For sponsor Asha Vale · Investment committee · 7 May 2027

Hold the final payment. Establish what the homes actually gained.

The defensible provisional gas effect is −3.48% to +9.83% across the same 48 homes. Even the favourable bound misses the 20% target. Authorise the specified £3,000 checks and return to the committee on 18 June 2027. [16-A] [16-C] [16-D] [16-G]

Synthetic evidence · No real-world certificationRecommendation, not approvalOriginal dated position retained below

The claim Asha can defend

£50,000 remains held

Initial assessment as at 7 May 2027 · Fixed evidence position; scenario controls do not rewrite it.

“For the same 48 refurbished homes, our provisional weather-normalised, project-attributable annual gas saving is bounded between −21,750 and +61,450 kWh, or −3.48% to +9.83% of their 625,000 kWh P1 baseline. The negative end represents a possible net increase. This is a deterministic sensitivity range under the committee’s convention, not a confidence interval. No permitted combination demonstrates the required 20% reduction. Comfort improvement is reported by 18 of 20 voluntary respondents, but portfolio-wide non-worsening is not established. We recommend retaining the £50,000 final payment pending the committee’s review of targeted monitoring.” [16-A] [16-B] [16-C] [16-D] [16-F]
Same 48 homes · Fixed P1 normalised baseline625,000 kWh20% target = 125,000 kWh/year
Full attributable effect bounds−3.48% to +9.83%Best case still 63,550 kWh short
Current-price bill equivalent / year−£2,392.50 to £6,759.50Modelled gas effect × £0.11; no standing-charge saving
Provisional avoided emissions / year−4.002 to +11.307 tCO₂e; negative = added emissions

The £360,000 work cost does not establish realised benefits. Comfort has no assigned monetary value. Asha recommends; only the committee can accept the evidence and release payment. Failure to demonstrate the target creates no automatic capital clawback. Deferral does not imply that monitoring is expected to produce a passing result. [16-A] [16-E] [16-G]

Audit the published claim

Original report 48 refurbished homes; gas fell 35%; annual bill savings £15,456. The original stored inputs are P1 625,000 kWh, P2 404,200 kWh and £0.07/kWh. They remain visible here after every scenario change. [16-F]

What the arithmetic actually compares
Workbook itemReconciliationAssurance conclusion
625,000 kWh P118 × 13,600 + 12 × 14,600 + 10 × 10,500 + 8 × 12,500All 48 treated homes; valid P1 total.
404,200 kWh P218 × 9,800 + 12 × 10,400 + 10 × 10,300Only 40 observed treated homes; T-D's eight readings are unresolved, not zero.
35% reduction(625,000 − 404,200) ÷ 625,000 = 35.328%Population mismatch; also raw gas, with no weather or attribution adjustment.
£15,456 “annual bill savings”(625,000 − 404,200) × £0.07 = £15,456Same population mismatch; P1 price applied to a false saving. Neither actual bill movement nor current-price attributable avoided bills.
Valid observed raw comparisonSame 40: 525,000 → 404,200 kWh; reduction 120,800 kWh / 23.01%An observed consumption change, not the 48-home target result or a causal claim.

Reconciliation from the cohort extract and billing rules. The eight unresolved homes account for 100,000 kWh in P1; omitting their P2 readings inflates the headline's subtraction by that full amount. [16-B] [16-E] [16-F]

Test the uncertain assumptions

The initial selection is the lower-benefit corner, not a best estimate. Adjust the three bounded inputs to inspect a hypothetical scenario. All eight T-D homes may share either bound; no independence, distribution or midpoint is assumed. [16-D]

Changes are local and temporary. They do not validate a meter, modify source records, authorise monitoring or approve payment. Reload or Reset restores the initial lower-bound selection. The full evidence range and original committee claim remain unchanged.
500–900 normalised kWh per replaced home. Eight of 12 comparison homes had replacements. Weighted add-back is 8/12 × this value; do not scale it for weather again. [16-D]
2,000–4,000 normalised kWh per home from additional occupancy/care needs. Added to the untreated counterfactual, not to observed meter use. Applies to 10 homes. [16-D]
8,000–14,000 actual kWh per home, including 2,000 non-heating. The control applies one illustrative value to all eight homes; their cohort mean is sufficient for the aggregate. Actual readings remain unresolved. [16-B] [16-D]

