Energy Price Forecasting Software: Day-Ahead and Real-Time Power Prices
What is power price forecasting?
Power price forecasting predicts day-ahead (DA) and real-time (RT) locational marginal prices (LMPs) in wholesale power markets. Amperon’s price forecasts operate at the hub and zonal levels, as changing zone-level dynamics increasingly drive trading and hedging strategies.
Power traders use price forecasts to identify arbitrage opportunities and access short-term trading instruments. Asset operators use them to know when to charge and discharge batteries and to optimize bidding into day-ahead, real-time, and ancillary services markets. Utilities and retail electric providers use price forecasts primarily to manage costs and risk.
72-79%
Of perfect-foresight returns captured in RT forecasts
40%
More high-price events detected vs. competition
25-30%
More battery dispatch optimization value vs. competition
ERCOT DA Price Forecast
Current forecasts for day-ahead LMPs for all ERCOT hubs and zones, refreshed hourly. Visualize forecasted prices and trends up to two weeks out, along with demand, net demand, and solar and wind supply forecast components.
14
Days of forecasted DA prices
ERCOT RT Price Forecast
Predicts real-time LMPs every 5 minutes for all ERCOT hubs and zones. Includes current, historical day-ahead, and 2-hour-ahead forecasts. Visualize forecasted real-time prices up to 3 days out.
75%
Of perfect-foresight returns captured
PJM DA Price Forecast
Predicts day-ahead LMPs for all PJM hubs and zones, refreshed hourly. Visualize forecasted prices and trends up to two weeks out, along with demand, net demand, and solar and wind supply forecast components.
14
Days of forecasted DA prices
Other grids coming soon
Why net demand accuracy drives price accuracy
Many vendors predict prices in isolation using public data alone. Amperon’s price forecasts are built on its industry-leading grid net demand forecasts, continuously updated to incorporate the latest weather and grid conditions into model re-training.
One platform
Demand and price forecasts on the same platform mean no vendor mismatch and no reconciliation.
Demand feeds price
The industry-leading accuracy of Amperon’s Grid Net Demand Short-Term Forecast is a primary input to the price signal.
Hourly retraining
Amperon’s AI-powered electricity price forecasts retrain hourly to continuously incorporate changing weather and load patterns, genstack data, and observed price behavior.
Energy price forecasting FAQs
Here are some answers about our platform, implementation process and pricing.
What grids and zones do your price forecasts cover?
Amperon’s price forecasts currently cover all zones in ERCOT and PJM.
What data does Amperon use to forecast power prices?
Amperon’s energy price forecasts consider weather and net demand from our industry-leading grid demand forecasts, along with observed LMP data and continuously recalibrated generation stack data such as fuel mix, fuel prices, fuel consumed, heat rates, available capacity, and outages. Data comes from public sources such as EIA and ISOs along with third-party and proprietary data.
How does a price forecast improve battery storage dispatch optimization?
Battery storage revenue depends on dispatching at the right hours: charging when prices are low, discharging when they're high, and bidding into ancillary services markets when conditions favor it.
A price forecast improves dispatch optimization by giving operators a forward-looking signal that demand forecasts alone cannot provide, so that charge and discharge scheduling and ancillary services bids are informed by where prices are going rather than where they have been.
A price forecast improves dispatch optimization by giving operators a forward-looking signal that demand forecasts alone cannot provide, so that charge and discharge scheduling and ancillary services bids are informed by where prices are going rather than where they have been.
How do batteries affect price forecasts?
In day-ahead price models, the effects of battery behavior on price formation are inherently inferred from supply stack information. In real-time price models, actual battery discharge behavior is reflected in the forecast in a matter of minutes.
How do ML-based price forecasts compare to traditional methods?
Machine learning-based price forecasting trains on observed market behavior, whereas traditional methods rely solely on bottom-up price constructions from generation stack data. Amperon uses both dispatch-based price models and multiple machine learning-based price models, dynamically re-weighted as new load, generation, and weather data become available. Each model contributes unique predictive benefits such as peak hour predictions, volatility response to extreme behavior, and holiday behavior.
How do you evaluate power price forecast accuracy?
Amperon uses Mean Absolute Error (MAE) to track our day-ahead and real-time price forecast accuracy. Request a backtest to see historical accuracy results for your region.
