The flash-lag effect, in which a moving object is perceived ahead of a colocalized flash, has led to keen empirical and theoretical debates. To test the proposal that a predictive mechanism overcomes neural delays in vision by shifting objects spatially, we asked observers to judge the final position of a bar moving into the retinal blind spot. The bar was perceived to disappear in positions well inside the unstimulated area. Given that photoreceptors are absent in the blind spot, the perceived shift must be based on the history of the moving object. Such predictive overshoots are suppressed when a moving object disappears abruptly from the retina, triggering retinal transient signals. No such transient-driven suppression occurs when the object disappears by virtue of moving into the blind spot. The extrapolated position of the moving bar revealed in this manner provides converging support for visual prediction.