How to Balance a Game of Werewolf or Mafia

Ensuring Balance is Important for Fun

Like most people, you probably like winning games. However, would you like to win all the time? You may be tempted to answer "yes", but would you eventually become bored if there was no challenge for you any more?

There is research to suggest that satisfaction comes from engaging in an activity that is challenging yet still within the grasp of your expertise. In a game context, this means there has to be a real chance of losing. Continually beating a three year old at chess is not really going to feel like an achievement for an adult.

Equally, however, there is little joy in continually being battered by someone that eclipses your skill. After the novelty of playing her had worn off, you are very unlikely to enjoy the prospect of Serena Williams consistently destroying you in a game of tennis.

Games that are too easy or too hard simply are not enjoyable. There is a gradient for this, of course, and some players will prefer more demanding roles and setups, whereas some players will prefer gentler ones. But in general, if one side has very little realistic chance, or if the outcome feels mostly predetermined, the game is worse for it.

A Dirty Little Secret About Game Enjoyment

There is a trap that a lot of moderators fall into when they start thinking about balance: they assume that if the village and the wolves each win about half of the time, the setup must be fair and therefore fun. That sounds sensible, but it misses something important about how people actually experience games.

Loss aversion hurts more than winning feels good. People generally feel the sting of a loss more sharply than the pleasure of an equivalent win. Games are no different.

In Werewolf and Mafia this matters even more because the factions are uneven in size. If you balance a setup so the wolves win fifty percent of the time and the village win fifty percent of the time, that does not mean half the players go home happy.

In fact, when the wolves win, usually two or three players feel very pleased with themselves and a much larger number feel the pain of having been lied to, outmanoeuvred, or dragged to the gallows. When the village wins, a much larger team gets to enjoy that win together. So from a pure "how did the room feel at the end?" point of view, identical faction win rates can produce very different emotional outcomes.

This is one reason I do not treat 50/50 faction balance as some sacred target. In plenty of groups, especially newer ones, a setup where the village wins a little more often will actually produce a better overall experience. Not because the village morally deserves that, but because there are simply more people sat on that side of the table.

Why 50/50 Win Rates Can Still Feel Bad

A setup can be fair on paper and still feel rubbish to play. If the wolves win by steamrolling the table, and the village spend the whole game confused and powerless, the fact that the spreadsheet says "balanced" is not much comfort to the eight people who just had a miserable evening.

This is why setup balance is not just about who won. It is about whether both sides felt that their decisions mattered. It is about whether losing players still had satisfying moments. It is about whether the result felt earned rather than arbitrary.

Good setups do not merely distribute wins. They distribute agency, information, tension, and payoff.

Letting People Feel Like They Still Won Something

Because loss aversion is real, one of the easiest ways to improve a setup without necessarily changing the raw numbers too much is to give players chances to feel like they achieved something even in a faction loss.

For the village, that might mean they correctly solved one wolf even if they lost to the second. It might mean they pieced together what the setup was doing. It might mean they forced the wolves into a desperate fakeclaim, or made the evil team work very hard for the win.

For wolves it can be similar too. Even in a loss, players often still enjoy the game if they pulled off a believable lie, manipulated a day well, or got away with something outrageous for far longer than they had any right to.

This is one reason reveal-on-death, partial information roles, and dramatic claims can all improve enjoyment when used well. They create little payoffs along the way. Players feel like the game is giving them something to work with, even if the final result goes against them.

Plenty of people can lose well and still enjoy themselves, so the aim is not to protect players from ever feeling bad about a defeat. What you can do is make sure a losing player still feels they took part in a game with texture, clues, turning points, and moments they can be proud of.

The Most Basic Variant - Mountainous 4:1 or 5:1 Villager:Werewolf

The easiest setup to explain is one that only contains villagers and werewolves, or town and mafia if that is your theme. No role madness, no special mechanics, no nonsense.

In this kind of game, reveal-on-death is often used as a feedback mechanism for players to get some concrete information as the game progresses. Without it, there is very little anchor for players to base their decisions on.

Ironically, although this kind of setup only includes two role types it may actually be harder for newer players to get to grips with, because they do not know where to start with their accusations. When playing in person, people will often begin by looking at body language, or drawing suspicion from who would be most likely to target the player that was eliminated during the night.

This is the game in its purest form. There is little to be learned from the game mechanics themselves and everything must be inferred from social reads, vote movement, and the narratives players are building during the day. There is no mechanical fakeclaim for the wolves to keep track of, so this kind of setup is favourable for players that are new to playing evil because they do not need to construct a believable role story at any point. They need to be smooth and reasonably consistent in how they present themselves.

