I need to find a simple game in which several people need to interact with each other. The game should be simple for an analysis (it should be simple to describe what happens in the game, what players did). Because of this last reason, video games are not appropriate for my purposes. I am thinking of a simple, schematic, strategic game where people can make a limited set of simple moves.

Moreover, the moves of the game should not be conditioned only by a pure logic (like in chess or go). The behavior in the game should depend on psychological factors, on relations between people. In more detail, I think it should be a cooperation game where people make their decisions based on mutual trust. It would be nice if players can express punishment and forgiveness in the game.

Does anybody knows a game that is close to what I have described above?


As requested, I would like to provide more details. I do want to use a game for experimental purposes. To elaborate, my hypothesis is that people play differently depending on who their partner is in the game. Or, in other words, a player's strategies depend on their interpersonal relations. It is a very general statement. At the moment I do not know what kind of interpersonal relations can exist. In the simplest case we can speak about "friends" and "strangers" (people who know or do not know each other). Another aspect of interpersonal relations I can think of is trust. People can play differently depending on the level of trust between them.

I am looking for a game that captures several aspects of interpersonal relations. The repeated Prisoner Dilemma game is close to what I need. But I am afraid that it is too trivial (people will just cooperate).

  • 2
    Have you considered multiple-round Prisoner's Dilemma? It's a game theory classic, and when played with people, has distinct psychological elements. Trusting one another leads to the maximum outcome for the group as a whole. Feb 16, 2011 at 14:10
  • It sounds like you may want to use the game for some sort of psychology case study or experiment. If that's true, perhaps if you gave us more insight into your hypothesis and intended experimental methodology we could help you design/tailor something. I admit my first thought was also a more complex Prisoner's Dilemma.
    – Adam Wuerl
    Feb 16, 2011 at 14:20
  • @ire_and_curses: based on the clarification above and some thoughts I had on simple modifications, I'm going to try and write up a PD-based answer. I feel bad though since it was originally your idea. Feel free to edit like crazy or make community wiki or whatever.
    – Adam Wuerl
    Feb 17, 2011 at 0:50
  • 1
    @Adam Wuerl - No need to feel bad at all. If you write up an answer, and I can't be bothered, you clearly deserve the rep! :) Feb 17, 2011 at 3:22

12 Answers 12


I cannot think of a better game to match what you describe than Diplomacy. It requires seven players, and it is lengthy. You can speed it up by having all the negotiation happen at the table in front of everyone, but that changes the nature of the game. However, there are plenty of places on the internet where you can play "by email" for free (for example, The Diplomatic Pouch) -- there, you can enter games all by yourself, but the time taken to play a game typically spreads out over days, weeks, or months (because you spend a few days emailing negotiation letters, before each time you "submit orders" for the next turn's actions).

Diplomacy's rules do include strategy, and the rules are quite simple. The game has been around for fifty years, so there are lots of online resources describing various strategies and tips and so forth.

Note that BoardGameGeek has a Negotiation style category for boardgames, and there are lots of other games listed in it; many would be suitable for fewer players and less playing time, but Diplomacy is pretty much the prototypical experience for what you describe in my opinion.

  • Diplomacy is an excellent answer I think. There are only 4 moves to be made for a small set of units: but strategy and negotiation play a huge role. For research you may want to use a variant map from here: diplom.org/Online/variants1.html Look for symmetrical maps. It depends on what you want to study. In standard diplomacy, some nations are disadvantaged and must rely on better negotiations to win. A symmetric map will put all players on even footing, meaning that victory relies entirely on negotiation and strategy. Feb 16, 2011 at 19:14
  • Best of all, Diplomacy has no dice, so luck is almost completely removed from general play.
    – Jon Hadley
    Feb 16, 2011 at 20:05
  • Diplomacy is a game where you can only win by trusting people at the right times... and betraying their trust at the right times, too. I know for a fact it has ended engagements, and possibly marriages too! Feb 16, 2011 at 20:57
  • Given the recent clarification that this is for a controlled experiment, I think Diplomacy is too long of a game. Multiple trials would be necessary to achieve statistically significant results. I'd say you could data mine an online database of Diplomacy games, but you wouldn't have any meta data on the relationships between the players, thus no experiment.
    – Adam Wuerl
    Feb 17, 2011 at 0:35
  • 1
    @Adam, generally I agree with you; however, I have a counter point. The play-by-email systems make it perfectly feasible to set up a game where (a) all players' real identities are known to each other, and (b) all players' real identities are not known. In both cases, the game itself is perfectly playable. Setting visibility options, and options for the scope of press, are easily done with Judge-moderated games, and do (in my experience) change the tone of the game (in interesting ways). Feb 17, 2011 at 14:17


is a party game I've played which I've discovered was actually designed by a Russian psychologist.

