The weight of evidence for an extraordinary claim must be proportioned to its strangeness.
The best decisions are made when you listen to both data and people.