Selected attributable gas saving · all 48
Current-price annual bill equivalent
Annual avoided emissions
Gas target status · 125,000 kWh required

Gas effect range, selected scenario and target 0125,000 target −21,750+61,450 Annual attributable kWh; positive = saving, negative = increase

Full deterministic evidence range. Green marker = selected assumptions. Red line = 20% threshold. The favourable bound reaches only 49.16% of the required 125,000 kWh saving.

Calculation bridge · normalised kWh per dwelling unless stated

N(P) = non-heating(P) + [total(P) − non-heating(P)] × 2,000 / HDD(P) C-A common reduction = 13,400 − [2,000 + (10,400 − 2,000) × 1.25] = 900 Corrected C-B P2 = 13,225 + (8/12 × boiler saving) C-B common reduction = 14,800 − corrected C-B P2 Attributable saving = P1 normalised − P2 normalised − common reduction + T-C counterfactual increment Portfolio % = Σ(homes × attributable saving per home) / 625,000 × 100

P1 is already at 2,000 HDD. P2 heating gas alone is multiplied by 1.25; non-heating is never weather-scaled. T-C P2 non-heating rises to 2,600. All calculations use unrounded values. [16-B] [16-C] [16-D]

Selected sensitivity: observed totals and assumptions are kept distinct
Cohort / homesP1 normalisedP2 normalisedP1 − P2Common reductionAdded untreated demandUntreated P2 counterfactualAttributable saving / homeAggregate saving

T-D's displayed normalised P2 is a model input derived from the selected bound, never an observation. The independently measured 2,000 non-heating kWh is retained. T-C's extra demand is treated as something that would have occurred without the project; its attributable saving can therefore be positive despite higher observed normalised consumption. [16-B] [16-D]

How each uncertainty changes the portfolio

Uncertainty across supplied boundsPer-home saving rangePortfolio range contributionDirection and interpretation
T-A · 18 homes950 kWh17,100 kWh fixedUnder the stipulated C-A common trend.
C-B boiler correction → T-B · 12 homes883.33–1,150 kWh10,600–13,800 kWh; width 3,200 kWh / 0.512 ppLarger non-project boiler savings imply a smaller common trend and a larger estimated T-B effect.
T-C untreated demand increment · 10 homes−625 to +1,375 kWh−6,250 to +13,750 kWh; width 20,000 kWh / 3.200 ppHigher counterfactual demand raises the estimated project effect; the meter observation is unchanged.
T-D meter bounds · 8 homes−5,400 to +2,100 kWh−43,200 to +16,800 kWh; width 60,000 kWh / 9.600 ppLargest uncertainty. Lower actual consumption yields larger savings; all eight may share an extreme.
Combined portfolioSame 48 homes−21,750 to +61,450 kWh; width 83,200 / 13.312 ppCompatible extremes combined directly, without probabilities or offset assumptions.

Here “pp” means percentage points of the fixed 625,000 kWh P1 denominator. No supplied constraint excludes these joint corners. Bounds describe these specified uncertainties; they do not capture every possible unobserved cause. [16-A] [16-C] [16-D]

What the two pre-periods can tell us

  1. P0 · Apr 2023–Mar 20241,800 HDD · pre-intervention
  2. P1 · Apr 2024–Mar 20252,000 HDD · target baseline
  3. Apr 2025–Mar 2026Installation · excluded
    Commissioned 31 March 2026
  4. P2 · Apr 2026–Mar 20271,600 HDD · evaluation period
Normalised use at 2,000 HDD · observed P2 shown only where validated
CohortHomesP0 normalisedP1 normalisedP0 → P1 reductionP2 normalised, before adjustmentsPre-trend comparison
T-A1814,222.2213,600622.2211,75022.22 kWh more pre-reduction than C-A.
T-B1215,322.2214,600722.2212,47522.22 kWh less pre-reduction than C-B.
T-C1010,911.1110,500411.1112,225188.89 kWh less pre-reduction than C-A; comparability weaker.
T-D813,111.1112,500611.11Unknown11.11 kWh more pre-reduction than C-A; post evidence still absent.
C-A2014,00013,40060012,500Stable; no boiler change or refurbishment.
C-B1215,544.4414,800744.4413,225Post-period contaminated by eight other-programme boilers.

