We Opened a Pre-Alpha. 70,000 People Showed Up.

I want to tell you what it feels like to watch 70,000 strangers play a game you built in a small office in Barcelona. It feels like losing control. In the best possible way.

The Week Everything Broke

Antihero Studios planned for 10,000 MISFITZ pre-alpha signups but 70,000 players arrived, overwhelming server infrastructure and exposing bugs in onboarding and character select. CTO Andre Parodi rearchitected matchmaking at 3am while the team shipped hotfixes to a live audience that kept playing despite placeholder art and missing features.

We planned for ten thousand signups. That was the optimistic projection. Andre had provisioned the servers accordingly, with what he called "reasonable headroom." Reasonable headroom was not seven times the expected load.

The first 48 hours were chaos. Not the romantic startup kind. The kind where your CTO is on a video call at 3am rearchitecting matchmaking because the queue times are unacceptable and players are already posting about it on Discord. The kind where you realize your onboarding flow has a bug that soft-locks new players and you're shipping a hotfix while people are actively trying to play. We had placeholder art, missing sound effects, and a character select screen that crashed on certain Android devices. It was, objectively, not ready.

Players didn't care. They played anyway. That was the first surprise.

The Stat That Changed Everything

MISFITZ players who died in a run retried at 80.5%, compared to 74.5% for players who successfully extracted. That 6% gap revealed that the extraction loop creates stronger retention through loss than through victory, because short mobile sessions turn "I lost everything" into "one more run" rather than frustration.

A week in, once the servers stabilized and we could actually trust the data, Frank pulled up the retry rates. He called me over to his screen without saying anything. Just pointed.

Players who died in a run were retrying at 80.5%. Players who successfully extracted were retrying at 74.5%. Read that again. Dying made people more likely to play again than winning.

In most games, failure is the thing you're trying to minimize. You smooth the difficulty curve, you add safety nets, you make sure the player feels successful often enough to stick around. Extraction flips that. When you die carrying a rare Relic, you don't feel defeated. You feel robbed. You had something valuable and you lost it and now you need to go get it back. It's not frustration. It's unfinished business. And unfinished business is the most powerful retention mechanic in gaming. It's the same principle behind the short-session design we chose from day one. You can't design it. The format creates it.

That 6% gap between death-retry and extract-retry is a small number that contains a very large idea: the extraction loop doesn't just work on mobile. It might be better on mobile, because the sessions are short enough that "one more run" actually fits into your life.

What We Watched Happen

During the MISFITZ pre-alpha, players invented their own trust rituals, dropping loot as peace offerings and using joystick dances to signal cooperation. Antihero Studios designed the betrayal mechanic, but the players built the social culture around it, creating emergent stories that no designer could have scripted.

Numbers tell you what happened. Watching people play tells you why. We spent most of that first week on Discord, in streams, reading every piece of feedback. What we saw was a social game inventing itself in real time.

Players developed trust rituals. They'd drop a piece of loot in front of a stranger as a peace offering. They'd do a little dance with the joystick to signal "I'm friendly." They created entire unwritten languages around cooperation. None of this was designed. We built the betrayal mechanic. The players built the trust culture around it. That distinction matters. When players create the social fabric themselves, it's stickier than anything a designer could spec.

And the betrayals. God, the betrayals. There was a clip on Discord that made the rounds internally. Two players teamed up for an entire run, shared loot, fought a boss together, healed each other. They're 20 meters from extraction. One of them stops walking. Turns around. Takes everything. The chat in the stream went completely silent for about two seconds before it exploded. That clip is why we're building this game. That two seconds of silence is the product.

The Session Pattern

MISFITZ pre-alpha players averaged 50 minutes of daily playtime spread across three sessions of roughly 15 to 17 minutes each. The 5-to-10-minute extraction runs drove return frequency that longer-session games cannot match, with power users logging up to 10 hours per day on a build with placeholder art.

