How to Play Mafia or Werewolf With 6 Players

For Mafia or Werewolf with 6 players, my group used a variant we called doubles. Each player received one card from an alignment pile and one from a role pile. Only two Villager cards together gave a village player two lives, lost either through execution or a wolf attack. This meant more phases could be played, with useful information coming from an execution that failed to kill anyone.

The progressive rounds of player elimination help to build tension in the later stages of the game. Having at least two evil players who know each other is also part of the mystery for the village to solve: who is working together as a team? Doubles was our attempt to keep more phases and uncertainty around claims with fewer people.

This was the inspiration for the Enduring aura on werewolv.es. These are the rules for our moderated card variant; they do not describe a ready-made doubles mode on the site.

Mafia with 6 players: roles and cards

Shuffle these piles separately and deal each player one card from each. There are six active players, plus a separate moderator to handle the deal and hidden night actions.

The Seer (Sheriff in Mafia) checks a player's alignment at night. The Protector (Doctor in Mafia) guards a player against a night attack.

What the combinations mean

Either wolf card makes its holder evil. If the two wolf cards go to different people, there are two wolves who know each other. If they go to the same person, there is only one wolf, but the Seer's power is useless for finding them.

A werewolf Seer was useless as an active role because the wolves already knew their allies. What mattered was that the wolf knew it was safe to fake. Claiming Seer did not prove someone was village.

The hidden Lycan

One of the double Villagers is also a Lycan, selected at random and unknown to that player. They keep their two lives but appear as a wolf to the Seer. The village Seer and Protector are not eligible for that Lycan selection.

A wolf result could therefore be a wolf or the Lycan. A village result is not conclusive either, because both wolf cards might have gone to the same player.

Night actions, extra lives and reveals

During the day, players vote to execute someone; at night, the wolves choose an attack target. The version we played allowed wolves to attack on the first night. The Protector can protect themselves, but cannot protect the same player on consecutive nights. A wolf dealt Protector is also called during that night phase and can bluff being the village Protector. Whether their protection actually works is left to the moderator's discretion; settle that rule before playing.

Only double Villagers have a spare life. Losing it to a wolf attack is not announced to them. If the werewolves had targeted them the night before, they would not know until they swung in the gallows.

An execution that uses the spare life is public because the player survives. If the wolves already took that life, the execution kills them instead. We played no reveal: when someone finally died, their cards and alignment stayed hidden. A survived execution therefore supplied useful feedback that a death did not. Under these rules, it identified a double Villager, though that player might still be the hidden Lycan.

Extra lives do not change parity

Standard surviving-player parity applies. Two wolves against two villagers is a wolfpack win, however many spare lives those villagers have left.

For example, suppose a six-player deal gives the village a Seer, a Protector and two double Villagers. On the first night, the wolves attack one of the double Villagers, Alice, while the Protector guards someone else. Alice silently loses her spare life and all six players reach the morning.

If the village then executes the Seer, there are two wolves and three villagers left. The next night, a successful second attack on Alice kills her and gives the wolves two-versus-two parity. Protecting Alice would prevent that death; protecting elsewhere would not. The extra life bought a phase, but did not guarantee another chance after a wrong execution.

Adapting the setup for 7 or 8 players

To compare larger groups, the simulations extended the six-player piles as follows.

Werewolf with 7 players: card lists and balance

For seven-player Werewolf or Mafia using this variant, add one Villager to each six-player pile. There are seven active players, plus a moderator:

• Set A: six Villagers and one Werewolf.
• Set B: four Villagers, Seer, Protector and Werewolf.

Keep the same extra-life rules and one hidden Lycan among the double Villagers. This is the extrapolated deck used in the simulations, not a confirmed list from our group's games. When the wolf cards go to different players, two wolves face five village players. One village death therefore leaves two wolves against four villagers, giving more room before parity than at six.

Werewolf with 8 players: card lists and balance

For an eight-player Mafia or Werewolf game using the same approach, add two Villagers to each six-player pile. There are eight active players, plus a moderator:

• Set A: seven Villagers and one Werewolf.
• Set B: five Villagers, Seer, Protector and Werewolf.

Again, retain one hidden double-Villager Lycan and the existing extra-life rules. This larger deck is also an extrapolation. When the wolf cards go to different players, two wolves face six village players. One village death leaves two wolves against five villagers. That gives more room before parity, but the simulations are needed to assess what happens over the whole game.

Comparing Werewolf and Mafia setups for 6, 7 and 8 players

For this article, 36,000 games were simulated across the three player counts. The comparisons used exactly two wolves and one hidden double-Villager Lycan. At seven and eight players, a Villager was added to each pile for every extra player. Those larger lists are extrapolations of the six-player deal, rather than confirmed lists from our group's games.

The ranges show the results of two bot models, not confidence intervals.

One version could mistake a night without a death for proof that the protected player was village, overlooking a double Villager absorbing an attack. A second version suppressed those claims. The models disagreed substantially about removing the opening kill, including whether it helped village or wolves.

With normal opening attacks, seven players came closest to even in both models. Six favoured wolves and eight favoured village. None of these games ended before the first daytime decision. At six players, one opening village death leaves two wolves against three villagers, so the single attack cannot produce immediate parity.

These are bot outcomes, not expected win rates for your group. Each combination of player count, opening rule, bot model and deal shape ran 1,000 games. The results weight the possible deals by how often they occur, but exclude the case where both wolf cards land on one person. The bots also do not fully reproduce doubles role-claim knowledge or human discussion. The numbers are a reason to test seven with normal opening attacks, not proof that it is balanced. They do not establish that removing the first-night kill fixes six-player doubles.

A six-player idea: one wolf and a Bloodmoon Cult supporter

One suggestion for the six-player setup is to substitute one of the wolves for a Bloodmoon Cult member who knows who the wolf is, but the wolf does not know them. For a deal that would otherwise produce two wolves, this would leave one wolf, one Cult member and four village players.

Before revealing allies, the moderator would tell the Cult member which player is the wolf, without telling the wolf who their supporter is. That one-way knowledge is part of this proposed variation. The Cult member wants the wolf to win and can help through votes and arguments, but does not add another wolf attack. The wolf might even kill their own supporter.

There is also a numerical difference. Under the site's Bloodmoon Cult parity rules, the Cult member counts on the village side, despite sharing the wolf's win condition. One wolf, one Cult member and one villager is therefore not wolf parity: the count is one against two. Eliminating the wolf lets village win even if the Cult member is still alive.

This keeps two players working towards an evil win while changing what they know about each other. It gives the village a different relationship to uncover, and gives the wolf a reason to judge who might be helping them. It is an untested suggestion for this doubles setup. It was not included in the simulations above, so those win rates do not apply to it.

Can you play Mafia or Werewolf with 5 players?

We never played doubles with five players. A proposed adaptation removes one Villager from each six-player pile: Set A becomes four Villagers and a Werewolf; Set B becomes two Villagers, Seer, Protector and Werewolf, without the added Lycan rule.

With two wolves and three villagers, an unprevented first-night kill of a village Seer or Protector ends the game at parity. Extra lives on other players do not stop it. A separate 24,000-game simulation study also favoured wolves with or without the opening kill, though it shares the behavioural limitations above. That five-player proposal does not have our group's play history and should not be treated as a balanced recommendation.

Where to go next

For the underlying design problem, read why small Werewolf and Mafia games feel different. For alternatives to adapting Werewolf, return to Mafia and Werewolf with 5 or 6 players. If you are designing your own variation, use the setup balance guide to think through its information and claim space.