adding extra logic so we can uniquely predict situations where a prediction has a much higher score than the others?

[dieuska]

This example gives 3 predictions (TO CHECK), with a prediction score that is much higher than the others.

We can search for a way to uniquely define this best prediction? Or does the user has to make this choice (programmaticaly?

[dieuska]

maybe excluding low prediction scores if there are much higher ones (but this can resulting in loosing a special prediction that the user is searching for). We can also augment the steps in relevent distances from 0.1 to 0.2, so the lowest prediction score will become 20 (and all predictions that were 10 will be excluded as these has a very low stanbility range of 10cm)