The content of this website is licensed under a Creative Commons CC BY-NC-SA 4.0 Attribution–NonCommercial– ShareAlike 4.0 International License. You can freely use and share everything non-commercially and under the same license as long as you attribute it to and link to:
Why DriveThruRPG? It’s the largest download store for role-playing stuff in existence and you’ll probably end up buying much more than just your copy of Ludotronics. Which would benefit game designers everywhere!
Why not Amazon? For one, illustrated non-fiction isn’t well-suited for the Kindle format. Also, at a €14.99 price point, Amazon’s cut amounts to €9.75. Well, no.
Level Five: Architectonics
Beat 4. Problem Space
Paths & Possibilities
Designing choices for the player is an important part of the design process, and making decisions as a player is a significant part of the playing experience. It’s so important and so significant that there are games that offer choices and demand decisions all the time. But not every choice is alike, and many games would be poorer for it if everything were just about choices and decisions. Gameplay with and without choices will be discussed in two successive beats: this beat, which focuses on agency, and the following beat, which focuses on participation. Lacking clear ideas about either makes design decisions with respect to empathic tasks more difficult than they need to be, and their results less compelling than they ought to be.
Autonomy and agency together form one of the four motivational drivers of our player motivation model that we developed in Level One: Integral Perspectives I. Now it’s time to breathe some life into it and expand on it at the same time. We will look at the interlocking machinery of autonomy & agency, choice, decisions, and problem space, in that order.
Agency is first and foremost about the question whether player actions and decisions make a difference in the game world. But autonomy also figures in. Having autonomy, the player can decide to a certain extent what goal to pursue and how to pursue it in the first place. Together, autonomy and agency cover the capacity to form intentions, establish goals to realize these intentions, and then make decisions from a set of choices that have different consequences to achieve these goals. A typical game provides the player with a goal, around which they can wrap their intentions, and with choices that require a greater or lower number of correct decisions to achieve this goal. So far, so good. Often enough, though, games attach one correct decision to the choices they offer, which looks innocuous enough but opens up a can of worms.
First off, if a choice has just one correct decision attached to it, that breaks the first two rules of our rule set for empathic tasks that we developed in the preceding Beat 3. Paradigms Lost. But it gets worse. If there’s a choice with just one correct decision, the player has neither a choice nor agency! Imagine you can choose between hot, medium, mild, or no salsa with your burrito and you’re abruptly shown the door if you don’t choose “hot” as the one right decision. (Which it is, but still.) So in a vast number of cases, the player has to make decisions, but literally has no choice, and there’s much less agency involved than we think it is.
This is something we can change. We can design games that give the player different things to want. We can design games that give the player different goals to pursue that match these wants. We can design games that give the player the choice to pursue their goal in different ways. And for all that, we can and must design choices that enable the player to make and enjoy more than one correct decision. The player might choose mild salsa instead of hot salsa and still achieve their goal of enjoying palatable food, which in turn realizes their intention not to starve. But this decision provides the player with a different experience than any other decision.
But there’s another thing we have to consider: how player agency relates to task type and game type. To do that, we need once again our task model from Level Two: Interactivity that we already used in the preceding beat.
This tells us two important things. Autonomy/agency is not directly aligned with either side of the story/narratology vs. mechanics/ludology divide, and autonomy and agency are not closely correlated but aligned in different measures to different game types.
Now that we have gained a basic understanding of how autonomy and agency relate to games, let’s proceed to the nature of choice and the question of designing excellent choices for the player. We will come back to our task categorizations and game type associations later on, to connect these characteristics to problem space.
In Game Design Workshop, Tracy Fullerton provides a great introduction to the nature of choices. From that, we will pick two important points and proceed from there. According to Fullerton, a choice is then, and only then, a true choice when the options on offer are neither obvious nor arbitrary from the perspective of the player. With a line of arguments quite different from ours, Fullerton’s analysis leads her to the same conclusion we arrived at above and in Beat 3. Paradigms Lost: if all decisions save one lead to failure, it’s not a true choice and therefore not true player agency. The second point, and that’s the one we will expand on now, is that a choice that offers different decisions with different consequences doesn’t accomplish anything if it isn’t an interesting choice, and that the way to make a choice interesting is to design it as a dramatic choice.
