R2D2 - a symmetric measurement of reactor neutrinos free of systematical errors

We discuss a symmetric setup for a reactor neutrino oscillation experiment consisting of two reactors separated by about 1 km, and two symmetrically placed detectors, one close to each reactor. We show that such a configuration allows a determination of $\sin^22\theta_{13}$ which is essentially free of systematical errors, if it is possible to separate the contributions of the two reactors in each detector sufficiently. This can be achieved either by considering data when in an alternating way only one reactor is running or by directional sensitivity obtained from the neutron displacement in the detector.