This strategy evaluates several hypotheses at each time step and eventually chooses the hypothesis that
has the overall highest probability for the entire sequence.