Image illustrating the risks which have just been described in the text.

I’m going to make an assumption at this point that as a GM, you’ve got a pretty solid grasp on success and failure. You understand that failure should not stop play, and that you should only turn to the dice to determine success and failure when both outcomes are interesting and fun to play. So what we’re going to talk about that other stuff that surrounds success and failure – specifically we’re going to talk about risk.

In this context, the risks of a situation are the things that could obviously go wrong, but which do not necessarily make success more or less likely. For example, if a character wants to kick down a door1, there is a risk that it will make enough noise alert the guards. This risk has no impact on the action, but it has a profound impact on the situation.

When games take risks into account, they often simply fold them into the difficulty of the roll, and assume success to mean the risks have been addressed, and draw upon the risks in the case of failure. More nuanced games, fold the risks into ideas like partial success or success with consequence, so there’s a middle tier of results between success and failure.

While I can’t pretend to cover the entire range of possible risks in play in one post, I would suggest th

at there are a handful of risk types which you will see over and over again. They are:

Cost: The most straightforward of risks is as simple as a price, usually in the form of lost resources.

Harm: Equally classic, winning but taking an injury is an iconic example of success at a cost.

Revelation: The acting character reveals some piece of information, whether it be a clue or their location.

Confusion: The acting character conveys something other than intended, creating opportunities for upset, bad timing, offense or more dangerous misunderstanding.

Waste: Functionally, this is akin to cost, but where cost is intentionally, waste is the result of misapplication of resources.

Ineffectiveness: Hitting the target may not mean knocking it down. A success without follow through may end up reaping limited (or no) rewards.

Spillover: Alternately, sometimes the problem comes from too much effort – fragile things break, pieces no longer fit, people are annoyed and other results of overkill can all be problems.

Delay: Sometimes things will just take longer than intended.


(This list is almost certainly not comprehensive, and I’m 100% open to suggestions for additions.)

Now, this list is very useful to a GM who is looking at a situation and thinking “ok, what might the risks be here?”. Running down a simple checklist is an easy prompt to the self to consider potential options before the dice hit the table. But, critically, almost no situation will call for all risks. If there’s no one watching, there’s no much risk of a reveal. If there’s no hurry, then delay is not much of a risk. This is not a problem – that different situations have different risks is a feature because risks drive player behavior.

That is to say, risks impact how characters approach a problem, and they provide an avenue of action and play that grows naturally out of the success and failure that are already afoot in your game. Two mechanically identical rolls can feel drastically different when presented with different risk profiles. That benefit is so profound that we’re starting today with laying the groundwork on thinking of risks as something distinct from difficulties.

Risks and Approaches

(This bit is fairly FAE Specific, but some of it can be more broadly re-used).

I love Fate Accellerated, but I think it’s generally understood that there are times when the question of which approach to use becomes more of an exercise in mechanics than fiction, which rather misses the point of using FAE in the first place. I’ve written about a few other ways to approach the problem of how to make it matter which approach someone choose, but I think the real secret sauce lies in risks, with one simple trick of perspective:

Approaches are less about success and failure than they are about mitigating risk.

Ok, maybe that sounds weird, but work with me here. Straight success is not hard to pull off in FAE, even with a ‘weak’ approach. Difficulties aren’t super high2, and aspects provide a lot of extra oomph when needed. But for all that, there is still incentive to explain why everything you do it clever and therefor gets your +3 bonus.

But suppose you looked at that list of risks as the inverse of approaches. Remove cost and harm – they’re always on the table as the situation demands it – but the rest line up suspiciously well.

If you are not quick then you risk delay.

If you are not clever, you risk waste

If you are not forceful, you risk ineffectiveness

If you are not flashy, you risk miscommunication

If you are not sneaky, you risk revelation

If you are not careful, you risk overkill

That is to say, the choice of approach can be a reasonable response to risk. In play, this means that the right choice of approach can nullify a risk.

