17 MEANEST Board Games EVER (these destroy friendships..)

There is a specific kind of silence that falls over a game table when somebody realizes they have just been completely, deliberately, and systematically destroyed by somebody they trusted. Not bad luck, not a bad draw. The person sitting across from them smiled, played their cards just right, and had been planning it for three rounds.

That silence is what we are talking about today. >> And mean games are a genre unto themselves. Not the same as hard games, not the same as long games. These are games designed at their core to let you reach into another player's position and take something from them. A resource, a vote, a round of safety, or their entire strategy.

They deliver a kind of satisfaction that cooperative experiences simply cannot replicate. We have 17 of them today, sorted by how much rules knowledge you need before the suffering begins. >> Starting with lightweight cruelty, these you can learn in under 20 minutes, but the damage lasts considerably longer.

And first up has to be Survive Escape from Atlantis from 1982. Still doing exactly what it set out to do over 40 years later. >> This is the one where the island of Atlantis is sinking one tile at a time. Your little wooden people are scrambling toward boats into the surrounding shores. The problem is that the ocean is not empty.

Sharks circle the waters, sea serpents lurk beneath, whales patrol the open ocean, and the person sitting across from you controls all of them. >> That is the part nobody really gets until they sit down with it. On the surface, it looks like a frantic family game about escaping a disaster. What it actually is, once everybody understands how it works, is a game about directing predators at your friends.

When you flip a tile and reveal a sea creature, you place it in the water. Then, at the end of your turn, you roll a die that determines which type of creature you move, and you choose exactly where that creature goes within that constraint. >> Sharks kill swimmers, sea serpents destroy entire boats and kill everyone aboard in the same movement.

Whales capsize occupied boats turning all passengers into survivors stranded in the open water. We had a session where somebody spent two turns rerouting a sea serpent across the board rather than making any progress with their own survivors. Sank a boat carrying three of an opponent's pieces and one of their own.

Watched their last surviving piece swim to shore. The opponent did not speak for the rest of the game. The sea serpent player maintained eye contact and smiled the whole time. >> The hidden point values are what seal it for me. Every survivor has a secret value only visible to its owner. You're never quite sure who is the most valuable target.

Do you send the shark toward the cluster near the boat? Or do you track that single swimmer making for sure who might be worth 10 points? [music] Pure theater. Nobody at the table is innocent. >> From there we go to Cutthroat Caverns from 2007. Dungeon crawl card game that starts cooperative and very quickly becomes something much darker.

Players are a party of adventurers making their way through increasingly dangerous encounters needing to work together to survive. The twist is that only the player who delivers the killing blow earns the prestige for that defeat. Most prestige wins. Survival is required but it is not enough. >> What that does to the table is incredible.

Players cannot attack each other directly. Instead, the game becomes an ongoing exercise in manipulation. You need your companions to weaken the monsters because you can't survive alone. But you absolutely cannot let them land the final hit. Constant micro sabotage. Hanging back, redirecting damage, disrupting a teammate at the precise moment they are about to finish a monster off.

Every player is permanently suspended between cooperation and betrayal. The tension is not a spike. It is a sustained condition that runs the entire session. We had somebody positioning for the killing blow on the final encounter for three full rounds. They had the cards, they had the timing, declared their attack, another player immediately played a card that had technically resolved on the previous turn, reducing the monster's health by one.

Boss died before the planned hit landed. The argument about sequencing lasted longer than the entire game had to that point. >> Hey, That's My Fish from 2003 is a different kind of mean entirely. Deceptively simple abstract game where penguins collect fish from hexagonal ice flows. Players move their penguins in straight lines across the board collecting the tile they leave behind.

Board shrinks turn by turn. Most fish wins. >> There is no luck, none at all. No dice, no cards, no random events. Every outcome is the direct result of a deliberate decision somebody made. When a player gets cut off and isolated on a shrinking island with two fish tiles while the rest of the board is comfortably divided between opponents, they can't blame fortune.

They were outmaneuvered deliberately with calculation. >> Players use their penguins not just to collect fish, but to physically block movement corridors, fracture the board into smaller regions, and trap opponents in the worst possible territory while keeping the richest areas exclusively for themselves.

A good player is not just thinking about where to go, they're thinking about where to ensure their opponents have no good moves at all. >> We had a seasoned player sit down with newcomers and spend their first four turns not collecting a single tile. Just moved and blocked and moved again. By turn five, two players had been cut off from the majority of the board.

