A multigrid algorithm for isotropic neutron transport is presented in x-y geometry, using discrete ordinates in angle and corner balance finite differencing in space. Like Diffusion Synthetic Acceleration (DSA), this algorithm employs transport sweeps, followed by a correction from a coarse angular subspace. While DSA uses a Galerkin P-1 closure to form the coarse-grid system, the method presented here uses a scaled least squares minimization principle. The goal of the minimization principle is to avoid the problems of consistent differencing associated with DSA. The algorithm is viewed as a two-grid scheme in angle, where sweeps are the relaxation, and the coarse-grid system is obtained from the scaled least squares minimization. The coarse-grid system is subsequently solved with standard spatial multigrid. Numerical results are presented.