Now, what is a dramatic choice and how do you go about designing one? A dramatic choice offers the player decisions that have different consequences in terms of advantages and disadvantages. One decision could have short-term advantages, another long-term advantages. One could be of material value, one of spiritual value. One can be purely selfish, one unselfish. One can be based on fear, one on trust. And so on. (Level Two: Interactivity introduces the Prisoner’s Dilemma as an example for non-transitive conflict design that fits such patterns well.) But that’s not enough. To design a truly dramatic choice, you need to flavor its options with different emotions. And that’s still not enough—to design a truly great dramatic choice, it must serve a function! Like any task, as discussed repeatedly, challenging the player with a choice should advance the plot, develop the character, or teach the player something new, i.e., advance one or more of the three development arcs that we developed in Beat 1. Partitioning. And if you want to dazzle and delight, design your choice in such a way that its dramatic function depends on the player’s decision!
That calls for an example. In Hamlet, Act III, Scene iii, the titular character is given the choice to kill or not to kill the King who murdered Hamlet’s father and married his mother, as Hamlet now knows for certain. Here, the function is contingent upon the character’s action. Had Hamlet killed the King at this moment, this would have advanced the plot, or story development arc, and the play would have proceeded and ended differently from the play we know. Hamlet’s decision not to kill the King at this moment, in contrast, advances the character development arc. Hamlet’s knowledge is deepened about the nature and requirements of revenge in this world, and the audience’s knowledge is deepened with regard to Hamlet’s character and motivation.
On a meta-level in this dark, unfathomable play, Hamlet’s fundamental choice throughout is to advance or not to advance the plot. Which, in turn, teaches us something else: that a choice can and should include the option of inaction as a decision that has consequences. According to Miles Tost and Nikolas Kolm from the aforementioned presentation “The Witcher 3: Crafting a Compelling Narrative in a Believable Open World,” this was indeed an important parameter for designing player choices in The Witcher 3: Wild Hunt. (But brand-new it is not: according to Michael Bhatty in Interaktives Story Telling, 1996’s interactive movie video game Star Trek: Klingon was the first game to offer the player such “in-action” decisions with consequences.)
We can go so far as to say that without a function, there can be no dramatic choice. But we have to qualify “function” a bit more thoroughly in this context. Collectable trinkets in action-adventure games like Tomb Raider or Uncharted or fetch quests in role-playing games, for example, do have functions that are tied to the player development arc. Collectibles can sometimes be used to unlock add-ons or can be sold in the game world to upgrade the player’s equipment. Fetch quests usually provide experience points to level up the player character. But while going after trinkets or following fetch quests can be dangerous, these are not dramatic choices. The trinkets and classic fetch quests themselves advance neither the story development arc nor the character development arc through deepening relations with NPCs or widen the understanding of the game world in meaningful ways, and no emotions to speak of are attached to the player‘s actual decisions to follow them or not. (Also, they often mess with the player’s intrinsic motivation, a problem we will look into in Level Six: Integral Perspectives II. What’s more, a problem rears its head in some games that is discussed in Level Two: Interactivity in the context of racing games. When players don’t advance by getting better at playing the game but by upgrading their weapons, damage values, stamina, etc., by way of experience points and leveling up, this makes it incredibly difficult to sculpt a player development arc that is intrinsically stimulating instead of just extrinsically rewarding.)
Ideally, it shouldn’t be about collectable trinkets and experience points, but about artifacts and quests that make sense in the game world. It takes superior writing skills to design meaningful side quests, but it can be accomplished. (Even on a vast scale, as The Witcher 3: Wild Hunt shows.) But that’s the job for a game writer, not for a game designer. When you, the game designer, sketch a side quest, or even just plan for one, your job is to attach a meaningful function to it beyond loot and experience points. It can advance the story development arc by feeding back into events from the main story line in interesting ways. It can advance the character development arc by deepening the player/player character’s understanding of the game world. And it can advance the player development arc by raising the player’s proficiency levels instead of merely accruing XPs.
Before we can wrap this up, we need to answer one more question. Every choice needs to have a function, but must that function always be known to the player? Well, decisions can have hidden consequences that the player cannot know, to support delayed plot points, surprise plot twists, and similar. Still, the player must be able to perceive different consequences, otherwise it’s not a dramatic choice! The good news is that the function as perceived by the player doesn’t have to be true, and that makes designing such choices a delightful exercise.
