I-SEM Daily Briefing — 18 May 2026

Price Profile #

Std dev €35.29/MWh · Median €150.39/MWh · Periods above €150: 25 of 48 (52%)

A seesaw. Tall morning scarcity, a long collapse, a contained evening. Three distinct half-days in one chart.

Elevated overnight (23:00 → 05:30): €144–156. Wind fell from 48% at 23:00 to 3% by 04:30 — the familiar EirGrid boundary artefact places Sunday’s 48% in periods 1–2, then the real decline runs through the small hours. The overnight floor lifted into mid-merit gas territory as renewables left the system. No period below €144; a floor that would have been a peak on last Tuesday.

The morning scarcity event (06:00 → 09:30): €169 → €200 → €240 → €255 → €240 → €243 → €231 → €232. Five consecutive half-hours from 07:00 to 09:00 averaging €240/MWh. Wind during this block: 1.0%–3.0% — effectively zero on the entire system at the height of the morning demand peak. The merit order ran out of mid-merit gas, climbed into peakers, and the next MW of demand pulled the next plant up the stack. €255 is plant-scarcity pricing: these are peaker costs, probably within €10–20 of the I-SEM’s structural ceiling at 4,000 MW demand.

The wind-driven recovery (10:00 → 15:30): €222 → €189 → €185 → €172 → €161 → €152 → €140 → €134 → €126 → €126 → €123 → €121. €101 of collapse across eleven half-hours, tracking wind climbing from 4.7% to 48.8%. The inverse of May 17’s evening decline — same slope, same physics, opposite direction. Each ~4 percentage points of wind added removed roughly €12/MWh from the clearing price.

Contained evening (16:00 → 22:30): €128 → €163 → €163 → €159 → €149 → €133 → €128. Evening peak capped at €164 around 18:00–19:00. Wind held 44–55% through the evening — the wind that disappeared at dawn was back by dinner — and the merit order found mid-merit gas comfortably throughout. The evening peak was €91 below the morning peak. Same demand profile, completely different wind regime by 18:00.

Price vs Wind #

Mean wind: 25.0% · Range: 1.0%–55.7%

Wind 1.0%–55.7%, mean 25.0%. The 55-point range is the analytical statement: today wasn’t a low-wind day or a high-wind day, it was both, in sequence. The wind chart reads as a staircase — close to zero through the morning (1–10%), climbing fast from 10:00 to 16:00 (7% → 55%), holding 45–55% through the evening, easing back to 46% late.

Wind’s capture price today was substantially below the daily mean — probably 65–70%, one of the worst structural days for wind revenue in the run. Wind generated almost nothing during the €240+ morning and abundantly during the €120–165 afternoon and evening. The paradox lands clearly: this was simultaneously the day that made the strongest case for wind build-out (look what happens to prices when there is none) and the worst structural day for wind revenue per MWh actually generated. The system-need signal and the capture-price problem point in opposite directions. A wind investor and a wind farm operator look at the same chart and reach opposite conclusions.

Week in Context #

Three comparisons, same structural point from different angles.

May 5 (Tuesday, windless) ran 8.5% mean wind at €174 mean. Today ran 25.0% mean wind at €163 mean. Three times more wind on May 18, only €11/MWh cheaper. The reason is timing: May 5’s wind was uniformly low (4.4–15.7% range, bad timing all day but no extreme moment). May 18’s wind averaged 25% but ran 1.0% during the morning peak and 55.7% in the evening. Mean wind doesn’t capture the timing of wind shortfalls. Same average, near-identical market outcome.

May 11 (Monday, 23% mean wind, €138) vs today (Monday, 25% mean wind, €163): demand barely moved (3,810 → 4,036 MW), but the wind range went from 17–32% (15-point band, stable) to 1–56% (55-point band, extreme variability). The wind variability — not the wind level — drove the €25 difference between two Mondays with essentially the same mean wind.

Mean wind is not the right statistic for pricing analysis. Intra-day timing is.

Price Duration Curve #

Periods above €150: 25 (52% of day) · Above €200: 8 (17% of day)

25 above €150 — more than half the day — and 8 above €200. A steep top half: the morning scarcity block (8 periods €200+, 5 periods above €230), extending down through the elevated overnight and recovery gradient into a broad middle around €150–190. Then a shallow tail of 23 periods below €150, mostly the afternoon and evening wind-rich lows.

