How it Works — Each Interstellar Object (ISO) is an agent on a 2D plane with neighbors. At each time step, agents may acquire new evidence cues (anomalies) based on set prevalences, influenced by observation intensity, observation-window loss, and social influence from higher-level neighbors. Evidence persists, and the agent’s Loeb level is updated per the rules on the right.
Social Influence — For undetected features, detection probability per step is multiplied by 1 + influence × (fraction of high-level neighbors). For large agent counts, neighbors are randomly sampled for performance.
Presets — Oumuamua ≈ strong NG + no coma + odd shape (~Level 4); Borisov ≈ natural (~Level 0); 3I/ATLAS ≈ mixed cues with retrograde alignment (~Level 4).
Tip — Use Step for manual advances; Start/Pause for continuous simulation. Download CSV exports current agent states and reasons.