As an example, let’s look at Garrus’s loyalty mission in Mass Effect 2. When Shepard and Garrus finally track down a guy by the name of Sidonis, who had betrayed Garrus and got the latter’s squad killed, Shepard can let Garrus kill Sidonis or intervene and save his life. It’s a situation full of tension and emotion. Does it feel right, for the player, to let Garrus kill Sidonis? Does Garrus’s loyalty depend on this decision? (And, crucially, does the option to romance Garrus later in the game depend on this decision?) As it turns out, there are no negative consequences, Garrus being Garrus. Which comes as a great relief to players who just couldn’t let Garrus kill Sidonis and made the decision to intervene.
As this example shows, a choice must not have a tangible function as long as the player considers it an important decision, as long as it feels consequential and meaningful. Indeed, there was even a delayed plot twist attached to this mission, to be revealed in Mass Effect 3, which didn’t make it into the final release. (That’s from unverified sources, however, so take it with a grain of salt.) What this mission also does is develop the player’s self-perception, and what they want Shepard to be, by forcing the player to make a moral decision.
Which, after the nature of autonomy/agency and the nature of choices, brings us to our next topic, the nature of decisions.
In the Nicomachean Ethics, Aristotle states that decisions are “most closely bound up with excellence” and “discriminate” character better than actions. What Aristotle almost certainly seems to have in mind for the operative verb in this sentence (κρίνειν) is to indicate that a person’s decisions let us “determine” the character or nature of that person. But, as a second layer of meaning, it might also convey the sense that a person’s decisions “determine” that person’s character or nature.
Why does Aristotle think that decisions determine character better than actions? What he refers to are, of course, dramatic decisions as moral decisions, not burrito sauce decisions. Imagine you have a daughter and you’re a reliable parent and never break your promises to spend time with her and to attend her important life events. That’s actions. Now imagine that your daughter has an all-important afternoon baseball match that you promised to attend, but things at work have been deteriorating as of late, and your colleagues have begun to fear for their jobs and become increasingly competitive. What do you do? Will you tell your daughter that you can’t come to her all-important game, or will you go to your boss and ask to leave early? And what will you tell your boss, if you do? That you’re going to watch your daughter play baseball, or will you tell some lie to make it look more urgent? The dramatic decisions you make embedded in that specific context reveal a great deal more about your character than your subsequent actions—because what you do is keeping or not keeping your promise or telling or not telling a lie, which isn’t very illuminating. Garrus’s loyalty mission’s no different from that. Your decision to save or not save Sidonis’s life and why you arrived at that decision is what defines or “determines” your character, not the action attached to it.
Another excellent game-related example how decisions “determine” character is relayed by Anna Anthropy and Naomi Clark in A Game Design Vocabulary. Early in Affairs of the Court, a choose-your-own adventure RPG, a symbol for good luck appears and the player character is asked by an NPC to make a wish. The context suggests, and the player accordingly expects, that this is a momentous decision that will greatly affect the course of the game somewhere down the road, similar to our Mass Effect example. Here too, no consequences follow, and the player has no way of knowing that. What’s different this time is that the player never learns that it had no consequences. Which, counter-intuitively, greatly amplifies its effect!
Anthropy and Clark call this unusual kind of choice “reflective choice,” which indeed should become part of your game design vocabulary. Similar to making Shepard’s decision, deciding on a wish in Affairs of the Court makes the player reflect on the wants and needs of their player character and what they want their player character to be. In other words, the player “determines” the player character’s character or nature in the Aristotelian sense through a dramatic decision. And it has consequences in the game world, after all! It affects the player’s playing style and later actions and decisions that have consequences in the game world. It’s hard to come up with something that advances the character development arc more strongly than a reflective choice.
If you dig deep enough, you can find reflective choices, and different types of them, in a substantial number of games, albeit somewhat more pedestrian than our examples so far. The first two Mass Effect games, for example, try to augment almost every dramatic choice with reflectivity by labeling some decisions “paragon” and others “renegade.” That doesn’t work out so well for a variety of reasons, and it appears heavily tweaked for the trilogy’s finale. Then there are certain forms of reflective choice in open world–games. In Mafia III, for example, the player can make the moral decision to keep the player character’s lieutenants happy or enlarge the player character’s personal wealth. But what qualifies even more strongly as a reflective choice is that the player character can roam around the city and rob cash registers with or without harming people, for negligible loot and with easily evaded repercussions. Yes, the player character is a criminal and a murderer, but ransacking corner stores or refraining from it is still a strong statement of character. Finally, in games where the player is given the choice to refrain from killing non-player characters, that’s also a reflective choice. But this one has developed: from games like the original Thief trilogy, where it was indeed purely reflective, to games like Dishonored and especially its sequel, where its consequences have broadened to affect the entire game world up to and including the game’s multiple endings.
