There’s a particular version of this problem that can occur in games that feature both strong narrative and strong gameplay: the game’s narrative replayability falls short of its ludic replayability. Ludic replayability is something which we accept can be prodigiously extensive. But narrative replayability is usually taken for granted as being inherently limited. You’ve seen it once; you’ve seen it all, right? Well, not exactly.
Typically, there are two basic directions one can go towards creating narrative replayability: vertical and horizontal. Vertical replayability deals with narrative length and excellence in presentation, either through a depth and richness of information (textual, visual, etc.) that requires thorough parsing to exhaust, or brilliance of expression (deeply felt acting, supremely communicative music, superior craftsmanship of the mise en scène, etc.). You can play the mind blowing opening sequence of Mass Effect 2 at least three or four times before it gets old, for example, and it took forever before people finally got over bullet time.
The concern of this post, however, is horizontal replay value, or the volume of new content that can arise from variations in the expression of a narrative. Horizontal expansion allows us to increase the amount of content in a game without unduly lengthening the narrative. Creating greater horizontal narrative space in games, however, is a conceptually tricky thing as the size of this expansion depends on the ludic mechanics which generate variance. That is, since all variations in a game’s narrative originate from ludic activity, the available horizontal space is directly determined, not by the narrative, but by the underlying ludic structure.
This requires us, then, to take an examination of how ludic systems have been incorporated into game narration.
Two Areas of Skill
For our purposes, there are two primary skill areas of interest: twitch/dexterity skills and resource management (i.e. strategizing and execution). In practice, there’s a lot of overlap between the two, and their efficacy usually requires other skills (e.g. pattern recognition, memory recall, or emotional command skills like bluffing, morale control, etc.). Nevertheless, a significant portion (if not majority) of skill tests through which a player exercises agency in games is derived from these two categories of skills.
Twitch skills deal with speed and accuracy of reactions. Resource management, on the other hand, pertains to making choices that best utilize the available assets to achieve specific ends (these ends may be ludic or narrative). This includes things like knowing when to blow skills with 60 second cool downs; using appropriate countermeasures at the right times; or effectively investing more obvious, numerical assets like money or skill points; etc.
A lot of us have had the misfortune of experiencing twitch skill tests in narration before through devices such as QTEs. The subject of the failure of these devices is slightly off the topic of replayability, but it merits a brief discussion here anyway because of its relevance to understanding ludic systems in game narration. In the past, these devices have basically failed because there has usually been no room for variation, no option, except do or die.
Actually, the lack of variation in these devices itself isn’t the problem so much as their inability to make use of the absence of agency (we’ll have to see how Quantic Dreams does with Heavy Rain). They could work if only they were properly interfaced with a narrative rationale that justifies the destruction of agency. It’s not enough to just toss in some arbitrary Simon Says overlay which blocks the progression of a cinematic and call it a day. That is to say, the failure of these devices so far shouldn’t be taken as some final indication that narrative and gameplay don’t mix.
Not to mention, there’s still that whole other set of skills which happens to be the main concentration of this post. How can we more effectively make use of resource management mechanics in game narration? Here, I say “more effectively” because, again, it is not as if this hasn’t been done before. But here too, the implementation has been pretty thin—not so much resource management as “get to point b by brute force spamming all the dialogue options.”
To be honest, the above assessment of resource management in narratives isn’t fair. Games like the original Fallout have done a pretty good job of introducing management skills into narrational navigation. Distributing points in such games to get the narrative you want while sustaining a viable enough build to get you through the game can be quite tricky. And, to a lesser degree of relevance, there’s always that so called “non-linearity”1.
The problem, however (at least, for replayability), is that in these cases, most real narrative variations are intimately tied to static or committed predicates. This means narrational navigation is usually more a matter of maintenance than management. Either there is only a single set of ludic options towards achieving a specific narrative end, or there is only a single narrative progression which is possible for a specific character build.
For instance, moral choices are usually boring binary decisions because taking a specific moral stance typically means there’s only one “choice” you can make. If you want to retain a Light Side playstyle and narrative while trying out a different class in Knights of the Old Republic, well, you’re going have to make the same narrative decisions2. Or in the other case, simplistically speaking, the only way to experience a variant narrative in Fallout is to use an entirely different build (perhaps a bad example). You might call these narrative divergences “static branches” in that the individual branches of a game’s narrative provide the same narrative experience in every playthrough.
In any case, to rephrase the above once more, there are two basic problems with static branches: 1) they mean that the number of possible narrative variations (the horizontal narrative space) is almost immediately restricted by or is exactly the same as the number of “big choice” directions available in a game; or 2) obtaining a varied narrative requires the player to change his playstyle (see addendum) so that the other branches can be accessed.
