Tabletop roleplaying games are generally about shared storytelling, and stories are generally pretty boring if the story ends because the protagonists don't find the clue, don't get past the locked door, or don't survive a battle. Failure needs to continue the story by leading to something else interesting, being only a partial failure, or being success at a cost.
I will detail some of the concepts and approaches used by different systems, including several open source alternative roleplaying games, that can be incorporated into your game, even if the specific mechanics are not.
A related concept is the treatment of character death, which is usually the ultimate failure.
Some systems, like early Dungeons & Dragons, or revival games like Dungeon Crawl Classics, treat the adventuring group more like a wargaming troop, where when one character dies the next steps up to continue as the protagonist. Other modern systems have more explicit success at a cost mechanics or alternative ways to handle death.
The discussion is organised by concept, with examples given from different systems, including Fate, Dungeon World (Powered by the Apocalypse), Blades in the Dark, and Gumshoe.
Success and failure
Many systems, irrespective of the dice used, reduce to a binary system. While the statistical distribution of the dice rolls may differ, for a given situation (with given modifiers), the outcome -- success or failure -- is binary. (For a longer discussion of this see 3d6 is not less swingy thand d20.)
The question then is what to do in a situation where the dice result is a failure?
Some example situations, used in the discussion, and a corresponding unwanted story ending:
| Example situation | Unwanted story ending |
|---|---|
| finding a necessary clue | if the clue is not found, the characters have nothing to do |
| opening a locked door that leads to the next section | if the door cannot be opened, the characters cannot progress |
| the final exchange of blows in a battle | if the characters are dead, they cannot continue the story |
Common structured approaches include:
- Alternative mechanics
- No dead ends
- Success at a cost
As well as structured approaches, these situations can also be resoled by deus ex machina, or GM fiat, where the GM simply decides on a successful outcome, although this can sometimes seem overly contrived.
Note that these examples are only relevant where the situation is story ending. If the locked door was to a treasure room that was only incidental to the main plot line, then it is not disastrous if the door remains locked (but is rather boring if they players never get to see what was behind it.)
Alternative mechanics
Some game systems use alternative mechanics where the unwanted story ending is not possible, or not relevant.
Clues are always found
Mystery focused systems, such as Gumshoe, have a mechanic where base clues are always found by the character with the relevant investigative skill (no roll needed). Additional points may sometimes be spent for more clue details.
The interesting part of a mystery is not finding the clues, but how they are interpreted and the conclusions drawn from the clues.
A similar approach can be used in almost any game. When a clue is necessary to progress the plot, then it (or an alternative) is always found by the characters, either by a character with the relevant trained skill, or a contest between characters with the best roll determining who finds the clue.
Death is trivial
Some games avoid death being a dead end by making it trivial or impossible, such as in Paranoia when your character dies a new clone turns up. With games like Dungeons & Dragons where there is resurrection magic, death can be trivial at higher levels. See the section below on character death for further discussion.
No dead ends
Whatever the outcome of the dice, it would be very boring if that ended the story. Many systems encourage you to only roll when there is something at stake, or to make sure failure is interesting.
Explicit stakes
Some systems use explicit stakes for conflict resolution that are set before each roll, for example Dogs in the Vineyard. All systems have stakes, but often the stake is whether the task is completed, not whether the reason (or goal) is achieved.
Example: You may want to open a safe (the task) in order to get to a secret clue inside (the why). The implicit stake in many systems is the task; if you succeed then the task is completed, and the safe is opened; but the goal may not have been achieved, e.g. if the clue is not in the safe.
Most explicit stakes systems focus on making the why the actual stake, i.e. succeed and you will get the clue. The discussion of stakes, and tasks vs conflicts, is complex (see Conflict Resolution vs Task Resolution, fourth article down).
Explicit stakes systems also often include what happens on failure, which makes sure that failure is interesting up front, before you roll.
