A simple way to generate low energy phase shifts for elastic scattering using bound-state calculations is postulated, validated, and applied to the problem of e -Mg scattering. The essence of the method is to use the energy shift between a small reference calculation and the largest possible calculation of the lowest energy pseudostate to tune a semiempirical optical potential. The ' 1 partial wave for e -Mg scattering is predicted to have a shape resonance at an energy of about 0.13 eV. The value of Z eff at the center of the resonance is about 1500. DOI: 10.1103/PhysRevLett.98.173001 PACS numbers: 31.25.Jf, 03.65.Nk, 34.80.Bm, 34.85.+x One of the most technically demanding problems in quantum physics is the scattering problem, i.e., the prediction of the reaction probabilities when two objects collide [1]. The underlying difficulty lies in the unbounded nature of the wave function. This leads to a variety of computational and analytic complications that are simply absent in bound-state calculations, e.g., the Schwartz singularities that occur in the Kohn variational method for scattering [2,3].One approach to solve scattering problems is to use bound-state methods. There are many examples of such approaches, one of the most popular being the R-matrix methods that use the solutions of the Schrödinger equation in a finite sized cavity to determine the behavior of the wave function in the interaction region [1]. The total wave function is then constructed by splicing the inner wave function onto the asymptotic wave function.However, despite the considerable activity in this area, there are a number of problems that are beyond resolution. The e -atom problem is a notoriously hard numerical problem since the atomic electrons tend to localize around the positron, thus giving a very slowly convergent partial wave expansion of the wave function inside the interaction region (this should not be confused with the partial wave expansion of the asymptotic wave function) [4 -7]. For example, the dimensionality of the equations to be solved to achieve a given accuracy are about 5 times larger for e -H scattering than for e ÿ -H scattering. At present, there are a number of positron collision problems that are simply inaccessible with existing approaches [7].This Letter had its origin in a particular scattering problem, namely, the determination of the near threshold phase shifts for positron scattering from magnesium. The dimensions of the secular equations for bound-state calculations on group II atoms are very large, for example, a configuration interaction (CI) calculation of the e Ca 2 P o state had equations of dimension 874 448 [8]. The idea behind the current method lies closest to the box R-matrix method [9] which is exploited in quantum Monte Carlo (QMC) calculations of scattering [10]. In the QMC calculations, one extracts the phase shift by comparing the zero point energy of a finite size cavity to the energy of the system wave function in the same cavity. In the present method, the phase shift is extracted f...