Published live (2026-07-04), in checkpoints — this header shipped before the evidence fleet returned, per the publish-as-you-go convention, and each section was committed as it was completed; the run finished the same day. The fourth definite treatment, and a dual-purpose run: it is the first treatment conducted from the treatment spec (NOTES-treatment-spec.md) rather than from its author — the spec's own first climb condition — so this run carries a second deliverable alongside the treatment: a spec-friction log, every place the spec was wrong, ambiguous, or missing a tool, printed at the end. Process honesty, printed first: this run is single-seat. The evidence fleet runs live (web tools available this time), but the pipeline's load-bearing property — adversaries independent of the synthesiser — cannot be satisfied: the same agent that synthesises will gate. Per the spec's own standing clause (the verdict is never stronger than the process that produced it), whatever verdict this run issues is capped at the strength a one-seat gate can confer, and the independent-adversary re-run is condition zero before the verdict is treated as load-bearing. A prior attempt at this run was killed externally mid-execution; nothing from it survives to bias this one.

Update, same day: condition zero is discharged — an independent second seat ran the adversary pass hours after publication. The verdict survives, further narrowed: passes-with-conditions — two-seat. Two rung-level claims are killed in place (rung 2's "immediately actionable"; rung 7 as written), three factual legs are corrected (Corris is not empty — the Machynlleth club covers it; the grant window is five months, not seventeen, and both schemes' application rounds are already closed; Llanfairfechan is on-substrate), one candidate discharge is regraded ("confirmed in split form" → consistent-with, untested) — and the re-level itself strengthened under attack: the organ's founding record is undercounted by this note's own audit. See the independent-adversary section below; the one-seat sections stand as the historical record.


The sphere, and what the treatment actually is

Next in backlog order: the other slate valleys — Blaenau Ffestiniog, Nantlle, Llanberis, Corris — flagged as pattern-replication targets. Which forces a question the spec does not answer (spec-friction item 1): what is the unit here? Four valleys are not a design. The design under treatment is replication itself: the ratchet cell's rung 8 — cell-in-a-box, the kit + framework + shared back-office sold to the next community — pulled out of its parent ladder and treated as a first-class template with its own rungs, its own gate, and one valley as the worked instance. The fourth treatment is the second cell.

That makes this run the sharpest available discharge test for the-cell-compounds, and the prediction is registered before the fleet launches, so the evidence can falsify it: a new valley is off-substrate by construction — no shared billing rail, no accumulated trust, no wrapper to amend. The candidate therefore predicts that founding costs recur in each new valley (the quasi-founding reappears: a share offer, a local champion, a constitution written not amended). The observer-architecture note predicts the split precisely: specification travels, continuity doesn't — the artefact kit, the legal templates, the metering runbooks should transfer at near-zero marginal cost, while the trust, the membership roll, and the institutional wrapper should have to be re-grown valley by valley at close to full price. If the evidence shows second and later cells forming dramatically cheaper across the board — continuity included — the candidate's off-substrate prediction fails and it should say so on its face. If the evidence shows the split — kit cheap, trust expensive — both the-cell-compounds and the observer-architecture frame gain their first measured case.

What the fleet found — and what it did to the framing

Every figure below from a source fetched 2026-07-04, confidence flagged. The fleet's first duty is contradicting the prior work, and it discharged it immediately: the backlog frames these valleys as future "pattern-replication targets." The live world says replication is not a proposal — it is mid-flight, a decade in, and the way it actually happened falsifies the naive reading of the parent design's rung 8.

The replication record — cell-in-a-box has never been observed; something else has

The four valleys — four different rungs of the same ladder

The regime — the election-contingent flag resolves, favourably and precariously

The slate voids — the research good is being bought retail, N times

No dedicated characterisation of the flooded slate-void system as a water or heat resource is discoverable — BGS holds general Wales hydrogeology references; coal has the Coal Authority's mine-water heat programme; slate has no equivalent. Green's premise survives at the strength "no discoverable dedicated study." But the sharper live finding is what exists instead: Blaenau alone has commissioned three overlapping feasibility studies in three years, each privately scoped, consultant-delivered, and non-cumulative. The characterisation is being purchased as retail fragments, valley by valley, instead of once as a regional commons.

