1: \begin{abstract}

2: We prove the exact worst-case convergence rate of gradient descent for smooth strongly convex optimization, with respect to the performance criterion $(f(x_N)-f_*)/\Vert x_0-x_*\Vert^2$. This rate was previously conjectured in \cite{drori2014performance,taylor2017smooth}.

3: \end{abstract}

4: