It must decrease performance from two reasons.
1. we must run synchronization mode (game thread and video thread are synchronized)
2. we look for pixels in big array. Let’s say game has resolution 320×200, this is 64000 pixels, each pixel has two bytes. So we search in 128k big array – every frame. This search is executed for every “trap pixel”.
Both operations are time hungry. Is possible skip point 1 and run traps without synchronization mode, but my rough guess is around 5-8% failure if we set one pixel in pixel trap and maybe 50% failure if we set two or more pixels.