Forecast
Loading chart…
Loading chart…
Loading chart…
MAE (Mean Absolute Error) is the average absolute deviation of the forecast from the price the day-ahead auction actually settled at. For every hour of the day the absolute difference between the forecast price and the market price is taken, and those differences are averaged across all 24 hours. A lower value means a more accurate forecast. The forecast compared is the last one published before the auction closed.
Every zone behaves differently, so every zone has a model of its own - the one for Czechia is not the one for Estonia. Keeping those models good is our problem, not yours. You pick the zone you care about and put its code in the API request.
Alongside the main price curve, the forecast carries a band bounded by the P10 and P90 percentiles, which estimates how uncertain it is. It means the model expects the price to land inside that band in roughly eight cases out of ten. If you run a BESS, uncertainty is exactly the thing you have to plan around.
The forecast for the days ahead is our product. Being able to look back at what we said is something far fewer people offer. Going through the past days and weeks gives you a good idea of how our models are doing.
Electricity prices are sometimes moved by events nobody saw coming. A fast-changing geopolitical situation, regulation and the consequences of a changing climate all bear directly on what energy costs. The models cannot stand still, then: the size of the error is measured every day, and the models adapt to the new conditions step by step.
Last updated
Modelv3