Then they spent the rest of the game leisurely harvesting the richest sections while the others argued about whose fault it was. There was no satisfying answer. Munchkin from 2001 is the one most people have probably encountered. Card game parody of dungeon crawling. Players draw cards, kick down doors, fight monsters, collect treasure.

Each defeated monster earns a level. First to 10 wins. >> Other players can help or hinder during fights. Add to a monster's power or play cards that change the result at the last second. >> The particularly nasty part is the pile-on can sustain indefinitely. A player can be one step from winning for 20 minutes while the table depletes their entire hand preventing it.

Then when the table finally has nothing left to throw, somebody else slips through a back door and wins on a turn nobody was watching. >> We had a player reach level nine with a full hand of powerful combat cards and declare an attack that should have ended the game in seconds. Four other players played 17 combined cards over the next four minutes turning a trivially easy fight into a catastrophic defeat.

The player put their cards face down on the table and said quietly, "I have been playing this game for three years and I have never won." They did not win that night either. >> Citadels from 2000 in lightweight section is one of the most replayable games ever made specifically because of its meanness.

Each round players secretly select one character role from a shared set then take their actions in role order. Characters include the assassin, the thief, the magician, and others. >> Players race to be first to complete eight buildings triggering the end of the game and earning a bonus that no one else can claim.

>> The assassin eliminates a chosen character for the entire round. The thief takes all of another character's accumulated gold the moment that character's turn begins before they can take any action. Before a single building is placed the game is already a web of reads, bluffs, and counter bluffs. >> Every round begins with a private selection where you're trying to predict what everybody else will pick so you can either neutralize their plan or exploit the gap it leaves.

>> The meanness here is cerebral and targeted. When the assassin correctly calls somebody's role and removes them from a critical round, it is not luck, it is a read. Knowledge of the person across from you applied with precision. That stings considerably more than random variance ever could. >> We saw a player spend three rounds accumulating gold quietly, playing nothing aggressive, establishing a reputation for caution.

The table forgot about them. On round four, they picked the architect and laid down three districts in a single turn, suddenly threatening to win. The next turn, every other player independently chose an aggressive character, creating a perfect chaotic standoff where all their plans canceled each other out.

The original player had switched roles entirely and quietly finished their eighth building while the table destroyed itself. >> Now, into medium-weight cruelty. These need a session or two to learn and possibly a therapist to fully recover from. Starting with Mantis Falls from 2020, compact cooperative card game for two to three players where witnesses to a terrible event are trying to survive and escape a dangerous small town.

>> Players manage their hands carefully, navigating escalating threats together. In some games, depending on the role draw, one player secretly holds a hidden role that changes their true objective entirely. >> The meanness here is quiet and intimate, which makes it particularly effective. At two or three players, there is nowhere to hide.

Every card played is visible and subject to scrutiny. Every slightly unusual choice becomes a data point. When something goes wrong, the table has an extremely short list of suspects. >> That small player count is the entire design choice. Other traitor games diffuse suspicion across eight people. Mantis Falls seats you directly across from one other person and asks you to trust them completely while systematically creating conditions that make trust extremely difficult to sustain.

The psychological discomfort is precision engineered. >> We had two players deep into a cooperative communicating well, making good progress. One of them played a card that was defensible but slightly strange. The other said nothing and noted it. Three rounds later, the pattern became undeniable. The first player had been the hidden role the entire time, making tiny, plausible-looking sub-optimal choices for 20 minutes.

The second player stared at them and said, "You were never on my side." The first player replied, "I was always on my side." That is Mantis Falls in two sentences. >> City of Horror from 2012 takes the meanness in a different direction. Zombie survival negotiation game where players control multiple characters across different locations in a besieged town.

Each round zombies move, and when they outnumber the survivors in a location, somebody gets eaten. Who gets eaten is not decided by any mechanic, it's decided by the vote among the players present, weighted by each character's vote count. >> Players offer in-game resources for votes, make multi-round protection agreements, and occasionally sacrifice their own weaker characters deliberately to improve their strategic position.

>> The savage detail is that each character has a special ability they can use once during the game, but using it permanently reduces that character's end-game [music] point value. So, every vote is a calculation, not just about who lives, but whose ability is still intact and worth protecting, and whose has already been spent.

>> We had a player assemble the most powerful character combination on the board over three rounds. Everybody knew [music] it. When a zombie wave hit and a vote was called, every other player voted to sacrifice that character, including two who had shaken hands on a protection deal earlier. >> The player looked at the results calmly.

