Baseline Methods for Demand Response Programs

Published: March 8, 2026

X-of-Y (Highest Usage)

Best for programs that prioritize high-load representative days near event timing. This method selects top historical candidates by event-window usage and averages interval values.

Day-Matching Average

Use this when stability matters more than peak sensitivity. It averages all qualifying day-matched candidates and smooths volatility in single-day spikes.

Median Candidates

Ideal when outlier suppression is required. Median-per-interval selection reduces sensitivity to sensor errors or one-off abnormal loads.

EWMA (Exponential Weighted Moving Average)

Useful for dynamic load profiles where recent behavior should dominate. EWMA increases recency weighting and can adapt faster than simple averaging.

Selection Guidance

• Use X-of-Y for capacity-focused events where high usage historical patterns matter.
• Use median when data quality risk is moderate to high.
• Use EWMA when load shape is changing seasonally or operationally.
• Always constrain candidate selection to event hours for consistency.