The waitlist scan — two queues, at two tiers

Run per the spec, function identified at the level the queuers experience it. Tier one, households: the Bethesda club's closed waitlist (verified in the ratchet-cell run) and Corwen's closed-redirect are queues for cheaper electricity in the match windows — bill relief, already banked as the domain's adoption experiment. Tier two, communities: Energy Local runs a standing register-interest mechanism for places without clubs (depth unpublished), and a new group formed the moment £10k and DEG support were offered (Llanfairfechan). The second queue is not for electricity; it is for the founding service — the grant-writing, feasibility and club-establishment labour that DEG sells. The spec's scan assumes one queue per domain; this domain queues at both tiers of the adopter unit, and the design below enters through both doors. (Spec-friction item logged.)

The re-level — replication is not a shipping problem

The expansion pass's first question, per the spec: is this the right altitude? It wasn't. The backlog's framing — and the parent design's rung 8 — treats replication as a shipping problem: package the kit, sell it to the next valley, collapse setup from years to weeks. The evidence says the region never did that, and the observer-architecture frame says why: specification delegates, continuity doesn't. The templates, checklists and runbooks transfer at near-zero cost — and do not suffice, because what a new valley lacks is not documents but trust, a paid pair of hands, and a grant pipeline. The region's actual answer, running since 2017, is to mutualise the continuity at a tier above the cell: DEG/Cyd Ynni as a standing regional organ that carries the founding labour from valley to valley, funded once, amortised across every cell it serves. The trade-off the naive rung 8 couldn't dissolve ("kits are cheap but don't work; consultancy works but doesn't scale") dissolves at the higher altitude: the kit travels inside a shared organism. The cross-domain rhyme is old and load-bearing: this is how credit-union and co-operative federations replicated (a central body carrying founding services to each new society), and its absence is why forest school's unforced precipitation stayed thin. So the design under treatment re-levels: not cell-in-a-box but the second-cell ladder — the completion order for a valley, with the regional tier as rung 0 already standing.

The ladder — the second cell

Eight rungs. The adopter unit shifts as the ladder climbs — a household, then a generator's board, then a valley's civic vehicle — each named at its rung, per the anti-gerrymander discipline. Every rung individually rational on day-one terms except where confessed; no rung requires the next; the institution does not arrive late this time because it already arrived — the consortium precipitated in 2017, and this ladder's constitutional event count is zero.

Rung 0 — the regional organ, already paid. DEG/Cyd Ynni standing; its per-valley output is the valley dossier — roof census, generator candidates, EPC clusters, grid headroom, the local figures that turn the parent template's placeholders into prices. Grant-funded (Ynni Cymru £10m; Ymestyn to Nov 2026), which rung 0 is allowed to be; no rung above cites a grant except where confessed below.

Rung 1 — the household rungs, inherited verbatim. The ratchet cell's rungs 1–3 (hermit-and-cylinder, party-wall dyad, free-roof anchor) transfer to any valley unchanged — pure specification, re-priced not re-designed. This is the measured half of the inheritance claim: what compounds across cells without shared continuity is exactly the paper.

Rung 2 — the club around the standing asset (Llanberis's move). Where a community generator already exists, the cheapest unclimbed rung in the region: an Energy Local club needs the generator's board to say yes (one decision-maker, better than export price from the first matched kWh), members who join on Bethesda's observed record (~24% average savings, clubs that fill and waitlist), and coordination labour bought from the regional organ, not volunteered. No share offer, no new capital, no founding act. Ynni Padarn Peris has been generating for nine years without this rung — a £0-capex completion sitting untaken, which is itself evidence the kit doesn't self-execute.

Rung 3 — join the consortium, don't found the federation. Cyd Ynni membership: shared grant-writing, procurement, back-office. Day-one rational (cheaper services immediately); the federation is consumed as a service, never founded as a cause.

Rung 4 — the civic-rail entry (Nantlle's move) — the confessed rung. Where a valley holds a civic CBS but no generation: dossier first, then the free-roof-anchor pattern across the civic estate (Ynni Twrog's 13-building shape). Confession, printed per the grammar: post-FiT, new community generation does not reliably clear on self-consumption alone; this rung leans on the live capital-grant windows, and its day-one rationality is bought by a subsidy regime with dates on it. In the FiT era this rung stood alone; today it doesn't, and a ladder that hid that would be marketing.

Rung 5 — surplus seeding (the Llanfairfechan move, observed live). Established cells fund the next valley's dossier from their own surpluses — £10k scale, already happening. Second confession: the self-interest here is thin (each new cell spreads the regional organ's fixed costs across more members — real, but nobody writes a cheque for that reason). This rung runs on the same civic-blended motive the energy run found in community-share capital. The design carries one civic rung and says so, rather than inventing a rationale.

