Sympathy for the Devil

“Know thy enemy, know thyself, know victory.”

As tactical advice, Sun Tzu’s famous maxim applies to a majority of videogames. Knowing that Piston Honda blinks just before throwing an uppercut helps you know victory. Knowing that the mothership fires two small shots before throwing up its shields helps you know victory. Knowing how many whip strikes it takes to defeat Dracula helps you know victory.

But what about the kind of knowledge that transcends the tactical – the kind of knowledge that lets you truly understand your enemy’s motivations and background, his hopes and fears? As far as most games are concerned, such knowledge is unimportant. The enemy exists only as a part of the environment – a set of pre-programmed rules to be figured out and bested. It’s enough to know that Piston Honda wants to send you a “TKO from Tokyo” or that Cats thinks you have “no chance to survive make your time.” We want to know the enemy, just not, y’know, personally.

This is in stark contrast to other forms of storytelling media, which routinely include antagonists that are known for more than malevolence. Conflict is inherent in every story, but most well-told tales are not just a simplistic battle of good vs. evil. A hit TV series like Friends might feature characters with competing goals, but there are no characters that are completely morally reprehensible. A hit game series like the Mario games, though, can get by for decades with an antagonist that kidnaps royalty and casts destruction upon the land seemingly out of sheer boredom.

Even in stories where there is a clearly-defined evil, we can usually understand the bad guy’s motivations, even if we don’t agree with their methods. Most viewers can at least relate to the revenge and greed driving Simon Gruber in Die Hard: With a Vengeance, even if we would never attempt murder and massive theft ourselves. Other stories actively encourage the audience to root for the bad guy, finding the underlying humanity in normally vilified characters like mobsters (The Godfather) or psychopaths (The Silence of the Lambs).

As a medium, games are different in this regard. In games, the bad guy is, almost by definition, the one you’re not controlling – the “other” that is trying to destroy or limit you. If you’re controlling a cop, the gangsters are the bad guys. If you’re controlling a gangster, the cops (and, sometimes, the other gangsters) are the bad guys. There is no moral ambiguity – most games are designed so it’s you and your character(s) against the world by default.

No wonder so many game makers create paper-thin, cartoonish justifications for their virtual enemies. No matter how well-defined and believable a game villain is, his motivations will almost always pale in comparison to that of the protagonist you’re actively controlling. Knowing the misunderstanding that causes Sephiroth’s psychosis and rage in Final Fantasy VII doesn’t prevent you from preventing him from destroying the Earth in the final battle. Knowing that Otacon will be crushed by the death of Sniper Wolf in Metal Gear Solid doesn’t give you the option to spare her life and sit down to a tea party.

Like a Greek tragedy, most game narratives march inexorably toward the final condition of “you win” regardless of what this might mean to the fate of a likable, non-playable bad guy. Given this inherent rule of standard game design, the question regarding enemies becomes not “Why are they doing this?” but rather, “Do we really want to know?”

How can a game designer/storyteller get around this problem? Open-ended game design is a solution, but often only a partial one. Yes, you can spend days being a law-abiding pizza-delivery boy in Grand Theft Auto III or a humble fisherman in The Legend of Zelda, but if you want to move the story along, you have to go down the relatively narrow path the game proscribes – i.e., defeat the bad guy.

Eliminating the predestined defeat of the enemy often means eliminating the story altogether, leaving the narrative and goal to be defined completely by the player, as in most simulation games. In theory, this opens the game up to unlimited scenarios, but in practice all it really does is add a “you lose” ending to the “you win” of more linear games. Allowing your people to be overtaken in Civilization does indeed subvert the traditional storyline, but not in a way that’s fully satisfying to most players. (“I’m glad the Romans sacked my capital. They obviously wanted it more.”)

But Civilization‘s multiple selectable nation-states demonstrate one way to make a sympathetic antagonist in a game – namely, making him the protagonist. While videogames are limited by forcing the avatar’s point of view on the player, they are also superior to other media in that they allow the player to truly experience a conflict from all sides. While a movie can let you into the mind of a villain, a game can let you truly walk a mile in his shoes.

