Abstract: Making predictions (even statistical) for future measurements becomes a highly nontrivial theoretical challenge if the underlying fundamental theory of consideration allows for more than one solution. String Theory with its large landscape of solutions is a prime such example. After explaining why it is difficult to construct a measure on the space of the outcome of future measurements, a possible measure for the multiverse is presented. It turns out that the beginning of the universe plays an important role for making predictions. An overview over old and new proposals thereof will be given. Event recording: