> You need just one contradicting observation to invalidate something No, that's wrong. There are several reasons why a negative result doesn't necessarily dis

> You need just one contradicting observation to invalidate something No, that's wrong. There are several reasons why a negative result doesn't necessarily disprove a theory or a hypothesis: measurement uncertainties/noise, experimental errors, software bugs, bit flips, ...