During grad school, when stuck on a problem, I used to stare at it, frustrated. 

Over time, I learned to instead peel away complexity, solve a simpler problem, and then build up.