Rung 6 — the second commodity, on schedule. Each valley that completes rungs 2–3 becomes the substrate the second-commodity treatment priced: mobility (Yr Orsaf's three EVs are a pre-adaptation), then others, at marginal founding cost. The parent treatments' verdicts flow through unchanged.

Rung 7 — the research commons. The slate-void characterisation bought once, regionally, instead of retail per valley: Blaenau's three overlapping studies in three years are the price signal, and valley N+1's dossier is cheaper the moment the shared study exists. Individually rational for the next commissioning valley (a share of a commons costs less than a fourth consultant), and the first standing artefact the regional tier owns beyond its back-office.

Rung 8 — the voids (the headline, pre-demoted). Quarry-water heat at Tanygrisiau scale, void storage at Glyn Rhonwy scale, "the slate valleys as one organism" — quotable, and not load-bearing. The ladder pays in full if no void is ever tapped. Filed as a free option per the grammar; every run's headline has been its weakest claim and this run declines to test the pattern.

The clocks — typed before doctrine, per the spec

Three clocks, two of known type, one new:

The walls ledger

Wall Class Residue
Winter and the seasonal store Physical floor Inherited from the parent treatment unchanged: winter import is bought, per valley, forever.
Grid headroom (Manweb) Physical floor (with a political overlay, logged) Flexible/ANM connections dissolve the queue but not the copper: the residue is a curtailment cost, recurring, and a DNO investment cycle nobody in the valley controls.
No adopter unit (Corris) Physical floor, at valley grain The ladder needs a standing civic vehicle and Corris has none discoverable: this domain's conceded cohort, carried — if at all — by the regional organ plus the underserved-targeting money, and printed as uncarried until then.
Post-FiT generation economics Political contest Rung 4's day-one rationality is bought by grant windows with expiry dates; the founding path the existing substrate used (FiT-era hydro) is closed. The distributive fight over what replaces it is not this design's to win — confessed, not dissolved.
The counterparty rebuild Political contest Programme discontinuity during the merger; residue is the doctrine above and the risk it prices.
Anti-rails targeting Political contest The £1bn deliberately routes past compounding regions toward empty maps — a subsidy design that fights the very mechanism this treatment documents. Residue: the compounding region self-funds its next rungs while its least-developed valleys become someone else's grant category.
Heat-network authorisation Political contest The Tanygrisiau design walks voluntarily into the Ofgem heat-network regime the parent treatment's "share electrons, never heat" rule routes around: formal heat is bought at full regulatory price or not at all.
Kit-doesn't-carry-trust; research bought retail; founding-needs-capital Frame-artefacts Dissolved by construction (continuity mutualised at the regional tier; the rung-7 commons; the £0-capex club completion) — displacing, respectively: dependence on one regional organ (a new single point of failure, named), the commissioning autonomy each valley gives up, and the share-offer civic capital the club path never raises (with the civic-blended motive it therefore never binds).

The gate — one seat, three lenses, and the verdict that survives

Degradation printed first, again: the spec's load-bearing property — adversaries independent of the synthesiser — is not satisfied. One seat ran three adversarial lenses in sequence; what follows is honest attack but not independent attack, and the verdict is capped accordingly. The independent re-run is this treatment's condition zero.

Lens one — the catchment. The whole ladder is conditional on rung 0 existing, and rung 0 is not a document — it is a standing, funded regional organ that took a subsidy era and a decade to accrete. Inside the catchment (a region with a DEG-shape and civic-CBS density: Gwynedd, and whatever counts as its peers — count unknown, a condition), the ladder passes. Outside it, "replicate the pattern" collapses back into founding a regional organ — a founding act the bridge test forbids — and the backlog's pattern-export line (South Wales Valleys, Cumbria, Pennines) walks straight into this wall: regions compound cells, and nothing yet observed compounds regions. The gate passes the ladder inside its region and fails the export headline the backlog was hoping for; that sentence is the discrimination working.

Lens two — the rationality audit. Two of eight rungs are confessed (rung 4 grant-priced, rung 5 civic-blended). The grammar's clause 1 admits no exceptions, and the spec gives no confession budget (friction item), so the lens rules narrowly: the ladder is a true no-faith ladder only for asset-holding valleys — Llanberis-class completions, where every live rung clears on day-one terms with zero new capital. For empty valleys the ladder is honest but time-boxed: its rationality is bought by grant windows with published expiry dates, and the verdict must say "no-faith for completions, grant-priced for entries" rather than pretend one grade covers both.

