A GitHub repository by skorotkiewicz presents a Three.js experiment comparing two search strategies for finding someone in darkness. The signal-in-the-dark simulator models a sweep searchlight where a rescuer moves through an expanding pattern to illuminate a missing person, versus a raise beacon where the rescuer stays stationary while a mobile person approaches an omnidirectional light.

Users can test presets: Mobile hiker scenarios typically favor the beacon approach, while Injured persons make stationary beacons ineffective since they cannot move toward the light. Dense forest environments increase line-of-sight occlusion and produce mixed results. A Monte Carlo comparison runs both strategies against identical random worlds and starting positions after each parameter change.

The author emphasizes this is a conceptual toy model, not validated search-and-rescue guidance. It intentionally isolates few variables while omitting weather, terrain elevation, battery limits, sound, radios, group behavior, panic, injuries, and official rescue protocols. The page imports Three.js from a pinned CDN URL requiring internet access on first load.