We give a worst-case to average-case reduction to solve BDD for certain parameter ranges by reducing the input to a random BDD instance of smaller dimension. The approach is quantum-inspired in the sense that we first give a quantum algorithm achieving this result. This algorithm is fairly elementary from a lattice perspective. For example, it does not need Gaussians. We also cover related classical algorithms.