Trust Layer
Deterministic Arena
Determinism is a trust layer for BAIS. The same state and the same inputs should produce the same outcome, making learning inspectable and replayable.
Quick Facts
- The core engine avoids hidden randomness.
- Tie-breaks and graph ordering are documented.
- Determinism supports debugging, replays, and future persistence.
- Presentation layers must not contaminate deterministic behavior.
Why Determinism Matters
A brain game needs trust. If the AI changes for invisible or random reasons, players cannot understand what they trained. BAIS protects deterministic behavior so changes can be inspected and repeated.
This means the same state and the same inputs should lead to the same result. That makes the brain feel like a system, not a magic trick.
What It Enables
Determinism enables useful debugging, reliable replays, fairer comparison between brains, and future persistence. A saved brain should restore into the same structure and behavior rules that made it meaningful.
It also helps future tournaments and spectators because people can analyze why something happened rather than arguing that the system simply rolled a hidden dice.
How BAIS Protects It
The current BAIS core keeps deterministic ordering, explicit ticks, documented tie-breaks, no hidden RNG, and separated presentation layers. Unity visuals should display the core state without rewriting it.
This is one reason the project is building the engine foundation carefully before expanding into full arena presentation, economy, or multiplayer systems.