In research, what you failed to arrive at matters more than what you did — dead ends map the problem space and prevent retracing the same steps.