Two experiments for the price of one? The role of the second oscillation maximum in long baseline neutrino experiments