Games Design and Mechanics

On Games, Part 2: Social Deduction Games

In this article, I give an overview of the social deduction game landscape and ground it in the market economy framework from Part 1 of this series.

Ankit Buddhiraju
The Ugly Monster

--

Previous article

Let’s dive deeper into the genre of game that is closest to my heart: social deduction games.

Mafia: The “First” Social Deduction Game

Some popular social deduction games

The release date of a particular game, while an indicator of when a game’s mechanics first captured the public imagination, is not definitive evidence that no version of that game existed prior to that date. For example, chess has existed since the 6th century AD, but modern rules around stalemate, pawn promotion, en passant, and castling — arguably some of the most important rules in the game — weren’t finalized or agreed upon until the 19th century.

So the disclaimer here is: I am not a historian.

What does seem to be clear is that before Mafia was created in 1986, the foundations had already been laid:

Social games and deduction games were already popular before social deduction games emerged as its own genre

Social deduction games are the marriage of the two — games where social interactions and deduction are both critical to playing the game. One could take the interpretation further — in social deduction games, there is some information to make some deductions, but not enough information to make game-winning deductions with certainty. The gap must be bridged through social interactions.

Mafia came onto the scene in 1986. It was created by Dimitry Davidoff, a psychology grad student at Moscow State University, and the rules were as follows:

  • There are two teams: the Mafia and the Villagers. Typically there are at at least twice as many Villagers as Mafia (e.g. in a 9-player game, 3 people would be on the Mafia team). The game is moderated by a narrator, who does not play.
  • At the start of the game, each player is privately told whether they are a Villager or a Mafia. Each Mafia member is told the identities of all the other Mafia members.
  • The game consists of two alternating phases: night and day.
  • During the night phase, everyone closes their eyes (they are “asleep”). All surviving Mafia members open their eyes (they “wake up”) and nonverbally agree on a Villager to eliminate from the game (“kill”). The narrator takes note of the Mafia’s choice.
  • During the day phase, everyone opens their eyes. The narrator announces who was just killed by the Mafia, and that player is eliminated.
  • The survivors then discuss who they think the Mafia might be. At the end of the phase, someone is democratically voted out of the group (e.g. by majority vote) and eliminated. The narrator will announce whether the person voted out was a Mafia or a Villager*. Everyone then closes their eyes and goes back to the night phase.
  • The game continues with night phases and day phases until all members of one team have been completely eliminated. If all the Mafia are eliminated, the Villagers win, and vice versa.

*in most variants

The game became immediately popular and spread throughout colleges in Russia, slowly making its way to Europe and eventually to other parts of the world. Andrew Plotkin popularized the game by rebranding the Mafia team as Werewolves in 1997, when the game reached the United States. Once the game became widespread, different groups began creating their own variants, some of which were spun off into their own games. Beyond plain old Mafia and plain old Villagers, special roles with special abilities were added to both teams, increasing the variety and volume of information and actions available to different players.

To this day, game designers frequently put Mafia in the list of “most culturally significant games”. Some countries like Russia, China, and now the US have proper Mafia clubs where you can go and play the game competitively with aficionados over a drink. Millions of people have played the game at least once, often in a social setting like a summer camp, a dinner party, or a company offsite; and millions still do, both in person and online.

The Innovations of Mafia

It is critical to understand what made Mafia different from the social games and the deduction games that preceded it:

  • Mafia demonstrated that it was possible to play a team game where you don’t know your team. We may take this for granted now, but previous hidden identity games either publicly revealed the members of the “evil” team upfront (e.g. the board game Scotland Yard) and designed mechanics to let them do their actions in secret; or there was only ever one member of the evil team, meaning there was no real team dynamic to speak of (the game felt more like a pure deduction game).
  • Mafia demonstrated that people create their own information. Mafia firmly put the “social” in “social deduction”. Previous hidden identity games used involved, information-obscuring game mechanics to give the evil team enough resources to counter the good team’s deductions. Mafia demonstrated that simply getting a group to eliminate individuals from the game creates enough perceived information to let players theorize about who the evildoers might be. Beyond the assignment of roles, the information in Mafia is generated entirely by the players.
  • Mafia consequently defined the paradigms for the entire social deduction game genre: 1) an informed minority against an uninformed majority, 2) roles assigned in secret, and 3) a game mechanic that allows the evil team to impede the progress of the good team in secret. Every social deduction game that came afterwards contained these key ingredients.

Understanding Mafia in the Market Economy Framework

In the previous article, I introduced the market economy framework for thinking about games, where every key ingredient of a game relates to the “currencies” that teams start off with:

  • Currencies measure the key quantities that teams start with, and that change throughout the game, so as to have a direct impact on the game’s outcome
  • Game mechanics specify how currency can be gained or lost through conversions and transfer
  • The win condition tells you what needs to be true about specific currencies for the game to end, and how the winner is determined

Let’s take a look at Mafia through this lens.

