Constraints breed creativity


I tend to find that too much choice is bad for me. When faced with the giant wide-open possibility space of unconstrained choice, two things tend to happen:

  1. I spend considerable time and effort trying to narrow down my options and pick the best choice, resulting in a form of analysis paralysis that I’ve come to think of as candy overload.
  2. Unless I am explicitly aware of it, and make an effort to avoid it, I’ll end up making a comfortable (and therefore typically boring) choice.

Here’s a quotidian example: the age-old question, “What shall we make for dinner?” Without further guidance, I will probably end up suggesting one of maybe fifteen meals that I’m familiar with. It’s much more exciting to face the question, “What shall I make for dinner from this magazine?”, or “What shall I make for dinner given that I need to use up this polenta and these tomatoes?” because I’m immediately narrowing the possibility space with constraints. These constraints are also likely to shift me out of the ruts of the everyday list of tried-and-true dishes, forcing me to cook outside my comfort zone, and making dinner an overall more interesting occasion.

So the role of constraints is two-fold:

  • First, to lessen the time and effort required to make a decisions in a large possibility space, and
  • Second, to force you outside your comfort zone.

The role of constraints in roleplaying

Without rules governing play, roleplaying games become an exercise in shared narrative, governed solely by the social contract existing between their players. Obviously there are very few constraints in such a game, and it generally takes a pretty skilled and motivated group to get a decent story out of such a setup.

Rules serve a number of purposes in any roleplaying game. One of these – the one I want to talk about – is to constrain play. Setting does a good deal of the initial constraint, and we can really think of setting as a collection of narrative rules. For example, if you’re sitting down to play a game of Star Wars, regardless of the mechanics1, you can be pretty sure that you’re going to be playing a cinematic Space Opera game, and you’ll probably have a good idea about the sort of characters, problems, and situations you’ll encounter. You might start thinking of characters that this setting allows or encourages: honour-bound Jedi, crafty smugglers, treacherous aliens, and so on. If you’re going to be playing a game of Apocalypse World, the mechanics constrain the types of characters too: you could be a hulking Gunlugger, a creepy psychic Brainer, an enticing Skinner, etc., but don’t give you the option to play, for example, a merchant whose abilities lie in haggling and general knowledge. In either case, our choice of characters is (implicitly or explicitly) constrained by the rules, either narrative or mechanical.

Aside: mechanical versus narrative

You’ll note that I’ve divided the rules immediately into narrative and mechanical. When I first thought of this divide, I went “Oh, this makes sense, of course there’s two different types of rules!” Now, as I try to divide rules neatly into two camps, I’m not so sure. Here’s my go at the two categories:

Mechanical rules are rules that govern the immediate here-and-now, generally putting constraints on what characters in the game do. They’re often characterised by quantifiable properties, dice-rolling, and determination of odds.

Narrative rules are rules that govern the overall flow of the story, often from a much higher perspective than mechanics. They may be implicit rather than explicit, and are often characterised by hand-wavey language and fuzzy borders.

That’s a rather terrible description, I know, but right now I don’t have much of a handle on the two. I think it’ll probably turn out that these are the two ends of a spectrum, with rules about five-foot steps in combat down one end of the scale, and ideas about shared social conscioussness down the other.

Rules in play

Once the game has started – once we have an idea of setting, the characters, and the general mood of the game – people still make a lot of decisions. Rules constrain our choices here too, hopefully to speed up play and encourage interesting choices. The majority of mechanics act as constraints for the GM: rather than having to decide if every action succeeds or fails, task- or conflict-resolution mechanics give the group a simple way of determining the outcome. In this way, mechanics make the job of GMing a lot quicker and more effortless.

Mechanics don’t just reduce the possibility space: depending on their application, they can also force interesting choices on us. Apocalypse World does this well and often. Consider the move Read a sitch:

When you read a charged situation, roll+sharp. On a hit, you can ask the MC questions. Whenever you act on one of the MC’s answers, take +1. On a 10+, ask 3. On a 7–9, ask 1:

  • where’s my best escape route / way in / way past?
  • which enemy is most vulnerable to me?
  • which enemy is the biggest threat?
  • what should I be on the lookout for?
  • what’s my enemy’s true position?
  • who’s in control here?

