Derive the maximum stable learning rate for a quadratic objective and classify convergence from the chosen rate.
