Accurate control parameter optimisation and power production assessment is essential to evaluate the performance of a wave energy converter (WEC). However, commonly used numerical models excessively simplify the power take-off (PTO) system of the WEC, which may strongly affect power production predictions. Therefore, the present paper compares a commonly used WEC model that includes nonlinear viscous losses and an ideal PTO system model, referred to as NLideal, with a highfidelity wave-to-wire model (HFW2W) model. Results show the incapacity of the commonly used NLideal model to accurately optimise control parameters, particularly using reactive control. Likewise, the annual mean power production (AMPP) predicted using the NLideal model is significantly overestimated, with differences of up to 160% with respect to the more realistic HFW2W model. More dramatically, the use of control parameters optimised with the NLideal model in the HFW2W model results in negative AMPP, meaning that the WEC consumes more energy than it produces.