Conflict¶
This document is part of the Frontiers Design Documentation, a set written to teach instead of to instruct. Each entry takes one design problem, explains what that problem is in the wider language of tabletop design, surveys how games have solved it, and then shows what Frontiers does as one worked answer among many.
It is written for several readers at once:
- Designers using Frontiers who want to understand why it works the way it does.
- Designers using Frontiers who want to change it, and need to know what a change will cost.
- Designers building something else entirely, who want the design space mapped before they choose.
Frontiers is a toolkit, and a toolkit is only useful if you can see what each tool is for. None of this is a rule; the rules live in the Rulesets, and what follows is the reasoning, the history, and the alternatives.
There will be terminology here that is unfamiliar to casual players. Treat this documentation as advanced content, written for designers interested in building. It will not teach you every term or phrase it uses, but it will do its best to define them as they arrive.
Conflict sits at the front of this set with Scope & Genre because both of them ask what your game is, where the rest of the set asks how a given rule works, and for that reason both are worth settling before you so much as pick a die. The difference here is that this one does have a Frontiers answer, and most of that answer turns out to be about what the engine deliberately leaves empty.
What is Conflict¶
Conflict is what your game is actually about once it stops being easy.
Ask a table what conflict means in Wizards of the Coast's 5e SRD and you'll have an answer before you've finished the question: a fight. Nobody sat down and decided that. The fight is simply where the rules are, and a rulebook ends up telling you what it's about by where it chose to spend its pages, whether or not anyone meant it to.
Ask the same thing about Chaosium's Call of Cthulhu (no, I don't have a bias thank you very much) and the answer comes apart in your hands. A conflict there might be a fight, though the fight will most likely kill you, which makes it a bad plan rather than a goal. It might just as easily be a chase down a stairwell, two investigators who have stopped trusting each other, a locked room with a book inside, or the mystery itself, which resists for an entire campaign and can't be shot.
Those are two very different answers to a question most designers never notice they're answering, because conflict and combat are different words and a good chunk of the field is written as though they were the same one.
The question has three parts, and while plenty of systems don't answer all three out loud, they end up answering them regardless:
- What counts as conflict? Which situations in your game pick up mechanical weight the moment they turn tense, and which ones stay a conversation.
- What are the players trying to do? Winning is too vague to build anything on. The real goal might be defeating something, surviving it, arriving somewhere, understanding something, getting someone to concede, or just holding together, and whichever one it is decides what machinery you're on the hook for.
- Where does the uncertainty live? Whether the open question in a tense scene is finding a fact, reading it correctly, lasting long enough, or talking someone into giving something up.
Why Conflict matters¶
Whichever goal you land on decides what you have to build, and the building is the expensive part.
Defeating something needs a durability number and some way to reduce it, which is to say hit points and damage, and most of a combat system grows outward from there. Understanding something needs clues, a way to be wrong, and a reason the characters can't simply ask. Arriving needs distance, supply, and exhaustion. Getting someone to concede needs leverage and some way to lose standing. Four goals, four almost completely separate piles of work.
The awkward part is that building the wrong one doesn't announce itself. Nothing breaks. The rules just sit on the page unused while the pressure your game actually generates runs on improvisation instead. A system with forty pages of combat and one paragraph on investigation, running a campaign about investigation, still functions perfectly well; it has simply handed its GM the job of writing the missing half, and the GM usually doesn't notice that's what they're doing.
Flow of Play makes a related point about tempo, which is that the mode with the most content attached is the one a table remembers, no matter what the introduction claims. Point that same observation at content instead of tempo and it says your page counts are a statement of values, one that tables read accurately even when nobody meant to make one.
Experienced designers get caught by a subtler version of this. A game can have excellent rules for its central conflict and still have picked the wrong conflict, because the designer never thought to ask the question and built whatever the last game they read had built. It plays competently and feels like somebody else's game, and by the time anyone works out why, the numbers are already in.
