Off The Script: The Problem With Storytelling In Open-World Games

Niko Bellic quietly contemplates the futility of war, of the violent life he has chosen. He mourns the friends stolen from him by his enemies, the tragic losses he has accrued through violence and crime. In the end, his experiences have taught him only this: that the American dream is an illusion, impossible for anyone to achieve.

The next day, he steps out onto his balcony and fires indiscriminately at the streets below, trying to see how many cops he can get to follow him. He spends five minutes carefully throwing grenades, trying to get the angle just right to put one under a car. Then, he takes the ruined cars, shoves them into a big pile, and ramps his motorcycle over it.

Niko Bellic's a complicated guy. Which is to say, he's the protagonist of an open-world game.

"Open-world" games came to prominence with the adventures of one of Niko's predecessors in 2001's Grand Theft Auto III. Obviously, games (especially RPGs) have been presenting players with large worlds to explore for years, but GTA III was one of the first to allow us to interact with and explore a large 3D environment in real-time*.

I love open-world games. Love the freedom they allow me, love the ability to let my id run rampant, love the thrill of exploring beautiful, interesting worlds. I'll play them happily for hours. There's just one problem, and, for a nerd with my particular interests, it's a big one. Because the vast majority of open-world games are terrible at telling a story.

It comes down to a simple question: How do you effectively tell a controlled narrative when your game's entire philosophy is built around letting the player do what they want?

If you're Rockstar, and most of their ilk, the answer is Missions. Tightly scripted narrative missions, where your amazing game world becomes nothing more than the backdrop for a shooter (or driver or brawler) level. Where you go out of your way to curtail player freedom so that he or she can't "break" the flow of your narrative.

In case you couldn't tell, I don't care for scripted missions. They downplay the genre's strengths (freedom, exploration, player choice) and emphasize weaknesses (see: any Grand Theft Auto game that forces you to shoot for extended periods). Even when missions are well-designed or allow the player to do cool things, they emphasize what CAN'T be done during unscripted gameplay. Worse than that, they lead to scenarios like the one outlined above, putting the player's portrayal of the protagonist directly at odds with the designer's. In a novel, everything a character does is part of who they are - in Grand Theft Auto IV, the character's actions only "count" when the designers dictate them.

In essence, missions divide open-world games in two - there's the wild anarchy of the sandbox portions, where the player is given complete control of who their protagonist is and what he or she** does. And then there are the missions, where the designers tell the story THEY have chosen to tell, through cutscenes, required player actions, and scripted events. The protagonists of these two games share a skin and (usually) the same controls... but that's it. Instead of using their open worlds to tell a story, these games deliver the story as a separate gameplay experience.


Generally, when developers try to combat these problems, they do so by curtailing the freedom of the sandbox portions. This sounds bad, but it works surprisingly well at times. Rockstar Vancouver's Bully, which takes place at a boarding school, limits the amount of violence and destruction protagonist Jimmy Hopkins can mete out (both by punishing certain actions, and flatly disallowing others). By cutting off actions like violence against young children, Rockstar a) quiets moral watchdogs, b) establishes fundamental, playable aspects of Jimmy's character, and c) sets a tone for the game - one significantly lighter than that of their flagship series. Essentially, Rockstar uses Bully's limitations on freedom to enhance the storytelling by cutting off avenues the player could take it down if given more choice.


This can be taken too far, of course - there's a point of limitation where the question becomes "Why did we build an open world in the first place?" Case in point - Team Bondi's fascinating-but-flawed L.A. Noire. The game features a beautiful, vibrant recreation of 1940s Los Angeles - and precious little to do in it. The player's entire time in the sandbox portions of the game is spent driving to locations related to cases, or answering calls (which generally lead to ill-advised shooter segments). The player is controlling a by-the-book L.A. detective - one so by the book, it leaves no room for freedom or fun. The game is all mission, no sandbox, an open-world in name only.


Developers have taken the balance to the other extreme, too. One of my favorite open-world games, Crackdown, developed by Realtime Worlds, forgoes missions entirely. Instead, the player is given a set of objectives (read: people) to accomplish (read: kill). Optional objectives are available to make these goals easier to achieve, and the game strongly suggests an order to follow to make things easier, but most of the game's choices are firmly in the player's hands. There is no scripting (and, unfortuantely, precious little story). The game feels, in many ways, like a trial run (albeit an extremely fun one) for a more interesting game, one where the player's choice of what areas of the world to explore and engage with informs the story***.


Lastly, there are those open-world games that embrace the destructive anarchy of player freedom, that tune their narratives to work synergistically with the chaos and mayhem the average player unleashes when given the chance. First and foremost among these are the Saints Row games, whose plots and missions strive to be crazier and more over-the-top than the stunts and nonsense that the players get up to when left to their own devices. Unlike Niko Bellic, there is no disconnect between the behavior of Saints Row: The Third's protagonist, "The Boss," in scripted missions or open sandbox play - he (or she, or it) is always an energetically sociopathic mass-murderer.


