Peer skill ratings
After a match, you rate the people you played with. One tap each, under thirty seconds. Those ratings settle into a profile that is genuinely yours rather than a guess.
Halbreven is a closed app for the group of people you already play with every week. It picks fair teams, remembers who actually turned up, and keeps a record of it that follows a player from season to season.
The problem
Somebody posts "Sunday 10, who's in?". Twelve replies arrive over three days, four of them are jokes, two people say maybe and never come back. On the morning, one person scrolls up trying to count. Teams get picked by whoever shouts loudest, and they are lopsided about a third of the time.
Nobody remembers that the same player has said yes and not shown up four weeks running, because a chat thread has no memory. It scrolls.
Halbreven replaces that thread, and adds three things a chat app cannot do.
After a match, you rate the people you played with. One tap each, under thirty seconds. Those ratings settle into a profile that is genuinely yours rather than a guess.
The app splits whoever turned up into even sides using those ratings and the positions people actually play. It takes a second, and nobody has to be the one who picks last.
A trust score that falls fast when you say yes and do not show, and recovers slowly when you do. Yellow and red cards for the things a group already argues about. Visible, so the group can decide.
Why build it this way
These are written down before features are, and they are the reason this is not another social sports network. If something breaks one of them, it does not ship — that has already killed ideas that were otherwise good.
No public profiles. No search for people or groups. You get in because somebody handed you a code.
A full post-match pass in under thirty seconds. Complexity kills participation, and a rating system nobody uses is worse than none.
Trust falls fast and recovers slowly, the way a credit score does. That asymmetry is the whole mechanic.
The app will not ban anybody. It shows the history and the group decides what to do about it. People, not software.
The timeline is for the game. Everything else belongs in the chat people already have.
No paywall on anything that matters. If a group has to pay to see who is coming on Sunday, the product has failed.
You see your own scores. You never see who gave you which one — not as a player, not as an admin, not in any export.
This app is for friend groups. If Ardit can see that Blerim rated him a four, it has manufactured a real argument out of a Sunday kickabout. The whole premise depends on aggregation providing cover, which is why it is enforced in the database rather than in the interface. How that works.
Who it is for
Anything where a standing group turns up, splits into sides on the day, and plays for a score. That is a much wider net than football, and the app draws each sport's real playing surface rather than a generic rectangle.