When a scene in your game stops being easy, what are the players trying to do, and what is it that pushes back?
The Approaches¶
No answer here is correct. Unlike scale, though, conflict does have a dominant tradition, so most of these shapes read as departures from it, which says a lot more about the history of the hobby than it does about merit.
Seven shapes come up again and again. As with Scope, treat them as landmarks rather than a complete map: plenty of games run two of them at once, and plenty more sit somewhere between.
Conflict is combat¶
The default, and it is worth taking seriously. The opposition is an enemy with a statblock, the goal is to bring it to zero, and the subsystem built to do that is the most refined thing in the book. The 5e SRD lineage is the version nearly every table already knows, and Massif Press's Lancer is the modern high-water mark, splitting its mech combat off into something close to a complete tactical wargame while keeping narrative play deliberately light.
Its strengths get undersold a lot. Combat is the most legible conflict a game can offer: the stakes are obvious, the turn structure tells everyone when to pay attention, progress shows up on a number anyone can read, and decades of refinement hand you a working vocabulary before you've written a word. It's also the only shape that reliably gives every player something to do at a predictable moment, which is worth more at a real table than it sounds on paper.
The catch is that everything else inherits no structure at all, and the page count ends up teaching the table where the game lives. A group that showed up wanting intrigue will keep drifting back to the fight, because the fight is the part with rules attached.
Conflict is the mystery¶
The opposition is a fact somebody is hiding, and the goal is to know it. This territory has split three ways, which is what makes it the most instructive corner of the whole field.
Call of Cthulhu puts the uncertainty in discovery: you roll to find things, you can fail, and the resource getting spent is the investigator's mind rather than their health. Pelgrane Press's Trail of Cthulhu and the rest of the GUMSHOE family decided that was the wrong place to put it. A character with the relevant ability is simply handed the core clue, no roll involved, on the reasoning that a mystery which can stall on a failed perception check has a structural problem to solve, and no amount of retuning the difficulty reaches it. The uncertainty moves to interpretation, and to what the investigation costs you on the way. The Gauntlet's Brindlewood Bay goes further and removes the answer altogether: the GM doesn't know who did it either, clues get gathered as inert facts, and the players build a theory whose odds of being true scale with how many of those facts it manages to account for. On a success, the theory becomes what happened.
All three produce a conflict that force can't solve, which gives you a very different kind of tension than a fight does. All three also leave you to manufacture your own pacing, because an investigation has no natural clock, and the usual fix is to bolt one on afterward.
Free League's Vaesen keeps players out of the fight by a route worth stealing. Rather than discouraging violence through tone, it makes violence mechanically inert; the creature can only be banished by the right ritual, so investigation is the only way through. An option that does nothing holds up a lot better than asking people not to take it.
Conflict is the journey¶
The opposition is distance, weather, supply, and time, none of which has a will of its own. The goal is to arrive with enough left to matter.
Free League's The One Ring gives the Journey a system of its own sitting beside combat, with assigned roles, events drawn by region and season, and Fatigue laid on by nothing more than the length of the road. Cross enough ground and your character is worse at everything before an enemy has even turned up. Burning Wheel HQ's Torchbearer is the purest case of the lot: its turn counter advances on success as well as on failure, so caution and thoroughness spend the same scarce resource that fighting does, and time itself ends up as the antagonist. Yochai Gal's Cairn gets somewhere similar for much cheaper by making Fatigue take up an inventory slot, so your exhaustion crowds out the supplies that would have prevented it.
Attrition creates stakes without needing an enemy at all, so you get pressure in every scene for the price of a single subsystem. It also makes preparation interesting, because what you chose to carry becomes the question the game keeps asking you.
The price is bookkeeping, and quite a lot of it. Every one of these games depends on the table actually tracking the numbers, and a group that starts hand-waving supply has removed the game's only source of tension without realizing that's what they did.
Conflict is the other person¶
The opposition is someone with wants of their own, and the goal is to get them to give something up.