Multiplayer games have exploited this advantage for years – while it’s possible to argue that StarCraft‘s Zerg or World of Warcraft‘s Horde are the “evil” side of those games, it’s not an easy argument to make to a devoted player of either race. Even simpler games can exploit this difference – M. Bison is an unplayable “bad guy” in Street Fighter II, but once you can control him in Champion Edition, he becomes just another potential avatar in the fight against all comers (even though his ending reveals that he wants to wrap the world in “the darkness of one man’s evil”).

Single-player games make balancing the morality calculus more difficult. While it’s easy enough to allow players to choose between a good or evil character at the beginning of a game, this choice again locks the player into a single point of view, requiring the player to replay the game multiple times to fully experience all sides of the equation. Games like Fable and Black & White partially fix this problem by allowing a player’s alignment to change throughout the game, but at any one moment the player is still only experiencing one side of the dichotomy.

How do you combine the personal experience of the game and the detached gaze of the camera? One of the most daring experiments in this regard is Indigo Prophecy (Fahrenheit to our European readers). Though players start the game as possessed murderer Lucas Kane, the point of view jumps quickly and often between him and police officers Carla Valenti and Tyler Miles, who are investigating his case. Actions performed as one character affect the success of future missions by the others, and in the beginning it’s unclear to the player which character, if any, he should be rooting for. Are the police the bad guys because they try to thwart you as Lucas, or is Lucas the bad guy because he’s trying to thwart you as the investigators? It’s impossible to choose, because they both represent you, and what self-respecting person thinks of himself as the bad guy?

More than multiple viewpoints, though, Indigo Prophecy succeeds in having believable characters because it is focused on human interactions rather than endless battles. Far too many games feature hordes of expendable enemies that are barely around long enough to form a wisecrack; good luck forming a believable character structure around them. The ones that do stick around longer are usually just more powerful versions of the throngs of chattel, similarly waiting to destroy or be destroyed.

To truly understand your enemies, in virtual life as in real life, you need to be able to engage them in conversation as well as battle. Games like Indigo Prophecy and Knights of the Old Republic use branching conversations to engage non-player characters, but this method inherently limits what you can say and how the characters can respond. Every path in the question-and-response tree is predetermined, each discrete branch penned beforehand by a writer.

To really introduce moral ambiguity into a game, you need a system like that in Facade, an art/research project by two artificial intelligence experts. The game invites you into the home of Grace and Trip, a couple in their 10th year of a deeply troubled relationship. The evening starts pleasantly enough, but the resentment between the couple threatens to destroy the civility and possibly the relationship.

You are forced into the viewpoint of the guest, but you aren’t limited in what you can say or where you can try to lead the conversation. You can take Trip’s side and harp on Grace’s insecure need for validation, or you can comfort Grace and defend her from Trip’s passive-aggressive barbs. Or you can strive for a balance, picking apart both parties for their petty concerns. Or you can make a pass at the hosts, earning a quick dismissal.

Facade is notable because neither non-player character is the clearly defined bad guy. You’re not caught in a battle between good and evil, but between two deeply flawed, deeply sympathetic people. The conflict is more awkward than that of most games, and also more real.

Why aren’t more games like this? Well, it took a team of people five years to develop the 20,000 lines of dialogue in Facade, and even then, the two main characters tend to repeat themselves after only a few plays. Apparently, it’s a lot easier to design a good gun than to design a good, free-flowing conversation.

But despite their limitations, games like Facade and Indigo Prophecy show that there is at least the potential for videogames to allow players to divine personal as well as tactical knowledge of an enemy. There is potential for a future where game villains aren’t just remorseless killing machines, where the bad guy is a sympathetic character that’s striving for acceptance and understanding, just like us. When the “love thy enemy” ideal becomes truly integrated in our games, we’ll be able to grok our villains so fully, we won’t be able bear destroying them.

It’s enough to give you preemptive nostalgia for the days when Soda Popinski’s only goal in life was to “make you feel punch drunk.”

Kyle Orland is a video game freelancer. He writes about the world of video game journalism on his weblog, Video Game Media Watch.