How to Balance a Game of Werewolf or Mafia

Balance Is About Fun, Not Only Fairness

When people talk about balance, they often mean fairness. That matters, of course, but it is not the whole point. Balance is also about fun.

Games become disengaging when the level of challenge is wrong. If one side is heavily favoured, the result starts to feel predetermined. If a game is too easy, it becomes dull. If it is too hard, it becomes frustrating.

A good Werewolf setup should feel like a proper contest. Both sides should have a real chance, and the players should feel that their decisions can matter. That is what keeps people switched on and wanting another game afterwards.

Start From a Baseline You Understand

You do not have to build every setup from absolute scratch. In fact, that is often a good way to make a mess.

It is usually better to start from a baseline you already understand. That might be a pure mountainous setup with no roles at all, or it might be a setup your group has played many times and which has settled into being roughly fair over repeated games.

The important thing is not whether the baseline is simple or role-heavy. The important thing is that you understand what is already doing the balancing work. What information is available? How much feedback do players get? How easy is it for evil to hide? How many miskills can the village survive?

Once you know that, you can start making changes deliberately. If you add something, ask what it strengthens. If you remove something, ask what it was quietly supporting before. Good setup design is often less about inventing clever parts and more about understanding what needs to be compensated for once you start meddling.

Feedback and Direct Intel Are Both Information

One of the easiest mistakes in setup design is to think only direct intel roles count as information. They do not.

Feedback is information as well. Reveal on death is feedback. A claim collapsing because of later events is feedback. A protection working, or clearly not working, can be feedback. Players learning that one of the worlds they were considering can no longer be true is feedback too.

This matters because the village's access to information is not just about whether there is a Seer in the setup. It is about the total amount of reliable information the game gives them over time.

As a general rule, feedback and direct intel checks should be inversely proportional. If you give the village strong, clean feedback, you usually need less direct intel on top of that. If you take feedback away, you often need to compensate elsewhere, either with more direct intel, more villagers, or a simpler game state.

This is one of the reasons ratios shift between different styles of setup. A village that gets very little trustworthy information early will usually need more room to be wrong. A village that has cleaner ways to narrow the world down can function with less raw numerical advantage.

Add a Small Amount of Intel

Mountainous is probably the easiest kind of Werewolf setup to balance, because the ratio is doing most of the work. Once you start adding roles, the balancing problem becomes less clean.

A small amount of intel can give players more to work with, but it also introduces more swing. A strong Seer, Protector, or similar role can have a large impact on the game, especially at lower player counts. That does not make these setups bad, but it does mean they are no longer balanced by ratio alone.

This is where you need to think more carefully about what sort of information the village is getting, how reliable it is, and how much room the evil team still has to manoeuvre. Once roles enter the game, you are not simply adding flavour. You are changing the shape of the setup.

The ratio of village players to werewolves can often come down a little once reliable intel is added, because the village no longer has to do all of its work through pressure and execution feedback alone. But that does not automatically make the setup easier to balance. In many ways, it makes it harder.

Add More Roles, and the Balance Changes

Once you keep adding roles, you are no longer merely adjusting faction power. You are changing how the game can be solved, what lies are believable, and how much room each side has to operate.

Roles add direct power, of course, but they also add claim space, rescue chances, mechanical pressure, and hidden constraints. A role does not only matter because of what it can do at night. It matters because of what it does to the rest of the table's reasoning.

This is where setup design often starts to go wrong. Moderators can fall into the trap of thinking roles equal fun, and so keep adding more of them. But more roles do not automatically make a game richer. Quite often they just make it tighter, noisier, or more oppressive for the side that has less room to hide.

Evil players need somewhere to hide. If the setup is full of one-of roles, very narrow claim space, or too many mechanics that quickly force hard contradictions, the evil team can end up feeling mechanically cornered rather than challenged in an interesting way.

So when you add a role, do not only ask whether the role itself is interesting. Ask what it does to the whole setup. Does it create worthwhile decisions? Does it give the village too much certainty too quickly? Does it leave evil with believable worlds to live in? Once roles start piling up, those questions matter at least as much as the role's headline ability.

