Day-ahead and intraday markets provide different opportunities to trade electricity for future delivery. For a renewable portfolio, the initial production schedule and subsequent revisions are linked to these opportunities. A better understanding of generation can support position management, but the effect depends on deadlines, available liquidity, commercial objectives and the decisions taken from the forecast.
Day-ahead establishes an earlier position
Day-ahead trading concerns electricity for the following day under the relevant market's products and procedures. A renewable participant uses its expected production profile alongside its contracts and trading policy to decide a position. The forecast must cover the required delivery intervals and be available before the applicable decision deadline.
An annual energy estimate is not sufficient for this task. Scheduling requires a time profile and a clear measurement boundary. Specify whether the forecast represents available generation or expected export, and whether known outages and operating restrictions are included. Those choices affect the volume that can reasonably support a position.
Intraday enables later adjustment
Intraday trading provides opportunities to revise positions closer to delivery through the relevant continuous or auction arrangements. New weather runs, telemetry and plant status can change the production expectation. An update is useful only if it arrives while the participant can still act and if the trade is consistent with the applicable products and constraints.
Do not assume one universal deadline, interval or liquidity condition. These details vary by market area and product and can change. Confirm the current venue rules for the process being implemented. A general forecasting article cannot substitute for the participant's market access arrangements or operating procedures.
A generation revision is not a price forecast
A renewable forecast estimates physical production. It does not by itself predict the price at which a position can be adjusted. Price formation depends on wider supply, demand, constraints and participant behaviour. Keep those inputs and their uncertainty separate when designing a trading policy.
A downward generation revision may indicate that a sold position exceeds the latest expected output. What to do next depends on exposure, available trading opportunities and the chosen policy. It is not enough to say that a forecast avoids losses. An economic result needs a defined strategy and evidence under the conditions in which it would operate.
Align forecasts with actual decision cutoffs
Archive every forecast issue and receipt time. For each simulated historical decision, select only information then available. Using a later revision to represent an earlier position gives the strategy hindsight. Likewise, using perfectly cleaned telemetry that arrived later can overstate the benefit of a near-term update.
Track forecast versions separately from delivery intervals. Several successive forecasts may concern the same interval; their differences describe how expectations evolved. The final forecast before delivery is not automatically the forecast used to create the original position. Evaluation should preserve both so teams can investigate which revision supported which action.
Evaluate physical and commercial outcomes separately
Start with interval-level forecast accuracy at the relevant lead times. Examine bias, large deviations and ramp timing, using consistent units and normalization. Then replay the defined trading policy with its actual assumptions about transaction opportunities and costs. The first analysis assesses production information; the second assesses how that information is used.
Keep a credible baseline. Compare with the existing forecast or policy under the same decision rules, rather than changing several assumptions at once. Report periods of missing data and late forecasts separately. A meteorologically strong forecast delivered after a cutoff cannot improve that earlier decision.
Portfolio aggregation requires care
Production errors can offset across assets, but common weather can also create shared errors. Assess both site-level performance and portfolio totals. A portfolio score can be commercially relevant while hiding a persistent operational issue at one plant. Neither view should replace the other without explaining the decision boundary.
Document who is responsible for reviewing a large revision and when automated action is permitted. Keep that rule stable during a comparison. Otherwise, differences in human or system response can be mistaken for differences in the underlying production forecast.
A practical implementation ends with a forecast-to-position workflow: asset mapping, issue schedule, receipt monitoring, decision cutoffs, revision handling and a documented fallback. That workflow makes renewable forecasts usable in day-ahead and intraday processes. It also makes the limits visible, including the fact that improved generation information alone is not a guarantee of a favourable trading result.