Illustrated set of which approaches counter which risks

Ok, that’s all well and good in theory, how does that look in play?

So, mechanically, when the GM looks at a situation in play, it should have one or more risks (if there are no risks, then definitely question why there’s a roll at all), and set difficulty, with the reminder that 0 is a totally reasonable difficulty. Then add the following 2 twists: First, every 2 points over the difficulty can cancel out a risk (if it makes sense). This makes for a sort of proxy difficulty increase with automatic success-with-consequence. Second, the approach chosen cancels out any appropriate risk.

To go back to the door example: Finn needs to get through a door to escape pursuit. It’s locked, and the guards are in pursuit, so it’s time to kick it open. It’s not a super robust door, and the GM is comfortable with a difficulty of 1, so she does a quick audit of potential risks:

Cost or Harm aren’t really in play directly, but they’re always on the table when things go pear shaped.

Waste isn’t much of a concern. There are no points for neatness in door kicking.

Delay on the other hand is a problem. If this takes too long, the guards may catch up.

Miscommunication isn’t really a concern, since ideally there’s no audience.

Revelation is borderline – the GM could say that one of the risks is that the guards will know which door Finn went out. However, he’s going to break down a door, so they’ll probably be able to figure it out however the roll goes, so the question is more whether they’ll find him soon enough to matter. From that perspective, this shades into the territory we’re already covering with the risk of delay, so the GM lets this one slide.

Overkill is almost certainly not a problem

Ineffectiveness, on the other hand, really would be. He cannot afford to be dainty here. But despite that, this merits a little thought too – the consequence of ineffectiveness is also that the guards catch up, so is it really that different? Wouldn’t ineffectiveness really map to failure in this case? Those are reasonable concerns, but they also need to be balanced against the sensibilities of the moment and the fact that the GM has already been generous about Revelation, and this feels right. However, double dipping on the guards catching up is unfair, so instead she considers the door not quite breaking all the way and him having to squeeze though, probably leaving some loot behind.3

So with that in mind, the GM figures the situation has risks of delay and ineffectiveness one top of the +1 difficulty. If he succeeds, he’ll get out through the door, but there’s a risk that the guards will be in hot pursuit if he takes too long, or he may have to leave some loot behind if he can’t kick the door all the way open.

In terms of pure math, this suggests a fairly large number of options:

If Finn tries to be quick (nullifying delay):

  • On less than 1, he fails, and has some guards to fight
  • 1-2: He manages to squeeze out the door but leaves some loot behind.
  • 3+: He kicks the door open dramatically and runs out onto the street.

If Finn tries to be forceful (nullifying ineffective):

  • On less than 1, he fails, and has some guards to fight
  • 1-2: He kicks the door open, but the guards catch up and the chase continues out onto the street.
  • 3+: He kicks the door open dramatically and runs out onto the street.

If Finn tries some other approach.

  • On less than 1, he fails, and has some guards to fight
  • 1-2: He squeezes through the door, dropping loot, and it pursued by the guards.
  • 3-4: The GM makes a quick judgement call based on which next step seems more fun, or if Finn’s description suggests a particular direction, and Finn either drops some loot, or is pursued.
  • 5+: He kicks the door open dramatically and runs out onto the street.

That looks complicated, but in practice, it’s pretty simple and logical. And critically, it makes the choice of approach meaningful. This is a situation where a Quick or Forceful character will have an opportunity to shine, but success is still equally within reach of all characters.

And, critically, there is still plenty of room for creativity and problem solving. If a player has a clever way to mitigate or transform a risk, or use an approach in an unexpected way, then awesome! That’s a good thing! The goal here is not to penalize “wrong” choices, but rather to give weight to the choices made.

Risks and Success

But wait, you might say: What if avoiding a risk is implicit in the action the character wants to take? What if I want to sprint, or sneak or do something else where triggering the risk would equate to failure?

