Auction surveillance

Select a delivery day and bidding zone to compare the cleared price with the reconstructed reference, inspect the published bid curves where available, and test changes to your own orders in the simulator.

The console covers 38 European bidding zones: DE-LU, FR, NL, BE, AT, CH, DK1, DK2, NO1, NO2, NO3, NO4, NO5, SE1, SE2, SE3, SE4, FI, EE, LV, LT, PL, CZ, SK, HU, SI, RO, BG, IT-North, IT-Centre-North, IT-Centre-South, IT-South, IT-Sicily, IT-Sardinia, IT-Calabria, ES, PT, GB. Reconstruction uses the available generation resolution and depends on generation and price inputs being available for the selected zone and day. Published auction prices and curves retain their delivery-period resolution, including 15-minute periods where published. Calibration is walk-forward, using only earlier days. The review shows out-of-sample model error for the day under review alongside the findings.

Request access to TradingSurv, or discuss your markets and review needs. Check coverage and historical availability before choosing.

A day-ahead price is one number standing in for thousands of orders nobody outside the exchange sees. When an alert points at an hour, the first question a reviewer has to answer is whether that price is what the available supply and demand would have produced, or whether something in the order book moved it. That question can only be answered against a reference: what the hour should have cleared at.

Regulators use the same idea. ARERA established economic withholding against A2A in 2026 by re-clearing the Italian day-ahead market with A2A’s offers replaced by each unit’s estimated variable cost and every other offer unchanged, and treated the difference as the measure of the effect. Read the A2A decision.

TradingSurv rebuilds a merit order for each delivery hour from published fundamentals, clears it against demand, and compares the result with what cleared. Screens highlight hours where the two differ by more than the model’s own error, and where structural context, such as a steep supply curve at the margin or a fleet needed to meet demand, makes a small change in offers matter more.

The model is calibrated walk-forward: each hour is judged against parameters fitted only on earlier data, never on the hour itself. The model’s out-of-sample error for the day under review is shown alongside the findings. A difference smaller than that error is not evidence of anything. Read how the reconstruction is built.

Where an exchange publishes aggregated supply and demand curves, the auction itself can be read directly: the crossing, the volume offered near the clearing price, and how far the price would move for a given change in volume. Curves show the aggregate book, not who submitted what, so they support statements about the market, not about a participant. Read the bid-curve guide.

Hydro scheduling, scarcity, network constraints, unplanned outages that were published late and information the model does not have can all explain a flagged hour. A finding is an indicator for a human to review, never a determination of abuse, and public data does not identify trading firms. The economic withholding guide sets out what explains a high offer innocently. Read the withholding guide.

The simulator re-clears an auction with your own orders changed or removed. Your inputs stay in the browser. The result assumes the rest of the market would have submitted the same orders, which is the assumption every counterfactual of this kind makes, including the regulator’s. Read the simulator guide.