Dry until May: the curve bets on a fast spring fade

What is the idea?

Oslo minus Stockholm is +31.30 for Q2-27 and +44.77 for Q1-27, so the net, Q2 minus Q1, is -13.47. None of the four legs traded on 15 October.

An EPAD pays the area price minus the Nordic system price, so the system leg cancels inside each Oslo-Stockholm pair and each pair pays the realised Oslo-Stockholm difference over its quarter. Q1-27 has 2,159 delivery hours and Q2-27 has 2,184, so at 5 MW a leg €1/MWh on the net is worth €10,795 to €10,920, depending on which leg moves.

The ledger's rules for the idea

• We close the paper idea at -3.0 or at -22.0 on any settlement: about +€114,000 or about -€93,000 from the reference.
• The target sits 10.47 away and the stop 8.53 away. With no drift the stop would be hit first about 55% of the time, so the idea needs a hit rate above 45% before costs to break even.
• Otherwise we hold it into delivery. Q1-27 cascades into January, February and March contracts at the end of December and Q2-27 into April, May and June at the end of March; each month settles against the realised day-ahead average.
• Each €1 of round-trip slippage on the net costs about €10,900, and the four legs are thin (see Liquidity below).
• We will add the idea to our ledger at the Euronext settlement on the day we commit it and publish its marks. The 15 October prints are this page's reference.

What is the curve pricing?

The premium is highest at the front of the curve: 56.20 for October 2026, 52.54 for November and 53.35 for December on the monthly EPADs, then 44.77, 31.30, 28.09 and 26.74 for the four quarters of 2027. Q4-27, next winter, carries the lowest premium on the board. The curve prices the drought premium fading through 2027, and the October spot backs the starting level: Oslo cleared above Stockholm on all sixteen days of the month so far, by €61.9/MWh on average (ENTSO-E day-ahead, prices published by Nord Pool, 1 to 16 October).

Prices roughly halve from winter to spring, so a premium that keeps a steady share of the price shrinks in euros. As a share of the Stockholm price, Oslo's premium is 42% in Q1-27 and 56% in Q2-27, a spring tilt of 14 points. In 2021 to 2026 the tilt was +16, +9, +19, +8, +55 and +15 points, and the three dry years were +9, +19 and +15.

The net was -10.43 when Q2-27 first settled on 1 April, rose to -1.05 on 14 July and fell to -18.53 on 14 September. Between 14 July and 14 September, Jan-27 TTF rose 30.1 and the net fell 17.5.

Gas moves the net. From 1 April, on daily changes, each €1 on the TTF front month moved the Q1-27 Oslo-Stockholm spread by €0.246 and the Q2-27 spread by €0.135, a net of -0.111 (134 changes, CME TTF futures). Daily changes understate the link, because most of these marks were untraded quotes. Over the first draft's window, 13 April to 15 October, which the review kept for these measures, the net's beta was -0.20 on non-overlapping weekly changes (correlation -0.62) and -0.24 in levels (correlation -0.72). The idea gains when gas falls and loses when it rises: at the level beta, a further €30 rise in TTF is worth about -7.2 on the net, roughly -€78,000.

How low are southern Norway's reservoirs?

NO2 holds the largest share of Norway's reservoir capacity, 34 TWh of 87, and it sits 31 points under its 20-year median for the week.

NO2's low comes in spring. In 26 of the 31 years from 1996 to 2026 its lowest fill of the first half-year came in weeks 14 to 18, with a median of week 16. The median low comes in week 15 for NO1, week 18 for NO5 and week 17 for Norway as a whole. In the dry years NO2's bottom was 18.6% in week 17 of 2022, 32.4% in week 18 of 2023 and 27.6% in week 16 of 2026.

The three southern zones price close together on the curve: the Q1-27 forwards are €150.54 for NO1, €150.15 for NO2 and €148.34 for NO5.

What did the dry years do to the spread?

Oslo minus Stockholm, quarterly mean of daily day-ahead averages, with the spring-to-winter ratio of the EU gas benchmark:

€/MWh. Day-ahead prices from the ENTSO-E Transparency Platform, as published by Nord Pool. Gas column: the IMF's monthly EU natural gas benchmark on FRED (in US dollars) for 2021 to 2026, and TTF Q2-27 over Q1-27 on the ICE Endex table Elexys publishes (in euros) for the forward. The 2027 row is the Euronext EPAD settlements of 15 October 2026.