"You all agreed beforehand," they said. "It was not a question." In another round, somebody replied, "You would have won and none of us could have stopped it." There was a pause. [music] "That is fair," they said. Stunning composure. >> Intrigue from 1994 might be the the nakedly mean game ever designed.

Players are wealthy families trying to place their scholars at each other's palaces, where those scholars generate income each round. The mechanism for deciding who gets placed where is bribery and promises, nothing else. The rulebook explicitly states that no agreement is binding. It is mean by explicit architectural design.

Every round, [music] players offer in-game money to secure placement of their scholars at opponents' courts. The players receiving those bribes can take the money and place whoever they want. There is no penalty for breaking the agreement. It is not a loophole, it is the entire intended experience. >> What makes this devastating is that Intrigue has nothing else to hide behind.

No combat, no card draw, no dice variants to absorb the social impact of a broken promise. There is only what you said, what somebody paid you, and what you actually did. The game strips negotiation down to its most mercenary possible level and asks you to operate there with your friends for an entire session.

>> Two players in our group negotiated a mutual placement agreement and exchanged significant sums. One immediately placed the other scholar in the least profitable position on the board, said nothing, made no excuse, moved on. The offended player maintained perfect composure for the rest of the session while methodically ensuring that not a single one of their opponent's scholars was ever placed in a profitable position again.

Revenge in Intrigue does not explode, it closes doors, quietly, [music] permanently. >> Nothing Personal, the revised edition from 2019, takes that same negotiation engine and dresses it in mob gaming. Competing crime families fighting for dominance within an organized crime syndicate. Characters move through the syndicate's ranks as players place influence on them to contest control.

Whoever commits the most influence to a gangster claims their allegiance in the income and respect that come with it, the negotiation around those placements is relentless. Every deal is durable as the moment it was made. >> The mob theming is not decorative. The experience it represents is operationally accurate.

Alliances are transactional. Loyalty has an exact price. The most reliable seeming ally three rounds ago may currently be calculating whether you are more useful as a partner or a problem. The most effective strategy consistently involves appearing trustworthy for a sustained period before acting decisively on the goodwill that has been accumulated.

Patience is the primary weapon. >> We watched a player spend the first half of a game building trust meticulously, always delivering on small promises, always being the measured voice when others escalated. In the final quarter, they called in every accumulated favor simultaneously, advancing their candidate to the top while having the other two leading players key figures eliminated in the same round.

Their explanation was three words, "I was patient." Nobody underestimated them in the rematch. They did it again anyway. >> Inis from 2016 is the elegant one. Celtic-themed area control for two to four players. Players draft cards each round and use them to move clans, build sanctuaries, and contest territories.

Three simultaneous win conditions exist: being present in six or more different territories, being present in territories that collectively contain six or more sanctuaries, or being the chieftain of territories that together hold at least six opposing clans. >> [music] >> Any player who satisfies a condition can declare victory, but opponents can contest it.

>> It maintains a state of sustained paranoia from round one because three different win conditions exist simultaneously. Every time somebody approaches one, the rest of the table has to choose between contesting that threat and watching the player who is silently closing in on a different condition entirely.

The card drafting allows targeted denial. >> [music] >> You can take cards you know an opponent needs, not because you want them, but specifically because you want them not to have them. >> Combined with the ability to contest victories at the very last moment, the game creates a condition where near wins happen constantly, and being stopped on the final step is a particular and recurring form of agony.

We watched a player build quietly toward the territory condition all game, while others fought loudly in the center. By the time the table noticed, they needed one more move. Three players responded simultaneously, flooding the key territory. In the resulting chaos, a fourth player who had been ignored for an hour quietly satisfied the sanctuary condition in a different region.

They had been watching the table look the wrong direction the entire [music] time. >> Old King's Crown from 2024 belongs in the same conversation. Two to four player area control set in a kingdom fracturing after the death of its ruler. Players place followers across regions, engage in card-driven standoffs for territory control, and compete for claiming the most valuable parts of the realm.

The card play system is built around commitment, timing, and reading opponents' intentions under incomplete information. >> It is mean with refinement. The card system forces decisions with incomplete information about what opponents are holding. You commit resources to a region, either reading a bluff correctly and winning decisively, or missing it entirely and watching an expensive investment evaporate on a perfectly timed counter.

The ability to redirect influence and turn regions against their current controller at the last moment creates the game's most demoralizing outcome. A player who appeared dominant finding the board entirely reshuffled in a single exchange. Nothing that happens is random. Somebody made a plan. Somebody anticipated your plan.

