Playing games of any kind, from tennis to board games, it is easy to notice that games are configured in space, often using stripes or a kind of map on a board. Some games are clearly performed within this marked border, while it may be difficult to pinpoint such a visual border in a game like hide-and-seek, even though this game is still spatially configured. The border (visible or not) both seems to separate and uphold the game that it is meant for. Johan Huizinga noted this “separateness” in his classic work “Homo Ludens” (Huizinga 1938, translated into English 1955). This has since been developed into the concept of the “magic circle” by Salen and Zimmerman (2003), as an understanding of playing games as a kind of alternate reality. When a person enters the magic circle of a game, the player suddenly finds himself in another world, where artefacts are given new meaning and where other rules apply. This makes sense, but also demands that play and non-play can be easily separated. I will examine how games make use of space, and show that the magic circle not only is a viable, though criticized, concept but should be understood as a spatial concept. In order to do this several games are examined, leading to introduce a spatial model of the game performance comprising a primary and secondary game space. I will show how new game genres can profit from using this model when designing new games.