Cardinal and ordinal only then. Derive the vector from screen centre towards mouse cursor and snap it to cardinal and ordinal values only. I also noted that you’re moving just the camera component - is this intended? One would usually have a camera pawn and move it around.