How do you create a fantasy faction?

A faction is a decision-making machine

The reason to invent one is so that the world reacts without you having to improvise on the spot. When the party burns the customs house, you should be able to ask each faction the same question (what do you do about this?) and get five different answers.

That only works if the faction has been defined by its behaviour rather than its aesthetics. A heraldic device, a colour scheme and a founding legend tell you nothing about what the group does on Tuesday. A goal, a constraint and a preferred method tell you everything.

Six questions that produce a usable faction

Answer them in order; each one constrains the next. Fifteen minutes is enough.

1. What does it want, specifically? Concrete enough to be achieved or lost. "Control the river tolls between the two bridges" is a goal. "Wealth" is a mood.
2. What is stopping it? The obstacle defines the story. A guild blocked by a rival guild behaves nothing like one blocked by its own oath.
3. Who joins, and why? The recruitment pitch, honestly stated. This is where a faction stops being an abstraction; someone had a reason to sign up last month.
4. What will it not do? The line it will not cross is more characterful than the ones it will. It also gives the party leverage.
5. What does it actually control? People, places, money, information, or a monopoly on something. If you cannot name the asset, the faction cannot be robbed of it.
6. What is its next move if nobody interferes? Write it down. This is the single most useful line on the page, because it turns the faction into a clock the party can run against.

Make it collide with something

A faction in isolation is inert. The interesting properties emerge from overlap: two groups that want the same asset, one group that depends on another it despises, a third that profits from the quarrel continuing. Two factions with an explicit relationship generate more play than five with none.

The cheapest way to get there is to define each new faction partly in terms of the ones you already have. Not "a thieves' guild", but "the people the Assize cannot prosecute, and why".

Avoid the tidy symmetry of a faction wheel where every group opposes exactly one other. Real organisations have lopsided relationships: grudges that only run one way, dependencies that only one side knows about, and those asymmetries are where a party finds purchase.

Test the faction before you use it

• Given a specific crisis, you can say what it does without inventing anything new.
• It wants something another faction also wants.
• A player could plausibly join it and have a reason to stay.
• It has something concrete to lose.
• Its next move happens whether or not the party shows up.