Evaluating the financial value and decision-making benefits of sequential learning algorithms (SLA) for medium range gas demand forecasting and hedging during UK winter periods: Case study for British Energy Markets

Natural gas remains the dominant source of winter heating demand in the UK, with 85–87% of homes relying on gas-powered central heating boilers. Accurate gas demand forecasting is therefore critical for ensuring supply reliability, optimising operational costs, and informing both short- and long-term infrastructure planning. In the medium and short term, temperature is the single most influential driver of demand variability relatively small forecast errors can translate into significant supply-demand imbalances and price exposure. As a result, improved weather forecasting has become a strategic priority for energy retailers seeking to manage price volatility and maintain supply adequacy. In this study, we examine the financial impact of incorporating sub-seasonal to seasonal (S2S) weather forecasts generated using a Sequential Learning Algorithm (SLA) and benchmarked against ECMWF outputs into demand forecasting and hedging activities for British Energy Markets, across lead times of weeks 3 and 4 for the winters of 2020–2025. SLA outputs are used to forecast Non-Daily Metered (NDM) demand covering residential and small business consumers using a Composite Weather Variable (CWV) that demonstrates greater skill than climatological benchmarks at weeks 3 and 4 lead times. Four demand estimation approaches are evaluated, spanning both deterministic and probabilistic SLA frameworks, to assess their relative value for hedging decision-making and financial risk management. The net cost associated with each approach is quantified by comparing gas procurement contracts across daily, weekly, and monthly time horizons. Thus, this study provides evidence that S2S weather forecasts, generated through a Sequential Learning Algorithm, offer measurable financial value for gas demand forecasting and hedging decision-making at sub-seasonal lead times with implications for how energy retailers manage supply risk and procurement costs during winter periods.