This work compares and quantifies the annual losses for three battery system loss representations in a case study for a residential building with solar photovoltaic (PV). Two loss representations consider the varying operating conditions and use the measured performance of battery power electronic converters (PECs) but differ in using either a constant or current-dependent internal battery cell resistance. The third representation is load-independent and uses a (fixed) round trip efficiency. The work uses sub-hourly measurements of the load and PV profiles and includes the results from varying PV and battery size combinations. The results reveal an inadequacy of using a constant battery internal resistance and quantify the annual loss discrepancy to −38.6%, compared to a case with current-dependent internal resistance. The results also show the flaw of modelling the battery system’s efficiency with a fixed round trip efficiency, with loss discrepancy variation between −5 to 17% depending on the scenario. Furthermore, the necessity of accounting for the cell’s loss is highlighted, and its dependence on converter loading is quantified.