Peephole interaction occurs when a spatially aware display is moved and acts as a viewport to reveal different parts of the virtual space that cannot all fit within the display at once. We investigate pointing within this peephole metaphor, where the targets may not be initially visible on the display, but are dynamically revealed by moving the display. We develop and experimentally validate a quantitative model for peephole pointing. Our results indicate that the model accurately accounts for peephole pointing for a variety of display sizes, both with and without users' having prior knowledge of the target location.