But what other option is there besides static branches?
Before considering a different paradigm, however, it needs to be recognized that static branches centered on “big choices” or character classes do serve a purpose—they reduce the possibility of redundancy. That is, simply introducing a lot of ways to do the same thing for the sake of having a lot of ways to do the same thing in no way contributes to replayability. Besides cheapening each individual challenge and decision, this can give rise to two possible scenarios:
What we need, then, are dynamic branches, or a narrative structure that retains the “big choice” directional guidance/hard divisions and progressive focus of static branches but also produces a different, randomized or dynamic experience for each individual branch (”What?” you say?). If you could call static branches “playstyle based divergences,” you might call dynamic branches “playstyle preserving convergences” (”What?” you say?). Ok, let me explain.
A) If there’s no reason not to take any or all the different ways to get to the same location, the demands of resource management dictate that the player waste a lot of time intentionally backtracking and taking the other paths to make sure he didn’t miss anything. Nobody wants to voluntarily pass up an item container or other ludic opportunity (recall the split pathways just before Saren’s base in Virmire in ME1).
Or B) If there are no hard divisions or directionality between the different paths, the player ends up lost and/or aimless, thus either unintentionally backtracking and covering a lot of the same ground, or losing interest and steam (because the sense of progress is so diffused). This is like the Normandy DLC in ME2 where the space is wide open and repetitive (directionally indistinguishable) enough that picking up all twenty dog tags is a lot more difficult (and ultimately, boring) than it needs to be.
Let’s consider an implementation example. Dynamic branches could be done through the use of narrative content packaged around randomized game objects such as NPCs and items. This is “randomized” in the sense of randomized item drops in hack and slash games, where the resources that are available to the player are dynamic. Because the narrative content would be tied to these random objects, such a system means the content that appears to the player depends on the object “drop”.
Actually, something close to dynamic branches has been done before in Westwood’s Blade Runner. But, as I mentioned, the amount of real narrative variance that can arise depends on the game's ludic structure, and Blade Runner had a very limited—that is, static—ludic structure. The same ludic problems/solutions (and thus the same narrative content tied to them) spawned in every game. Whether or not the NPCs were replicants might have been random, but in practice the lack of ludic variance meant that there was hardly any narrative divergence at all until the end of the game.
To understand, then, how this might work, we could apply this to the genre of the Hong Kong style deep undercover police drama. This genre is particularly useful as it provides us with a wide array of variables which can be randomized while maintaining narrative cohesion. We have: suspects who may or may not be undercover agents themselves, agents who may or may not be dirty, evidence or drugs deals which will “spawn” in some games or won’t in others, etc.
In this hypothetical game, the two main “big choice” directions (the two “dynamic branches”) that are available to the player pertain to the player’s decision to remain a cop (albeit an undercover one), or his inability to resist the lure of drugs and seemingly limitless power, resulting in the player going dark and actually becoming the criminal he is impersonating.
Each key narrative event, encapsulated into “cases”, occurs or unfolds based on the above mentioned variables. Since the player mostly won’t be able to farm a narrative in the same way he can farm end act bosses for items to make up for an unlucky resource output, we’ll have to implement something like quest rewards in WoW which guarantees that a player will at least have a reasonable resource pool. In this example, it could be something like a “get out of jail free” card received at the end of every completed case which allows the player to call in some favors (from judges, major criminal players, informants, etc.).
So, let’s flesh out a possible scenario for a case that might occur earlier on in the game. The case premise/variable ludic resource in this scenario is an informant who provides a tip that a certain deal is going down.
Sample Progression A: the deal turns out to be a set up; a firefight breaks out from which player narrowly escapes with his life; afterwards the player recovers the informant to salvage what’s left of his cover; the player is now tasked with taking down the major players behind the busted deal.
Sample Progression B: the deal goes down as promised, but the informant and drug bust is compromised; the player “works the streets” (about as close to “farming” as this example will get) or calls in a favor to figure out where the informant is being held; the informant is recovered after a considerable firefight, but the major players get away, etc.
In the above scenario, the two main “big choice” directions play out in the following manner.
At any rate, this type of structure would allow the player to retain his specific playstyle (staying clean or going dark) while experiencing a fairly different narrative within the same case. It also requires the player to make-do with the “drop” he gets, thus engaging his resource management skills and generating greater ludic replay value. The cases, then, would accumulate as the player goes deeper and deeper under cover, with the final outcomes being either that the player busts a “major-major” cartel, or becomes the head of one himself.