Example of bad failure stakes: Player: "I open the same", GM: "Why?", Player: "To find the clue where to go next", GM: "Okay, but if you fail you don't find the clue". Better failure stakes might be "if you fail, the safe is damaged and they will know someone was here", which avoids the dead end.
Picking the stakes that are interesting is still a subjective and manual process, and has the downside that only half of them will ever be used.
Still, adding explicit stakes to your existing game can be a useful tool to avoid dead ends (an old discussion is here: https://forum.rpg.net/index.php?threads/rant-i-houserule-explicitly-setting-stakes-into-every-game-i-run.315006/).
Adding explicit stakes can also be used to make physical conflicts more interesting than simply a fight to the death, but changing to explicit alternative outcomes like "you are captured" or "they drive you off".
Story branching
In Mouse Guard, conflict has a explicit story branching approach:
- Determine the skill, challenge factors, any help or other modifiers, and then roll.
- If your roll succeeds, the challenge is overcome, and your goal achieved. END.
- If your roll fails, then either:
- A new challenge (that you need to overcome) gets in the way to your goal. Return to step 1; or
- The goal is achieved, but at a cost (usual a status, e.g. Tired, Hungry, Angry, etc). END.
This can continue with as many new challenges as needed, but ultimately will end with either a rolled success, or success at a cost. There are no dead ends.
Example: You are trying to pick lock, but fail the roll. The GM decides you are unable to pick the lock, but guards turn up and you notice the leader has a ring of keys on his belt. You try and fight the guards to get the keys, but fail that roll. The GM decides you manage to knock out the leader and get the keys, but are Tired after the fight.
I like the story branching approach as it can be applied to most systems, and does not require explicit up front setting of stakes.
It can be difficult to improvise a meaningful new challenge that involves and alternative path to the goal (the keyring on the guards belt). A simple fallback is to have the task succeed and then a separate new challenge, e.g. on a fail you successfully pick the lock (maybe taking a long time) but then the guard turn up.
If you can't think of a new challenge, success at a cost is an easy alternative, e.g. on a fail you successfully pick the lock but are Angry because it took a long time.
This doesn't work so well with arbitrary off-plot actions, e.g. after describing a forest a characters wants to search for edible mushrooms. If it is not really connect to the plot, then the outcome of "you don't find any" may be fine.
The difference is probably best explained by Blades in the Dark concept of Action Rolls vs Fortune Rolls. Action rolls are to accomplish a goal, like opening the door. A Fortune roll is simply a tool the GM can use to disclaim decision making -- rather than deciding if the character finds the mushrooms or not, they use a random roll to decide, but there is no other consequence of failure.
Success at a cost
The most widely implemented, and most varied, is where the goal is still achieved (the story continues), but there is a cost.
Last person standing
Most traditional combat systems, e.g. Basic Role-Playing/Open Basic or Dungeons and Dragons, are a fight to the death, with each individual attack resolved, and the last team standing the winners.
The complexity of a battle almost always boils down to success at a cost.
Having the die is an unwanted ending of the story, so the characters eventually win every battle, with the variable being how much it cost them to win. Resolving battles one attack at a time can be a lot of fun for those who enjoy tactics and strategy, and demonstrating skill in the game system rules. The better you do, the lower the cost (less resources) are used.
In theory, resolving one attack at a time also gives a chance to abandon the action (retreat), although this requires failure (death) to be likely (which is an unwanted ending), and players often don't want to voluntarily back down, so this is rare. Some game systems also make it difficult to accomplish through tasks.
Some game systems make outcomes over than a fight to the death difficult. Portraying stories where you succeed in driving off the enemy, fight your way past them, or capture them; or failures where you are driven back or captured (but not killed) are difficult in many systems.
Example: An ambush situation where the ambushers goal is to capture the characters, and the characters goal is to drive off the ambushers, can be difficult to accomplish in a traditional fight to the death based system.
If you want to add these alternative outcomes to a traditional combat system, you may need to look at adopting explicit stakes (see above) and possibly add some alternative mechanics to support this.