We could easily substitute this move for something a little more generic:

When you read a sitch, tell the MC what you’re after and roll +sharp. On a hit, the MC will tell you what you wanted to know; on a 10+ they will also tell you something else that might be useful.

In practice, these moves give the same rewards for the same roll; but consider what the player sees. When using the latter move, the player has to choose what they’re interested in, and that could be anything: “Have I seen these guys before?”, “What sort of bikes are they riding?”, “Do I know anything about these plants?”, “Where would someone hide a weapons cache?”, etc, etc. While the scene will naturally limit what the player can ask, there’s still a wealth of options that they need to think about. When using the former move, however, the player has to pick from one of five questions. As long as these five questions are relevant to the scene at hand, they can speed up play and may even give the player some guidelines as to what they should be doing. In other words: it’s limiting our possibility space, and also may force us outside of our comfort zone.2

Too much of a good thing

The question of relevance and meaningful choice is one that I think merits addressing in slightly more detail. The risk of introducing constraint is that it can curtail player agency, that is, the ability of the player to affect the story. What do I mean by this? If the constraints introduced (by the rules, the setting, or something else) mean that I have to choose between two options which both suck or just don’t interest me – if I can’t even begin to approach my goal because of constraints – these constraints are limiting my agency. Alternatively, the choices may both be interesting to me, but it may not matter at all whether I pick option A or B: both lead to the same result with minor cosmetic differences.

The example of Read a sitch above is an interesting one to examine in this light. In order to give the player agency, the designer must ensure that the choices they provide are generally relevant and varied: the player must be able to acquire meaningful information as a result of the move, and the information they receive must vary depending on the choices they make. This is doubly interesting because it indicates that by moving from unconstrained choice to constrained, but meaningful, choice, the game designer has not simply eliminated the time and effort that would usually be invested in making in-game choices: instead, they have transferred it from the player, during the game, to themselves, in the writing of the system.

(And this isn’t a bad thing! The game designer, of course, can spend this time and effort over weeks if they need to. Meanwhile the elimination of the time- and effort-sink of unconstrained choice from play greatly speeds up the game and generally makes the play experience better for everyone involved.)

Case studies

I started thinking about constraint after playtesting a “Powered by the Apocalypse” game where a lot of the constraints on character creation (name, look, etc.) had been torn out, and lamenting the opportunities for flavour that the creator had given up as a result. After that, I started looking at some of my other favourite games, and I started noticing either where constraints sped up and shaped play, or where lack of constraint slowed it down.

FATE: aspects

FATE (specifically FATE 3rd edition, and more specifically, Spirit of the Century) is a classic example of the “freeform game”. Your aspects can be anything you want! You have ten of them! If you can tell everyone how they apply to a situation, you get a bonus! If they hinder you, you get more FATE points!

The big problem that I’ve always had with aspects in FATE is that they’re just too flexible. Ten aspects, with only a few guidelines on how to assign them, tend to lead to a mix of about 80% redundant, boring aspects and 20% workhorse aspects that get called all the time. This great flexibility also starts to eat up time when the player finds themselves missing a roll: they start looking down their sheet, trying to find an aspect that applies to this situation - or worse, trying to wrangle one of them into convoluted shapes that might fit the situation. I’m not saying that this guaranteed to happen, but it’s happened enough times in games I’ve played in or run, that I think it suggests something at fault in the mechanics of the game.

Later iterations of FATE doubled down on constraints for aspects. The Dresden Files Roleplaying Game introduced the idea of each aspect being tied to a specific part of your character, and this constraint suddenly leads to much more evocative and constructive aspects. I think this is a great improvement: not only are you limited in what you turn into an aspect, but the labels on the aspects also help you to make creative, interesting choices about your character.