The answer, counterintuitively, is that it changes nothing, except that it clearly communicates the approach that you want to use in this situation, and in doing so, may call into question the necessity of the roll.

That is, if there is only one risk (say, getting spotted or not), then that risk is obviated by using a Sneaky approach, at which point, why are you rolling? The answer might be “more or less because rolling is fun” in which case I refer you to the next section, but ideally it’s because “Oh, right, there should be more going on for this roll than this simple binary – what other risks might be in play?”4

All Risk, All The Time

Ok, dirty truth. I have occasionally found myself in situations where I’ve called for a roll and I’m not really prepared for failure. I should know better, but sometimes the situations just feels like a roll is the right call at the moment, and I might need to fake my way out.

Risks are a great tool in this situation because, frankly, I can drop a 0 difficulty and assume success, with the question being one of risk. See, these situations almost always have multiple risk vectors in play simultaneously, and that is what the instinct to roll is responding too. I’m not really looking for success or failure, I’m just looking for how well things stay under control. “Failure” in this situation means all the consequences come home to roost, even the one the approach was supposed to mitigate.

Communicating Risks

There might be a temptation to list off risks as bullet points so players can respond to them, but I would recommend against it. The categories of risk we’ve listed are a shorthand, not an actual description – they are for your convenience as a GM, as a placeholder for the actual fictional risks you will describe (or not describe in some cases) to your players.

As a general best practice, use risks as cues for describing the situation. You don’t need to elaborate on each risk, but when you think of descriptive elements for the scene, take a moment to think about each risk and see if it might contribute to the description.

If you do start literally laying out risks as Mechanical constructs, I’ll be curious how that goes for you, though.

Risks vs. Consequences

This is a bit of bonus content. If a risk is something that might happen, a consequence is something that will happen as a result of action. Beyond that, they are structurally very similar, and once you have gotten the hang of thinking about risks, you can apply the same sort of thinking to consequences to use one of the most powerful tools of scenario design out there.

That is, if you want a really solid session, one trick for doing it is to present a single, simple problem or task which the players can absolutely accomplish, but which has numerous dire consequences. The adventure then becomes a matter of identifying the consequences and figuring out how to nullify or redirect them before performing the main task.

Which is to say: It’s how to design a heist.

Icons used are largely from game-icons.net

  1. Because by the law of RPG examples, it must be a door. ↩︎
  2. In fact, they can feel “too low” to the GM at times, which can lead to some undesirable behaviors. Adopting the risk model to FAE allows the GM to embrace low difficulty numbers, because there’s still plenty of room for low difficulty being accompanied by plenty of complicating risks. ↩︎
  3. This would technically be a cost, which is fine. You can absolutely borrow across risk types when it makes sense, and both cost and harm are sort of generic resources. ↩︎
  4. Interestingly, this reveals the real problem with one of the great RPG bugbears, Stealth. We have been trained by skill lists to think that “sneaking” is an action in its own right, but really it’s just moving with intent. ↩︎

19 thoughts on “Risks

    1. Erik Dewhurst

      Thank you so much for this article.
      It started a discussion between myself (DM) and my players about the fail forward approach.

      We run a podcast (Dungeons and Tangents), so we had said discussion in our podcast.

      My biggest takeaways were: A) don’t make your player roll if there’s no risk involved. B) Being aware of what could is at risk makes it easier to come up with consequences to “failure”.

  1. Matt Popke

    Candidate for another risk:

    Unexpected Outcome: The character succeeds, but the results are not what they intended.

    Yes, you intimidated that guard, but you didn’t realize he was related to the local magistrate who now thinks you and your party are bullies, not to be trusted.

    Sure, you cracked that safe, but what you find inside are sensitive documents that clear your intended blackmail target rather than incriminate them, and now you have to decide if you’ll retrieve them for your employer or not.