Burning Wheel HQ's Duel of Wits shows this can carry combat-grade machinery. An argument gets a pool standing in for hit points, both sides script their moves in secret volleys, and the winner has to accept a compromise scaled to how close the thing was. It almost never ends in an outright win, which is the point. Buried Without Ceremony's Monsterhearts uses Strings, countable units of leverage one character holds over another and spends to push them, so the live question at the table becomes who owes whom. Pelgrane Press's Hillfolk goes furthest of all, defining conflict as one person wanting an emotional concession from another and handing the decision to the player being asked instead of to the dice.
This makes other characters the most interesting obstacle in your game, and the content comes free, since every relationship is a conflict you didn't have to sit down and write. It also produces outcomes a fight can't, because a concession leaves both parties still present and still wanting things.
It does need buy-in to work. A mechanic that compels a character to feel something is pressing on a player's sense of ownership, and the games that do this well all spend real page count on consent and on the right to refuse. Burning Wheel's mandatory compromise is that same caution written as procedure instead of being offered as advice.
Conflict is the place¶
The opposition has no statblock and no will, and it can't be beaten in any useful sense of the word. The goal is to last.
Hedgemaze Press's Trophy reduces this to a single dial: Ruin, a counter from one to six that only ever rises in its darker mode, because there's no healing to be had. Every scene is one step closer to the forest finishing what it started, and the forest never needs statistics. Tuesday Knight Games' Mothership uses Stress, which builds from failure and from the environment and gets checked against rather than reduced, so lethality follows how long you've been exposed more than it follows what hit you.
The threat is unlimited and costs nothing to maintain. A place needs no stats, no scaling, and no balance pass, and the dread comes for free because no version of the scene lets the players feel safe.
In exchange, victory is off the table, which some groups will experience as the game refusing to be played with them. Designs in this shape tend to run short for exactly that reason, and a long campaign built on one needs something else for the table to want.
Several kinds, each with its own procedure¶
Some games decline to choose at all, building parallel subsystems and letting the fiction pick which one you're in.
Burning Wheel runs three of comparable weight: Fight for melee, Range and Cover for closing and withdrawing at distance, and Duel of Wits for argument. Evil Hat's Fate does the same job with a decision rule instead of three rulesets, sorting a scene into a Challenge, a Contest, or a Conflict by asking whether anyone opposes you and whether anyone gets hurt. Torchbearer takes a third route, keeping one engine and typing it by declared goal, so Kill, Convince, Flee, and Trick all share a structure with different skills attached and different starting numbers.
Nothing ends up second-class, and a table that would rather argue than fight gets rules of equal quality for arguing. Fate's version is the cheapest by a wide margin, since a taxonomy is far less work than three subsystems and gets you most of the same effect.
The trouble is volume. Each procedure has to be written, balanced, taught, and then actually remembered, and they all compete for the same session time, so one of them usually becomes the real game while the others gather dust.
Nothing pushes back¶
Included because it proves the question is a real one. Possum Creek Games' Wanderhome is built around travel and has no dice, no GM, and no resistance of any kind. Moving on produces new places and new people rather than hazards, and the only economy is tokens you earn by letting your character be vulnerable.
The game becomes about whatever the players feel like looking at, and no session can be derailed by going badly. For the register it's aiming at, opposition would only do damage.
Nothing supplies momentum, though, so the table has to. Without a group willing to make its own stakes, play drifts pleasantly along and then stops.
Rowan, Rook and Decard's Heart falls slightly outside the list, and is useful for exactly that reason. It has no combat subsystem at all: one roll, and Stress dealt to one of five named arenas (Blood, Mind, Echo, Fortune, and Supplies). Because its attrition is generic, every kind of conflict, physical, social and supernatural alike, can share one engine without the physical swallowing the rest. That answers the usual argument for giving combat its own rules, which is that combat needs attrition and nothing else does.
