Next-Gen Storytelling Part One: What Makes a Story?

Warren Spector Legacy Author

Published: Apr 19, 2007 09:45 pm

This article is over 17 years old and may contain outdated information

“The stories we tell reflect and determine how we think about ourselves and one another. A new medium of expression allows us to tell stories we could not tell before, to retell the age-old stories in new ways, to imagine ourselves as creatures of a parameterized world of multiple possibilities, to understand ourselves as authors of rules systems which drive behavior and shape our possibilities.”

–Janet Murray, “From Game-Story to Cyberdrama,” First Person

The quote above really speaks to me-and to the importance of game narrative. Murray’s comments speak to the personal and cultural uses of stories… to the new stories afforded us by a new medium… to the ways in which narrative structures remind us that we live in a world of infinite possibility… even to the ways in which each of us is the master of his or her fate. I can’t read that quote without feeling like I’d be wasting my time if I didn’t at least try to bridge the gap between game and story.

I draw further inspiration from yet another Janet Murray quote, this one from “From Game-Story to Cyberdrama”:

“[W]hy are we particularly drawn to discussion of digital games in terms of story?…[I]t is a medium that includes still images, moving images, text, audio, three-dimensional, navigable space-more of the building blocks of storytelling than any single medium has ever offered us.”

How could anyone not look at the state of the art and wish for more?

So, how do games tell stories, and what does the future hold for us? In this four-part series, I’ll discuss the elements of story, how those elements are (and can be) employed to tell stories in games and how the advent of next-gen, PC-like console hardware can make our jobs as storytellers harder, and what we can do to even the odds in our favor.

What Makes a Great Story?

If you’re going to consider the value of a powerful story in your game, you need to start by understanding what features are common to all good stories.

Change

For starters, the state of the world, of the main character and of the characters the hero meets along the way must change as a result of the hero’s actions.

Pacing

Next, the pacing has to be right, which is to say varying, depending on where you are in the story. Typically, for me, you want a pretty quick, accelerating build-up at the beginning when players are learning how the world works; you want to slow things down a bit during the middle stages when things get complicated and the player has to figure out how to resolve the world’s problems; and then you want to race to the climax once the player knows what he or she has to do to fix everything. Your mileage may vary but, regardless of how you vary the pace, the pace must vary to hold players’ interest.

Compelling Characters

You have to care about the hero and about the people or creatures who live in your story, or what’s the point? Your characters better not be cardboard cutouts. They better have their own beliefs, goals, needs, desires and attitudes.

Subtext

Your story should be about something – and not just the surface level series of events that make up the plot. You need a subtext. What your story’s really about, regardless of what the box blurb says it’s about.

The Game Scorecard

Look at that list of “Great Story” features and it’s apparent why game stories tend not to be so good:

We have a lot of work to do before we can hope to create a great game story.

Narrative Game Structures

What about Game Structures? How have we tried to bring narrative to gaming?

I see five more or less distinct ways of approaching storytelling in games:

As a player, it’s useful to know which approach you find most interesting and appealing. As a developer it’s vital to know which approach you’re going to take before you home in on the details of your design.

The five approaches listed above offer different advantages and disadvantages, and require different kinds of game systems and even different approaches to the development process.

Rollercoaster (aka the “String of Pearls”)

By this I mean a game with a predetermined narrative from which players can’t deviate. Player input in such games consists entirely of things on the level of “which weapon will I use to kill this thing?” Players run, jump and shoot their way through one mission after another – no interesting choices, no significant options, no unique experiences.

Don’t get me wrong; I love some games like this. I’m a well-known Zelda freak. (A proud Twilight Princess completer!) I also like Half Life and Ico and God of War and countless others. You can tell a great story in this format because, as a developer, you have complete control over, and complete knowledge of, what the player experiences.

I don’t think games like this will ever – or should ever – go away, but I see their numbers decreasing in the years to come and their influence on design and development diminishing.

Retold Stories

By “retold” I mean a game that has no story at all, other than the highly individual, remembered narrative recounted by each player after the fact. MMOGs fall into this category, as do sports games and puzzle games.

My experience of playing Tetris is inevitably going to be different than yours. You know, I might describe a Tetris game like this: “I kept waiting for that long straight piece, but it never came so I had to rotate an L and got myself all screwed up. It sucked!” It’s unlikely your description of your Tetris game would match mine.

Unless you’re a hardcore academic, you probably don’t think about this as a “story,” per se. And you’re probably right. So let’s move on.

Sandbox

This is what I call the “Will Wright School,” but it could be described as a class of games whose story is “player-generated.” In these games, the developers provide players with tools to construct their own stories. It’s not just that you create a story after the fact – in the retelling – but that you use tools provided by the developer to create the story as you play. That’s what the game is all about.

There’s no predetermined, overarching narrative, but there is a strong sense of a story being created through the individual choices you make. And exceptionally talented, creative players can generate some remarkably powerful narratives – “real” stories, if you will. They do this through play and/or through the use of an array of expressive tools and easily manipulated rules – tools that can be used to generate new content, and rules that offer gameplay possibilities that allow stories to emerge seemingly naturally.

Shared Authorship

The sandbox game isn’t the only way to offer players something more than emotional rollercoaster rides or highly personal, remembered adventures. Thankfully, the “Shared Authorship model” of Origin/Looking Glass/Ion Storm/Irrational and others still seems promising, or I’d be outta work!