The village is likely going to execute the wrong people during the first few rounds as it starts to find its feet. In order to compensate for this, the ratio of villagers to werewolves is advised to be a little higher than in role-heavy games. For this kind of setup you usually want around four to five villagers per werewolf.

You may need to tweak those numbers as your group develops. If the wolves are getting demolished every game then perhaps give them a little more help, but do not overcompensate. If the village are never getting close, give the town another body and try again.

Why Roles Change the Maths

Once you start adding roles, the game stops being purely about raw numbers. Roles add information, fakeclaim space, mechanical constraints, and rescue chances. They also change how easy it is for the evil team to hide.

This is where some moderators go wrong. They develop an opinion that roles equal fun and start trying to cram as many as possible into the setup. That can very quickly create an oppressive game for the players who wind up evil.

Evil players need a place to hide. If your setup is made up of several one-of roles, anyone claiming a role that is already genuinely in the game is immediately caught in a 50/50 conflict, and that tends to end very badly for the evil team mathematically.

So when you add a role, do not just ask what exciting thing it does. Ask what it does to the claim space. Ask what sort of intel it puts into the game. Ask whether it creates real choices, or merely narrows the game into a small number of mechanically correct lines.

If you want that point in full, read why evil players need somewhere to hide.

Too Many Mechanics Can Be Worse Than Too Few

If a game has a lot of players who are newer to the roles available on werewolv.es, I typically try to keep the number of different mechanics down. This is not because simple has to mean boring. Quite the opposite.

Elegant roles are often simple roles that enable lots of different scenarios to play out. Simple and complicated are not really the point on their own, though. A complicated role is not automatically bad, and a simple role is not automatically good. The real question is whether the role creates interesting decisions and worthwhile scenarios.

For example, if I am putting a lot of role-view intel into a game through seers or adjudicators, I might avoid visiting intel like stalkers or harlots entirely. I prefer to keep to families of mechanics in games for newer players, so the number of interactions they need to consider stays manageable.

A player learning one family of information roles can start to build good habits. A player who is expected to juggle role views, visit tracking, protection layers, redirections, and edge case interactions in their first few games is much more likely to get overwhelmed and make random decisions.

Balance Is About Decision Quality, Not Just Outcomes

One of the best tests for a setup is not merely "who won?" but "did the important decisions feel meaningful?"

If the village lost, did they still have enough information to form sensible worlds? If the wolves lost, did they still have believable lies available to them? If a player made a good read, could it actually matter, or was the setup so clogged with intel anyway that it did not really matter?

I tend to distrust situations where there are too many "this is the correct way to play, no matter what" moments. If a setup contains too many of those, either the game is not very interesting, or the role design is doing too much of the thinking for the players.

Interesting games let players be clever. They also let players be wrong for understandable reasons. That is part of what gives the game texture.

How To Judge Whether a Setup Was Bad or Just Had a Rough Run

I prefer to avoid swingy setups for the most part, but sometimes swing creeps in without you quite realising. That does not mean every ugly game was caused by a bad setup.

When a game goes badly, I look at the random elements first. Usually the earlier night deaths are down to a more random choice from the killers, and as the game progresses their decisions become more targeted. If the early deaths had gone differently, would that have materially changed the game? How a game plays out is not how every run of that setup would play out. I try to imagine different outcomes.

This is where it helps to distinguish between an imbalanced setup and a poorly designed setup. They might be different things. A setup can be broadly balanced in win rate terms and still be badly designed because it feels rotten to play. Equally, a setup can be reasonably designed and still produce a daft, lopsided game because several random events lined up in a nasty way.

If you want that distinction unpacked properly, read swing, imbalance, and bad design in Werewolf setups.

Listen To Complaints, But Not All Complaints Equally

Post-game feedback matters, but not all feedback should carry the same weight.

If a player who usually does not get salty is voicing objections, I will take that with more consideration. If it is a player who always moans when they lose, and they cannot offer solid reasons as to why they think the game was imbalanced, I usually ignore it and move on.

You are not looking for raw volume of complaints. You are looking for good reasons.

A Sensible Starting Point

Traditional setups of one seer and one or two protection roles have been a staple in many communities for years for a reason. They are often my go-to example setup when explaining the game, and they are also often the argument I use against passive villager play: the seer might be dead already.

Information in Werewolf does not only come from what a role can reveal at night. It also comes from people spotting things, reconstructing narratives, and forming collective judgements on who is lying. A balanced setup still needs the village to participate in that process, not sit around waiting for a power role to save them.

Final Thought

The point of balancing a game is not to create a sterile 50/50 spreadsheet result. It is to create a game where both sides have room to play, where the losing side still feels they were in a proper contest, and where the players leave with stories rather than just irritation.

If your setup creates hard choices, believable lies, interesting claims, and a decent chance for the village to feel they achieved something even in defeat, you are probably on the right track.