Demand Forecast Misses and Day-Ahead Prices in Europe, 7 October to 4 November 2026
No reliable link between forecast misses and prices showed up. Across 44 European bidding zones, 7 October to 4 November 2026, the demand forecast was off by a median 1.9% a day. In 23 of 44 zones the quarter of days with demand furthest above forecast had a higher price than the quarter with demand furthest below it; the median zone's price on the demand-above-forecast days was 0.8 percentage points higher than on the demand-below-forecast days, measured against its typical price level. The sample is about 28 days per zone, which is too short to separate a real effect from chance, and this is correlation, not cause.
Day-ahead wholesale prices and demand forecasts from ENTSO-E for 44 bidding zones priced in euros, about 28 days per zone, 7 October to 4 November 2026. This page reports an association measured in a short window. It is correlation, not cause, and the window is too short for a firm conclusion.
Key Facts
- No reliable link between forecast misses and prices showed up. Across 44 European bidding zones, 7 October to 4 November 2026, the demand forecast was off by a median 1.9% a day. In 23 of 44 zones the quarter of days with demand furthest above forecast had a higher price than the quarter with demand furthest below it; the median zone's price on the demand-above-forecast days was 0.8 percentage points higher than on the demand-below-forecast days, measured against its typical price level. The sample is about 28 days per zone, which is too short to separate a real effect from chance, and this is correlation, not cause.
- Sample: 44 zones, about 28 days per zone, 7 October to 4 November 2026. Each zone's own correlation rests on about 28 points, and a correlation must reach about 0.38 in size to beat chance at the usual 5% level.
- The median zone's correlation between forecast error and price deviation was 0.00; 22 zones were positive and 22 negative. 9 of 44 zones passed the 5% bar (2 positive, 7 negative); about 2.2 would pass by chance alone.
- On weekdays only, 26 of 44 zones showed higher prices on the days with demand furthest above forecast; the median zone's price on those days was 2.9 percentage points higher than on the days with demand furthest below it.
- Combining zones into one point per day (28 days), the correlation was -0.18 against a bar of about 0.38.
- 44 zones is not 44 independent samples: neighbouring zones share weather and trade power. All prices are day-ahead wholesale prices from ENTSO-E and demand is ENTSO-E actual load against its day-ahead forecast.
What Was Tested and What Was Found
44 zones, about 28 days per zone, 7 October to 4 November 2026- What is tested: whether days when demand came in well above or below its forecast also had day-ahead prices well above or below the zone's usual level. The day-ahead price is fixed the day before delivery, so a forecast miss cannot cause it. A link, if there is one, would mean that the conditions behind the miss, such as a cold snap or a holiday, also moved prices.
- Per zone, the correlation between the daily forecast error and the day's price deviation ranged from -0.65 to +0.81 with a median of 0.00. With about 28 days a zone, a correlation has to reach 0.38 to be distinguishable from chance. 9 of 44 zones did: 7 with higher prices when demand came in above the forecast and 2 with higher prices when demand came in below it. About 2.2 would be expected by chance alone.
- Comparing, in each zone, the quarter of days (7 days) with demand furthest above forecast against the quarter with demand furthest below it: the day's price was higher on the demand-above-forecast days in 23 of 44 zones. The median zone's price deviation was +0.6% on the demand-above-forecast days and +0.2% on the demand-below-forecast days. An exact sign test on the split gives p = 0.88.
- Weekends are cheaper and carry different forecast errors, which can create a link on their own. Repeating the contrast on weekdays only: higher prices on the demand-above-forecast days in 26 of 44 zones, the median zone 2.9 percentage points higher on those days, p = 0.29.
- Averaging across zones for each day (28 days with at least 10 zones), the correlation was -0.18, against a bar of 0.38.
- Reading: the three checks do not agree on a direction that clears chance, so the data does not support saying that forecast misses move prices. It does not show that they do not: a real but modest effect would be hidden by a sample this short and by everything else that moves prices day to day, such as wind, solar, gas prices and imports.