Obviously, this approach limits the kinds of stories that can be told - but it HAS been used to great effect to tell interesting stories about bad people. Far Cry 2, for instance, uses player freedom to force you into the shoes of a potentially amoral mercenary in war-torn Africa. If your character is a heartless killer, it's because YOU chose to play him that way. By allowing the player to be a monster, it more effectively throws the consequences of monstrous behavior in the player's face in a way that would be significantly less interesting in a more scripted game.

Video games allow the player's choices to affect the course of the story in ways unprecedented by other formats. Open-world games in particular embrace freedom. I look forward to seeing games utilize that freedom in the way they tell their narrative, instead of confining it to the sandbox.




*I'm glossing over a lot of games here, because GTA III was the first game to pull this off successfully on a console - opening the genre to the vast majority of players. Bethesda, for instance, published Daggerfall, which contains a huuuuuuge (and mostly featureless) open-world in 1996, and the very weird Terminator (described, entertainingly as always, by PC Gamer columnist Richard Corbett here) in 1990.



**I wracked my memory for examples of female sandbox game protagonists, but the pickings were pretty slim. The only ones that come to mind are (optionally) The Boss in the second and third Saints Row games, and Jade from Beyond Good and Evil****. If I'm leaving anyone out, shoot me an e-mail or leave me a comment.



***The S.T.A.L.K.E.R. games generally also fall under this category*****, which made me wonder... Why are so many open-world games set in third-person? If I had to guess, I'd assume it has to do with the genre's console roots (at least as it's thought of today), and the discomfort in previous console generations with first-person views. But there also might be the fact that sandbox games emphasize exploration, and it's easier, I think, to engage with environments if you can see the figure you're moving through them. Just a random thought.



****EDIT: Twitter chum @patchworkearth, who posts excellent game criticism at the site for his soon-to-be-published, certain-to-be-awesome web comic Project: Ballad, pointed out quite rightly that Beyond Good and Evil hews much more closely to the Zelda model than the GTA one. Which begs the question: is a game that features a large, open, explorable world, with the player funneled into meticulously designed dungeons, a "sandbox" game? Or does the term refer more specifically (at least, in the methods I've been using in this article) to the urban exploration games that descend from Grand Theft Auto III? (Bonus reading: The slightly-surprising Wikipedia page for "Grand Theft Auto clone").



*****Okay, maybe I'm stretching a little bit here to include S.T.A.L.K.E.R., since they break the game's world up into individual zones. But I think they still count, since they're non-linear real-time 3D games based on exploring areas. Which.. I'm just going to keep expanding this genre until it includes Skyrim. And Super Mario Galaxy. And everything, ever. Then, and only then, will I be free.

But, seriously, the goal of this essay was to look at how you handle telling a story in a genre built around non-linearity. If I strayed from dead center of that topic to look at how other games handled the problems, or how they reflect on the sandbox genre... I can live with that.

Over The Flagpole: Escape as Art

I got into a conversation the other day with a friend of mine, a comp lit grad student. Her familiarity with gaming being somewhat limited, and my love of pontificating being extreme, I gave her a quick re-hashing of the old Roger Ebert "Video Games Can Never Be Art" discussion (a conversation I usually shy away from, since Ebert is a) obviously wrong, and b) probably arguing in bad faith for his own amusement). That argument/conversation morphed into a more interesting one, which I'll try to summarize here.

For me, a game becomes art when its mechanics and interactivity - the things that define it as a game as opposed to a series of pictures or a movie or a novel - make a meaningful statement about reality. Jonathan Blow's Braid is the obvious example, with its themes of perfectionism and regret amplified by the game's time-manipulation mechanics. (More recently, Molleindustria and Jim Munroe's game, Unmanned, examined the disconnect between Americans and the consequences of their actions by forcing players to divide their attention between two screens, one displaying dialogue choices and the other arcade-like mini-games related to the life of an American drone pilot). It's obvious to me that games can be designed to be art.


But the question that nagged at me while thinking about the issue was this: Can games generate art (or at least, artistic moments) without designer intent? Not just in the sense of unintentional meanings being found in the design, but as an expression of player freedom (hypothetically one of the key tenets of game design?)

One of Ebert's major contentions against video games is that, by giving up control to the players, a game's designer can never express a singular artistic vision the way a director or an author can. What he doesn't understand is that game designers don't give up control - they incorporate it into the stories they tell. (Most) games are not designed to play themselves - they expect certain actions from players and respond in turn. The design of a game is not just the world the player moves through, but the actions the game allows the player to take. In fact, the whole point of much of game design is to constrain player actions to the ones the designers want. Going even further, pieces like The Stanley Parable* make the case that ANY freedom in a game is an illusion, since the world is entirely constrained by the designer's intent. The only choice is to play, or not to play.


