Your P Values Are Wrong

摘要

Typical solar photovoltaic (PV) industry methodologies for calculating energy production probability of exceedance values (P values) for PV power plants do not provide reliable or actionable results for risk management purposes. By excluding common sources of major plant underperformance, such as availability and curtailment, and by ignoring asymmetries and non-linearities in system performance, typical P value methodologies practically guarantee an overly optimistic prediction of PV plant performance. While we suggest ways to improve the accuracy and usefulness of P values, we emphasize that long-term PV plant owners should prioritize contractual management of tail risks over predictive analyses.