Zone by Zone
Highest correlation first, 2026-11-04 (4 November 2026) latest day| Bidding zone | Days | Median daily miss | Correlation | Beats the 5% bar (size of bar) | Price vs typical, demand furthest above forecast | Price vs typical, demand furthest below forecast | Difference (points) |
|---|---|---|---|---|---|---|---|
| Lithuania | 28 | 8.1% | +0.81 | Yes (0.38) | -60.8% | +70.1% | -130.8 |
| Finland | 28 | 3.6% | +0.53 | Yes (0.38) | -3.2% | +303.6% | -306.8 |
| Italy South | 28 | 8.9% | +0.35 | No (0.38) | -11.6% | +5.0% | -16.6 |
| Italy Calabria | 28 | 6.0% | +0.34 | No (0.38) | -12.5% | +1.9% | -14.4 |
| Denmark East (DK2) | 22 | 0.3% | +0.31 | No (0.43) | -9.4% | +8.2% | -17.6 |
| Portugal | 28 | 1.3% | +0.25 | No (0.38) | -4.4% | +8.1% | -12.5 |
| Sweden South (SE4) | 28 | 6.9% | +0.24 | No (0.38) | -21.7% | +26.9% | -48.5 |
| Montenegro | 27 | 3.9% | +0.24 | No (0.38) | +1.1% | +10.0% | -8.8 |
| Austria | 28 | 1.9% | +0.24 | No (0.38) | -6.9% | +1.9% | -8.8 |
| France | 28 | 0.8% | +0.23 | No (0.38) | -16.6% | +7.3% | -23.9 |
| Poland | 28 | 1.1% | +0.22 | No (0.38) | -18.9% | +7.4% | -26.3 |
| Denmark West (DK1) | 22 | 0.3% | +0.20 | No (0.43) | -36.7% | -4.5% | -32.3 |
| Albania | 27 | 2.3% | +0.19 | No (0.38) | -1.9% | +10.4% | -12.3 |
| Slovenia | 28 | 1.9% | +0.14 | No (0.38) | -2.6% | +0.7% | -3.3 |
| Switzerland | 28 | 0.4% | +0.12 | No (0.38) | -2.6% | +0.4% | -3.0 |
| North Macedonia | 24 | 4.2% | +0.10 | No (0.41) | -7.0% | -1.9% | -5.2 |
| Croatia | 28 | 1.5% | +0.10 | No (0.38) | -2.7% | +1.1% | -3.8 |
| Sweden South-Central (SE3) | 28 | 0.9% | +0.08 | No (0.38) | -5.9% | +13.8% | -19.7 |
| Italy North | 28 | 1.3% | +0.06 | No (0.38) | +1.8% | -0.8% | +2.5 |
| Norway South-East (NO1) | 28 | 1.5% | +0.06 | No (0.38) | -13.4% | -2.7% | -10.7 |
| Serbia | 28 | 2.0% | +0.04 | No (0.38) | -5.1% | -3.5% | -1.6 |
| Greece | 28 | 2.2% | +0.02 | No (0.38) | -6.9% | -11.0% | +4.1 |
| Sweden North (SE1) | 28 | 5.2% | -0.02 | No (0.38) | +47.9% | +3.8% | +44.1 |
| Belgium | 28 | 1.7% | -0.03 | No (0.38) | +4.2% | +3.3% | +1.0 |
| Norway Central (NO3) | 28 | 1.6% | -0.03 | No (0.38) | +0.0% | -2.3% | +2.3 |
| Bulgaria | 27 | 2.3% | -0.04 | No (0.38) | +2.2% | +4.7% | -2.4 |
| Czechia | 28 | 0.6% | -0.06 | No (0.38) | +8.4% | +7.7% | +0.7 |
| Kosovo | 28 | 1.5% | -0.13 | No (0.38) | +2.3% | -5.8% | +8.1 |
| Italy Sardinia | 28 | 1.9% | -0.16 | No (0.38) | +10.1% | +0.1% | +10.0 |
| Slovakia | 28 | 1.1% | -0.21 | No (0.38) | +4.7% | -5.5% | +10.2 |
| Germany and Luxembourg | 28 | 2.1% | -0.22 | No (0.38) | +13.0% | +0.3% | +12.7 |
| Norway North (NO4) | 28 | 2.2% | -0.25 | No (0.38) | +15.0% | -2.7% | +17.7 |
| Romania | 28 | 0.8% | -0.27 | No (0.38) | +6.9% | +0.4% | +6.5 |
| Estonia | 28 | 3.6% | -0.30 | No (0.38) | +40.5% | -13.5% | +54.0 |
| Italy Centre-North | 28 | 1.5% | -0.34 | No (0.38) | +6.8% | -0.1% | +6.9 |
| Sweden North-Central (SE2) | 28 | 12.2% | -0.36 | No (0.38) | +107.1% | -14.7% | +121.8 |
| Italy Sicily | 28 | 2.6% | -0.38 | No (0.38) | -0.2% | -13.6% | +13.4 |
| Netherlands | 28 | 9.6% | -0.39 | Yes (0.38) | +17.5% | -31.8% | +49.2 |
| Hungary | 28 | 2.9% | -0.41 | Yes (0.38) | +7.1% | -6.6% | +13.8 |
| Italy Centre-South | 28 | 2.8% | -0.46 | Yes (0.38) | +8.7% | -3.6% | +12.3 |
| Latvia | 28 | 2.5% | -0.46 | Yes (0.38) | +29.2% | -56.3% | +85.5 |
| Spain | 28 | 0.6% | -0.47 | Yes (0.38) | +7.6% | -17.0% | +24.6 |
| Norway West (NO5) | 28 | 1.9% | -0.47 | Yes (0.38) | +8.3% | -19.1% | +27.5 |
| Norway South-West (NO2) | 28 | 1.9% | -0.65 | Yes (0.38) | +15.6% | -31.2% | +46.9 |
The forecast error of a day is the day-ahead demand forecast minus actual demand, over actual demand, summed across the day's hours (days with at least 20 hours of both readings). The correlation is Spearman's rank correlation between that error and the day's average price against the zone's median day, so a positive value means higher prices on days when the forecast ran too high (demand came in below forecast) and a negative value higher prices on days when demand came in above forecast. The last three columns compare, within each zone, the quarter of days with demand furthest above forecast and the quarter with demand furthest below it; the difference is the first minus the second, in percentage points of the zone's median day. A zone whose median daily average was under EUR 5/MWh is left out (1 zone here).