Lens three — the evidence audit. The re-level's central claim (the regional organ is the replication mechanism) rests on one region's history, partly from secondary sources, with the primary page password-protected. And the honest timeline cuts against the strong reading: the first-generation cells (Ogwen, Padarn Peris, Anafon) were FiT-era foundings that predate the consortium — Cyd Ynni precipitated from the cell wave in 2017, and only then began seeding. The organ-as-mechanism claim therefore has exactly one post-window mechanism observation (Llanfairfechan, £10k, outcome unknown). The structure survives; the evidence for it is n=1 with the result not yet in.

Verdict as issued by the one seat: passes-with-conditions — one-seat. (Superseded the same day: the independent adversary ran, and the two-seat verdict — passes-with-conditions, further narrowed, with two rung-level kills and three factual corrections — is in the adversary's section below. This paragraph stands as the record of what the one seat claimed, so the delta is auditable.) The narrowed claim, stated exactly: bridge-clean, capital-free, and immediately actionable for valleys already holding a community generation asset (the club-completion move); grant-priced and window-dated for civic-rail valleys entering cold; conditional throughout on a standing regional organ whose replication power has one unresolved observation; and false as a portable export to organ-less regions, where the founding problem this design routes around reappears at full strength. Conditions: (0) independent adversary re-run; (1) ask Ynni Padarn Peris the nine-year question below before any claim about rung 2 ships; (2) track Llanfairfechan to an outcome — it is the design's only live test of rung 5; (3) count the catchment — how many UK regions actually hold a DEG-shape?; (4) have Nantlle and Corris dossiers ready before the £1bn window opens, or print that the doctrine wasn't followed.

The sharpest objection the run could not answer — the nine-year untaken rung

Llanberis has held the asset, the precedent one valley north, consortium membership, and access to the kit for nine years — and the club rung sits untaken. The design reads this as "the cheapest unclimbed rung in the region." The objection reads it as evidence against the design's core: if a rung is individually rational, capital-free, and served by a paid regional organ, and it stays untaken for a decade in the region that invented the pattern, then rationality-plus-kit-plus-organ is demonstrably not sufficient — something unnamed is binding. One candidate answer is disquieting for the whole series: the FiT contract. A generator already collecting FiT payments on every unit may gain little from a club's matching arrangement; what looks from the ladder like an unclimbed rung may be, from the generator's side, a rung not worth climbing until the FiT expires. If that is the explanation, the region's map of "incomplete cells" partly redraws as "contented cells," the addressable substrate shrinks again, and the completion move quietly acquires a clock of its own (FiT expiries, 2035–2040). The run cannot resolve this from a desk; it is condition 1, and it attacks the design's strongest rung, which is exactly where the sharpest objection should live.

What this run gave the candidates

The run obligations, discharged explicitly — five standing candidates this time:

the-cell-compounds — the run this treatment was built to serve, and the prediction registered in the header comes back confirmed in split form. Off-substrate degradation happened exactly as predicted: every first-generation valley re-paid its own continuity costs in full (Ynni Padarn Peris ran its own £280k share offer with Ogwen's precedent standing one valley away), while specification travelled free (the household rungs transfer verbatim, re-priced not re-designed). The observer-architecture split — kit cheap, trust expensive — now has its first field case. And the candidate gains a recursion it must now carry: the compounding unit has a second tier. Regions compound cells the way cells compound commodities — DEG/Cyd Ynni is to the next valley what the energy cell was to mobility — and the internal-enclosure objection gains a concrete face at the new tier: a regional organ holding all founding labour is a chokepoint-in-waiting, and the line between a compounding region and a regional monopoly on founding is the same line the candidate already owes at cell grain. This moves the candidate; it does not climb it — the degradation was read from the historical record, not measured in a controlled run, and the seeding mechanism has one observation with the outcome open.

bridge-catchment — the catchment was drawn at region grain for the first time: inside, regions holding a standing paid organ plus civic-CBS density; outside, everywhere the backlog's pattern-export line points. The anti-gerrymander discipline met a new case — a ladder whose adopter unit legitimately shifts rung by rung (household → generator board → civic vehicle) — and the rule that survived is: name the unit at every rung and re-apply the median test at that rung's own unit (friction item: the spec assumed one unit per design). The conceded cohort at valley grain is Corris: named, priced at "carried by the regional organ plus the underserved-targeting money, or uncarried."