    There’s a risk that using success with complications as a way to introduce plot twists will feel manipulative or make the players feel like you’ve pulled the rug out from under them so it’s definitely something to use sparingly (and probably only for minor plot points rather than major ones). But while running an Edge of the Empire campaign for a little over a year, I found myself resorting to it on a couple of occassions. Those damnable despair symbols kept popping up at the worst possible times. Eventually you run out of clever ways to succeed with major complications. I started prepping my adventures with optional plot twists built into them at critical points because I knew I was going to get at least one awkward despair symbol per session.

    The odds of getting those rolls increases over time as the characters level up, too. The math on that dice system is incredibly swingy with high level PCs and high level challenges. The odds of just success or success with some complications shrinks over time in favor of more extreme dice pool results (failures with triumphs became pretty common too).

    1. Rob Donoghue Post author

      This does an excellent job at articulating something I couldn’t quite put my finger on. I had been pondering the possibility of what I’d been calling “Surprise” risks, but pulling on that thread revealed an entirely different ball of yarn, so I quietly shut that tab and walked away. 🙂 But you’re right, it’s a fantastic tool.

  2. Loz

    Good tool. I could easily see this in the inner sides of a GM’s Screen (And since I’ve recently ordered a a customizable GM’s screen… 🙂 )

  3. Alpo

    Also applicable to PbtA. Not every miss means harm, just look at the approach followed by the player and you have the most likely or appropriate cost. Good stuff!

  4. Dave Maple

    How does this compare with the four actions of normal Fate Core/Accelerate Actions?

    Breaking a door would be an Overcome Action which you would normally succeed at a cost on a tie and even if you screw up the roll badly you elect to succeed at major cost.

    So you might say kicking a door in while being chased is an fair task, +2

    On less than 2, he fails, and has some guards to fight or succeeds but the guards are “On your tail” (they get a boost on the next chase roll) and he gets a “bruised shoulder”.

    On a roll of 2, he succeeds but the guards are “On your tail”.

    On a roll of 3 you get out the door.

    On a roll of 5 you get out the door and “Hit the ground running” giving a boost to the next chase roll.

    Does this replace that system so you now can actually fail on a low roll, succeed but with a risk (kind of like succeed at cost) at lot of the time, and pass completely on high roll but with no additional benefit like you use to get. How does it work with Create Advantage/Attack or Defend rolls?

    Generally I find the choice of approach meaningful anyway, they could Sneaky try and pick the lock, but that’s going to take time. They Carefully pry it open if they had the right tools or perhaps they Cleverly remember the guard using the access code before, so that has the advantage of not damaging the door so the guard might not notice you went this way. I kind of like the idea but I think if you are going to add complexity, there has to be a noticeable gain. Not sure what you are gaining adding this to a system like Fate that already makes a GM think about multiple possible outcomes.

    I think an example of it applied to something other than Fate would help.

  5. Random_Phobosis

    Here’s one more potential risk for your system:
    Path is closed. Whatever approach you used to get what you wanted, it’s not applicable in this situation anymore. The target may hide in building, impossible to hit from your sniping position across the street, eyewitness might get too scared to tell anything, the barkeeper closes the tavern for tonight, so no more rumors.
    This may be similar to cost and waste, but instead of supporting the same action with different resources, it actually allows using the same resources, but only if you come up with another way to make them useful. Should help dissuading players from spamming their best moves.

    As for the system itself, I’ve tried to decouple success from risks by building a resolution system on top of Fate, which curiously was very similar to the one described above. I’ve made two separate axes and called them Effect (do you do what you wanted to do) and Cost (what do you have to pay for that). The system was meant to give players a choice between succeeding at a considerable cost or cutting losses and accepting lesser effect. Basically it was PbtA 7-9 all the time, but the players get to choose, which 7-9 result they get.

    What I found out though, is that having a table with various conversion costs of successes to actual consequences became cumbersome. Even worse was the feeling that instead of choosing between drastically different courses of actions players “go shopping” and try to optimize their shopping cart of optimal outcomes.

