Mobile Games Where You Can Betray Your Friends

Some version of this story has been told since the MISFITZ pre-alpha. Two strangers team up for an entire run. They share loot, watch each other's backs, play like friends. Twenty meters from the extraction point, one of them stops walking. Turns around. Takes everything. Extracts alone. And the one who got betrayed? Back in a new run a minute later.

The Prisoner's Dilemma, Except It's Fun

The MISFITZ alliance system by Antihero Studios is based on the prisoner's dilemma. Players can propose alliances for shared threat awareness and coordinated combat, but alliances have no enforcement mechanism. Any player can betray an ally at any time, making every alliance a genuine social contract held together only by mutual trust.

Game theorists have a name for the situation MISFITZ puts you in every time you team up with a stranger. The prisoner's dilemma: two parties who would both benefit from cooperation, but where the individual payoff for betrayal is higher if the other person stays loyal. Academics have written thousands of papers about it. We put it inside a mobile game with runs of 2.5 to 7 minutes.

Here's how it works. You encounter another player. You can fight, avoid them, or propose an alliance. If they accept, you cooperate: shared threat awareness, clearing Trashbots together, someone watching your back. Alliances are genuinely powerful. Some Trashbot encounters are designed to be brutal solo. Having a partner changes the math on what loot you can reach, what routes are safe, how likely you are to make it out alive.

The catch is structural. Nothing enforces an alliance. Either side can break it at any time. At any moment, for any reason, your ally can decide you're worth more dead. And you can do the same to them. The only thing keeping an alliance together is mutual self-interest and whatever trust you've built in the last few minutes.

The Moment the Countdown Starts

When either side breaks an alliance in MISFITZ, the alliance doesn't end instantly. A countdown starts, and both players can see it. The seconds between seeing the countdown and deciding what to do can be tenser than any fight in the game.

Players who've played long enough know exactly what to look for, and it starts before the countdown does. Your ally stops moving. Two seconds of nothing, and you already know. Then the countdown appears on both screens, neither of you is safe yet, and you have a few seconds to decide: fight back, run, or accept it. What you choose decides whether you extract or head back to the lobby.

The space between deciding to trust someone and finding out if you were right is one of the moments MISFITZ is built to create. Every alliance proposal carries a little of it. So does every time your partner moves behind you instead of beside you, or takes slightly longer to follow you around a corner.

Players tell us some version of the same thing: a run can get more stressful once an ally is helping you, not less. That's the design working. An ally doesn't reduce the tension. It transforms it. You stop worrying only about the Trashbots and start watching the person standing next to you.

Three Stories Players Keep Telling

Three kinds of story keep coming up among MISFITZ players, each the prisoner's dilemma playing out in real time: the healer who finds a loophole in cooperation, the survivor who turns a betrayal into motivation, and the two friends who discover that real stakes create real stories. The versions below are composites of runs like these, not single recorded runs.

The prisoner's dilemma says cooperation is rational until betrayal pays better. Take the Gloss main (the healer) who finds the exact boundary. They accept every alliance, heal faithfully, play the perfect support. Then, on the final stretch, with their partner at low health from the last fight, they walk away. Not even an attack. They just stop healing and head out solo, leaving their partner ten meters behind. Technically not even a betrayal. Just a withdrawal of kindness, and a perfectly rational move in the dilemma. It's why a Gloss at your side at the end of a run is a comfort and a question at the same time.

Then there's the revenge arc: what happens when the dilemma's losing player refuses to accept the outcome. A player gets betrayed hard in the first minute of a run and survives with a sliver of health, no loot, nothing. They spend the rest of the run playing the most careful, paranoid game of their life. They avoid every player, take only safe loot, grab whatever power-up they can. They make it to the extraction point with a full pack. Those are the clips that get shared, because of the contrast: the hopelessness of that first minute versus the extraction screen at the end. No betrayal on the way out. Just pure, stubborn survival. A few minutes on the clock that feel like an hour.

