The Spring-Loaded Inverted Pendulum (SLIP) model poses a challenging control problem, important for the development of legged robots, due to the difficultly in solving the stance phase of the dynamics. Multiple attempts have been made to approximate these dynamics to allow for an analytical control method; here four of these methods have been compared for controlling agile hopping, where there are large changes in forward velocity across a single stance. In addition, a new, empirical, approach has been demonstrated. In this, a simple control law is formulated, based on some simple approximations, which allows the parameters to be selected empirically through simulation. This has led to a controller able to offer similar performance to the best analytical approximation but with a much simpler form. This empirical controller may present new opportunities for controlling more complex dynamics and the development of a self-tuning method in future work.