GRIDFLOOR
House price forecast — accuracyGridfloor's optimisers plan against published Agile rates where they exist, and a house price forecast beyond that (docs: the cost/carbon curve). This page is the honesty check on that forecast: how wrong is it, by how much, and does the error grow the further out it looks? No model is perfect — the useful question is whether it says so plainly.
Model status
Model not yet trained — needs at least 96 half-hourly samples (~2 days) of history before it can start forecasting. Until then, the optimisers plan only on published Agile rates.
Mean absolute error, by forecast horizon
Mean absolute error (MAE) is the average size of the gap between what the model predicted and what the price actually settled at, ignoring direction — an MAE of £5/MWh means predictions are, on average, about £5 off in either direction. Lower is better, and MAE should rise with horizon: predictions further into the future have less to go on and are expected to be less accurate.
No accuracy figures published yet — the model needs to run for a while before there's enough history to score it.