DFS system tests: what did the prices signal?

The short answer: NESO’s three late-summer, system-tagged DFS tests show accepted capacity at indicative average prices close to the stated guaranteed acceptance prices. They do not show the full offer curve, a marginal clearing price, or delivered response. That distinction matters before treating the results as a verdict on whether the price signal was strong enough.

This article examines the three system-tagged DFS tests listed in NESO's public dataset as of 24 September 2026. D001

Put the guarantee beside the accepted results

NESO’s August market-information report proposed three tests to examine flexibility across directions and zones ahead of locational constraint management. It specified a £180/MWh guaranteed acceptance price for upward, demand-turn-up tests and £500/MWh for downward, demand-turn-down tests. One proposed test was open to all zones; two targeted zones 4 and 12. That plan alone does not identify the zone of each accepted offer in the aggregate results. D007

Observed system-tagged tests	Published requirement and accepted capacity	Accepted cost and indicative average accepted price	Stated guarantee
Upward: two events, four 30-minute windows, 30 August–13 September	2,000 MW-per-window summed requirement; 305.7 MW-per-window summed procurement, or 15.285% of requirement D006	£27,404.12 accepted cost; £179.29/MWh indicative average D006	£180/MWh D007
Downward: one event, two 30-minute windows, 10 September	1,000 MW-per-window summed requirement; 66.7 MW-per-window summed procurement, or 6.67% of requirement D005	£16,675 accepted cost; £500/MWh indicative average D005	£500/MWh D007

These are sums of MW across 30-minute windows, not MWh; the percentages compare summed accepted procurement with summed published requirement. NESO defines procured MW as accepted volume and settled MW as delivered volume available once settlement data is received. D001

The comparison is striking but narrow: the upward indicative average sits just below its stated guarantee, while the downward average equals its guarantee. D006 D005 D007 These are averages derived from accepted cost and volume. They are not a reported marginal clearing price and do not show individual bid prices.

What can a market reader conclude?

The accepted quantities establish that procurement was below the published requirement in both directions in these tests. D005 D006 That is evidence about accepted quantity, not proof that the guarantee failed to attract available flexibility. The supplied aggregates do not contain an offer stack, rejected bids, or offer-level locations and prices. An average across accepted procurement cannot show whether bids clustered at the guarantee, what a marginal offer cost, or why the remaining requirement was not procured. D001 D005 D006

Nor should procurement be mistaken for delivery. None of the six system-test windows has a non-null settled-volume observation in this extract: coverage is zero of two downward windows and zero of four upward windows. The reported settled-volume sums of zero therefore do not establish zero delivered response; settled delivery for these windows is unknown here. D005 D006 D001

For a fair test of price formation, the next evidence is window- and location-level offered capacity and bid prices, acceptance and rejection decisions, and subsequently settled delivery. Those would let readers separate the incentive to offer from NESO’s procurement choices and from physical response. The three-event sample cannot settle a market-wide participation question. D005 D006

Supported by National Energy SO Open Data