I wanted two temperature sensors to be combined to get a median value.
Maybe it is my bad habbit, but in most of these situations, I create a template sensor for this.
Then I thought it might be more efficient to use a pre-defined integration like min/max sensor, because the hard work is not done by jNinja?. So i looked how to do this. it seems the min/max sensor is kind of replaced by the statistic sensor? Oh wait it seems now available as group sensor!
All implementations however support some sort of setting precision, because the calculated result might have many decimals. However, the group sensor does not have a precision setting for the sensor type. I would suggest to add that!