Which brings us to our final topic in this beat, the problem space where all these choices and decisions “live.” This problem space exists in the game, but it also exists in the player’s mind as an internal representation (the term problem space is adopted from cognitive psychology).
The problem space comprises all the subproblems that have to be solved, all the choices these subproblems present, and all the player decisions these choices allow. When a game advances along its three development arcs toward its ending or one of its endings, this space can be affected in a number of ways. It can remain more or less stable; it can widen to accommodate a greater number of choices and decisions; or it can shrink and leave less and less room for possible choices and decisions.
Earlier in this beat, we mapped player agency to different game types from the perspective of physical, cognitive, and empathic tasks. Now, we can extend this map to include problem space.
As empathic tasks are the focus of our interest in the context of this level, let’s have a look at the mechanisms of diminishing problem space in story-driven games in more detail.
Diminishing problem space is a function of suspense. If you have a story development arc, that story should be gripping. To make a story gripping and take the player on a wild emotional ride from start to finish, at least in the Western dramatic tradition, your story development arc needs to have a structure that includes an inciting incident, plot points, ups & downs, sudden reversals, and so on, and especially ever-rising stakes. It’s this structure that leaves the player/player character fewer and fewer options until everything hangs on one final option that offers one particular choice with one particular decision for one particular heroic action whose success or failure will lead to triumph or tragedy.
This has been recognized by several writers, among them Brenda Laurel in her seminal Computers as Theatre and Richard J. Gerrig and Allan B. I. Bernardo in their Poetics article “Readers as Problem-Solvers in the Experience of Suspense.” In Laurel’s terminology, there is the potential space, largely equivalent to our problem space. Initially, everything is possible in this space. When the story progresses (her primary example is Hamlet, befittingly), less and less is possible until only the probable is left. This, in turn, is diminished further and further until there’s only one course of action left, the necessary. That’s the climax of the story. The more possibilities and probabilities fall off, the greater becomes the suspense. And when the necessary finally kicks in, the audience must have the impression that it couldn’t have happened otherwise. Gerrig and Bernardo define suspense in similar ways as the audience’s belief that the quality or quantity of paths through the hero’s problem space is diminished.
Classical and modern dramatic structures play with this diminishing problem space in wonderfully effective ways. One technique worth mentioning is the dramatic moment of delay (more intriguingly called das retardierende Moment in German). It’s the promise of a different, non-tragic solution that, to employ Laurel’s terminology, suddenly springs from the depleted repository of the probable as a glimmer of hope, only to give way to the necessary, the tragic solution, in short order. (The equivalent for comedies is the “everything is lost!” moment of delay shortly before the ending, which then gives way, conventionally, to a merry mass wedding.)
Where does this leave us for story design in games and the structure of the story development arc? If we want the story to be gripping, we must give it a strong dramatic structure that provides ever mounting suspense through a diminishing problem space, complete with dramatic delays, calm before the storm–moments, and all that jazz, and ends in a climax. How can we square that with player agency, meaningful choices, and decisions that make a difference? In other words, how can we integrate such a dramatically structured story development arc with the player development arc (and, to an extent, the character development arc)?
Here’s what you should do. To design empathic tasks that fuel and deliver a tight dramatic structure but integrate with player agency, you need to differentiate between “what” and “how” and become more ruthless with the former and more liberal with the latter. Let’s look at both propositions in turn.
This is about the rules of empathic tasks that we developed in Beat 3. Paradigms Lost. Let’s recap them briefly and add some techniques. Player decisions can be right or wrong; there should be more than one right decision and more wrong decisions than right decisions; the outcomes should have different advantages and disadvantages; and different decisions should be coupled with different emotions. Then there’s the special rule that a truly climactic choice is allowed to have one, and only one, right decision, which then must excel in plausibility and inevitability. What’s already baked into these rules is the option to diminish your problem space! When the story development arc advances over the course of the game, your choices can provide fewer and fewer correct decisions, up and until the very last choice, the climax, that provides only one correct decision, the dramatic action that leads to triumph or to defeat and defines the game’s ending, as discussed in Beat 2. Prime Positions.
