The goal of this MR is to streamline the reasoning for the last optimization and improve the notation.