The two teams — Mafia and Villagers — start off with different amounts of different currencies. These are:

  • Power currencies: the power to vote, the power to kill (at night)
  • Influence currencies: influence, but you could use another word — basically, the ability to persuade people to vote a certain way and to believe you’re a Villager
  • Information currencies: information (the Mafia know the full state of the game but the Villagers only know their own roles)
The initial currencies of both teams in a game of Mafia

The Villagers start off with more players, and therefore more votes; as compensation, the Mafia starts off with more information and the power to kill. This asymmetry is what drives the entire game. The balance of the game lies in how well these currencies can be exchanged.

Different game mechanics allow for currency transfer and exchange:

  • During the night phase, the Mafia deplete the Village’s vote currency, but the Village’s information currency increases when the narrator announces who has died
  • During the day phase, all players compete for influence currency; some players have information which they try to convert into more influence
  • When someone is voted out at the end of the day phase, one team’s vote currency is depleted, and the Village’s information currency increases

The Mafia’s information never depletes because they start the game with complete information. Both team’s votes shrink throughout the game, but the Mafia can deplete the Villagers’ votes faster with their kill power, at the cost of giving the Villagers more information. The win condition is hit when the votes currency of one team is completely depleted, making it a race to the bottom.

Influence: The Currency that makes Mafia Mafia

In Mafia, influence is the meeting point of power and information — the intermediate, nebulous, contested currency where the game actually happens. Everyone tries to persuade the group to use their votes in a particular way, and there is only so much influence to go around because only so many people can be Villagers. The day phase is where influence gets transferred between players as they accuse and defend one another in the public sphere.

In most other types of games, you have to channel your personality through the currencies. We can talk about Federer being an “all-court player”, Djokovic being an “aggressive baseliner”, and Nadal being “the king of defense”. We can talk about the romantic, the classical, and the modern eras of chess, and the players who embodied them. Your personality affects the way in which you play the game, but it does not define it. You may certainly have preferences for how to play the game, but every dimension of every currency is ultimately endogenous to the game, allowing you to experiment with different playstyles if you so choose.

Mafia takes this to an extreme by blurring the line between the game and its players. In Mafia, you don’t channel your personality through the currencies: your personality is the currency. Half your influence in Mafia is precisely your ability to not be suspected of wrongdoing — a component that is hugely influenced by factors exogenous to the game. If your group deems you the kind of person who talks too little, talks too much, makes controversial jokes, doesn’t make enough jokes, smiles weirdly, has a reputation for being “good at games”, “screwed someone else over last time”, or whatever else it comes up with, your influence has been compromised. Mafia relies on players using flimsy heuristics to decide guilt in the absence of information.

This gets to the core of why Mafia resonates so deeply with the people who play it. Mafia, lacking in sophisticated power or information currencies by design, make the players — with all their hunches, biases, quirks, missteps, and all-too-human irrationalities — the focus of the game. Your personality is put on full display for others to dissect, and feelings carry over from game to game to create a rich and chaotic metagame in which past alliances are immortalized and past betrayals are seldom forgotten.

The influence currency takes the familiarity of human interaction and makes it strange and interesting. Mafia lives in its own special category as a celebration of the capacities within ourselves that we deem distinctly human — emotional intelligence, heuristic reasoning, and social calculus. It is a game whose finale is intimately tied to the personas of its protagonists. It is a game about you, as a person and thinker, trying to convince others of your intentions in the marketplace of ideas.

Critiques of Mafia

Critique 1: Player Elimination

In the previous article, we saw that there are three flavors of win condition: races to the top, races against time, and races to the bottom. I made the claim that the win condition of a game is not what gives a game its unique identity, because games with the same win condition can test very different skills.

However, in order to have a meaningful race to the bottom, you almost always need one core game mechanic: player elimination. In other words, not every player gets to participate for the entirety of the game. In Mafia, as in all elimination games … getting eliminated can suck.

Races to the bottom can be high-variance for individual players. Even when the game is a team game, the team may be consistently good at the game, but it’s hard to predict when an individual on that team will be eliminated. And even when a specific game is NOT a race to the bottom, meta-games like tournaments, brackets, leagues, or weekly episodes for that game are often structured as races to the bottom. A team either advances to the next round for winning enough games, or faces elimination. (Contrast this with race-against-time meta-games like a round robin or a Swiss system.)

In games where the mechanics allow individuals to directly eliminate each other through their own actions, races to the bottom can also be intensely personal. In other games with different win conditions, the only way to beat someone is to legitimately play better than them. But in a race to the bottom, my entire strategy could revolve around eliminating one specific person, even at the potential cost of my own victory.

