The future smart grid (SG) will be populated with high penetrations of plug-in electric vehicles (PEVs) that may deteriorate the quality of electric power. The consumers will also be seeking economical options to charge their vehicles. This paper proposes an overnight maximum sensitivities selection based coordination algorithm (ON-MSSCA) for inexpensive overnight PEV charging in SG. The approach is based on a recently implemented online algorithm (OL-MSSCA) that charges the vehicles as soon as they are randomly plugged-in,while considering SG generation, demand, and voltage constraints. In contrast to the online approach, ON-MSSCA relies on inexpensive off-peak load hours charging to reduce the cost of generating energy, such that SG constraints are not violated and all vehicles are fully charged overnight. Performances of the online and overnight algorithms are compared to the modified IEEE 23 kV distribution system with low voltage residential feeders, populated with PEVs.