The trade sitting underneath all seven is precision against coverage. A procedure built for one kind of conflict resolves that kind very well and says nothing at all about the rest; one engine stretched across every kind treats them all alike, which is fair and a little flat. The more exactly you model the conflict your game is about, the more of the remaining play runs on improvisation. What you're really choosing is which parts to hand back to the table, and there's no wrong amount to hand back.
Conflict in Frontiers¶
The engine takes no position here either, and this time the refusal is built into the structure rather than left as an absence.
The Encounter is a tempo. Everything rests on that one decision, and Flow of Play is the document that actually makes it. An Encounter is the mode where time gets counted in seconds, because the situation turns on split-second choices. A brawl qualifies, and so does a cable giving way, a bomb counting down, or an argument in front of a crowd where every sentence closes a door behind you. Leaving violence out of the definition is deliberate: the moment a counted mode gets defined as combat, every other kind of conflict becomes a special case of it.
Combat is a Module. It ships as Basic Combat, outside the core, alongside Currency & Exchange and Undertaking Projects. Nothing in the core depends on it, and pulling it out leaves a working game behind. A system about courtroom argument and a system about boarding actions need very different combat rules, and the engine's job is to be the part those two share.
What the core does supply works for any form of conflict you like. The Gradient reports how far past the line a roll landed, which means as much for a negotiation or a day's march as it does for a strike. Favor expresses circumstance without caring what the circumstance is. ⊖ Fatigue is a cost that any kind of pressure can impose. None of them asks what sort of trouble you're in.
The caveat from Scope & Genre applies here too, and it shows up in the same place. The vocabulary leans physical: ⊗ Wounds, ⛊ Physical Defense, ⛉ Mental Defense, ✚ HP. A game whose conflicts are wholly social can rename every one of those and the arithmetic won't notice, but somebody does have to do the renaming, and the engine should admit it isn't starting from nowhere.
In place of an answer, the engine offers you three worked ones. The Starter Sets run on a single chassis and disagree completely about what a conflict even is.
d20 Fantasy¶
Conflict is a fight, and the game is comfortable saying so out loud. Something is in the way, it has intentions of its own, and the characters are strong enough to go through it. It's the d20's home register, and the one a new table needs explained to them least.
d100 Sci-Fi¶
Conflict is competing interest, and it takes more shapes here than in either of the other two sets. The crew are good at their jobs, and the trouble is usually that somebody else wants the same thing, or that the people who hired them were lying about something. Fights do happen, and so does hacking a system, finding a route through the vents, and talking your way past somebody who outranks you. It sits nearest the middle of the chart because the set asks for a bit of everything, and a fight is only one of the ways a job goes wrong.
d18 Wasteland¶
Conflict is the journey, and this set is the one that makes the case for the whole document. The setting is ruined, and the central difficulty is moving through it: what the distance costs you, what the weather does to you, what you failed to bring. Raiders are a complication on the road rather than the reason for it, and a session can carry its full weight with nobody fighting anybody. Set beside the Fantasy set on the same chassis, it's the strongest evidence in the project that conflict is something a designer picks.
Three answers, one engine, and no conflict rules in the core at all. The whole difference lies in what each set's content assumes trouble looks like.
Variants¶
Variants keep the philosophy intact: the Encounter stays a tempo, the form of conflict stays a matter of content, and the core stays neutral. All that changes is where you put the weight.
Give your central conflict a Module¶
This is the pattern the engine was built for in the first place. Decide what your game's conflicts are, write the procedures for that one thing, and leave everything else to the core.
A game about investigation gets a clue economy and a pacing clock. A game about travel gets supply, navigation, and wear. A game about standing gets leverage and reputation. Each time, you're doing what Basic Combat does, at the same layer, with the same freedom to be ignored by anyone who doesn't need it. It costs you a module's worth of writing and nothing else, and it's the cheapest route there is to a game that feels like it's about something.
Make a non-physical track the primary one¶
⊗ Wounds and ⊖ Fatigue are both five-step tracks that stay the same size no matter how far a character has advanced, so either one can carry your game's main pressure without you having to rebalance anything.
