This is true for the toy example. In the actual problem, the loop guards are quadratic equations or even higher order, and his first method generalizes better (

This is true for the toy example. In the actual problem, the loop guards are quadratic equations or even higher order, and his first method generalizes better (and for small ranges is hardly slower).