Somebody played around it. We had a player who controlled three high-value regions at the midpoint and appeared comfortable. An opponent spent two rounds playing defensively, apparently struggling. Then, on the final scoring round, they played a sequence of cards that flipped two of the three regions, converting a dominant position into a loss.

The dominant [music] player reviewed the previous turns afterward and understood, "You were setting this up since round three," they said. The opponent nodded and said nothing. That nod was the meanest part of the entire game. >> Rounding out the medium-weight section is Game of Thrones: The Board Game, second edition from 2011.

Three to six-player area control set in Westeros, where players control great houses fighting for dominance. Players write orders for each territory simultaneously and reveal them all at once, producing explosive interactions between plans nobody else could see. Three influence tracks determine key advantages.

Wildling attacks periodically threaten the entire board. Alliances are honored for exactly as long as they remain useful. >> The simultaneous order system is the engine of everything. You cannot see what your ally planned. You negotiate in one phase, write orders privately, and reveal them at the same time.

The moment between writing and revealing is where every agreement is tested and frequently found to be worth exactly nothing. An ally who pledged support may have written a completely different order. The board tells you what happened. It does not tell you why. >> We had a six-player game reach its final rounds with two players tied, one castle each needed to win.

A third player out of contention held the critical border territory both needed. Both leading players negotiated privately and both emerged claiming they had secured the deal. The third player collected both bribes, considered the situation, and then marched their own army into their own castle rather than handing it to either claimant.

Game ended in a draw. The third player appeared genuinely at peace with this. Both leading players did not speak to them for the rest of the evening. They had not expected this level of commitment to the bit. >> Now into heavyweight cruelty. These demand significant investment to learn and deliver maximum sustained carefully engineered suffering in return.

Starting with Nemesis from 2018, semi-cooperative science fiction survival where players are crew members aboard a spaceship infiltrated by alien intruders. Each player has a secret personal objective that may or may not align with the rest of the group. The ship is deteriorating, the intruders are hunting, and nobody can be certain whose goals actually match their own.

>> The alien intruder system creates genuine sustained physical threat throughout the game, but the deeper tension lives in the human relationships. Because personal objectives are secret and can conflict, every crewmate's decision is potentially evidence of something. Why did they take that corridor?

Why did they trigger that noise event when a different path was available? Why did they lock that door before you got through it? >> has formal trader mechanics in certain configurations, but even without them, structurally misaligned objectives create natural conflict. A crew member whose personal objective is incompatible with the group's survival has no mechanical incentive to announce this.

The game simply creates the conditions for quiet, cold, entirely plausible abandonment. >> a player spend four rounds making entirely credible cooperative decisions, patching systems, fighting intruders, communicating efficiently, built genuine trust. Then they made a series of choices that were each individually defensible, but collectively moved the ship toward an outcome that served only them.

By the time the table understood, it was irreversible. You were never trying to save us, one player said. I was always trying to save myself, came the reply. That is a different thing. It is, in fact, exactly what Nemesis is designed to produce. >> Feif France 1429 from 2015 takes the meanness into historical territory.

Medieval political simulation set in 15th century France, where players control noble families competing for land, titles, and influence, territorial control, military conflict, dynastic marriages between players' families, and the ability to elect both a king and a pope from among the players at the table, both of whom receive significant powers and victory point advantages.

Everything is negotiable. >> Fief operates on the principle that every power structure can be turned against somebody sitting at the table. The elected pope controlled by a player can excommunicate opponents. The elected king holds additional institutional powers. Both offices are held by actual players with their own victory conditions to pursue, creating a permanent conflict of institutional interest.

>> system is particularly well designed for producing betrayal. Marriages create shared victory point structures, genuinely tying your progress to your allies, until the moment one of you needs those points more than the relationship is worth. Every military action is then a statement about where loyalties currently lie, made in the context of these intertwined personal and dynastic stakes.

>> Two players in our group maintained a marriage alliance that had dominated most of the game. A third player, trailing significantly, quietly arranged support for their preferred candidate for pope. The new pope, now in that player's control, immediately excommunicated one of the allied nobles. Unable to act, the excommunicated player watched their own partner seize the opportunity to take their most valuable territory.

Alliance dissolved in a single round. Both former partners finished worse than when the alliance began. The third player who engineered the entire collapse finished second, extremely calm about it. >> Then we have Dune, the board game, the Gale Force Nine edition from 2019. Up to six players control the iconic factions from Frank Herbert's novel, House Atreides, House Harkonnen, the Fremen, the Spacing Guild, and the Padishah Emperor.

