Track hard evidence first: who was on the mission that failed, who was near the body, who voted for what. Then let gut reads break the ties.
A plain guide to social deduction games like The Resistance, Werewolf and Among Us: why a clean mission clears nobody, how to count the possible spy pairs instead of chasing one suspect, and how to accuse the play rather than the person. With a tracker that shows what each mission result really proves.
Stop trying to read faces first. The players who get good at social deduction games track hard evidence: who was on the mission that failed, who was near the body, who voted for what. Then they let behaviour break the ties. And they remember the rule that trips up almost every newcomer: a clean result clears nobody, because the traitors are allowed to behave.
Social deduction games (also called hidden-role games) split the table into two teams, and only one of them knows who is who. A small informed side (spies, werewolves, impostors) knows its own members. The larger side has to work them out from what happens and what people say. The Resistance (for 5 to 10 players), Werewolf and its party-game cousin Mafia, and the online game Among Us all run on this idea. In the language this site uses for difficulty, the hard part is Competition: the difficulty is other people. But there is a quieter Systems load underneath it, and that part can be learned in one evening.
A "read" is a gut call about someone's behaviour: they went quiet, they defended a friend too fast, they laughed at the wrong moment. Reads are noisy: everyone is a bit nervous, and a newcomer looks guilty mostly because they are new. Hard evidence is different: it is a fact about the game that rules certain teams in or out, no matter how anyone acted.
The Resistance makes this unusually clean, so it is the best place to practise. With five players there are two spies, and the spies know each other. Each round a leader proposes a team, everyone votes on it, and the team members each secretly play a Success or Fail card. The cards are shuffled before they are revealed, so you learn how many Fails were played, not who played them. A single Fail sinks the mission. Here is what that information is actually worth.
Press "Example". On mission one you went with Ana and there was one Fail. You know you played Success, so Ana is a spy: that is certain, and the tool shows her in every pair left. On mission two you, Ben and Cleo came back clean. Now tick the box. Under the "spies always fail" assumption the tool clears Ben and Cleo and convicts Dev. Untick it, and Ben, Cleo and Dev are back to one chance in three each, which is the truth.
The rulebook lets a spy on a mission choose either card. So a clean mission tells you only that the spies on it, if any, chose to pass. Experienced spies often let an early mission through to look trustworthy. If you treat every clean mission as a clean bill of health, you will happily put a spy on the third mission.
Failed missions are where the hard information lives. One Fail on a two-person team that includes you is as good as a confession. One Fail on a three-person team that includes you puts at least one spy among the other two.
The habit that makes this click is to stop asking "who is a spy?" and ask "which pairs of spies still fit everything that has happened?" With five players there are 10 possible spy pairs. From your own seat you can rule out the 4 that include you, leaving 6. Each failed mission crosses some off. A player who appears in most of the surviving pairs is your top suspect; a player who appears in none is someone you can trust with the next team. You do not need the tool at a real table: two or three short lines on a scrap of paper ("M1: me + Ana, 1 fail") do the same job, if your group is fine with notes.
Votes are evidence too, just softer. In The Resistance everyone votes openly on each proposed team, and five rejected teams in a row in one round hands the spies the win. Who keeps approving teams that go on to fail, and who keeps blocking teams that would have been safe, is worth a line in your notes.
Every game in the family has its own version of hard evidence. In Werewolf it is mostly the record: who claimed which role, who voted for whom and, in groups that reveal a player's role when they are out, what they turned out to be. In Among Us, a meeting starts when someone reports a body or presses the Emergency Meeting button, then the group discusses and votes someone out. The hard evidence is location and timing: who was in that room, who came from that direction, who you saw doing a task. If the host has turned on Visual Tasks, some tasks play a visible animation, and watching someone complete one is about as close to proof of innocence as the game offers.
The same arithmetic tells you what not to do. If both spies end up on one mission and both play Fail, the table now knows the exact pair. If you play the only Fail on a two-person team, your partner on that mission knows it was you. Good spies think about what their card will let the others deduce, not just whether the mission fails.
Accusations are the game, but they should land on the play, not the person. Say "the Fail came from mission two, and you were on it", not "you're acting weird". If you are introducing the group to one of these games, the tips in how to teach a board game to new players apply: play one open practice round before anyone's role is secret.
For: anyone who has played one or two rounds of a hidden-role game, felt lost in the arguments, and wants something solid to say. Skip it if: your group plays these games purely for the chaos and the laughs. That is a fine way to play, and keeping notes may feel like homework.
Next step: at your next game of The Resistance, or a lobby of Among Us, keep a three-word note for every failed mission or body found, and before each vote ask which suspects that note actually rules out.
No: in The Resistance a spy may play a Success card, so a clean mission only tells you that any spies on it chose not to fail it.
Ask your group before you start; notes on public facts (teams, results, votes) hide nothing and make the game fairer for newer players.
Two, out of five players, and they know each other from the start.
The Resistance on video: the rules, a full game and the Avalon variantSee the clips ▸
The video game take on hidden roles: meetings start with a body report or the emergency button, and the evidence is who was where. Turn on Visual Tasks to give crewmates a way to prove themselves.