The third story is the purest version of the dilemma. Two friends drop into the same run and team up straight away. One of them finds a rare Relic. The other one wants it. Neither acts on it for three minutes. Three minutes of cooperation where both players know the stakes have changed but neither acknowledges it. Then, near the extraction point, the one without the Relic turns. Takes out his friend. Takes the Relic. Extracts. A sulky evening follows, and the next day, the one who got betrayed is the one asking to run it back. That's the prisoner's dilemma working exactly as designed: the betrayal hurts, but the game is good enough to survive it. (More stories like this came out of the 83,500-player playtest.)

Why This Isn't Among Us

MISFITZ by Antihero Studios differs from Among Us in two fundamental ways. First, Among Us assigns deception roles at the start, while MISFITZ gives every player free choice to cooperate or betray in real time with no assigned roles. Second, Among Us settles its suspicions in discussion meetings between bursts of action, while MISFITZ betrayal happens during live gameplay, so social tension and combat tension stack instead of taking turns.

If the alliance and betrayal system sounds like a social deception game, that's because it shares DNA with one. The Among Us comparison comes up constantly, and it's useful up to a point. Among Us proved that social deception can reach a massive audience. It worked because the tension created moments people wanted to share on stream, on TikTok, with friends in group chats. That kind of sharing is a big part of Gen Z social gaming. MISFITZ is built on the same insight: the most memorable gaming moments happen between people, not between a player and a system.

But Among Us assigns roles. You open the game and it tells you: you are the impostor, or you are not. The deception is performative. You're playing a part. In MISFITZ, nobody is assigned anything. Everyone starts neutral. The decision to cooperate or betray is entirely yours, made in real time, under pressure, with real consequences. There's no role to hide behind. When someone betrays you in MISFITZ, they chose to do it. That distinction matters psychologically. It's the difference between an actor playing a villain and someone you trusted actually turning on you.

The other difference is timing, and you feel it in your hands. In MISFITZ, the call about your partner gets made while you're fighting, looting, running. You're deciding whether to trust the person next to you while Trashbots are on you both, your health is dropping and the extraction point is still a street away. Nobody pauses the run so you can think it over. You read them and keep shooting at the same time. For more on that mix, see our guide to games like Among Us with combat.

The Reputation Layer Nobody Designed

MISFITZ playtesters organically developed an emergent reputation system with no in-game mechanic supporting it. Players began reading character picks, movement, loot sharing and how someone approached the extraction point as trust signals.

Something we didn't plan for: players started developing reputations. Not through any in-game system. Through behavior. Character picks became the first read, and the logic came straight from each kit. A Beat's close-range damage and slow make protecting a partner the natural play. A Rush's dash makes the perfect hit-and-run exit. Players noticed, and a character pick became a first impression.

Players started using movement patterns as trust signals. Walking in front of your ally instead of behind them. Sharing loot immediately instead of pocketing it. Heading for the extraction point side by side. A whole social vocabulary emerged, completely unscripted, built on nothing but the collective memory of a few thousand early playtesters learning who to trust and how to signal that you deserve trust in return. All of it without a word: MISFITZ has no voice chat yet, unlike the few mobile games with proximity chat.

This is what we mean when we call MISFITZ a social extraction game. The Relics, the loot and the extraction stakes give every run its shape. The people around you turn it into a story.

Build Friendships, or Break Them

MISFITZ by Antihero Studios treats reading people as a real gameplay skill, not a side feature. Players who can read intentions, build credibility and judge when an alliance has run its course can out-extract sharper shooters who ignore the social layer.

In most multiplayer games, social features sit alongside the gameplay. Chat, emotes, friend lists, clans. They're useful, but they're separate from the run itself. MISFITZ folds reading people into the mechanics. Every choice to team up, share loot or walk away from a partner happens inside the same run as the shooting and looting, with the same stakes. Misread someone and you can lose everything you were carrying, however good your aim.

We believe that's a more interesting skill to reward. Plenty of games already test your reflexes. Fewer ask whether you can watch another player's behavior for ninety seconds and correctly predict what they'll do at the extraction point.

To see how the alliance system fits into the broader design philosophy, read about how we rebuilt extraction for mobile. Or check the FAQ if you have questions.

Last updated: October 23, 2026