I worked as a research assistant in physics 20 years ago and the opinions then was treating simulation, experiment, theory as 3 pillars for physical science. Each one has its own advantage and disadvantage in dealing with real world problem. But most of time, experiment is the only way to disapprove theoretical model/computer simulation.
Human beings tend to become overconfident in thoughts. And I think scientific methods tries to make sure that overconfidence won't mislead our judgement.