Over the six years the mean was +5.0, the range -3.10 to +14.14 and the standard deviation 6.7. NordLink opened to the market in December 2020 and North Sea Link in October 2021, so the 2021 row predates the British cable; starting in 2022 moves the mean to +4.8 and leaves the minimum unchanged. Earlier years were different. In 2015 to 2019 Oslo's quarterly premium over Stockholm never exceeded €4.0, and in the wet year 2020 Oslo traded as much as €20.4 under Stockholm. Starting in 2015 would add two springs below 2023: 2016 at -3.2 and 2020 at -7.1.

Scaled to 2027 prices, the answer depends on the scale

Applying each dry year's Oslo/Stockholm price ratios to the 2027 Stockholm forwards gives a net of -20.65 for 2022, -9.73 for 2023 and -4.11 for 2026, around the forward's -13.47. Scaling the same years by the gas price instead gives 0.0, +3.4 and -6.9, all above it.

The idea leans on the gas-scaled view. The curve's shape has moved with gas since April, gas itself falls by a fifth from Q1-27 to Q2-27 on the ICE Endex table, and the reservoir low arrives in April whatever gas does. On the Stockholm-scaled view the forward is about fair, and there is no trade. To reach the -3.0 target with the Q1 spread unchanged, Oslo would have to average 1.75 times Stockholm through Q2-27. No second quarter since 2021 has: the highest was 1.68 in 2025, and the dry-year Q2s were 1.61, 1.61 and 1.41.

Month by month, the dry years split

January to April and February to May favoured spring in all three dry years, March to June favoured winter in all three, and 2023's April edge was 0.2. Gas held near flat into spring in 2022 (0.96) and rose in 2026 (1.14), which props up spring premia on its own. 2023, the one dry year when gas fell by a third into spring (0.67), is also the one with Q2 below Q1. Three dry years make a thin sample.

By dry years we mean the winters that began with NO2 under 70% full in week 41: 2021 (56.8%), 2022 (62.6%) and 2025 (68.9%), which set up the spreads of 2022, 2023 and 2026. 2025 qualifies by 1.1 points. The next-lowest fill since 2020 was 80.5% in 2024, so any threshold from 69% to 80% picks the same three winters.

What pushes the spring premium up or down in 2027?

For the idea

• The refill. Once the melt starts, producers in the south hold water back to rebuild storage for the next winter. In the dry years the premium held into May, at €64.9 in 2022, €32.7 in 2023 and €27.6 in 2026, and dropped in June, when the flood peaks (26.6, 17.2 and 4.0).
• Stockholm's spring. Heating demand falls, the northern Swedish melt runs south and solar output climbs: Stockholm averaged €51.7 in April 2026 against Oslo's €100.2.
• SwePol. The SE4 to Poland cable is filed 600 MW out from 20 December to 27 February, which tightens southern Sweden in Q1.

Against the idea

• Swedish nuclear. Vattenfall's annual outages fall mainly between May and September, and 2026 showed how far they overrun: Ringhals 3, planned out from 4 May to 23 July, stays out until 31 October, and Ringhals 4, planned for 2 September to 15 October, now runs to 30 November. With Ringhals 3, Ringhals 4 and Forsmark 1 due back by 12 December, Stockholm should be well supplied in Q1-27, which widens the Q1 spread against the short leg. Q2-27 has Ringhals 3 out from 19 May.
• Wind. Swedish wind capacity grew from 12.1 GW at the end of 2021 to a forecast 19.5 GW for the end of 2026. Nordic wind output runs higher in winter, so the extra capacity most likely lowers Stockholm more in Q1 than in Q2.
• Norgespris. About two-thirds of households in southern Norway have opted into Norgespris, a fixed 40 øre/kWh before VAT on the first 5,000 kWh a month: 67% in NO1, 79% in NO2 and 68% in NO5 on 1 October 2026. The 2027 budget proposes 45 øre, Rødt, SV and Senterpartiet want 40, and households must re-order for 2027. A fixed price removes demand response mostly in the winter heating season, which supports Q1, the leg this idea sells.
• Continental solar. The cables that tie southern Norway to continental prices also bring in the continent's spring. Rising solar output pushes continental prices down from April to June relative to January to March, which cuts the export value of Norwegian water in Q2.
• Maintenance. Transmission work clusters in the summer half. The 2026 filings show heavy SE3-to-NO1 and SE2-to-SE3 work in May and June; cuts on SE3 to NO1 widen the spread, while cuts on SE2 to SE3 and on Norwegian export cables narrow it.

