ENERGY PRODUCT // 2026
ENERGY PRODUCT - ROMANIAN DAY-AHEAD MARKET

Day-ahead price forecasting for the Romanian market. 96 intervals a day, checked against reality.

The platform forecasts electricity prices on the Romanian Day-Ahead Market at 15-minute intervals, then automatically compares the estimates against the prices that actually settled.

PT15M RESOLUTION
P10-P90 BANDS
--:--:--
Product overview

Forecast the day. Then check the forecast.

Every delivery day is published as 96 fifteen-minute intervals, each with a P50 estimate and a P10-P90 uncertainty band. A chart shows the shape of the day; a table shows the numbers behind it.

Once the market settles, realised prices are placed next to the forecast and every interval gets a signed error. The day is scored with MAE against a baseline, so accuracy is visible instead of assumed.

PT15MP10 / P50 / P90ENTSO-E 12.1.DBacktestingMAE vs baselineCET/CEST
96
Intervals published
for every delivery day
15m
Market resolution
matched end to end
3
Quantiles per interval
P10, P50 and P90
1
Automatic backtest
against realised prices

The forecast, and the
error right next to it.

Representative screens from the live product, showing how a delivery day is published and then reviewed.

From market data
to a scored day.

The same five steps run for every delivery day.

01

Ingest

Market and system data for the delivery day is collected and aligned to the 15-minute grid.

02

Forecast

Each interval receives a P50 estimate and a P10-P90 band instead of a single number.

03

Publish

The full day appears as a chart and a value table, with average price, forecast peak and band width called out.

04

Compare

When realised prices arrive, they are loaded next to the forecast and each interval gets a signed error.

05

Score

The day is summarised with MAE and measured against a baseline, so drift shows up early.

Need price forecasting that
reports its own accuracy?

The same approach adapts to other markets, other resolutions and other price signals.

Start a project