Source: ENTSO-E Transparency Platform (transparency.entsoe.eu), CC BY 4.0. Aggregated and converted by World Power Monitor; ENTSO-E does not endorse this site. (ENTSO-E Terms of Use plus CC BY 4.0 for listed data items) Data as of 2026-11-04. Forecast errors are computed by World Power Monitor from ENTSO-E actual load and its day-ahead forecast; prices from ENTSO-E day-ahead prices.
Why This Cannot Say More
- Sample size: about 28 days per zone (20 weekdays), 7 October to 4 November 2026. That is one autumn month. The export holds no longer record of the demand forecast, so nothing here shows whether the pattern holds in winter or summer.
- Not independent: neighbouring zones share weather, trade power and often share prices. 44 zones are not 44 samples, which is why the page also reports one point per day across zones (28 days, each with at least 10 zones).
- Correlation is not cause. The day-ahead price is fixed the day before delivery, from bids placed before the forecast miss is known, so a miss cannot change it. Weather, holidays and wind or solar output can move demand, the forecast error and the price together.
- Weekends are cheaper and carry different errors. Every contrast is repeated on weekdays only.
- Prices are day-ahead wholesale prices in euros, not converted and not what a household pays. Demand is ENTSO-E actual total load; readings outside 30% to 300% of a zone's own median hour are treated as reporting faults and ignored, as on the forecast accuracy pages.
- The price used is the day's average. The cost of a miss shows up in balancing and intraday prices, which this site does not hold, so this page says nothing about what a miss costs once it happens.
Forecast Misses and Prices: Questions and Answers
Do demand forecast errors change electricity prices?
Not in a way this data can show. Across 44 European zones, with about 28 days per zone, 7 October to 4 November 2026, the zone by zone correlations went both ways and the cross checks did not agree on a direction that clears chance. That does not prove there is no effect; the window is too short to tell.
How is the forecast error measured?
For each whole local day with at least 20 hours of both readings, the error is the day-ahead demand forecast minus actual demand, over the total actual demand, in percent. A positive error means demand came in below the forecast. The median zone's typical daily miss was 1.9%.
What does price deviation mean here?
The day's average day-ahead price against the zone's own median daily average over the same days, in percent. A zone whose median day was under EUR 5/MWh is left out because a percentage of a near zero price means nothing (1 zone was left out for that or for too few days).
Why is a short window a problem?
With about 28 days a zone, a correlation has to be about 0.38 or larger in size to beat chance, and most days sit in one season. Prices move for many reasons at once, and forecast misses tend to arrive with weather that also moves wind, solar and heating demand, so a link could not be pulled apart from them even with more days.
Related
Cite this page
World Power Monitor. (2026). Demand Forecast Misses and Day-Ahead Electricity Prices. https://worldpowermonitor.com/prices/forecast-error/
World Power Monitor. Demand Forecast Misses and Day-Ahead Electricity Prices. Data as of 2026-11-04. Data from ENTSO-E Transparency Platform. https://worldpowermonitor.com/prices/forecast-error/
@misc{wpm_prices_forecast_error_2026,
title = {{Demand Forecast Misses and Day-Ahead Electricity Prices}},
author = {{World Power Monitor}},
year = {2026},
note = {Data as of 2026-11-04. Data from ENTSO-E Transparency Platform.},
url = {https://worldpowermonitor.com/prices/forecast-error/}
} Permalink: https://worldpowermonitor.com/prices/forecast-error/
The Copy buttons add today's date as the access date.