This could be something like the Quick Combat rules for Savage Worlds, which is more abstract than tactical combat and could easily handle win and loss situations such as "drive them off" or "get captured".
Separate success and cost mechanics
In some systems, such as the Cortex family (for example, Firefly), and also the Genesys system (used in Star Wars), there are separate results for success/failure and cost/opportunities.
For example, in Cortex Prime you might roll a total of 15 and beat the opposing total of 12, so succeed, but you may have also rolled 2 ones (hitches) which the GM can turn into a complication. Genesys is similar in that Success/Failure and Advantage/Threat are separate results.
This creates a range of results that are success at a cost: In Cortex you can be taken out if you fail, but also taken out by complications even though you succeed.
By itself this can still have outright failure as the outcome, however there may be other rules to manage this. For example in Cortex Prime outcomes should always change the status quo, i.e. you attack either succeeds and damages the enemy (possibly at a cost), or the situation gets worse. It is never just a miss.
Cortex Prime is strongly based on TV tropes, so complete failure means you may be taken out of the current scene, but doesn't take you out of the story.
Success at a cost as a result
Systems such as Fate have an explicit result, when there is a tie, that is success at a cost. Due to the low number of discrete outcomes ties are fairly common (around 20% of the time).
A tie can also mean a partial success, if that makes sense for the scenario. This is kind of a success at a cost, where the cost is that you don't fully achieve what you wanted. You need to be careful with this, to make you don't negate the actual success; often it is better to keep the cost separate.
Fate still has the possibility of outright fail, but even then the rules are such that this could still mean success at a serious cost, for those situations where success is necessary to not get stuck in a dead end (such as finding a needed clue, or opening a needed locked doorway).
Deciding on what the cost, or partial result, is still requires some improvisation, but there is some guidance in the different action types in Fate Core, or general approaches in Fate Condensed. For example Fate features strong use of aspects, so it is easy to make a cost be a negative aspect or boost to the opponent (or stress or damage).
Success at a cost even appears in systems such as Dungeons & Dragons, as a tool in the Dungeon Masters Guide (page 242) to make things more interesting: when an action misses by only one or two, you can rule that it succeeds but at some cost, e.g. your sword gets stuck after the attack.
Aside: Is separate partial success result useful?
In the story branching approach, each roll is either a success or success with a consequence; there is no outright failure.
Similar with the examples such as finding a clue or opening a locked door, where the eventual success result is required for the story to continue (the variable being how many challenges it takes, and what the cost is). This is also reflected in the Fate failure option for success at a serious cost.
From this point of view a critical action only needs two outcomes:
- Success
- Success at a cost
Adding a separate dice roll result for partial success changes this by adding a third option in between, but doesn't change the overall nature:
- Success
- Success at a minor cost
- Success at a cost
In this view success has a very broad meaning of keeping the story going, even if the direction is different, i.e. it simply means no dead ends. E.g. Getting the keys from the guard after knocking them out is still a success at getting through the door, even though you failed to pick the lock.
So, a separate option for partial success can make things more varied, but is not necessary. To some degree adding an explicit success at a cost complicates things, and you still have total failure that you sometimes need to be failure at a serious cost. Fate recognises this as an option, but in the Blades in the Dark rules a poor result on a gather information check gets nothing. This is a problem if the information was necessary, and where you want the quality of information to vary based on the fortune roll, then maybe add a rule that a poor results still gets any minimum necessary information (which was probably the intention).
Explicit options for partial success
Systems such as Dungeon World (Powered by the Apocalypse) have a very explicit list of options for partial successes or success at a cost that make it very easy to use, and add more variety, expanding the list of options even further.
For example on the Volley action, as well as 10+ success, and 6- failure, if you roll 7-9 you choose one cost from the three options (giving total 5 different outcomes):
- you end up in danger,
- reduced damage, or
- consume ammunition.
Note that rolls in Powered by the Apocalypse serve double duty as both the character's action and the enemy's response, e.g. in the above 'you end up in danger' could be the enemy advancing to attack. This is explicit in actions like hack and slash where for the success at a cost option you both achieve your result (damage the enemy) and take damage from them.