The comparison to May 15 is instructive. Similar plateau height (€240+ today, €244+ on Friday), similar width at the top (8 vs 9 periods above €200). May 15’s plateau was clustered tightly near its ceiling; May 18’s extends further down through the morning recovery gradient. Both are scarcity fingerprints — peaker pricing concentrated in a short block — but May 18’s bimodality is more extreme: a €134 range with the floor at €122 versus May 15’s €113.

Peak / Off-Peak Spread #

Peak avg (07:00–22:00): €169.8/MWh · Off-peak avg: €150.95/MWh · Spread: €18.85/MWh

+€18.85. A positive spread, and for once the metric earns its fee.

The morning scarcity falls cleanly inside the peak window (06:30–09:30, all above €169). The off-peak window catches the elevated overnight (€144–156, wind leaving the system through the small hours) but no scarcity. The peak window runs expensive at both ends — morning scarcity and a mild evening shoulder — while the off-peak window has only the elevated floor. €19 spread is the mechanical result of a morning-aligned scarcity event.

Compare to May 17: same order of magnitude of scarcity hours (6 vs 8), spread +€2.48 vs +€18.85. The difference is entirely which side of 22:00 the scarcity fell. Morning scarcity aligns with the peak window; the metric works. Late-evening wind-decay scarcity crosses the boundary; the metric breaks. Same instrument, different events, very different readings.

BESS Dispatch Signal #

Simulated 1MW/2MWh battery, one optimal DAM cycle. Gross before network charges and capacity costs.

€168 gross — a new high for the run. Captured spread €121 — widest of the series. The dispatch pattern is the operational note: for the first time since May 5, the optimiser chose morning (07:00) for the discharge window rather than evening.

The reason is mechanical. The four highest consecutive half-hours today were 07:00, 07:30, 08:00, 08:30, averaging €244.78. The evening peak’s four-hour cluster (18:00–19:30) averaged €162.50. €82 of difference per half-hour. The optimiser discharged into the morning scarcity block and recharged in the afternoon wind-recovery trough (€124 at 14:00) — a reverse sequence compared to every other day in the run. The operational implication is direct: optimal discharge follows the wind shortfall, not the demand peak. Most days the two coincide. On May 18 they didn’t, and a model that defaulted to “evening discharge” would have left €82 per half-hour on the table.

The 15-day BESS series now stands at €1,518 cumulative. Top 4 days (May 7 €139, May 15 €156, May 17 €161, May 18 €168) = €624 — 41% of cumulative revenue from 27% of days.

Commentary #

Monday’s mean of €162.73/MWh is the highest of the 17-day run. The peak hit €255 at 07:30 — clearing May 15’s €244 — and eight half-hours cleared above €200, all in the morning block. The driver was wind at 1.0% during the morning peak: effectively zero on the entire system while demand was at 4,036 MW. The merit order ran out of mid-merit gas, climbed into peakers, and the next MW of demand pulled the next plant up the stack. Five consecutive half-hours from 07:00 to 09:00 averaged €240/MWh. €255 is plant-scarcity pricing — the marginal-plant costs of the cheapest peakers in the stack, probably within €10 of the I-SEM’s structural price ceiling at this demand level.

The most analytically important fact of the day isn’t the peak — it’s the comparison to May 5. May 5 (Tuesday): 8.5% mean wind, €174 mean. May 18 (Monday): 25.0% mean wind, €163 mean. Three times more wind on May 18, only €11/MWh cheaper. The reason is timing. May 5’s wind was uniformly low (range 4.4–15.7%): bad timing all day but no extreme moment. May 18’s wind averaged 25% but ran 1.0% at the morning peak and 55.7% in the evening. Mean wind doesn’t capture the timing of wind shortfalls. Same average, near-identical market outcomes. This has direct implications for capacity adequacy analysis: models using mean wind capacity factors will systematically understate how often the system runs into May 18-style morning events.

For storage, a new high and a structural first. €168 gross, captured spread €121, ROI 67.7%. For the first time in 13 days, the optimiser chose morning for the discharge window — the 07:00–08:30 block averaged €245, €83 above the evening peak. Optimal discharge follows the wind shortfall, not the demand peak. Most days they coincide; on May 18 they didn’t.