The Shadow of Yesterday (and Lady Blackbird, which is influenced by it) use a similar technique (which they call “keys”), but in a very different manner. For example, one of the characters in Lady Blackbird has the Key of the Guardian, which would be represented in FATE by an aspect like “Loyal Guardian”. The key, however, provides explicit instruction on how to “hit” it and regain dice/XP:

You are Lady Blackbird’s loyal defender. Hit your key when you make a decision influenced by Lady Blackbird or protect her from harm.

By putting constraints around exactly when and how keys are triggered, these systems remove a lot of the uncertainty that I find hovering around a table whenever anyone eyes their stack of FATE points and goes fishing for a compel. In my experience these constraints mean that play is a lot faster and more directed than in FATE3-based systems.

Society of Dreamers: Character creation

I’ve written about the crazy mad love affair I have with Society of Dreamers’ character-creation system before. To recap:

  1. Every character is defined by four categories: gender/sexuality, nationality, profession, and age.
  2. Every player writes down two possibilities for each category, on pieces of paper.
  3. The piles are shuffled. Each player draws a possibility for each category.

This effectively dictates your character for you: it’s definitely on the draconian end of the constraint spectrum, but in one blow it rids you of the time you usually spend agonising over exactly what your character will be like, and jolts you straight from your rut of “usual characters”. Handily, the reliance on group input means that your characters will still all fit the template of a “typical character” for the game, through the lens of your group. The last time we played Society, I ended up with a young lesbian Scottish travelling con-artist: never in a million years would I have made this character up of my own accord.

Interestingly, Society of Dreamers also suffers from lack of constraint in its conflict-resolution system, by which I mean: it doesn’t have one. I’ve seen sessions of this where someone fires a gun and suddenly either the director or the table have to decide if they hit the target. This is the sort of thing that can break you out of a scene and slow down the game, but Society is designed to evoke a very particular theme, and extraneous rules aren’t part of that theme. If I were to run it again, I would seriously consider getting some form of rule system to cover these situations though. It doesn’t need to be complex: it just needs to be some form of constraint that gives you an out every time something unexpected and uncertain happens, and lets play continue to flow.

Enter the Avenger

Enter the Avenger is a tiny delocalised-GM game about sorcery and revenge, and it relies on the players for a lot of creative input. Every scene occurs in a new location, with a new villain, so people are constantly introducing new characters. Any time where players have to introduce something new to the narrative, the author has handily provided a “jumping-off” list. For example, to start the scene, the protagonist addresses the narrator thus:

“In order to exact my due vengeance, I will travel to…”

  • “…the vice-pits of Abtsibea, where all sorts of unwholesome pleasures can be enjoyed for a high enough price.”
  • “…the Towers of Silence in Seebharim, where the dead are exposed to the birds of the sky and the heat of the sun.”
  • “…the Court of Fools in the city of Mezelith, where a new king rules each day and at nightfall is taken down to be sacrificed.”


…or any other wondrous place of the player’s invention.

Here the author has spent some time thinking up some nice evocative places with a bunch of hooks for the other players to use – but they’ve also given the player explicit permission to throw off the constraints when they don’t need them. I think it’s worth noting, as well, how all that mental effort determining “what is good for the story here?” has been transferred back in the cycle, from the player to the designer. Much like we saw with Apoc World’s Read a Sitch, the designer spends a lot of time during the design process, to remove the amount of decision time and effort required during the game proper.

To conclude

For me, a lot of this post has been exploratory. It’s about mapping the boundaries of constraint in gaming, and finding out: how do we use constraints? How have people used them in novel ways? Are there any patterns here?

Introducing constraints in a good way is, I think, hard work. The designer has to have a clear idea of the game as a whole, and how constraints on play can push towards that idea. Without that vision, and that work, constrains become just another mechanic or system that the players need to grapple with, to no good cause. Whether your constraints take the form of rules and tables or situational starting-off points, the goal is not to add enough of them to cover every situation, but instead to provide just enough well-thought-out constraints, to keep your players making awesome choices.

  1. That is, the rules governing dice, ability scores, XP, etc. 

  2. This idea of the move having to be well-designed - that is, the fact that the designer must ensure that those five questions are applicable to almost every situation - is something we’re going to see popping up again and again, incidentally.