The Things We Learned on Liberty Island: Designing Games to Help People Become Competent Game Players

Oliver and Pelletier's study examines how players learn to become competent at video games, using a methodology grounded in Activity Theory and detailed transcripts of play rather than players' self-reports. Its central argument is that competence develops through play itself, and that careful analysis of how players construct strategies can reveal which design features help or hinder playability. The work positions this kind of observational, empirically grounded analysis as both feasible and valuable, an approach the authors argue is largely absent from games studies research. The paper was presented at the 2005 DiGRA conference and centers on a small, qualitative case study rather than a broad population survey.

The core investigation involves two players of the 2000 action-adventure game Deus Ex, one of whom completed the game's training level while the other skipped it. Across these sessions the researchers identified ten strategies for the training level and twelve for the first mission, showing that strategies evolve continuously during play and that Deus Ex is considerably more complex than the earlier pilot game Harry Potter and the Philosopher's Stone, which required only six. The player who trained transferred some skills successfully, such as searching bodies, jumping, and aiming, while failing to transfer others, such as disarming gas bombs and noticing datacubes, and misapplying certain tactics in fast combat. The player who skipped training developed a 22-rule strategy set by the end of the session, gradually shifting from direct assault toward stealth, though this path slowed learning and left basic controls unmastered.

A consistent finding across the analysis is that prior gaming experience both helps and hinders play. Experienced players bring useful expectations but also apply habits that do not transfer to a new game's object rules. From these observations the authors draw several design recommendations: genre-sensitive game openings, visual or behavioral cues that signal shared rules for objects, risk-managing features such as quick save that support trial-and-error learning, and empirically grounded metrics for classifying games and measuring difficulty. The authors further suggest that the strategies players actually use could be encoded as more plausible behaviors for non-player characters. Because the evidence rests on two participants, the conclusions are best read as illustrative of the method's potential rather than as generalizable claims about all players.