Prices in this section are monthly means of daily day-ahead averages from the ENTSO-E Transparency Platform, as published by Nord Pool.

Which readings did we test and drop?

One reading was that October's spike is a line outage. Svenska kraftnät has filed works on Swedish lines (Ramsele-Midskog and Hall-Hedenlunda) that cut the Stockholm-to-Oslo direction by 1,095 MW from 12 October to 3 November and by up to 1,795 MW from 2 to 15 November (Nord Pool UMM). The same works cut SE2 to SE3 by 2,000 to 2,100 MW, which lifts Stockholm. JAO's Nordic flow-based figures give the grid story no support either way: daily SE3-to-NO1 capacity explains almost none of the day-to-day variation in the spread, with a correlation of 0.08 over 96 days from 1 July to 16 October (4 to 15 August missing).

The other was a plain long on the Q1-27 spread. At €44.77 it sits where January and February 2022 printed (39.6 and 46.3) with gas at a similar level. This idea expects the Q2 spread to rise toward Q1 by delivery instead.

Why is this our second Norwegian idea?

On 24 September we opened a paper idea, long NO2 against DK1 for Q1-27 at -6.90. It was stopped on the 5 October settlement at -16.68, through its -13.00 stop, for a paper loss of €105,575 at 5 MW.

The reservoir read was right: NO2 printed 52.1% in week 41, below the 56.1% median path the idea was built on and under that path's low band of 52.2%. The loss came from the DK1 leg. Over the seven sessions to the stop the Århus Q1-27 EPAD rose from 35.83 to 46.28 while Kristiansand moved from 28.93 to 29.60, and TTF was flat to lower: the CME TTF front month closed at 75.11 on 24 September and 73.51 on 5 October. It rallied afterwards, from 73.5 to 87.1 between 5 and 14 October.

This idea keeps the reservoir work and drops the DK1 leg; both halves are Oslo against Stockholm. It still carries gas exposure. On weekly changes and in levels the net has half to three-quarters of the NO2-DK1 pair's gas beta (-0.20 against -0.39 weekly, -0.24 against -0.33 in levels), with a similar correlation.

What would make us wrong?

Which dates should a desk watch?

How will we mark the idea?

We will add the idea to our ledger at the Euronext settlement on the day we commit it and publish its marks on every settlement after that, losses included. Delivered months will be valued at the realised day-ahead average.

Where do the numbers come from?

• Forward prices: Euronext Nordic power futures settlements (system price and EPADs, quarterly and monthly), 16 March to 15 October 2026. Euronext's daily settlement procedure (9 March 2026) takes the volume-weighted price of the last five minutes, then the mid of the best bid and offer, then the last traded price including block trades, then the previous day's settlement or a theoretical price.
• Day-ahead prices: ENTSO-E Transparency Platform (prices published by Nord Pool), daily averages on Oslo local days, 2015 to 16 October 2026. Monthly and quarterly means only; no daily series is shown.
• Reservoirs: NVE Magasinstatistikk, weekly, 1995 to week 41 2026 (NLOD licence).
• Capacity: JAO Nordic flow-based publication, max exchange SE3 to NO1 (a flow-based maximum on UTC days); Nord Pool REMIT UMM for filed transmission and nuclear unavailability.
• Gas: ICE Endex TTF via the public Elexys forward table, read on 15 October 2026; the IMF's monthly EU natural gas benchmark on FRED (PNGASEUUSDM) for 2021 to 2026; CME TTF front-month and Jan-27 futures from public delayed quotes for the betas and the TTF closes in section 07.
• Scaled comparisons: each year's quarterly Oslo/Stockholm price ratios applied to the 2027 Stockholm forwards, or the spreads scaled by the gas price, recomputed in an independent review of this note on 16 October 2026.
• Policy and supply: NVE and regjeringen.no on Norgespris; Elhub uptake figures as reported by Fortum and Vibb; NRK, TU and VG on the 2027 budget proposal; Vattenfall and Nord Pool UMM on Swedish reactors; Energimyndigheten (2021) and Svensk Vindenergi (April 2024 forecast) on wind capacity.