A version of this article first appeared in the February 2011 issue of Game Developer magazine.

As mentioned in last month’s column, there are many ways for games to leverage story, ranging from passive background information to being the primary driver throughout all the game content.  However, in most of the examples that we talked about, story was passive, useful for guiding people through the content but giving the player little avenue to actually change the flow of the story.  Many great story games, such as Uncharted 2 and Starcraft, present stories that the player might find deep and engaging, yet give the player very little agency to make changes.

However, some games try to go farther, and let players actually make choices that shape and change the narrative of the game.  The patron saint for these games are, of course, tabletop RPGs like Dungeons and Dragons, where a room full of dice-rolling adventurers are free to take their adventurers in any direction they choose, while a frantic dungeon master tries desperately to get them to the front door of the dungeon he brought to the table.

This freedom is one of the hallmark of these tabletop RPGs, and it should come as no surprise that developers trying to bring the tabletop experience to life on the PC.  I work at one of these companies today, and seeing Bioware put these games together up close has given me new appreciation for the remarkable design intricacies in the construction of making these games.

How Much Choice?

How much choice can you actually give the player?  In a tabletop game, the players can go any direction they want to seek content and narrative, with a dungeon master there to drive players to content and to decisions.  In a tabletop game, the answer may not necessarily be to save the princess or sacrifice her to save the village, but instead to find a middle road, an ingenious grey area where both might be possible.  This kind of emergent problem solving provides some of the most magical moments of the tabletop experience.

In video games, this third path is likely one you  can’t afford to provide the player.  Designers have to account for and predict all of the player’s likely choices – if they provide a third door to open, there better be content behind it. Each choice you allow the player to make is content that you have to create, support and QA.  And when choices are stacked on other choices, you’re creating a spiderweb of content, of which your average player will only see one path.  Each mediocre or lukewarm choice you offer reduces the odds that the players will see your best stuff.

And choice must be constrained for other reasons.  In a tabletop game, the players can feel free to kill the king that grants the quest – the gamemaster is there to rescue the narrative, perhaps spinning the story towards a lesser Duke or rival who might have similar reasons to send the players into the Dungeon of Whatever.  Providing this level of responsiveness, and still providing content delivery (cinematics, VO) at a high level of polish, is nearly impossible in the CRPG.

Invalid Choices

In the tabletop game, the gamemaster could just let the game be broken, and declare the dungeon lost to the sands of time.  The players chose to invalidate the quest, so let them, right?  Players are less patient with electronic narrative that stalls because he made some unfortunate choice, whether it was accidentally or not.  I suspect the $45 dollar price tag might be a factor.

Most writers who aspire to join the ranks of Bioware understand that narrative choice is a cornerstone of our game design philosophy.  Still, many applicants stumble with the pitfall of ensuring that all choices are valid.  Letting the princess die or the village burn might be an unfortunate choice.  It might have deep ramifications that affect the players further within the game, or require the player to perform supernatural acts to atone for his decisions.  But a choice cannot leave the player’s game in a broken state.

Even if this were somehow desirable, it wouldn’t particularly be good game design.  Sid Meier once said that a game is a series of interesting choices.  Choosing between a viable story path and a narrative dead end is not a particularly interesting one, because there is so obviously a right choice to be made.  If you let the princess die or the village burn, there must be value to that choice.

Compartmentalize

Choice is hugely important, but it only really truly begins to shine once the player can see the results of their actions in-game.  Choosing to save the princess and let the village burn doesn’t carry much weight if the village is still there after the fact.  Seeing the results of your actions bear fruit, whether it is inside the narrative (i.e. turning an enemy into an ally for the endgame) or through mechanics (such as when Mass Effect 2 conversations and quests would increase your standing with your companions) is what truly makes these choices matter, rather than being transparently inconsequential.

But having those reactions appear in the world can be tricky, especially when they can compound.  Consider that you’ve completed three quests for the King, and in one you saved his son, in another you killed his treacherous daughter, and in a third you erred and let a village burn to the ground.  Even if each choice were only binary, there are 8 possible outcomes the player could choose, which means even crafting a return conversation with the king is likely to be a heroic task.  Writing dialogue when you’ve done one good thing, one terrible thing, and one tragic thing is almost entirely doomed to be a contradictory, muddled narrative mess – especially if it must be spliced together at runtime through a dialogue tree.

One way to work around this is to compartmentalize your big decisions – instead of having one king grant all three quests with momentous decisions, divide them up beyond three nobles, each of whom reacts strongly to the results of the quest of most interest to themselves, but is relatively complacent about the troubles of his other nobles.  One can see this in action in Dragon Age, for example: the Elves are quite concerned about their own problems in the forest, and care only for how the player handles their woes.  They care little about the player’s interference with the likes of wizards or dwarves.  The narrative is compartmentalized.

Too much compartmentalization can undermine the sense that the player’s choice matters all that much if overdone, but it doesn’t take all that much for the designer to grant that sense of accomplishment.  A couple of well-placed NPCs or the pronouncements of a town crier can create the sense that the player’s actions have consequences.  In Dragon Age, they go a step further by making the choices you make helping the elves, dwarves and wizards affect which allies you have by your side in the final confrontation.

Player Alignment

It’s impossible for NPCs to react to every combination of choices that the player has made, especially if the design of these stories has been heavily compartmentalized.  As such, many designs centering on narrative agency include a summary score that describes the player’s actions so far.  Dungeons and Dragons had alignment, of course, but it was meant to be more prescriptive than descriptive.  Since then, dozens of games have had similar systems.

The Ultima series is one early notable game – its virtue system tracked the actions of the player thus far.  However, this system was not really about narrative agency – the player had little choice but to be virtuous in order to beat the game.  Since then, though, many games, mostly RPGs, have attempted to provide some sort of moral choice, quantifying all choices combined into a single meter.

Knights of the Old Republic, for example, gave the stark choice between the Light Side and the Dark Side expected from fans of the Star Wars movies, but the choices are frequently more interesting when they aren’t strictly good or evil.  Vampire: Bloodlines allows the player choose between embracing their humanity or their budding bestiality.  Mass Effect allows the player to choose between playing as a do-gooder Boy Scout, or a Jack Bauer renegade bent on getting the right thing done no matter the cost.  Red Dead Redemption allows the player to choose between honor and dishonor.

These systems must be designed with care, however.  Reaction to both sides of the meter must be given by the game or the NPC inhabitants within or the system will be seen as a waste of time.  However, if spiking one of these scores give tangible, powerful rewards of some sort, then players will likely choose to ‘game’ the system, choosing distasteful choices for material gain (whether or not this is a good design result is one that will vary from game to game).  Worse, if spiking a score gives strong, tangible and important results, the question quickly arises as to how to quantify and reward the ‘grey’ player who walks a middle path.

The Story Thus Far

If games truly are about providing meaningful choices to the player, then allowing players to determine the flow of the narrative itself is one of the most powerful and effective ways that games can truly claim to be interactive.  However, interactive fiction is not about merely allowing the player to make a couple of inconsequential choices inside of a dialogue tree that all leads to the same place.  Interactive storytelling should make the player feel like he’s not just an agent in the world, but actively making decisions that shapes it.