It's also commonly known as Werewolf, Assassin or Witch Hunt.

One person is a moderator. Everyone else is dealt a card that gives them their secret role as a player.


  • Townspeople (The innocents)
  • Mafia (Who win by killing off the innocents)
  • Sheriff (An innocent who can detect mafia members)
  • Nurse (An innocent who can protect other innocents from attack)

There are two phases of the game, night and day. During the night the moderator tells everyone to close their eyes (the villagers are asleep.) There is then some general tapping by everyone involved to mask movement. Then the moderator tells the Mafia to wake up and acknowlege each other. Wordlessly, they must choose someone to "assassinate" (i.e. remove from the game.) They silently gesture until they show unanimity.

Next the Sheriff wakes up. He is asked to point out someone as a suspect. If the suspect is indeed Mafia, the moderator nods.

The Sheriff is told to go back to sleep, then the Nurse wakes up. She is told to point out someone to protect.

Then the day phase begins. Everyone is told to "wake up", open their eyes, and the moderator reveals who was assassinated (unless that person was protected by the Nurse.) For the enjoyment of all, this is embellished with narrative detail, etc.

Then the players deliberate over who should be nominated for execution. Once the deliberation is over the moderator asks for a vote and whoever gets the most votes is "lynched" and removed from play. In some variants of the game (the way we played we did this) that player's card is flipped over so everyone knows the accused's true role.

The game ends either when the last Mafia member is killed (Innocent victory) or the Mafia members outnumber or equal the Innocents during the day (Mafia victory).

When I played it with several families, I found especially fascinating the interaction between spouses, and between spouses and their children.

It doesn't take that long to play, and I think it might be close to what you're looking for. The Wikipedia article I linked in the heading has many different variations including additional roles that you can tweak to investigate what you want. One variant we played was a Spy role, an innocent who was able to peek at any one person's role card each night.

  • I have no idea if this would make for a good experiment, but it simply sounds so awesome I need to up-vote immediately.
    – Adam Wuerl
    Feb 17, 2011 at 1:35
  • Sounds like an awesome game. Doesn't get more psychological than this.
    – Hackworth
    Sep 16, 2011 at 12:44

Modified Iterated Prisoner's Dilemma

I think an ideal co-operative game for your experiment is the Iterated Prisoner's Dilemma:

  • There is substantial literature in the field of experimental economics on this game, and ample theoretical and experimental bases for determining optimal play strategies. These experiments can form a baseline control group (as they are typically played with strangers) for you to compare your results to.

  • You mentioned that you think the game is too simple and that people will simply co-operate. However, the experimental results show that for strangers this is not true. Tit-for-tat is one of the best strategies, yet even this only works if both players use it and if the game has indeterminate length. So I think it would be significant if the experimental results demonstrated alternative optimal playing strategies given non-stranger relationships.

  • The game is simple to play and players can quickly gain an intuitive feel for how they're doing. Compare this to a game like Diplomacy, where teaching your subjects to play with any skill is going to consume enormous amounts of time (unless you drew from a pool of experienced gamers, which would probably introduce unacceptable bias into the results anyway).

