With the 1-3 mixing angle measured at reator neutrino experiments, Daya Bay and RENO, there are still three unknown oscillation variables, the neutrino mass hierarchy, the octant of the atmospheric mixing angle, and the leptonic CP phase. Of these three, the CP phase is the most difficult to measure precisely and important for distinguishing flavor symmetries. In this talk I will first talk about residual symmetries, which takes the role as low energy effective field theory for neutrino mixing and is analogy of custodial symmetry in the electroweak sector, to predict the CP phase from the measured mixing angles and then introduce a new type of neutrino experiment to measure the CP phase. Currently, T2K and NOvA are designed for this measurement. Nevertheless, they suffer from degeneracy and efficiency problems. The situation can be improved by adding a muon decay at rest (muDAR) source. With T2(H)K running in neutrino mode and muDAR in anti-neutrino mode, both using the same detector, the CP measurement becomes more precise and can break the degeneracy between \delta and 180^o - \delta. The same configuration can also be applied to next-generation medium baseline reactor neutrino experiments like JUNO and RENO-50, enhancing their physics potential. With only one source and no extra detectors, this dessign is much better than DAEdLAS which requires 3 sources, but only 20% duty factor and 4 times higher luminosity for each.