Early-Stage Resource Economy Calibration in a Shelter Survival Game through Scripted Simulation

Abstract

Aims: This study examines resource economy calibration in an early-stage shelter survival game through scripted simulation. It asks whether different resource support configurations can produce distinguishable survival outcomes as environmental pressure increases, before formal player testing is introduced. Study Design: A 3 × 3 scripted simulation design was used. Three resource support modes, Baseline, Ability Weighted, and Facility Supported, were combined with three survival pressure levels, Low, Moderate, and High. Place and Duration of Study: The scripted simulation was implemented and executed in a Python computational environment during 2026 before manuscript submission. The study did not require a physical laboratory, field site, or human participant setting. Methodology: The simulation focused on three persistent shelter variables: Food, Morale, and Security. Each of the nine experimental conditions was evaluated with the same set of 20 fixed random seeds. Each run continued for a maximum of 25 simulated days, producing 180 runs in total. Survival Days, Completion Rate, and Collapse Cause were retained as the main outcome measures. Results: Under Low Pressure, all three resource modes completed the 25-day simulation in every run. Under Moderate Pressure, Completion Rates were 75%, 85%, and 95% for Baseline, Ability Weighted, and Facility Supported modes, respectively. Under High Pressure, Baseline and Ability Weighted modes recorded no completed runs, with mean survival durations of 17.95 and 19.30 days. Facility Supported Mode reached a mean survival duration of 24.75 days and an 85% Completion Rate. Collapse Cause showed a further pattern: increasing Food gain and Security mitigation removed Food depletion as a direct failure cause in Ability Weighted Mode, but Morale became the dominant bottleneck. Conclusion: Scripted simulation provided early feedback on resource balance in the developing shelter game. The tested parameters should be read as a provisional demo configuration rather than a finalised game economy. Even within this limited setting, the simulation distinguished stable, fragile, and bottleneck-dependent configurations. It also showed that strengthening selected resources may shift the dominant failure mechanism rather than eliminate instability. Scripted simulation can therefore serve as a practical calibration step before formal player testing.