    After wrestling with the system for some time, I came up with following solution.
    To do away with “convert number of successes to descriptive quality of success” step, I replaced dice with a little deck of cards. One had 12 cards for various degrees of success-failure spectrum, the other had low cost-terrible cost spectrum cards. Each card has a word, so you don’t have to calculate or reference anything.
    When the character attempts a risk, the player draws a card from each deck. Each pair of Effect + Cost becomes one of the distinct paths that can be taken, and players may spend Fate Points to draw more pairs. This removed “shopping list” feeling.

    Then I merged two stack of cards into one stack of two-sided cards, which also somewhat balanced randomness of the outcomes. Eventually I stole idea of using rough Risky/Controlled/Desperate positions from Blades instead of exact numerical margins, and came up with rules for Controlled and Desperate draws.

    The resolution process turned out to be quite slow, so it shouldn’t be used for trivial tasks. But for dramatic moments it works pretty well.

    1. Ron Stanley

      Hey, Random_Phobosis. The one-page RPG Ghost|Echo by John Harper has a very simple mechanic that is similar but not as granular as your method. You roll two dice. The player assigns one die to Danger and one die to Goal. Low rolls fail the Goal and bring on the Danger. High rolls succeed at the Goal and avoid the Danger. So if you roll high on one die and low on the other… you have an ugly choice to make. Do you want to fail and avoid the danger, or succeed and succumb? It’s a simple, brutal, and wonderful system, plays kind of like Die Hard meets Blade Runner.

  6. Ps

    Based on this article I noticed a correlation to the various attributes I use in my game and came up with this table.

    Inadequate (Brawn)
    There is a chance the hero might be physically thwarted, overwhelmed, overpowered or outmatched. Basically hes not physically up to the task. “He was as useless as titts on a bull.”

    Excessive (Finesse)
    There is a chance the hero may cause unintended collateral damage to people, possessions or places. “He Was like a Bull in a China Shop.”

    Stressful (Resolve)
    There is a chance the hero might suffer fear, panic, uncertainty or hesitancy. “A Coward dies many times before their death.”

    Confusing (Presence)
    There is a chance the hero might miscommunicate, misrepresent or cause unwanted offence. “He couldn’t help but put his foot in his mouth.”

    Inefficient (Intellect)
    There is a chance of hero’s actions of decisions having unintended or unforeseen consequences, Side effects or be generally wasteful. “He couldn’t see the forest for the trees.”

    Cost (Cunning & Presence): There is a chance of financial cost, social or reputational cost.

    Harm (Brawn & Resolve): There is a chance of physical harm or personal injury.

    Delay (Intellect & Finesse): There is a chance the Hero might be delayed in reaching an objective.

      1. Loz NEWMAN

        Minor point, Fabio:
        Many = countable
        Much = uncountable (e.g. abstract things )
        Mnemonic: Did you have much luck? Hos many Euros did you win?

        1. Antonio

          What’s the ettiquette on correcting a years-old correction? Whatever, here i go:

          Fábio’s error doesn’t seem to be misunderstanding of many/much, it’s that “people” sounds like it should be uncountable and abstract.

          If I say “persons” you obviously understand what i mean, even if you would probably be tempted to correct me to “people”. But portuguese does use both, “muitas pessoas” and “muita gente”, and while “people” doesn’t sound all that similar to “gente”, it is very clearly “the one that’s not the plural of person”, so it’s not unreasonable to assume the same distinction would carry through.

          That’s what i would assume happened here.

  7. Pingback: Wofür stehen die Turbo-Fate-Methoden? - FateRpg.de

  8. Sami Liukkonen

    I think this is a great way to think about this, and a great way to include risks to make every check matter and change things.

    However, I wonder if this would apply to Attack or Defend, or other opposed checks?

  9. Pingback: The Court Procedure from Lavender Hack // Take on Rules

Leave a Reply

Your email address will not be published. Required fields are marked *