Φ(x) = y for a monotone Φ, evaluate the midpoint, see whether it lands above or below the target height, and throw away the half of x that can't contain the answer.This is bisection: the same halving idea as before, generalized from integers to real numbers. It only needs Φ to be monotone — it never needs a formula for Φ-1, which is exactly why it works for functions (like this logistic curve) that have no closed-form inverse.