Embracing Failure in Game Narrative

Notes from GameLoop 2011

GameLoop is an “unconference” on the topic of games (both digital and non-) organized by Scott Macmillan and Darius Kazemi, based on the BarCamp spirit of enabling direct, spontaneous discussion not possible within the typical “draft, submit, obtain approval, dictate to audience” model of traditional conferences. The fourth annual session was held this past Saturday in Cambridge, MA, where it originated, though with the welcome spread to other cities (Philadelphia, Montreal) we refer to this particular gathering as GameLoop Boston.

The format is built on the morning session, where each of the 150 attendees takes a turn giving their name and a few “tags” to let others know what they are interested in, followed by an open call for session topics/titles which are then subject to voting and finally arranged to form the day’s schedule. These are overwhelmingly “roundtable” sessions where all are welcome to contribute their own insights, or questions they would like others to address. Conversations also thrive in the hall and on the patio of Microsoft’s NERD space, supplementing the scheduled events with further social and topical discussions.

This was my third year in attendance (having missed the founding) and I’d like to share some of what I took away from the sessions and unscheduled discussions. This post is based on the first session I attended:

“Embracing Failure in Game Narrative”

An open discussion of player/player-character failure and how it can be used to enrich the experience instead of being seen as valueless or a loss. In this context, ‘failure’ covered a range of meaning, including:

  • Not achieving a goal stated or implied in the game‘s design (solving a puzzle, defeating an opponent) or narrative (rescuing the abducted party, preventing invasion/destruction of the world)
  • Not meeting player’s own measure of success e.g. 100% completion, combat victory with no damage/losses incurred
  • Reaching an end-state which implies a death/loss, both traditional “Bad End” out of multiple possibilities and cases where no alternative exists, as in waves of enemies stopping only when the player-character is defeated.
  • A turn of events the player did not intend, whether considered a “failure” by the designers or not.

Failure lies mostly in the eye of the player, but designers have power to inform that perception. A strong example was provided in discussion of L.A. Noire, where bumbling through an interrogation or case will not prevent you from proceeding or meaningfully constrain the remainder of your journey through the game. Many traditional gamers nevertheless reload save states based on performance metrics provided in case summaries or the cues of mishandled questions mid-interrogation. This was contrasted against anecdotal cases where less experienced players would continue along regardless of performance and viewed the statistics as a curiosity, if noticing them at all. Might moving on to new content regardless of these insignificant missteps sate the need to feel successful, if not for the enumeration of possible successes and recorded failures?

To encourage an acceptance of perceived failure, at least two approaches were proposed: setting the expectation ahead of time that there is value in diverse outcomes, and timely post hoc indication that the perceived failure state offers new goals and experiences to the player. One participant recounted loading a prior save almost immediately when his player-character was abducted and became a mutant in the original Fallout. He regrets limiting the range of his play experience in retrospect, but at the time was provided no sufficient cues that this new state of his player-character was a valid one. He wanted to get back to “the game,” not this pitfall he had stumbled into.

This was a recent experience—Fallout was approached with expectations set by games of recent years, taken out of the game’s original context of PC titles in the ’90s. Games of this ilk (retroactively labelled “immersive sims”) are praised for the diversity of possible play styles and game states arising from systems-based design. This is in contrast to the model more commonly used today, where decisions are made at points and within constraints that were planned just so. Obvious exceptions exist in modern games: sandbox titles like Just Cause 2, lightly-scripted/AI-heavy shooter S.T.A.L.K.E.R., deep simulations like Dwarf Fortress, and the fever-dream multiplayer improvisation of Sleep is Death all diverge from that tightly plotted, beat-driven design that aims to emulate film, television or comic books more than statistical war games or pen & paper role-playing games.

In systems-based design, the rules of those systems allow you to determine the value or limitations of a given state and react accordingly; there is an implication that any state you can enter is valid to the degree that you are still able to perform desired actions. That is: if I, as a mutant, can still navigate and interact with the game world, I will do so. Only the cruelest designer will provide a level of player-character control comparable to that of standard play when in fact the game has nothing further to offer from the current state. Indefinite imprisonment of the player-character, able to move and act within a cell with no possibility of escape, is sadistic and should be reserved for design where player trust is being willfully betrayed.

Working within the context of intentionally crafted experience, timely reassurance that the player’s perceived failure does offer compelling content is essential. This can be done through effectively foreshadowing an interesting turn of events resulting from an apparently negative outcome, or prompt recognition after perceived failure that provides an incentive to continue down the current fork in the road instead of backtracking and pursuing the intended outcome. If the player can recruit a party member, for instance, they should know or suspect that this character will continue to be relevant and interesting even if the recruitment fails, rather than dying or existing meaninglessly out of view for the remainder of the game.

Crafting a narrative that allows the player to alternate between success and failure while providing motivation and compelling content in each case is no mean task. Interesting content only accessible through a particularly unlikely success or failure can be especially rewarding to those who encounter it, but lost on the majority who may be unsatisfied by the alternative they experience. Creating a scenario where results are identical but immediate feedback provides a sense of success/failure may lead to any number of frustrations if encountered again via loading a save or on a subsequent playthrough, revealing their choice or performance as irrelevant.

A common solution is creating a relatively short game that can be replayed to experience paths not taken in previous attempts by altering decisions and behaviors. A more novel approach is employed in the PSP remake of Tactics Ogre: Let Us Cling Together, where the “Wheel of Time” system allows the player-characters to time-travel and explore alternate branches of the story without the player maintaining separate saves or experiencing redundant content across multiple playthroughs.

One exceptional case of failure being central to narrative is From Software’s Demon’s Souls. If you’ve heard one thing about Demon’s Souls, it’s probably that you fail—often. This is not only embraced mechanically (as in roguelikes, brutal platformers like Super Meat Boy, and so on) but as a defining trait of its environment and story. The player-character must fail before the game begins, and the continued death and resurrection cycle of the game (based on the player’s failure and success) is at its core. Beyond that, the game tells of a ruined society, fallen heroes, the ways in which one’s success creates loss for others, and how achieving goals is no sure guard against failure. You can read Matt Weise’s “Inhabiting Demon’s Souls – My Memories of a Haunted World” for more on how effectively this is accomplished.

So how do you embrace failure and perceived failure in your game as a designer: use of an auto-save system, removing the ability and temptation of a player to incrementally reverse their state? Obfuscation of smaller successes and failures, minimizing quantitative performance analysis that may detract from the qualitative success players achieve? Provision of novel content that keeps a player engaged with whatever path they find themselves on? Design built around the inevitability of failure, either on the path to success or as an end in itself?

It is clear from prior works that diverse approaches can be used to great effect in creating a game where failure adds value to the player’s experience. You aren’t required to make “good” experiences as rewards for exemplary player performance, with “bad” or non-existent feedback as punishment for poor decisions or execution. In fact, to do so is to constrain both the design of and interaction with your work.

By rewarding what the player may consider failure, you open the world of your game to be interacted with in a free and playful manner. You instill a sense of possibility, and even wonder, while removing player frustration and stress associated with slavishly doing what they guess you intend as the “right” thing, to what they guess is your standard of expected performance.

By providing a context in which failure is acknowledged or even inevitable instead of being scorned or ignored entirely, you free yourself to create play that is challenging, without making the player feel inadequate or hopeless. These types of experiences can be incredibly memorable and rewarding to those who endure them.