Players collect spice across Arrakis, bid on treachery cards, navigate the moving desert storms, and fight for control of key strongholds. Each faction has dramatically asymmetric powers and a fundamentally different path to victory. >> The traitor system is the knife at the game's heart. Each player holds cards representing leaders from other factions.

If a leader from an opposing faction is committed to combat and you hold their traitor card, you can reveal it and win the combat instantly, regardless of any other factor. No fortification helps, no numerical advantage matters. Your own leader turned against you the moment they walked onto that battlefield.

>> This creates a game where committing your best leaders to crucial battles is an act of faith that the player across from you is not holding their card. You cannot know for certain. The bidding for treachery cards creates a secondary arms race of paranoia that runs parallel to the military campaign.

The faction asymmetry ensures the first several games are spent discovering just how unpleasant certain matchups can be. >> We had a player commit their most powerful leader to a critical stronghold battle backed by significant spice and two treachery cards. They had been building toward this for four rounds, were confident, an opponent revealed the traitor card.

The leader defected, the spice was lost, the stronghold fell to the opponent. The player stared at the board for a long time. Then they said quietly, "I held that card for two hours." >> [music] >> The opponent nodded. "I know," they said. "That is why you committed everything." >> Diplomacy from 1959 is the godfather of this entire genre.

Seven European great powers in 1901, each player controls one. Units move across a map of Europe seizing supply centers to expand. Units can support each others movements or defenses. The game is played in simultaneous order writing rounds with a negotiation phase where all conversations are private and players agree to anything they want.

[music] There are no dice, no cards, no luck of any kind, only what you said, what your opponent said, [music] and what both of you wrote down. >> It is the meanest game ever designed because it is the only major board game where the primary mechanic is human trust and the primary strategy is its violation.

Every alliance is a tactical choice with a measured shelf life. Every agreement is useful information about somebody's intentions given in exchange for carefully calibrated false information about yours. >> The particular cruelty is not in any single betrayal. It is the accumulation of small, entirely reasonable seeming acts of cooperation that lead precisely to it.

An ally supports your attack one round and the next and the one after that. Then on the round that matters most, they do not. They have been building toward their own victory for three rounds and your trust was the resource they were quietly, patiently depleting. >> Diplomacy is 67 years old. In our view, its design has never been surpassed in its specific niche.

We had a session that lasted 7 hours. Two players maintained an alliance that controlled the western and central boards throughout. With the end game approaching, one of them passed a note to a neutral player who had held a stable position for five rounds. The note said, "Attack the east on this move and I will allow you three supply centers." The neutral player agreed.

The eastern player was eliminated. On the following turn, the note writer absorbed the neutral player's now weakened forces. They won the game. Their former long-term ally and the betrayed neutral player stood outside the venue afterward for 20 minutes not speaking to each other or to anyone else. That [music] is diplomacy.

>> And our final entry is Food Chain Magnet from 2015. Brutally efficient economic simulation where players build competing fast food empires on a shared suburban map. Players hire employees, develop their corporate structure, create menu items, run advertising campaigns to generate customer demand, and race to fulfill that demand before competitors do.

Most money at the end wins. [music] Very little luck. A great deal of carefully calculated pain. >> Food Chain Magnet is mean at a systems level, which is a different and arguably more lasting form of cruelty than most games deliver. It does not stab you in a dramatic moment. It locks you out structurally, efficiently, and often before you have fully understood what has happened.

>> The game features permanent milestone bonuses. Advantages that the first player to reach a specific condition claims for the entire rest of the game. One player will always have lower production costs. One player will always control a particular advertising channel. These advantages compound. A player who secures the right combination early can build an [music] economic lead that becomes mathematically insurmountable within a few rounds.

>> The price war mechanic is equally devastating. When two players compete in the same area, the lower price captures all the customers. A well-timed price reduction can absorb an opponent's entire revenue for that round, and force them to match it, compressing both margins until somebody is operating below cost.

We had a session where a trained accountant leaned back from the table mid-game and said quietly, "This is showing me things about markets I did not want to understand." >> A first-time player in our group spent three rounds building a beautiful marketing campaign, driving enormous demand for lemonade near their restaurant, felt good about it.

An experienced player who had spent those same three rounds quietly building their employee structure without advertising a single item, opened a competing stand in the same block on round four, priced $1 below, and captured every customer in a single turn. The newcomer looked at the board for a long moment.

"You planned this," they said. The response was honest and brief, "Since round one." The worst part was not the loss. The worst part was understanding that it had never been a competition. It had been a lesson on a schedule the teacher set before the first card was played.