institutions-precipitate — a third regime case, and a counterweight to the occlusion wrinkle. Cyd Ynni precipitated from the FiT-era cell wave unforced (nothing legally required a consortium) but fast — two years from cell wave to wrapper, against forest school's nineteen — suggesting a refinement: unforced precipitation speed scales with the density of live trades among the precipitating parties (cells sharing one grant regime and one DNO federate quickly; dispersed practitioners barely federate at all). And where the FSA guards its definition against the widest catchment, Cyd Ynni seeds new members — the precipitate that shares a service opens; the precipitate that guards an identity closes. That distinction, if it holds, is the occlusion wrinkle's scope condition.

find-the-waitlist — the scan returned queues at two tiers of the adopter unit (households for bill relief; communities for the founding service), both at function level per the curriculum run's precision clause. Lens upgrade offered back to the spec: run the scan once per tier of the adopter unit, not once per domain.

the-closing-window — the typology grew and held. The opening-window type recurred (the £1bn launch — second observation, so the curriculum run's type is no longer a singleton), the closing type recurred (Ymestyn, seventeen months), and one genuinely new type surfaced: the counterparty rebuild (the merger — the rail re-laid mid-run, doctrine: convert programmes into contracts before the churn). Four runs, four types, and "type the clock before writing the doctrine" has now earned its place as the candidate's operating rule.

Corrections owed to earlier notes, logged for reconciliation: the second-commodity treatment's "~100 Energy Local-style clubs" reads against a live listed count of 36; and the ratchet cell's "Ynni Cymru — election-contingent" flag resolves to "funded, and being merged into a national energy company" — right risk, wrong direction: the programme survived the election and may not survive its own reorganisation.

The spec-friction log — where the spec was wrong, ambiguous, or missing a tool

This run's second deliverable: the first field test of NOTES-treatment-spec.md, operated from the document, logging every place it failed its operator. The spec's own climb condition asked for exactly this.

  1. The sphere→design gap (missing tool). The spec defines the treatment unit but gives no rule for deriving a design from a backlog sphere entry. "Other slate valleys" is four places, not a template; a pre-step had to be invented (name the template the sphere implies) before stage one could run.
  2. Recursion silence (missing rule). When the design under treatment is a rung of a prior treatment (this run gates the ratchet cell's rung 8), the spec says nothing about whether the parent's conditions flow down. This run treated parent verdicts as flowing through rung 6 untouched; a different operator could defensibly re-litigate them. The spec should choose.
  3. No grade for a one-seat gate (ambiguity, load-bearing). The spec grades fleet degradation (print first, weaken verdict) but not adversary degradation. Its three verdict grades are precedents, not a closed set, and it never says whether the set is open. "Passes-with-conditions — one-seat" had to be coined; the spec should either admit new grades explicitly or close the list.
  4. The scan's single-queue assumption (wrong, mildly). "Scan the domain for queues" assumes one queue per domain; this domain queues at two tiers of the adopter unit with two different functions. Upgrade: scan per tier.
  5. The clock typology's open/closed status (ambiguity). A fourth clock type was needed. The spec presents three types as an inventory without saying whether new types are spec amendments or candidate moves. Treated here as a candidate move on the-closing-window; the spec should say which.
  6. Time-indexed walls (ambiguity). Grid headroom is a physical floor this afternoon and a political contest over a DNO investment cycle. The three-class ledger has no rule for walls that change class with horizon; classified as floor-with-overlay and logged.
  7. The shifting adopter unit (gap in the gate). The median test assumes a fixed adopter unit per design. A completion-order ladder legitimately changes unit per rung; the anti-gerrymander guard needed the invented extension (name the unit at each rung, median-test at that unit) to stay operable.
  8. "The institution arrives late" has no branch for "already arrived" (minor). On a pre-precipitated substrate the clause is satisfied vacuously; the spec should say that consuming a standing institution counts as compliance, not as skipping the clause.
  9. No confession budget (gap, exploitable). Clause 1 admits no exceptions; clause 4 licenses confessing what can't be paid for. Nothing says how many confessed rungs a ladder may carry before it stops being a no-faith ladder. This run carried two of eight and split the verdict by valley class — a rule the spec should own rather than each run improvising.
  10. Evidence-log granularity (missing convention). How much raw fleet material to print is carried by house precedent, not the spec; a spec-only operator wouldn't know what the reader is owed.
  11. Instance multiplicity (gap). The template/instance dichotomy assumes one worked instance. This sphere offered four; resolving them as entry points at different rungs worked well — worth promoting from improvisation to instruction.
  12. What the spec got right, confirmed under field conditions: the fleet's contradict-the-prior-work duty (it killed the shipping frame inside the first hour); registering a candidate's prediction before the fleet launches (it made the the-cell-compounds discharge clean and unfudgeable); print-degradation-first; and the pre-demoted headline (the voids were this run's most quotable material, and pre-demotion kept them from bearing any load). None of these needed the author present — which is the delegation property the spec exists to have.