I would be fascinated to see this kind of experiment with various types of uni-lateral and bi-lateral relationships:

  • strangers (baseline)
  • uni-lateral stranger: player A knows and trusts player B, but B thinks that A is a stranger
  • false pretenses: player A plays many games with a fake player B who always co-operates, but then B is switched out part-way through the game for someone more nefarious
  • perfect collusion: players who know each other very well or who are required to evenly split their winnings
  • perfect collusion with mutation: same as above, but sometimes, randomly, the computer switches their choice
  • sympathy: player A is given back story on B that B is a nun or an orphan or something else sympathetic and told that B gets to keep their winnings but A will not (hypothesis that A might as well co-operate because they have nothing to lose but has the ability to help B in the "real world")
  • If this game is too transparent, you may be able to find a way to disguise it so the coop/defect decision is layered away. A simple stock trading game may work well enough; I would be surprised if no one has done this yet but I cannot think of an example.
    – MrHen
    Mar 2, 2011 at 19:04
  • Do you have an example of a unilateral stranger relationship? I can only think of fan-celebrity, but that would be hard to build an experiment around.
    – user848
    May 15, 2011 at 14:35
  • The easiest way to think about a unilateral stranger relationship is if the game were not played face-to-face (i.e. via a computer). One player could be told (or perhaps shown) that they were playing against someone they trusted--a friend, their spouse, etc. The other player could be told they were playing against a stranger.
    – Adam Wuerl
    May 21, 2011 at 1:37

In the same vein as the Mafia and Werewolf suggestions, I'd suggest The Resistance. It doesn't have the player elimination that Werewolf or Mafia has, which means you have more ongoing relationships throughout the game to study.

In The Resistance, the players are a resistance cell fighting against the empire. Some number of the players (1-3, depending on the total number of players) are spies trying to sabotage the resistance in their fight. There are 5 possible missions in the game. Either side needs to win 3 of them to win. For each mission, the leader (a role that rotates frequently during the game) first selects his proposed team (which is a subset of the group) to go on the mission. The entire group debates the proposal and then after the debate there's a vote. If the majority of people approve the mission, then the mission is a go. If the proposal fails, the leader role passes to the person on the left who then makes a proposal. The cycle repeats until either a mission passes or 5 proposals in a row fail (at which point the spies automatically win).

When a mission is a go, each of the people on the mission are given success and failure cards. They secretly pick one (each person who is a loyal resistance member always votes for mission success, while spies can vote for success or failure) and the chosen cards are shuffled together. Most missions fail if there is a single failure card played, though one requires 2 fails. With each mission, the entire group analyzes the information they have (who voted for or against a proposal, whether a mission passed or failed, who was on a given mission) and try to determine who the spies are so they can make proposals guaranteed to pass.

There are a lot of elements of trusts and interpreting behaviors in this game, and the playing time clocks in around 20-30 minutes per game, so I think it might be an ideal choice for a psychological experiment like you describe.

  • Agreed. If a Mafia-style game is what you want, The Resistance serves your purposes much better than Mafia itself. The wikipedia article gives a good description, but if you want a detailed run down, this video is good, if a bit slow-paced.
    – warbaker
    Feb 23, 2011 at 16:59
  • I played The Resistance for the first time last night, and really enjoyed it. One question - you say above that the leader role passes if a proposal fails. The group I was playing with had the leader role pass automatically each time, regardless of whether their proposal succeeded. Is that the correct rule? Sep 14, 2011 at 16:25
  • It's been a while since I played the Resistance, but as I recollect, yes that is the rule. At the start of each mission, the leader rotates normally. If a proposal fails, the leader rotates immediately for the current mission. Oct 28, 2011 at 19:32

Shadows Over Camelot or Battlestar Galactica both have interesting player interactions similar to Mafia/Werewolf. The group is working together as a team with a potential for one player to secretly be working against the group. I wouldn't classify either game as simple, however.


Cosmic Encounter (BGG link) is a simple game that offers plenty of opportunities for players to ally and attack each other. This BGG review offers a quick description of the game. The relevant mechanic:

Encounter cards can either Attack or Negotiate. Attack cards have numeric values that players add up to count their strength. Attack is very simple with the player with the higher score winning the battle. Defender wins ties. Negotiate auto loses but here's the kicker, the one who Negotiates gets to draw cards equal to the number of ships he had involved in the Encounter, from the winner. If both sides Negotiate, they have one minute to deal (meaning trade cards or colonies), or they both lose 3 ships.

  • to add to that, the fact that you play the game differently depending on what power you initially draw does provide a way of determining how people play against each other - ie the game decides what traits a player has that affects how the others react to you.
    – gbjbaanb
    Mar 4, 2011 at 18:45

The Castle of the Devil is a great game where you use support, attack, and trade to try and reveal who's on your side by exchanging information.


