Overview
You are responsible for a realm you cannot see clearly.
Trust is a strategy game about exploring decisions under uncertainty. Every piece of information is treated as a claim with a confidence level. That includes reports from within a player’s own realm as well as claims from institutions and other realms. There is no private source of unquestioned truth.
The goal is to keep a realm from descending into chaos or becoming unstable, then help it prosper. Stability is a measurable part of the world, but the game does not prescribe how to achieve it. How players respond to uncertainty, pressure, and one another is the interesting part.
The loop has three verbs: read, say, do. Players interpret the reports, state what they believe, and decide how to act. When the round ends, the world moves and those claims are eventually tested against what actually happened.
The challenge
Trust creates room for questions that do not have predetermined answers:
- Can you separate what you know from what you merely suspect?
- When is an uncertain report enough to justify action?
- Which sources become useful or reliable over time?
- How does stated confidence change the way others respond?
- What kinds of coordination emerge when realms need one another?
The game does not define a correct strategy. Players may share what they know, withhold information, bluff, cooperate, or act alone. The point is to discover which patterns lead toward stability and prosperity, and how those choices look once more of the world becomes known.
I hope integrity and collaboration prove valuable, but that conclusion belongs to the players. The setting remains abstract and non-partisan, and meaning comes from the simulation and the post-mortem rather than from a prescribed lesson.
The approach
Each player acts for a realm whose stability, security, welfare, infrastructure, legitimacy, and capacity are under pressure. Players can make promises, share resources, report what they believe, cooperate, or pursue their own strategies. Those choices shape both the realm and whether others continue to trust and work with them.
Play is asynchronous across days or weeks. Rounds resolve on a clock whether or not everyone acts. Human players can share a match with scripted or agent-driven realms, allowing a game to begin without waiting for a full group and continue when someone drops out.
The simulation separates the world’s actual state from the claims available to each player. Every internal report and external message belongs to the same information model: a claim, its source, and a confidence level. This allows the game to reveal an uncertain world consistently without giving any player information their realm could not know.
Trust is an Elixir and Phoenix LiveView application backed by PostgreSQL, built around a pure functional simulation core and a thin effectful shell. Claims form an append-only history; world state, obligations, trust, and the final post-mortem are derived from that history.
The domain owns the rules and deterministic simulation. OTP processes coordinate live matches and round deadlines, while Phoenix LiveView presents each player with only the information their seat is allowed to see.
What I’m building
- An asynchronous strategy game built around uncertain information
- A claim and confidence model for all internal and external information
- A simulation that separates ground truth from what each player is told
- Stability, chaos, and prosperity as outcomes shaped by player decisions
- Systems for cooperation, commitments, shared resources, and trust
- An append-only history for comparing claims, confidence, actions, and outcomes
- Matches that can include both human and agent-driven players
Trust is currently in development.