The independent adversary — the second seat (2026-07-04, same day)

Condition zero, discharged the day it was set. This section is written by a second seat, independent of the synthesiser, under the standing instruction to form its own attack from the corpus before adopting any of the note's framing — and to use the run's own evidence against it wherever possible. Process, printed per the house rule: this is a gate pass, not a second evidence fleet. The attack was armed from the prior notes (the ratchet cell, the second commodity, the cell-compounds candidate, the forest school, the observer architecture, the backlog) plus four targeted live checks run 2026-07-04 — the Ymestyn and Ynni Cymru guidance pages on gov.wales, the Energy Local clubs page, and two secondary passes on Cyd Ynni/DEG — each flagged where it bears load. Per-club founding dates and per-club DEG attribution could not be verified from a desk and are carried as conditions, not claims. The pass found the note wrong in both directions — including wrong against itself.

The attack that inverted — the organ's record is undercounted, by this note

The adversary's opening attack was the obvious one: lens three graded the organ-as-mechanism claim at n=1 with the outcome open, so kill the re-level as a hypothesis dressed as a finding — nine years of organ, £435k of grants, and no completed replication in evidence. The attack died on contact with the clubs page. The live list carries a North Wales cluster of roughly ten clubs — Aber (covering Abergwyngregyn and Llanfairfechan), Glaslyn (Beddgelert), Gwyrfai (Bro Gwyrfai and Caeathro), Pendraw'r Byd (Llŷn), Wnion (Bala to Barmouth), Machynlleth, Dyffryn Banw, Llanidloes, Corwen with its Glyndwr overflow, and Bethesda — roughly a third of the national 36 in the one region where the organ stands, and the consortium's 2017 Lottery award was explicitly mandated to establish Energy Local clubs. The Community Energy England town list this note flagged as garbled maps, town by town, onto that club geography (Llithfaen and Llannor → Pendraw'r Byd; Bethel, Llanrug, Caeathro → Gwyrfai): the garble was signal. If even half of those clubs are post-2017, organ-assisted foundings, the mechanism has several observations, not one, and the note undercounted the evidence for its own central finding. The re-level strengthens under independent attack — with per-club attribution still owed (condition below).

The same map, turned around — two kills

Kill one: rung 2 as claimed, and the verdict's strongest clause. The one-seat verdict's opening words — bridge-clean, capital-free, and immediately actionable — do not survive the club map. If the organ has carried club foundings across half of Gwynedd since 2017, then Llanberis is not a valley the mechanism hasn't reached; it is a valley the mechanism has stood beside for nine years — DEG's own supported-towns list includes Llanberis, YPP is a consortium member, and clubs formed in the adjacent valleys — and no club exists. The base case flips: not unclimbed, declined. And the note's own exculpatory hypothesis cannot carry the flip alone: the FiT-contentment reading must also explain why FiT did not content Ogwen, whose FiT-era 100kW hydro anchors the Bethesda club in the parent treatment's own record. Meanwhile the rung's one positive instance — Bethesda — was the UK's first club, a nationally supported pilot with the model's originators in the room: an assisted founding, not a bridge-crossing. Rung 2 therefore has zero unassisted observations and one nine-year counter-observation, and there is a squeeze behind the kill: inside the catchment, ten clubs' worth of completion has already happened, so the flagship move may be addressing a residue of hard cases rather than a frontier of cheap ones — the treatment's own export-kill, turned on its own flagship. The rung survives as a designed move not yet offered (nothing in evidence says the organ ever priced this specific completion to YPP's board); the claim that it is the region's cheapest unclimbed rung, available today, dies. "Immediately actionable" comes off the verdict.

