This post contains a video, which you can also view here. To support more videos like this, head to patreon.com/rebecca!
As my patrons know, each month I do a special livestream where we chat and I answer any questions they have. Last month, one of the Michaels asked about the game Werewolf, linking to a video hfe had seen a while back and asking if it was true that the game proved that “an uninformed majority will always lose a battle of information against an informed minority.”
He apologized for not asking a better question, but in fact that question was my absolute favorite kind: the kind that sends me into a research frenzy, where I learn some really interesting things that I never would have even thought to look into.
I’ll also admit that I liked the question because I like games, including Werewolf, and I LOVE the reality tv show “Traitors.” That show has been around for a few years now and there are different versions in different countries (starting with the Netherlands), but the US version is now on its third season and literally the only reason I started watching it in its first season is because it’s hosted by Alan Cumming, along with his otherworldly outfits and his adorable little rescue dog, Lala. I mean, really, come on.
But I kept watching as soon as I realized the show was just a new take on the game of Werewolf, which itself is a reskinned version of a game called “Mafia.”
If you aren’t familiar with any of these games, allow me to describe the simplest version: let’s say you have a group of eleven people. One person is chosen to be the all-knowing moderator or narrator. They hand each of the ten remaining players a card. Eight cards say “villager” on them, and two cards say “mafioso.” The game begins “at night,” when everyone closes their eyes. The moderator instructs the mafiosos to open their eyes and acknowledge one another. The two then silently choose a villager to “kill.” They close their eyes and the moderator announces the sun has come up and reveals who has died. The remaining players then discuss who they think the mafiosos are. When the moderator announces that the sun is setting, the players vote for one person to kill. The remaining players close their eyes, and the game continues with the mafiosos making their kill.
The game ends when either the villagers successfully kill both mafiosos, or the mafiosos manage to kill all but two villagers, at which point there would be no way for the villagers to vote to have them killed.
And that’s the basic gameplay, though in “Werewolf” you have villagers and werewolves instead of mafiosos, and you can also play with a “detective” who is able to get additional information, or a “little girl” who can keep her eyes open to witness the killers at night, or a “medic” who can protect one player each day, and on and on and on.
But the central premise always remains: an uninformed majority versus an informed minority.
Back to Michael’s initial question: was the video he saw true? Does the informed minority always win? I’m going to play a clip of the video now and I will give you a content warning for this incoming jump scare: the person who makes the claim is, in fact, Graham Linehan. The TV writer who was rich and successful and had a family but he gave it all up to hate trans people. Literally, he got so obsessed with hating trans people that his wife divorced him.
So yeah, Linehan brought up Werewolf because in his mind, trans and nonbinary people are the “informed minority” who are manipulating the general public. Trans and nonbinary people, who are at higher risk of being assaulted, murdered, denied healthcare, et cetera et cetera. Sigh.
With that, you may not be shocked to learn that Linehan is wrong, about everything. Let’s get into it.
First of all, the game was not, in fact, invented “by a student of sociology in Russia who wanted to prove his thesis that an uninformed majority will always lose a battle of information against an informed minority.”
Mafia was invented in 1987 by Dimitry Davidoff, a psychology undergrad in Russia. He didn’t have any such thesis to test–in addition to his own studies (completing his junior and senior years at the same time), he was teaching advanced psychology to high school students, so the actual hypothesis he was testing was “what’s the easiest way to make kids waste an entire class without requiring my input and while still offering some kind of psychological lesson?” Which, honestly, is very relatable. The game was a smashing success in that regard, and the kids he taught went on to travel around the world teaching it to others. Eventually, a programmer/writer/gamer/nerd named Andrew “Zarf” Plotkin learned the game and changed it to Werewolf, because he thought werewolves are “niftier.” Which they are, so I will continue to refer to the game as Werewolf from now on.
Okay, so there wasn’t a thesis, but does the game still prove that “uninformed majority will always lose a battle of information against an informed minority?”
No! If the werewolves won every game, it wouldn’t be much of a game, would it? But do they usually win? Does having information that the other team doesn’t give them an advantage, despite their smaller number? That’s where the science comes in.
