We present an analysis of the accuracy and cost of energy rating of photovoltaic modules. We identify the prominent sources of uncertainty and demonstrate that good estimates of energy rating can be made with a reduced set of measurements, thereby reducing cost. The energy rating standards IEC 61853 parts 1-4 provide a method for differentiating the expected performance of photovoltaic modules under real-world conditions. It combines a comprehensive set of measurements on modules with a numerical model to produce performance ratings for different reference climates. We have developed a simulation tool to explore the sensitivity of energy rating to various factors, including an uncertainty model developed from a survey of accredited test and calibration laboratories. The set of 22 performance-matrix measurements required by the standard do not span the full range of conditions experienced in the reference climates. Nevertheless, we find that extrapolation and interpolation of the measurements is sufficiently accurate for different module types. Indeed, the number of measurements can be significantly reduced without adversely affecting accuracy. We find that the overall accuracy of energy rating is similar to that of power output at standard test conditions. A sensitivity analysis identified the most important sources of uncertainty to be the measurement of irradiance and the nominal module operating temperature.