For progression A’s “clean” direction, the player can make a lawful arrest of the informant at the end and get some real information which compromises one of the major players, allowing the player to use that leverage to get back under cover. Or for the “dark” direction, the player could torture the informant before finally silencing him, using his mangled corpse as a means to salvage his cover (“would a cop do this?”). Alternatively, the player could have bluffed his way out of the set up earlier, making the informant out to be the cop and serving him up to the mercy of the criminals.
For progression B, similar decisions could be made about how the informant resource is “expended” in either a legal (the informant get his promised legal break, but he gives the player another tip as a parting gift) or illegal (the same tortured corpse option which provides an “in” back into cover by passing him off as an undercover agent) way.
If there are maybe ten possible cases total and eight cases per game (that is, there’s some controlled cycling between which cases show up and which don’t per playthrough), and each possible case has two or more real variations, that’s quite a lot of combination besides the two “big choice” directions the player can go for each variation. Hopefully, this translates into lots of narrative replay value.
Similar systems to the above would also work equally well in any genre with a large quantity of unpredictable variables or duplicity: time travel (Quantum Leap style), political spy thrillers, medieval court intrigue (think Lois McMaster Bujold’s Curse of Chalion), etc.
Anyway, the point of this post was not to pitch game ideas but to demonstrate that narrational navigation is already a firmly ludic activity, if one whose depth has yet to be fully explored. We can increase narrative replay value, then, by developing more sophisticated systems of resource based narrational navigation.
Ultimately, though, the amount of content that can be implemented in a game is limited more by production costs than a game’s narrato-ludic structure. Plus, all these “narrative replay” possibilities don’t mean a whole lot if the gameplay doesn’t hold up.
Still, it’s worth considering that we might get more mileage out of the same amount of narrative content by distributing it horizontally rather than vertically. This way, we can pack a lot of narrative without necessarily burdening the player with too long a progression per playthrough, thus helping to encourage the player to complete a run and go back in for another one.
1 It should be noted here, though, that non-linearity in the way it has been implemented in games such as Mass Effect or Dragon Age is a purely ludic—and not a narrative—device. It does nothing to add to narrative replayability since the order in which these events are done essentially has no impact on the narrative.
This is typically intentional. While narratives can take advantage of non-linearity, this means that certain narrative outcomes will no longer be possible depending on what the player does first (and, as I mentioned in another post, you can’t respec a narrative progression). Obviously this is problematic since the player will generally have no way of knowing this until it’s already four hours worth of gameplay too late to do anything about it, for instance.
2 The Mass Effect games do a better job at this than most, however, in that the morality isn’t a sliding scale but separate point pools. This allows the player to toss out a couple Renegade reactions here and there, but still remain Paragon for all the major decisions, for example. In other words, the two static branches aren’t as static as usual in that they’re not strictly exclusive, allowing the player to mix between the two quite easily. Nevertheless, the amount of new content that can arise in different playthroughs still depends on the player’s willingness to alter his playstyle decisions (see addendum).
Addendum: The Role of Emotional Reasoning
Narrational navigation, like most ludic choices (or general decision making overall for that matter), has a second component besides resource management: emotional reasoning. Emotions are frequently dismissed as irrational, but in analytical psychology, emotional processing (that is, feeling) is classified as a rational function. It makes personal sense—is even appropriate and expected—to be angry in some situations, happy in others, etc. There’s a degree of flexibility which corresponds to an individual’s emotional intelligence, but the point is that it’s an internalized, individualized system of reasoning which generally produces the same or similar reaction for a given situation.
For instance, in many cases where the player is given the choice of Renegade or Paragon dialogue decisions in ME2, there’s no real ludic consequence of going either way. But just as it makes emotional sense to some to deal out retribution, it makes sense to others to distribute mercy. Or better still, there is frequently a negative ludic (and even narrative) consequence for making a “morally upright” decision in Dragon Age, and still some players will feel a great deal of discomfort not taking that decision. A similar situation can be seen in players who refuse to leave behind fallen teammates in Left 4 Dead—it just “doesn’t feel right.”
At any rate, emotional reasoning is a fairly consistent, persistent, and insistent thing. Thus where narrative branches hinge around emotional choices (morality, or sometimes even a player’s build, etc.), relying on the player to make a different choice to achieve horizontal replayability is basically the same as asking the player to ignore or change this internal system of processing just to squeeze out a little bit more “new” content. But unless you’re quite flexible or aren’t really invested in the narrative to begin with, you’re probably not going to enjoy yourself as much playing a goody two shoes when you’re already used to being an ultimate badass.