Why several observations?
Statistical observation

It is rare for the result of an observation – of a measurement, for example –, to be exactly equal to the "theoretical" value calculated in advance by the engineer. Frequently, it is essential to explain the origin of the discrepancy. The measurement uncertainties described at the beginning of every elementary physics textbook (those that are experienced and those that are tolerated) do not explain everything.

To establish a physical law, it is customary to require that measurements carried out by two different experimenters lead to the same result. Observations are understood to be "all other things being equal". This banal expression is highly charged with meaning: all phenomena that might occur at second or third order are assumed to be fixed. In industrial and commercial practice, observation procedures are designed to minimize the influence of these second-...