Islands in the Machine: A Systematic Geographic Assessment of Insular Territories in Video Games

Islands occupy a privileged position in human imagination and in the geography of the real world. As circumscribed, ecologically distinct, and often isolated territories, they concentrate geographic diversity into bounded spaces that invite exploration, narrative, and control. These same qualities make islands recurrent settings in video games, where the island serves simultaneously as a spatial container, a design convenience, and a cultural signifier. Yet despite the ubiquity of island environments in interactive media, no systematic attempt has been made to assess how faithfully video games represent the specific geographic properties that define insular territories. This paper presents the first comprehensive geographic analysis of island representation in video games. Drawing on a catalogue of 125 titles spanning multiple genres, platforms and release decades, a seven-block variable framework is developed, covering insularity, geomorphology, climate, hydrography, biogeography, human elements, and spatial scale, and computing an Island Geographic Fidelity Index (IGFI) for each title. Restricting the main analysis to 96 genuinely insular games and using 28 continental titles as a comparative reference, a mean IGFI of 42.2 (SD = 10.8) is found on a 0–100 scale, with Microsoft Flight Simulator 2024 serving as a benchmark reference at 98.8. Results show that insularity recognition (IGFI-A: 69.1) and human elements (IGFI-F: 43.2) are substantially better represented than island-specific climate (IGFI-C: 32.8) and biogeography (IGFI-E: 28.5), revealing an islandness paradox: games capture the form of insular geography but not its ecological and climatic substance. Kruskal-Wallis analysis reveals significant differences by genre (H = 22.04, p = 0.0002) but not by island type or geological origin, implying that gameplay mechanics shape geographic fidelity more than referential accuracy. A significant positive Spearman correlation (ρ = 0.366, p < 0.001) between spatial compression and IGFI challenges the intuition that scale distortion deteriorates geographic quality.