keep my sanity (I’m not really sure how the knuckleheads at Failbetter Games chart Fallen London and I don’t want to know) and add what I feel are important “directorial flourishes” to the overall experience. Think more Valve’s AI Director and less Dwarf Fortress. This won’t be a sandbox open world simulation. There’s already enough of those in early access.

I should probably avoid ending this post with a dig on other game developers. Those open world games are fun, guys. Keep making ’em. Don’t listen to old jerks like me.

Anyway–long story short. tldr; as the kids say:

Flowcharts are a game designer’s best friend.