There is a kind of seemingly nonsensical play behavior found in the simulation sandbox game genre. This behavior is very spontaneous and impulsive and associated with self-initiated learning, and here the author seeks to better understand what it is, and why it seems associated with simulation sandbox games. That is: What purpose does it serve, and what might provoke it? This requires a review of the literature on this play behavior, and that of the simulation sandbox genre, respectively. From this review the author concluded firstly that there is strong support that exploratory play is for discovering the structure and behavior of systems, and secondly that the observable characteristics of exploratory play make it a highly probable candidate for the bizarre behavior observed in simulation sandbox games. Moreover, several hypotheses were generated by identifying many characteristics of the genre (e.g. system complexity and responsiveness) that are directly relevant to the theorized motivations for exploratory play, suggesting some directions for future research into what conditions and designs might encourage exploratory play. Knowledge of the relationship between this genre and this form of play could prove invaluable for designing games for learning, because despite being centrally relevant to many studies on game-based learning, exploratory play has been neglected. It has been needlessly isolated in distinct strands of research on its components, which will be unified here to provide a comprehensive account of this behavior and its importance to future research in this area.