The similar T-A/C-A and T-B/C-B pre-reductions modestly support the chosen comparisons, as do the stable-household records. A single pre-period change in cohort means cannot establish parallel trends or causal identification. T-C's weaker pre-alignment and changed care/occupancy needs weaken an unadjusted comparison; the independently documented increment addresses only the specified change. Correcting C-B removes a bounded known intervention effect, not all confounding. T-D remains unobserved after commissioning. [16-B] [16-C] [16-D]

This brief follows the committee's P1–P2 convention. P0 is a diagnostic only; it is not averaged into the target baseline or used to assert statistical significance.

Three different benefit claims

1. Actual bills increased for the observed 40

Same T-A + T-B + T-C homesP1P2
Homes4040
Actual gas525,000 kWh404,200 kWh
Energy tariff£0.07/kWh£0.11/kWh
Energy charges£36,750£44,462
Standing charges · 40 homes£3,520£4,480
Total actual bill£40,270£48,942

£8,672 higher (+21.53%), despite 23.01% lower raw gas use. Per home: £1,006.75 → £1,223.55. T-D is excluded from both periods; its actual post bill remains unknown. [16-B] [16-E]

Inspect the bill movement bridge
Volume effect at P1 price: (404,200 − 525,000) × £0.07 = −£8,456 Price effect on P2 volume: 404,200 × (£0.11 − £0.07) = +£16,168 Standing charge effect: 40 × (£112 − £88) = +£960 Actual total movement = −£8,456 + £16,168 + £960 = +£8,672

This is a billing decomposition for the same 40 homes. The volume component is not a project-attribution estimate.

2. Current-price avoided bills are conditional

The full provisional 48-home annual effect, valued at P2 prices, is −£2,392.50 to +£6,759.50. This compares modelled normalised consumption with a hypothetical untreated counterfactual; it is not an observed refund or reduction in actual payments. [16-C] [16-D] [16-E]

Current-price bill equivalent = attributable kWh × £0.11 Avoided CO₂e (tonnes) = attributable kWh × 0.184 / 1,000 No avoided standing charge: all homes remain on gas.

Rates and conversion are supplied exercise constants. This is a one-year P2-normalised equivalent, not a forecast of future tariffs, persistence or lifetime return. A negative amount means a potential extra cost/emissions equivalent. No comfort valuation is added. [16-A] [16-E]

3. Comfort evidence is encouraging but selective

18 of 20 voluntary respondents report improvement; replies cover 20 of 48 homes (41.67%). Ten of the 18 improvements are T-C, the cohort with changed needs. This is not 90% of all residents, and the remaining two replies are not established here as “worsened”. [16-F]

CohortReplies / homesImproved
T-A4 / 18 (22.22%)Not separately supplied
T-B4 / 12 (33.33%)Not separately supplied
T-C10 / 10 (100%)10
T-D2 / 8 (25%)Not separately supplied

There is no representative temperature series or nonrespondent assessment. “Comfort not worsened” across the portfolio is not demonstrated. Care and occupancy changes may contribute to reported comfort; the responses alone do not establish attribution. [16-F]

Decision requested: a bounded £3,000 assurance programme

Recommend all three specified checks now, with a six-week return on 18 June 2027 if authorised on 7 May. The priority is credible realised-benefit reporting and comfort assurance; narrowing existing bounds alone cannot make their current upper end reach 20%. The committee may defer its payment decision. [16-G]