In fact, where I used to make a clear (and probably ego-driven) differentiation between games like Ultima, Thief, Deus Ex and the stuff coming from Bioware, Bethesda and Rockstar, now, I tend to think of all of these as kissing cousins.

They’re linked by the goal of giving players significant freedom. They’re separated by how “coercive” the narrative is in terms of mission ordering and player role, and by how shallow or deep the player choices are. But all of these games and developers are after more or less the same thing.

I still think of this game/story model as the most promising for the near term, particularly in the commercial game space, but, longer term, I’m finding yet another game story approach more and more intriguing, even if I don’t yet know quite what to do with it.

Procedural Story Generation

There seems to be movement toward a fourth way of telling stories in games, exemplified in my admittedly limited experience by Façade (Michael Mateas and Andrew Stern), Storytronics (Chris Crawford) and, I suspect, others: “Procedural Storytelling.”

These folks are serious about giving players incredible freedom to explore game worlds and interpersonal relationships, to “grow” stories through their choices. The level of player freedom and control these folks are talking about scares the heck out of me. And my fear seems at least somewhat justified – early efforts have been marginally successful, at best.

But even if these efforts fail, I believe the new thinking they represent, the new ideas and possibilities they reveal to us, will ultimately help us achieve higher levels of storytelling in our own work. These guys may not have The Answer, but they’re pointing us in interesting directions.

What can we do now to structure games in more interesting ways?

Someone may take a big leap, but what can we do right now, with the knowledge, experience and hardware we have today?

At Junction Point Studios, story discussions often start with soul-searching about whether we should simply refine what we’ve always done or try something really new and different. (“Replace” is often the desired option but “refine” usually carries the day because, you know, you have to ship.) And when we talk about refinement, we usually end up talking about how much control we can – and want to – give players over the overall plot.

We know we want to give players ownership of the minute-to-minute (what Mateas and Stern have called “local agency”). But how much control can we give them over the plot itself? In other words, can or should we give players more “global agency”? Mateas and Stern seem intent on giving players lots of global agency, as much freedom over the way the overall plot unfolds, as they can.

Given the commercial realities – the fact that games must ship and sell – we tend to gravitate toward a middle ground. I mean, as much as I admire and respect Façade, it doesn’t quite pull off the goal of offering players real control over the plot, and that’s in the context of a 15-minute experience. Imagine doing that over the course of several hours, and imagine doing it with the kind of graphics our consumers expect. It’s downright terrifying.

So, at JPS we’re playing around with different mission ordering schemes as a possible compromise approach to nonlinearity. We can, with some effort, give players control over which missions they take and when they take them. We can attach consequences not only to what happens to players in a mission but to when they undertake a mission at all.

The idea of consequence is key (as it always is, at least in my thinking) – the player-driven ordering of missions must represent a critical choice. The consequences of doing Mission A before Mission B must be a real, significant difference compared with the player choice to do B before A.

OK, now, to be completely honest, we had a whole game mapped out around the idea of player choice of mission order, but, as soon as we got to scoping the project, overall, mission order choice was among the first cuts. The payoff just didn’t seem to be there, relative to other game features.

So we fell back to something related but with, we think, more bang for buck:

We knew we were going to offer players as much local agency as possible – lots of play-style and problem-solution choices. And we knew we wanted to build much of our game around aligning yourself with one or more of several factions in the game, each with its own goals, needs, desired mission outcomes and so on.

With that in mind, we decided to ditch the idea of calling out specific player choices and redirecting the game as a result of whether you killed someone or snuck past them, or which choice you made from a menu of conversation options, or which of three doors you chose to use to leave a particular room.

That was the granular approach we took to player choice in the first Deus Ex games. Instead, we started working on the idea of tracking the general trend implied by countless small player choices. We can open or close paths, options, situations, factional responses, character behaviors, etc. as a result. I guess it’s kind of like what Fable did, but with less of a Black/White, Good/Evil or Light Side/Dark Side feeling. And we want to really drive the plot in different directions as a result of those myriad small decisions, rather than just changing the PC’s appearance or ability set.

We feel we can leverage this “Trend Tracking” approach to move away from “on rails until the end and then open up several endgame possibilities” and toward more tangible and consequential player choices, earlier.

For example, in one game we’ve concepted out at JPS, you interact with a variety of factions and your decisions about whose goals to pursue actually result in one of those factions being, in essence, destroyed as part of the Act 2 climax. We don’t care which faction goes away – that decision belongs to the player – but we have to craft a story that can work regardless of which faction ceases to exist.

As simple as this sounds – many of you are probably playing with similar ideas – this poses a lot of development and gameplay problems:

That’s a ton of work, but I believe that the moment, in mid-game, when the player’s choices result in a dramatic story turn will be a stunning and marvelous moment.

… To be continued

In the next installment, I’ll discuss the elements of story telling as they relate to game, and the importance of character and conversation.

***

Warren Spector is the founder of Junction Point Studios. He worked previously with Origin Systems, Looking Glass Studios, TSR and Steve Jackson Games.

Next-Gen Storytelling is a four-part series. Part Two will appear in this space on Friday, April 20th. Part Three will appear, also in this space, on Monday, April 23rd. Part Four, including Warren’s conclusions, will be published in Issue 94 of The Escapist, available on Tuesday, April 24.