A similar system is Blades in the Dark (Forged in the Dark). This list of options is not so specific per action but generic based on the position. For example a partial success (4/5) for a risky roll means you succeed but you may suffer one or more consequences out of harm, complication, reduced effect, or ending up in a desperate position.
Although both systems add more structure to the success at a cost results, there is still the situation of outright failure (1-3), which needs additional techniques to avoid unwanted dead ends, although the structure of both is very narrative making it easy to include such techniques.
Blades in the Dark also makes a distinction between action rolls, which may have consequences (including harm), and fortune rolls, such as gathering information where the random result determines how much information you learn, but consequences don't directly apply.
A final note on character death
Character death is a complex topic.
Many people are familiar with the basic approach, taken in games like Dungeons and Dragons, where once characters accumulate a certain amount of damage, usually from the random rolls of combat, they die. If you are only familiar with that one system, then a concept like characters never die may seem like breaking the rules.
But there are many different systems with different approaches, and sometimes the rules are that characters can't die:
- In Paranoia, a sci-fi comedy, characters regularly die, only to be replaced by their clone a few moments later.
- In Firefly, and other Cortex games, like a TV show, characters can be taken out of a scene, but not killed. You have to explicitly make death one of the stakes for it to occur.
- In Pendragon, a generational game, each session is a year long, and when a character dies (if not in game, then from old age) you play their descendants.
- In Eclipse Phase, a near future setting, character's are backed up before going on a mission, so they can be reloaded into new bodies if needed.
- Old school games, and revival systems like Dungeon Crawl Classics may simply use large adventuring parties, when a player simply continues with an alternative character.
- In horror systems, such as Cthulhu, character death is commonly expected, with players simply creating new investigators (maybe investigating the suspicious deaths of the previous group).
- In Kiss Her Before the World Ends (a fairly obscure indie game that ran at a local convention, Go Play Brisbane) there is a fixed end of the world (i.e. you die) and the game is about what you do in the time left.
- In Dogs in the Vineyard you can accumulate damage during the game, but don't check if you die of the wounds until the end of the session.
- In systems like Blades in the Dark you are encouraged to have multiple characters in the crew, as a character will often be out of play due to injury, in jail, or retired due to trauma.
- Ars Magica is an older system where you alternate playing a main wizard character and followers or groups of followers (like a troop of guards, all with a similar personality), so you also always have an alternative ready.
In one of my favourites, Polaris: Chivalric Tragedy at Utmost North, eventually either characters die, turn into a demon, or the world ends (it is a tragedy), however for the first half of the game they cannot die (stakes cannot include the above until characters are veterans).
Even in Dungeons and Dragons, once you get to a certain level permanent character death is almost impossible with the ready availability of resurrection magic.
As you can see, there is a wide variety of ways to handle character death, from it never happening (characters can't die) through to it always happening (the game always ends in tragedy). You can even adapt the concept you want to a different game system.
One important thing is that player expectations are set correctly, e.g. A new player may assume their character is like a TV protagonist and can fail, but not actually die, and it could create a social misunderstanding if they were playing something like Dungeon Crawl Classics and get killed in the first half hour.
The other important thing, for this article, is that they way death is handled should not be a dead end for the story being told, any more than failing a check to spot a needed clue or failing a check to open a necessary door. While this can always be handled by GM fiat, e.g. "you all awake in a forest grove, having being saved by friendly druids that passed by", it may be better if you have an explicit way to handle it up front.
For example Savage Worlds has an explicit optional rule "Heroes Never Die"; and it's not just falling off a cliff and you can't find the body; an extreme example that comes to mind is Darth Maul got cut in half on screen, and still came back.
Also the death of a single character does not necessarily end the overall story; in fact sometimes the death of a character may be just a cost paid as part of an overall success of the adventuring party. That is why many systems where characters can't die have an exception of unless the player allows
Note that it is still good to have something for the player to do when, or while, their character is removed from the story.