Arthur Koestler, athinker about the process of thinking, suggested that "bisociation," an ability to "zoom in" and "zoom out" on a problem, is critical to discovery.
Richard Foster, whose paper "Who Designed Brunelleschi's Dome?" I recently read, suggested that elite chess players have this ability. They can expertly marry these two processes: "zooming in" on a large number of remembered chess moves from a storehouse of games in order to deduce the next logical move, while "zooming out" to pursue a strategic goal, which requires being able to recognize emerging patterns with very little data. I wonder how much of this is talent and how much can be incorporated as a deliberate and ongoing strategy.