Proposed authorised workCostWhy this uncertainty mattersProposed ownership and return evidence
Repair all eight T-D meter mappings; validate annual readings£1,500Addresses 60,000 kWh of the current 83,200 kWh range width (72.12%). Most valuable for distinguishing possible net increase from saving.Ben coordinates the specified checks; Mei reconciles the eight readings and recalculates the same 48-home register. Retain raw readings, mapping checks and exceptions.
Independent representative comfort follow-up£900Addresses the separate no-worsening condition and voluntary-response bias. No numerical energy-range reduction claimed.Ben arranges the independent follow-up; committee return records coverage, comfort outcomes, nonresponse and limitations. Do not infer guaranteed representativeness or findings in advance.
Individual C-B boiler histories and commissioning dates£600Refines the other programme's contamination and timing: up to 3,200 kWh / 0.512 pp of current range width. Records may leave some uncertainty.Ben obtains the specified histories; Mei checks the eight replacements and correction basis, retaining unresolved timing or saving limits.
Total requested · within the one-off budget£3,000£0 unallocated. These are proposed assignments and authorisations, not decisions already made.

Scope and availability come from Ben's monitoring options. Coordination and reporting assignments here are proposed. Consent covers these checks and the next year of meter access; this brief proposes no extra data acquisition. The T-C increment remains a documented 2,000–4,000 kWh/home sensitivity unless authorised evidence supports a defensible revision. [16-D] [16-G]

Owned return, not an open-ended hold

Ben: own the benefits register and quarterly review. Preserve baseline versions, all 48 identifiers and denominators, meter status, counterfactual assumptions, comfort coverage, payment status and source links.

Mei: use the available one analysis day per quarter for a reproducible update and exception note. Plan the initial reanalysis for the June return within that allowance; do not assume unlimited analyst capacity.

Asha: present the dated recommendation on 18 June. Committee: accept or reject the updated evidence and decide the held £50,000.

What would change the decision?

A validated 48-home result must be assessed against the unchanged 125,000 kWh threshold and a supportable no-worsening comfort finding. Narrowing the supplied range alone will not establish the gas target; any new contrary result needs transparent source reconciliation and committee scrutiny.

At the June review, record target demonstrated / not demonstrated / unresolved, the accepted claim and payment resolution. If evidence still fails, retain an explicit committee decision rather than implying automatic payment or clawback.

Proposed quarterly review dates after the return: 18 September and 18 December 2027, then 18 March 2028, within the supplied next-year monitoring scope. These are planning dates, not scheduled actions. [16-A] [16-G]

Original evidence and assumptions

The following supplied fictional records are the source of this brief. Open a record to inspect the underlying figures. No external observations, statistical significance or approval have been added.

16-A · Benefits agreement · 1 March 2025

Portfolio is the same 48 treated dwellings throughout. Target: at least 20% reduction in weather-normalised total gas attributable to refurbishment, with comfort not worsened. Divide aggregate attributable kWh by the same 48 homes' P1 normalised total gas in every scenario. £50,000 final payment is held until the committee accepts the evidence. Asha recommends; the committee releases. No automatic capital clawback if target is not demonstrated. Work cost £360,000. Comfort was not monetised. No homes disconnected from gas.

16-B · Validated cohort extract · Mei Tan · 30 April 2027

Rows are distinct homes; all values are arithmetic cohort means in kWh per home. Non-heating was independently measured, is included in total gas, and is not weather-scaled.

CohortHomesP0 totalP1 totalP2 totalNon-heating P0 / P1 / P2Status
T-A1813,00013,6009,8002,000 / 2,000 / 2,000Stable; matched C-A
T-B1214,00014,60010,4002,100 / 2,100 / 2,100Stable; matched C-B
T-C1010,00010,50010,3001,800 / 1,800 / 2,600Occupancy/care needs changed
T-D812,00012,500Unresolved2,000 / 2,000 / 2,000No validated point estimate
C-A2012,80013,40010,4002,000 / 2,000 / 2,000Stable; no refurbishment or boiler change
C-B1214,20014,80011,0002,100 / 2,100 / 2,100Eight other-programme boilers

Periods: P0 April 2023–March 2024 (1,800 HDD); P1 April 2024–March 2025 (2,000 HDD); installation April 2025–March 2026 excluded; all commissioned 31 March 2026; P2 April 2026–March 2027 (1,600 HDD). Evaluation 7 May 2027.

16-C · