Discuss if ComparerCollection is a good way to evaluate inner metrics. #5
Description
Currently, the inner metrics are calculated by combining the simulation-observation pairs in a modelskill ComparerCollection.
However, this could potentially cut off data of observations, if one of more of the included observations are available in a shorter time periods than the simulation time. In that case the ComparerCollection will evaluate only on the mutually matching time range.
We have to think about, if this is a behavior that we want, or if we should switch to individual evaluation of the simulation-observation pairs.