LAUSR

When participating in electricity markets, demand-side responses (DRs) must forecast market prices to optimize their bidding strategies and operational schedules. However, conventional forecasting methods that prioritize metrics such as Mean Squared Error (MSE) or R-squared do not necessarily maximize the financial benefits for DR participants. To address this limitation, this paper introduces a novel profit-oriented forecasting criterion, Extremum Timing Accuracy (ETA), which evaluates the accuracy of capturing the timing of large price gaps while accounting for charging duration, an aspect overlooked in existing metrics. Building on ETA, a practical DR management strategy is proposed that integrates profit-oriented evaluation into a Seasonal AutoRegressive Integrated Moving Average (SARIMA) forecasting framework. Unlike complex machine learning methods that often require extensive external data, the proposed method relies solely on historical price data, making it both accessible and transparent to analyze. Using clearing price data from Japan’s day-ahead wholesale electricity market, the results demonstrate that the ETA-based strategy consistently improves DR profitability across different operational constraints and regional markets.