Pandas Rolling Conditional Function

Question

I\'m having some trouble to use .apply or .aggregate in pandas on a rolling basis (assuming of course that it is the right way to solve my problem)