Players averaged 50 minutes of play per day, but not in one sitting. Three sessions. Roughly 15-17 minutes each. Morning, afternoon, evening. Like checking in on something they cared about. This is exactly what we designed for but hadn't dared expect from a pre-alpha. The short extraction runs (5-10 minutes each) weren't a compromise for mobile. They were driving a return frequency that longer-session games can't match. When a run takes 7 minutes, "one more" is always possible. When it takes 45 minutes, it's a scheduling decision.

Then there were the outliers. Players logging 10 hours on their phones. On a pre-alpha. With placeholder art. We didn't know whether to be proud or concerned. (Both. The answer is both.) But it told us the skill ceiling was real. The game had enough depth to sustain obsession, not just engagement.

The Shutdown Protest

When Antihero Studios closed the MISFITZ pre-alpha servers between test windows, players flooded Discord with complaints demanding the game back. This response signaled that MISFITZ had crossed from novelty into routine, with players integrating the game into their daily lives despite its unfinished state.

We ran the pre-alpha on a schedule. Servers up for a test window, then down. Standard practice. When the first window ended and we sent the "servers are closing" notification, the Discord channel lit up. Not with feedback. With complaints. "When is it coming back." "Why did you take it away." "I was in the middle of a run."

In game development, there's a specific signal you're looking for that no metric can fully capture. It's the difference between a game players enjoy and a game players miss. Players protesting the removal of a pre-alpha with missing features and broken UI is that signal. It meant the game had crossed from novelty into routine. They'd made it part of their day, and we'd taken it away. That's the kind of response you can't buy with any monetization model.

What the Creators Showed Us

Over 750 creators enrolled in the MISFITZ creator program through a partnership with KairosTime, producing 250 pieces of content within 10 days. The alliance and betrayal mechanics generated natural story arcs in every run, making MISFITZ inherently content-friendly without requiring creators to manufacture moments.

750 creators enrolled through our partnership with KairosTime. Within 10 days they'd produced 250 pieces of content. But the number isn't the interesting part. The interesting part is that they barely had to try. MISFITZ generates stories on its own. Every run has a narrative arc: will they cooperate, will they betray, will they make it out? Creators just had to hit record. The alliance and betrayal mechanics did the rest.

This is what "games worth sharing" looks like in practice. Not a marketing slogan. A design philosophy where the game itself produces moments people want to show other people. When your players are your distribution channel, growth compounds in ways that paid acquisition never will.

What We Got Wrong

The MISFITZ pre-alpha revealed two key problems: new players who had never played an extraction game churned in the first two minutes due to a tutorial that assumed too much, and two characters dominated the meta, reducing match variety. Antihero Studios identified onboarding and balance as the top priorities for the upcoming beta.

Not everything worked. The new player experience was rough. We lost people in the first two minutes because the tutorial assumed too much. Players who'd never touched an extraction game didn't understand why they should care about Relics or what extraction even meant. We watched new players wander around the map aimlessly, get killed by monsters, and quit. The ones who survived that initial confusion became the ones logging 50 minutes a day. So the game worked. The on-ramp didn't.

The character balance was also off. Two of the characters were dominant and everyone knew it. Matches started to feel samey when half the lobby was running the same build. This is fixable, and we expected it, but it reminded us that even a game built on social unpredictability needs mechanical variety to stay fresh.

What It Means

The MISFITZ pre-alpha proved that extraction gameplay works on mobile, with data, player behavior, and community response all confirming the thesis. Antihero Studios is now building toward a beta with improved onboarding, more content, and tighter character balance, while the core loop remains validated.

The pre-alpha answered the question we'd been carrying since that dinner in Helsinki. Can extraction work on mobile? It can. The data says yes, the behavior says yes, and the angry Discord messages when we turned off the servers say yes louder than any chart.

We're heads-down now on the beta. Better onboarding, more content, tighter balance. The bones are right. The bones were always what we were testing. Everything else is craft, and craft is what this team does.

If you want to read about why we started this studio in the first place, that story is here. And if you want to see how MISFITZ compares to other extraction games on the market, we put together a comparison.

Last updated: April 1, 2026