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

More Complexity Does Not Automatically Mean More Fun

Complexity is not a problem in itself. It can bring novelty, and novelty can be good. A more complicated setup can create situations that a simpler one never could.

But complexity for its own sake does not necessarily create a fun game. The real question is whether the added moving parts produce interesting decisions and worthwhile scenarios, or whether they merely make the game harder to parse.

This matters even more when the players are new to the mechanics involved. A setup that is perfectly manageable for experienced players can become an unreadable mess for newer ones, not because the setup is evil, but because they do not yet know which interactions matter and which can be ignored.

This is one reason I tend to think in families of mechanics. If I am already using a lot of role-view intel, I may avoid piling on visiting intel as well. If players are still getting used to one kind of information, I do not necessarily want them juggling three others at the same time. That is not because complexity is bad. It is because there is a difference between depth and overload.

I have a bit of a horseshoe theory of Werewolf setups here. In mountainous, you have almost nothing apart from people's narratives and world view. In setups with a small amount of intel, you can often test sensible worlds quite cleanly. But in very complicated setups, with intel, counter-intel, kill power, protection and everything in between, you can end up back where you started. You are no longer even fully sure whether your own intel means what you think it means, so once again you are leaning heavily on narratives and world view.

The difference is that in mountainous the uncertainty is clean, whereas in a bloated setup the uncertainty is muddy. That does not mean complicated setups cannot be good. They can. But the complexity has to earn its keep.

This matters especially in low-player-count Werewolf games, where a single extra role can distort the entire setup much more dramatically than it would in a larger lobby. If you are designing for five or six players, start even simpler than you think you need to.

How To Tell If the Balance Is Actually Good

Once you have built the setup, the obvious question is whether the balance is actually any good. This is where a lot of people reach for faction win rate alone, but that is only part of the picture.

A setup can produce respectable results over time and still not be especially good to play. Evil might technically win often enough, but spend those games feeling mechanically cornered because the claim space is too cramped. Or the setup may produce too many situations where the "correct" line is fairly obvious, so the game feels solved rather than played.

One of the best tests is whether the important decisions actually felt meaningful. Could good reads matter? Could the village form sensible worlds from what they had? Could evil still tell believable lies? Did the players have room to be clever, or did the setup do too much of the thinking for them?

I also think it helps to distinguish between swing, imbalance, and bad design, because they are not the same thing. A setup can be broadly fair in win rate terms and still be badly designed because it feels rotten to play. Equally, a reasonably designed setup can produce a daft result because several random events line up in a nasty way.

So when you look back at a game, do not only ask who won. Ask whether the setup gave both sides room to play, whether the key decisions had weight, and whether the result felt like it came from the players more than from the setup collapsing under its own structure.

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

If In Doubt, Start Simpler

If you are not sure whether a setup works, the safest answer is usually not to add another mechanic and hope for the best. It is usually to simplify.

That does not mean every game should be bare-bones mountainous. It means that if you are struggling to judge what is doing the balancing work, there is a good chance the setup is already becoming harder to reason about than it needs to be.

This is why known setups with a lot of play history are so useful. They give you something solid to work from. You can make one change, watch what it does, and then decide what needs compensating for. That is much easier than inventing half a dozen moving parts at once and trying to guess which one broke the game.

Traditional setups with one Seer and one or two protection roles have been staples in many communities for years for a reason. They are not the only good way to build a game, but they are a sensible grounding point. They give players something to work with without immediately disappearing into a swamp of edge cases.

And even in those setups, the village still has to play properly. The Seer might already be dead. Information in Werewolf does not only come from what a role reveals 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 produce a sterile 50/50 spreadsheet result. It is to create a game where both sides have room to play, where the information structure makes sense, and where the important decisions feel like they matter.

If your setup creates meaningful pressure, believable lies, and interesting decisions without collapsing into either obvious lines or complete mush, you are probably on the right track.