Kill two: rung 7 as written. "A share of a commons costs less than a fourth consultant" is true only for the fourth commissioner. For the first, a regional slate-void characterisation costs more than the narrow feasibility study they actually need — the rung is day-one irrational for exactly the mover it needs first, which is the collective-action shape the bridge test exists to forbid. And the run's own evidence is revealed preference against the commons: Blaenau had three consecutive chances to commission cumulatively and bought retail all three times. The rung works only if the regional organ fronts it on grant — which makes it the ladder's third grant-priced rung, unconfessed, in a rationality audit that counted two. Confess it or restructure it (the organ fronts, valleys buy dossier-shares priced below marginal retail — which is rung 0 wearing another hat, and should say so).

The factual corrections

  1. Corris is not empty. The Machynlleth club's own coverage line reads "Machynlleth, Pennal, Corris, Eglwysfach & surrounds" — the fleet saw the cluster and placed it five miles south of a village it includes. The walls-ledger entry ("no adopter unit; conceded cohort; printed as uncarried") falls as written, and falls to the run's own two-tier lens, never applied to its own control valley: at household tier Corris is carried today, by an existing club; what Corris lacks is the generation-asset and civic-vehicle tiers. The control-valley reading, and the "underserved" framing that routes £1bn-window money at Corris, both need re-arguing at the right tier.
  2. The clock arithmetic, and worse behind it. 30 November 2026 is five months from this note, not seventeen. And the live guidance is harder than the confession: the application windows of both schemes rung 4 leans on are already closed (Ymestyn closed 6 February 2026; the Ynni Cymru 2026–27 round closed 21 May 2026, projects to complete by 31 March 2027). As of publication day, rung 4 is not window-dated — it is windowless, pending a next round that the counterparty rebuild makes uncertain. The rung's confession was honest and still understated its own condition.
  3. Llanfairfechan is on-substrate. The Aber club's coverage includes Llanfairfechan. The design's only live test of rung 5 is therefore a seeding inside an existing club catchment — evidence for on-rails extension, structurally incapable of testing the off-substrate founding cost the header's prediction was registered about.

The grading corrections

"Confirmed in split form" is overgraded. The continuity half rests on the FiT-era foundings each paying full price — but those foundings were contemporaneous (Ogwen's raise and YPP's raise ran within the same window; neither had a decade of standing substrate the other could have inherited), so they cannot test a prediction about what a second cell inherits from a first. The specification half rests on household rungs said to transfer verbatim — but no second valley has run them from the kit, and the note's own fleet found the kit does not exist as a self-serve artefact. Add correction 3 above and the honest grade is consistent-with, untested at the tier the prediction was registered for — the discharge section's "first measured case" and "first field case" claims are withdrawn; "moves, does not climb" stands. The registered-prediction discipline held; the grading was done by the seat that registered the prediction, and it graded generously. One more wrinkle the run did not flag: its own re-level makes the registered prediction partly ill-posed — the prediction declared a new valley "off-substrate by construction," and the run's central finding is that no Gwynedd valley has been off-substrate since 2017, because the substrate has a regional tier.

The export-kill overreaches. "Nothing yet observed compounds regions" is an absence claim from a fleet that searched one region. The corpus itself carries counter-evidence: Egni Co-op — the parent treatment's rung 3 existence proof — operates the free-roof-anchor model at scale across South Wales, which means the backlog's first export target probably already holds a DEG-shape; and Energy Local itself is a national body whose club list demonstrably extends across regions (Dyffryn Banw and Llanidloes sit in Powys, outside the consortium's patch). What exports may be neither the kit nor the region but the national specification body plus a local continuity organ — which is the observer-architecture split operating at one more tier up, and exactly what the catchment count (condition below) should now look for. The kill survives against genuinely organ-less regions; the absolute sentence does not.

Smaller findings, carried

Rung by rung

Rung Verdict The condition, or the kill
0 — regional organ pass-with-conditions Mechanism better-evidenced than the run's own audit; per-club founding attribution still owed.
1 — household rungs pass-with-conditions Parent's five undischarged conditions flow down and must be restated, not waved through.
2 — club completion fail as claimed Zero unassisted observations; one nine-year counter-observation; base case is declined, not unclimbed. Survives as a designed move not yet offered — priced, to a named board, with the refusal reason recorded.
3 — join consortium pass Observed (members exist); trivially day-one rational.
4 — civic-rail entry pass-with-conditions Confession stands and understated: no application window is open today; re-dated to the next round or printed as dormant.
5 — surplus seeding pass-with-conditions n=1, on-substrate, outcome open; establish whether the £10k decision sat with member cells or with the organ.
6 — second commodity pass-with-conditions Inherited verdicts and conditions; substrate correction (36, not ~100) already logged.
7 — research commons fail as written Day-one irrational for the first mover; revealed preference runs against it; grant-fronting must be confessed or the rung restructured.
8 — the voids pass Properly pre-demoted; bears nothing; the pattern holds a fourth time.

