How a settled imbalance price is decomposed
What balancing cost on a delivery day, and what the published price was actually built from. Every number here is served by the system operator; the only arithmetic performed is one subtraction, and it is shown on the page so it can be checked.
The decomposition
A settled imbalance price is not a single published quantity. It is a marginal price plus a penalty component, and the view separates them.
- Base price
- The marginal price of whichever direction actually answered in that quarter-hour. Where no direction was called, there is none.
- Alpha
- The penalty component the operator applied, signed by whether the system was short or long. It is what moves the settlement away from what the offers alone would have set.
- Published
- The settled imbalance price as the operator published it.
- Residual
- Published minus the parts. It is ZERO wherever the arithmetic on this page is the arithmetic the publisher used, and the column exists so you can see that rather than take it on trust. A non-zero residual past the tolerance is flagged.
Minute by minute, and quarter by quarter
Two resolutions, deliberately not averaged into one. System imbalance is the position the whole zone was in and ACE is the area control error the operator was steering against; negative is short. Both are read at the resolution the operator publishes them at rather than averaged up to the settlement period, because the shape within a quarter-hour is what explains the price it settled at.
A delivery day with no per-minute rows published leaves that chart empty and says so - and says explicitly that the quarter-hour settlement below is unaffected, because it comes from a separate publication.
How settled the day is
A quarter-hour is restated as it validates, and only the latest version is served. Nothing here holds a copy, so a figure read today can differ from the same figure read tomorrow, and the view reports how many periods are validated, awaiting validation, or flagged by the publisher as a data issue.
That is why an exported number should carry the date it was read. It is also why this view makes no claim about a day that is still validating beyond what the publisher currently says.
What this view cannot say
Who was on either side of an imbalance. Settlement is published for the system, not per participant, so the cost of a day is a system cost and never anyone's settlement figure. Nothing on this view is screened or thresholded either: it is a replay, not a detector.