Preprint of journal submission. Abstract – The most established key performance indicator for (PV) system performance is the performance ratio (PR) metric, which is the ratio of the actual to the expected specific yield of a PV system. Accurate PR calculations are crucial for the PV systems industry, as errors can potentially lead to hidden underperformance, unfair financial penalties and uncertainty in system value. In systems with high DC:AC ratios, the regular occurrence of inverter clipping causes problems with PR assessment which include masking degradation and increased seasonal variation. A common practice is to exclude clipped data from PR calculations, but this can bias the remaining dataset and masks problems with inverters. In this work, a new improved clipping-corrected PR (CCPR) metric is introduced and evaluated. CCPR takes into account clipping and weather variability, and allows the use of all datapoints during a PV system’s operation. We introduce the new metric and evaluate it for different locations around the world using synthetic data. Different types of clipping ratios and faults of the system are tested, demonstrating that the new PR metric is more robust compared to the current standardised PR metrics, and is not affected by weather conditions. In addition, we show that the new metric is robust against the use of low temporal resolution Typical Meteorological Year data for calculating expected PR, in contrast with standard PR metrics. The widespread adoption and potential standardisation of this new metric can lead to more accurate PR assessment for PV systems, reduced contractual risks for companies in the sector and increased in confidence of PV generation by consumers.