The verdict — two-seat

Passes-with-conditions — two-seat, further narrowed. The independent seat confirms the ladder and strengthens the re-level: the regional organ is better evidenced as the replication mechanism than the one-seat audit dared claim. What it does not confirm is the verdict as worded. The two-seat narrowed claim, stated exactly: a completion-order map for valleys inside a region holding a standing paid organ, whose household rungs inherit the parent treatment's verdict and conditions; whose flagship completion move (rung 2) is a coherent design with zero unassisted observations and one nine-year counter-observation, and ships as a move to be offered and priced, not a move available today; whose cold-entry rungs are grant-priced against windows that are currently closed; and whose export claim fails against organ-less regions while the question of what compounds regions — a national specification body over local continuity organs — is opened, not closed, by this run's evidence. Two rungs' claims are killed in place (2 as claimed, 7 as written); three factual legs are corrected (Corris, the clock arithmetic and window status, Llanfairfechan's substrate); one candidate discharge is regraded (below); and the design's structure — organ over kit, entry-by-rung, the typed clocks, the pre-demoted headline — survived an adversary that arrived intending to kill the re-level and could not.

Conditions, renumbered for the two-seat verdict:

  1. The nine-year question, extended. Ask Ynni Padarn Peris — and any answer must also explain why the same consideration did not stop Ogwen's FiT-era hydro anchoring the Bethesda club. Record the refusal reason; it is the design's missing wall.
  2. Name the foundings. Per-club founding dates and attribution (DEG/Cyd Ynni, Energy Local nationally, or self-founded) for the North Wales cluster. This is what turns the re-level from strengthened-hypothesis to measured mechanism — and it also answers whether rung 2's addressable substrate is a frontier or a residue.
  3. Track Llanfairfechan to an outcome, including where the £10k decision sat — noting it tests on-substrate seeding only.
  4. Count the catchment at region grain, starting from probable n≥2 (Gwynedd; South Wales via Egni Co-op) and looking specifically for the national-body-plus-local-organ split.
  5. Rung 7 restated with the first mover's day-one arithmetic printed, or confessed as the third grant-priced rung.
  6. Rung 4 re-dated when a 2027 window opens — or printed as dormant, with the dossier doctrine (condition on Nantlle unchanged; Corris re-argued at the right tier) carried to whatever window the rebuilt counterparty actually offers.

What this establishes

The gate has now run four times and behaved a fourth distinct way: narrowed a claim (energy), split a design (mobility), weakened itself for process honesty (curriculum), and here failed the sphere's own framing while passing the design underneath it — the backlog asked for pattern export, and the honest answer is that the pattern exports one rung down (cells, inside a region) and not one rung up (regions, between which nothing compounds). The spec survived its first authorless outing with twelve friction findings and no structural failure: the stages ran in order, the grammar held under two confessions, and every invented patch was a refinement, not a repair. And the treatment series' quiet arithmetic continues: the first run built a cell, the second run stacked a commodity on it, the third tested the pattern off-domain, and the fourth finds the tier above — each run discovering that the unit it studied was standing on a larger one.


Not yet promoted into propositions.ts — capture, not climb. Forming register; scaffold name. Climbs when: (1) the independent adversary re-run confirms or breaks the one-seat verdictdischarged same day: the second seat confirmed the ladder, killed two rung claims in place, and strengthened the re-level (see the adversary section); (2) the nine-year question is answered from Llanberis, not a desk — is rung 2 unclimbed, declined, or not worth climbing, with any FiT explanation also covering Ogwen, and the addressable-substrate count corrected accordingly; (3) Llanfairfechan reaches an outcome, giving rung 5 its first completed observation — noting it tests on-substrate seeding only; (4) a Nantlle dossier exists on the day the £1bn window opens (Corris re-argued at the right tier first) — the counterparty-rebuild and opening-window doctrines followed in the world, not just written down; (5) the foundings are named — per-club dates and attribution for the North Wales cluster, turning the re-level from strengthened hypothesis into measured mechanism.