Catch That Train! — A Journey Under Pressure

Movement, handling, waiting and metro travel have different time rules. The engine separates actual input time from fixed task costs, then derives recovery from time already charged. Route forecasts and journey budgets reuse these rules to keep the interface and outcomes consistent.

Six entrance scenes in the Hongqiao student chapter require the player to reach a destination while avoiding crowds and railings. Small motion steps prevent crossing obstacles. Forecasts and simulations use the same collision geometry; other travelers and stations retain their existing event interactions.

Held movement ends on pointer cancellation, lost capture, keyboard release, blur or a hidden page. Completed entrance actions are guarded against repetition, and event revisions reject stale choices. These boundaries keep interrupted gestures from continuing into the next task.

English and Chinese presentation share canonical event data and action payloads. Language state is separate from game state, so switching preserves the active task and progress. Dynamic costs, results, challenge text and exported images follow the selected language.

Run IDs prevent duplicate achievement awards. Stored profiles are parsed, migrated and merged so an older browser tab cannot remove earned unlocks. Records stay in the current browser; shared links contain the station, title and outcome without uploading the player’s saved history.

An injectable random source supports seeded tests and journey simulations. Reports record the character, station, mode, input model and decisions, making rule changes easier to compare. Simulation results describe the tested strategies and assumptions, rather than measured human win rates.