A lot depends upon how many total players there are, and how many werewolves there are. Plotkin ran a bunch of simple simulated games to try to figure out at what point it made sense to add a third werewolf, concluding that it should happen around 17 players. But he points out the obvious flaw: to simulate a game, you need villagers who are very stupid, and who choose a player to kill totally at random.
But he did identify some interesting facts. For instance, a game with an even number of players (including the moderator) favors the werewolves, because an even game will always end at night, giving the werewolves an extra kill.
He also found that in a two-wolf game, you’d need to have 18 villagers for the villagers to have an even chance of winning, which DOES suggest that having information gives the werewolves a huge advantage.
In 2006, researchers created a more advanced model that assumed the villagers played an optimal randomized strategy, and found that for there to be an equal chance of either team winning, the number of werewolves should equal the square root of the total number of players, meaning that in a game of 10,000 people, you’d only need 50 werewolves for a fair fight.
That’s an absolutely tremendous advantage to simply have some information. But again: all these models must, by design, exclude the human factor. Using an optimal randomized method of voting for the werewolf is great and all, but it absolutely pales in comparison to good old human lie detection. I’ve stated in the past that humans, in general, are terrible at figuring out when strangers are lying–yes, even those guys on YouTube who claim to be body language experts–but we’re actually pretty good at figuring out when people we know are lying. Davidoff has said that the game is meant to be played with at least three rounds in a row, because as players take on different roles, they get better at those roles and they also get better at observing how other players act in those roles. That’s why if you’ve played this game before–and I have played it, a lot–you’ll notice that in fact, the innocent villagers tend to win more often than the werewolves, statistical analyses be damned.
All of which brings us to the psychological aspect: how do you win a game of Werewolf, Mafia, or Traitors? Obviously, being a good liar helps–as a psychologist pointed out to Wired UK in 2010, lying puts a huge amount of stress on the body, especially when you’re trying to fake emotion. Most people can’t pretend to be shocked for very long, so at some point they may fall back on simple aggression. Looking out for switches like that, or other more individual signs of stress in each player, can help spot a lying werewolf.
But there’s more to it than just that. Japanese researchers examined the game logs for 431 online Werewolf games and found that the werewolves’ chances of success increased substantially the more they talked, and also the more they secretly communicated with one another.
Developing trust, whether you’re a villager or a werewolf, is also an important aspect of the game. A different group of Japanese researchers used a hugging robot called Hugvie to see if it would help players win a modified version of Werewolf in which everyone was a villager, so to win they just had to trust each other and not kill anyone. Not only did Hugvie not increase the players’ chances of winning, it actually significantly decreased their chances, meaning that hugging a robotic pillow is NOT an effective way to convince anyone you’re a normal human.
One of the best ways to win seems to be making observations and keeping track of what’s happening: who is voting for whom, who succeeds in finding a werewolf, who makes lots of accusations? The game’s wiki page quotes a fan in an old CNet article, who says “If the villagers are allowed to keep a pencil and paper, they always win.”
And that, of course, speaks to the most basic, fundamental way to overcome the disadvantage: gather information. The minority has the advantage because of information, and you can overcome it by gathering more information. The better you are at that, the better your observational skills and your memory, the better your chances of winning.
To conclude, no, Graham Linehan was wrong about the game’s origins and the “proof” that an informed minority will always beat an uninformed majority. Not for nothing, but he’s also wrong about trans and nonbinary people being comparable to actual monsters killing humans in the dead of night. The game of Werewolf absolutely tells us things about human psychology and trust and behavior, but it isn’t simply about an informed minority. It’s an informed minority that actively seeks to harm and in fact destroy the uninformed majority, and it’s very telling that he thinks trans and nonbinary people are doing that. Because he is full of hatred for trans people, he can only see a shadowy conspiracy. Were it not for that irrational hatred, Linehan would be able to observe the world as it truly is. He would be able to see that in order to win the game, we have to trust one another, care for one another, protect one another. Because in the real world, there are no evil werewolves secretly killing in the dead of night. There are only humans, and they aren’t hiding the horrible things they’re doing. And if we don’t vote them out, then we all lose.
