Disorientation in first-person, screen-based 3D games breaks flow, yet most evidence comes from head-mounted display Virtual Reality or abstract tasks. We ask what minimal in-world information stabilizes orientation in visually repetitive worlds. In a within-participant mixed-methods study (N=20), a symmetrical game environment orthogonally varied landmark geometry and featural polarity, plus a no-landmark baseline. In-situ pointing and orientation ratings during learning and subsequent relocation, together with trajectory analytics and interviews, show that coupling geometric structure with strong featural polarity is decisive: relative to baseline, localization error drops (approx. 79%) and pointing accuracy rises (approx. 64%), whereas geometry alone helps less and smooth, featureless forms least. Players anchor on discontinuities (edges/corners; black--white boundary) and triangulate with distances and angles; without an anchor they revert to ground/light patterns and report symmetry-induced confusion. We establish a minimal-cue testbed for desktop play and derive actionable guidance: break rotational symmetry in a persistent landmark to enable legible, low head-up display environments.
ACM CHI Conference on Human Factors in Computing Systems