In dodgeball, the entire team can choose to lob balls at the most athletic player on the other team. In Mafia, the Mafia can just kill off players they don’t like, or the group can vote out players who have a reputation for being good (or bad) at the game. In The Bachelor, a group of contestants can go out of their way to undermine one specific contestant in front of the suitor. Even though these strategies may not be “optimal”, they do exist and they do get implemented, often leading to resentment, broken friendships, and lifelong grudges.

In short, races to the bottom create high stakes for individuals, for whom the cost of failure is not getting to play the game anymore; and for audience members, who are often rooting for one player or one team. It’s no wonder that TV shows have a fascination for them…

TV shows love races to the bottom

While races to the bottom boost audience engagement, they are often a sticking point for gamers and game designers. What fun is a game if you don’t get to play for all of it? Game designers will go out of their way to put delay or resurrection game mechanics in races to the bottom, to push the inevitable elimination of most of the players further out:

  • Most variants of Mafia are played with special roles that can prevent, or delay, people from dying. In most variants, the Mafia members are also prohibited from killing before the first round of discussion has taken place.
  • Most variants of dodgeball allow you to bring back an eliminated player from your team if you catch a ball that was thrown at you by the opposing team.
  • Most variants of Freeze Tag (when you’re tagged, you are eliminated and must stand still) allow you to free tagged members of your team by running under their arms (Melting Candles) or under their legs (Stuck in the Mud).
  • TV shows like The Bachelor, Survivor, and America’s Got Talent often bring back eliminated players, either within the same season or across seasons.

Critique 2: Lying to people who lack information

While enthusiasts (like myself) will describe Mafia as a transformative game about reading people, convincing others, and coming out of your shell, its critics point to two unavoidable consequences of the game design:

  • The Mafia has to lie in order to win games. They not only have to lie about their own identities (which is a trivial sort of lie) — they also have to intentionally misconstrue innocuous, meaningless comments and gestures from other people as evidence of those people’s guilt in order to avoid suspicion themselves. They may even have to throw their Mafia teammates under the bus to establish their innocence in the eyes of the group.
  • The Villagers, as we’ve established, have no information to go on and have nothing “special” to do during the entire game besides participate in the day phase. This can make the exercise of playing the game feel pointless or arbitrary, even if it is winnable. In more complicated variants of Mafia, most players are still Villagers and start the game only knowing their own role (and nothing else). Villagers have to survive long enough in the game to actually start making inferences about who the Mafia might be.

These two criticisms stem from the same source: Mafia emphasizes the “social” in “social deduction”. The lack of information is its key feature. This turns off players who would prefer to start the game knowing literally anything.

With these critiques in mind, we turn to the broader landscape of social deduction games.

Many Social Deduction Games Try to “Fix” Mafia

Plotting some popular social deduction games across two axes: social vs. deduction (a function of how pivotal the influence currency is to a particular game), and the flavor of win condition.

One chart cannot do justice to the many subtle differences between these games, but at a glance we see that social deduction games creatively try to solve the issues of Mafia in different ways:

  • Keep player elimination, but give every player more information (Werewolf)
  • Keep player elimination, but soften its impact, and give every player more information and something “special” to do (Witchhunt, Blood on the Clocktower)
  • Keep player elimination, but give every player more resources so that they survive longer and have more agency over when they get eliminated (Bang!, Coup)
  • Effectively remove player elimination by making the game just one round long (One Night Werewolf)
  • Effectively remove player elimination by creating two concurrent halves of the game that players flit back and forth between (Two Rooms and a Boom)
  • Remove player elimination and disperse more information from individual rounds between groups of players (Avalon) or pairs of players (Secret Hitler)
  • Remove player information and give players lots of useful and extraneous information (Deception: Murder in Hong Kong)

One of the designers of Secret Hitler, Tommy Maranges, has two great articles about hidden information and the policy deck in that game. These articles explain the design team’s thought process around effectively fragmenting information, restoring some of the “deduction” aspect to “social deduction”.

Mafia emphasizes the influence currency above all else. Over time, its successors have put more weight on the power and information currencies by introducing game mechanics that give players more information and more to do (trending from social to deduction), while mitigating the stress of potential elimination (trending from races-to-the-bottom to races-to-the-top).

Before next time…

Games are similar when they share similar currencies. Regardless of their specific win conditions or game mechanics, all social deduction games share two characteristics:

  • They emphasize how your personality is perceived in a group in their influence currency.
  • They start with an imbalance between the power and information currencies for the two main teams (“our team has more players” vs. “our team has more information”).

People who don’t enjoy social deduction games can find good justification — the stress of being eliminated, the unwanted scrutiny on one’s behavior, and the need to lie. People who do enjoy them enjoy the particular flavor of influence that they offer: the emotional rollercoaster of reading, and being read by, other people.

In the next article, I’ll dive into games of perfect information.

Next article

--

--

Ankit Buddhiraju
The Ugly Monster

MBA ‘21 at Stanford. I like uncovering how different fields of knowledge draw on the same universal principles.