The larger the required reliability level the larger are the sets that constitutes the possible measurement results.
How to select the reliability level depends on the consequence of wrong measurement results.
The predictions must be determined in a way that on average the measurement results have a minimum size.
Pictured above, a typical measurement result is shown.
Accounting technique, i.e. how measurement results are produced, can differ.
According to it the measurement result of an observable depends on the arrangements made to measure it.
If the object is not homogenous, then the measurement result may include changes in both quantity and quality but their respective shares will remain unclear.
A disadvantage is their low precision: measurement results may differ 10% and more.
Only the first measurement result is (potentially) random.
If some of these requirements are not satisfied then the measurement results will be inadequate or fail completely.