I would suggest a game called Warewolf, it is more a role game, where after each 'night' you have to discuss with everyone and find killers in the village, there are different roles in the game, each player gets its role on the card and has to keep hold of it till the end of the game, it is really fun game. as to board games, maybe Pandemic, where all the players battle with four different viruces spreading around the world, it is simple team game.


I do not believe that any of the games given so far will be quite what you're looking for. You've stated two things are important, one that strategy will depend on pre-existing trust and that there's some notion of 'partnership'. Here's a quick game off the top of my head which might address this:

You need 5 players, seated at a table in a pentagon shape. The two across from you are 'enemies' and the two to either side are 'friends'. Each player is given 3 cards or tokens (identical). The only requirement of the tokens is that they are easy to conceal, and can be played in one of two ways as you will see. A random player is chosen to go first.

No private or public discussion is allowed (this ensure that decisions are made from pre-existing biases about the people, per your experiment's needs); the starting player chooses a person to attack. This may be a friend or an enemy. Then all the other players choose as many token as they'd like played either in support of the attack or the defense, and kept concealed, but unchangeable. You MAY choose support an attack on your friend, or support the defense of your enemy.

Once selections are made, all players reveal and the attack is resolved. Each support of a friend is worth twice as many points, likewise each attack on an enemy is worth twice as many points.

The attacker himself cannot play any tokens. Likewise the defender cannot play any tokens (the success or failure is based only on the support of others).

whichever had more points win, the other is eliminated from the game (but retains his position on the table for friend-enemy purposes). A tie goes to the attacker.

Play continues with the next player making an attack, so one person is eliminated each time. Obviously, at some point two 'friends' must attack each other. The winner is the last person remaining (NOTE: on the last round, no token may be played, and the attacker immediately wins).

Optionally, you can have eliminated players gift any remaining tokens to others players of their choice. However, this will have drastic consequences on your results and vastly complicate their analysis.

In addition, if you can afford it, I suggest letting players play on a computer simulator first only a few times to get the idea of the game. This eliminates the learning factor from affecting your experiment. The computer players may act randomly, and only feedback like (you friend supports you! Or your friend betrays you!) is necessary.


The easiest game I know that matches your needs is Ghosts!. It's a 2 players game. Each one has 8 ghosts, 4 are good, 4 are bad. The goal is either to eat the opponents 4 good ghosts or to let him eat your own 4 bad ghosts.

A strategy would be to send bad ghosts to the opposite side of the board, expecting the other player to eat them. But the opponent knowing that, you could try and tempt him with good ghosts, making him think that they are bad. But him knowing that, you could...

For your information, such a mechanism is called "Second guessing".


Strange no one has proposed Munchkin. It's some sort of D&D and MtG hybrid, but unlike former it has strict rules and is card-based (much easier to analyze in-game situation) and unlike latter players' interaction with each other (trading, forming temporary alliances, doing bad stuff to other players etc.) play a big role in it. The rules, however, could not be called very simple.

Other good game: Life boat. Pretty simple, has strict rules and is actually about players' interactions. It also sort-of enforces in-game friends and enemies (each person randomly receive two cards specifying his "friend" and his "enemy". These cards can't be shown to other players, and if "friend" is alive (and "enemy" is dead) when game ends, one get additional points).

And also vote up for Mafia (I don't have enough rep. to actually vote up), it seams to suit your purposes excellently.

  • I don't really understand how anyone could call the rules of Munchkin "strict" compared to either D&D or MtG, but apart from that I like your answer! Sep 6, 2011 at 14:33
  • Well, by 'strict' I meant that hmm... the in-game situation is easier formalized. But I've played D&D only twice, so I might just not grasped the concept, or it was just the way we've played. And I wasn't saying Munchkin's rules are stricter than MtG's.
    – aland
    Sep 6, 2011 at 15:45
  • I see what you mean - I think D&D has very strict rules, but it's also definitely more open-ended than Munchkin. Sep 6, 2011 at 15:50

Bang! is a relatively simple, fast-playing card game with hidden, conflicting roles.

Not the answer you're looking for? Browse other questions tagged .