Even open-world games, touted for their openness to player choice, exist as shallow ponds, surrounded on all sides by the inflexible nature of the rules that make up the system. Most games that talk about freedom are actually talking about adding more rules, more systems into the game. Oh, you can own your own property in this one! - Once we've defined what property is, how it's purchased, and what benefits and effects it conveys in game. The system either tightens - you can romance many of the cast in Bioware RPGS, each of whom requires extensive scripting, animating, and writing - or becomes painfully vague - you can romance almost anyone in Fable III, but only by making the process generic enough to be described with extremely flexible rules. In either case, all actions are still defined by the rules supplied by the designer, and all "art" is the product of the designer's intent. But can the interaction of the rules themselves, divorced from the artist's goals, produce art?

All this got me thinking about Langton's Ant.


The Ant is a very simple (hypothetical) mechanism that produces very odd results. Existing on a grid of black or white squares, the ant moves according to two rules: If it encounters a white square, it turns right, flips the square to black, and moves forward. If it encounters a black square, it turns left, flips the square to white, and moves forward.

As you might expect, for the first several thousand iterations of these rules, the Ant produces a chaos of black and white squares. But then... (And this happens regardless of the initial black-and-white makeup of the grid, although "obstructions" may delay the process), the Highway emerges. The Ant begins to build a long, diagonal pathway away from the central pattern, made up of identical 104-step loops. In the most basic version of the Ant, this occurs at around 10,000 steps. Which is to say that, coded in those two very simple rules is hidden information about the incredibly complex behavior that occurs in their 10,000th iteration.

Video games have a lot more than 2 rules.

Obvious news flash: Game designers don't physically build the games we play. Instead, they lay out rules that describe the world they've imagined. If the rules are well crafted, the world is constructed to the designer's specifications, and the player is "trapped" within that design. They can only have the experience the designer WANTS them to have - any art that occurs in that scenario has been dictated by the designer, with the player acting, essentially, as a prop. But, as we just learned, rules can be tricky... And, really, what's a bug? A rule that's not doing what it's meant to.

Sometimes a bug in a game is there because someone miswrote a rule - a variable was mis-set, a step was left out. But sometimes it's because the rules are interacting in strange ways, ways the designer never intended.** Rules related to how player movement works intersect with rules defining how surfaces behave at corners... and suddenly you're on the other side of the wall.


There are whole communities built around these glitches. Speedrunners, especially, delight in finding ways to "break" the game to improve their times. But it occurred to me that there's something artistically meaningful in these sudden bursts of freedom - perhaps the most meaningful message games can convey. The artist loses control of his or her art - or, maybe, the rules themselves become inadvertent artists, working in concert with the players.

The example that comes to mind, that maybe this whole discussion builds out of in the back of my mind, is from the first game I ever owned: Super Mario Brothers. Every non-castle level of SMB ends the same way: the player guides Mario to the end-of-level flagpole, touches it, slides down, and control is taken away until the start of the next level. The flagpole stands as the edge of the designer's intended play area.

Which makes for a disconcerting experience when you realize you can jump over it...

Now, there are very few areas in the game where this is possible without cheating (although, now that I think about it, the use of cheat devices like the Game Genie to alter gameplay is ALSO a way for players to defy or circumvent the designer's intent by interacting directly with the game's rules). The easiest is in World 3-3, where there is a player-movable platform directly before the goal (probably significant that this trick is only possible in one of the rare parts of SMB where the designer gives control of the environment to the player). With good timing and a proper running start, Mario can leap OVER the flagpole, escaping the intended bounds of the playable area. Moving past the castle, Mario can now run along a featureless brick wall, unable to backtrack... until time runs out and he dies. Technically, it's pretty non-climactic.

But, as a kid, it was an incredible moment. I had ESCAPED the level. I was free from the designer's plan. I had done what wasn't expected.



Now, I'm not trying to suggest that this is some horrid violation of the universe's physics that would make Shigeru Miyamoto rip out his hair in a dark rage. It's clear that Nintendo knew about this glitch, and added a placeholder area behind the flagpole so that the game wouldn't crash. But the rules of SMB are exceptionally clear: the flagpole is the edge of the world. You can't jump over it, you can't escape. In the language of the game's design, it's a solid brick wall.

But by colluding with the rules, I beat the design. Experienced a moment of authentic-feeling freedom in a world entirely bound by constraints.***

It felt like art to me.



*The Stanley Parable is an excellent freeware mod for Valve's Source Engine. You can download it here, and I strongly recommend it. I'll probably have a post about it up in the next few weeks, but it's a game that deserves to be played before you read about it.

**Interestingly, combos, the backbone of most competitive fighting games, stem from a glitch in Street Fighter II that allowed players to chain moves together. Glitches like these often end up becoming part of high-level gameplay in competitive games, basically allowing the interactions between the rules, and the player's ability to take advantage of them, to trump initial designer intent.

***One of the factors behind the success of Valve's game Portal (besides being excellent), is how well it simulates this feeling. The entire first two-thirds of the game function as a metaphor for the game player, being forced to complete ridiculous, arbitrary tasks by an inflexible machine. The moment where the player breaks free of GLaDOS's control, escapes the fire, and breaks into the "backstage" is an in-design attempt to recreate the same freedom I'm trying to convey in this post. Of course, it's fake freedom... But it talks to the same part of the gamer mind, I think.