Running ⊖ Fatigue as the primary consequence suits journeys, sieges, long investigations, and anything else where the real question is endurance rather than injury. Characters wear down without getting hurt, and recovery turns into a pacing tool. ⊗ Wounds then has very little left to do, so either repoint it at something or say plainly in your text that your game doesn't use it.
Add a mode for it instead¶
Flow of Play lists a Travel or Social mode among its own variants, and that is the other place this decision can live. A mode gives your central conflict its own tempo rather than its own procedures: a lighter change, and sometimes the more accurate one.
It pays off when your conflict runs at a time scale the three existing modes handle awkwardly. When Exploration already covers it, skip it, because a fourth mode is competing for session time with three that already work.
Name it, even with no rules attached¶
The smallest useful change in this document. In your own text, write down what counts as a conflict in your game and what the players are trying to do about it.
No mechanics required. It costs you a paragraph and heads off the most common failure in this whole area, which is a table defaulting to the fight because nothing ever told them there was another option on the table.
Mix these freely. A survival game might run ⊖ Fatigue as its primary track and add a Travel mode on top; an intrigue game might write a leverage Module and never touch Basic Combat at all.
Alternatives¶
Alternatives replace the philosophy outright. Each one below changes what the engine assumes conflict is, and each reaches past this document, since Flow of Play and Attributes & Derivatives are both holding pieces of the current answer. They're listed roughly in order of how far they travel.
Pull combat into the core¶
Move Basic Combat inside the engine and treat the Encounter as a combat mode outright.
Most of the field does exactly this, and it's efficient for an obvious reason: if every game you intend to build is about fighting, then the separation is bureaucracy and nothing more. You get to tune Attributes, Defenses, and ability costs against a known combat baseline, which is a real advantage that the neutral version gives up. In return you take on precisely what the split exists to avoid. Non-combat conflict becomes a special case, abilities start getting written as working in or out of combat, and any designer building something else inherits an engine with an opinion already in it.
Parallel subsystems of equal weight¶
Build two or three full conflict procedures, in the manner of Burning Wheel: a fight, an argument, and a chase, each with its own structure and its own numbers.
Your game then supports more than one kind of pressure, and the table believes it, because the rules are visibly of equal quality. The price is the heaviest in this document. Each procedure needs writing, balancing, and teaching, they compete with each other for play time, and the engine's single resolution spine starts carrying structures that no longer much resemble one another. Expect to revisit Attributes & Derivatives while you're at it, since parallel subsystems usually end up wanting parallel durability numbers.
One roll resolves the conflict¶
Skip the exchanges and settle a whole conflict at once. Evil Hat's Agon does it by making the roll decide the heroes' standing against one another rather than wearing down the opposition, and that's the trick of it: take away the question of attrition and the exchange structure isn't needed any more. Tomkin Press's Ironsworn gets there a different way, marking progress through play and then cashing in the whole track with one roll, so the final odds are the sum of everything that came before them.
It's fast, it keeps conflicts from swallowing an entire session, and it scales to any fiction without new rules. What it gives up is the moment-to-moment tactical texture of an exchange, which is most of what players enjoy about combat in the first place. The Gradient also has to work a lot harder, since a single roll is now expressing a whole outcome.
Remove resistance entirely¶
No opposition, no attrition, no failure state: Wanderhome's shape.
For the right register this is the whole design, and it strips out most of the machinery: no durability numbers, no consequence tracks, arguably no dice. The cut does go deep, though. ⊗ Wounds, ⊖ Fatigue, both Defenses, and most of ✚ HP stop meaning anything, the Encounter has very little left to be a tempo for, and the Gradient loses its subject. You'd be using the engine for its vocabulary more than its mechanics, which is a perfectly legitimate thing to do as long as you know that's what you're doing.
Whichever of these you go with, write it down. Of everything in this set, conflict is the decision most often made by default, and it's the one a table notices fastest when nobody made it.