When the problem space diminishes and correct decisions become more rare, these correct decisions must become more plausible. Hence, correct decisions have to be prepared more carefully, and the player has to be primed more thoroughly. To prime the player for correct decisions, you can employ techniques like agenda-setting or priming that are discussed in Level Four: Narrativity. To prime the player against wrong decisions, you can employ warning signs, even fairly prominent ones at that. Either they make the player wary, or they make the player realize in hindsight that they should have known better all along. When you have pestered your player/player character repeatedly with flocks of Crebain, for example, that’s your warning sign. When they nevertheless decide to flag down a great eagle taxi to Mordor that can be easily spotted, they can then be discovered by such a flock and fried by the Evil Eye’s minions without further ado, basking in the belated realization that this was an ill-advised decision to begin with.
While you have to diminish the problem space that leaves the player with fewer and fewer correct decisions, you can and should give the player more options with regard to how they can realize these correct decisions.
You can realize such options through greater abstraction. Let’s illustrate this by assuming that, at some advanced point along your story development arc, there’s one correct decision the player/player character has to make. This can be to try and conquer an enemy base, or to try and escape from a prison, or to try and kill the antagonist in a duel. All three examples are very specific and leave little or no room for player decisions, let alone meaningful ones. On the most abstract level, each decision is about the player/player character overcoming an obstacle to get what they want. That’s as abstract as it gets! Your task now is to find a proper middle ground between the specific choice and its abstract intent. In the first scenario, the enemy base has to be taken, all right, no way around that. But you could leave it to the player to decide how to achieve that—a full-scale assault, a commando operation, bribery, blackmail, perhaps negotiation. In the second scenario, the more abstract version could be that the player doesn’t have to escape the prison but to regain their freedom, and escaping the prison is just one among several options to achieve that. (Use your imagination!) The more abstract version of the third scenario, finally, could be to remove the antagonist’s power and influence from the game world. Again, this could be achieved in different ways, and some of these ways might be more interesting than killing the antagonist off.
[Alert: spoiler in this paragraph for The Witcher 2.] A great example for our third scenario is The Witcher 2: Assassins of Kings. It ends with a climactic duel, but the player is given a choice after defeating the antagonist: to kill him, or to let him fade into obscurity. Good reasons are attached to both options, and also different emotions. Here, you have it all. The possibilities and probabilities of the problem space are utterly depleted in favor of the necessary, but the player is still able to make a decision that counts, along their own reasoning and emotions and in line with their self-perception. The dramatic climax has to happen (the what), but the player has agency and makes a meaningful decision (the how).
So far, we have only engaged the story development arc’s problem space. What about the other two arcs, the character development arc and the player development arc? Luckily, the former doesn’t have a problem space, and the problem space of the latter is of a different nature. Let’s have a look at them.
Whatever structure you prefer for your character development arc, either the hero’s journey, medieval quest patterns, or nostos as the return home, or something else entirely, these journeys are already supremely abstract. And, while the player/player character can certainly make really bad calls, the journey itself is not about choices and dramatic decisions—it is about stages the character goes through for their internal development. In the hero’s journey, some stages look like decisions, like the Refusal of the Call or the Refusal to Return, but these aren’t decisions because they aren’t choices! Then, Odysseus doesn’t “decide” to leave Troy, meander about in the Ionian Sea for ten years, lose his companions, get enticed by nymphs and sorceresses, and so on. The same applies to the quest. Here, the hero is called to a quest, which is superficially similar to how the hero’s journey begins, and seeks something valuable, which is superficially similar to the nostos structure. And while everything else that happens during the quest appears very bizarre from a modern perspective, these events again proceed along a succession of stages, not choices. Stage-based structures do not have problem spaces where the possible or the probable can be depleted over time.
The player development arc, in contrast, does have a problem space that can cause trouble. It is controlled by four of the game’s learning conditions, the skill spectrum, the skill threshold, the skill progression, and the skill maximum, all of which are covered in Level Two: Interactivity. If the learning conditions are off, quite a number of things can happen to the player development arc’s problem space, all of them detrimental to the playing experience. Among other things, the problem space can be depleted too early, or immoderately stuffed toward the end, or simply oscillate haphazardly between too much and too little like a drunken bandoneon.
This phase’s reading order is fixed for the Introduction, Level One and Level Six.
Level Two through Five are non-linear: you can tackle them in any order you like.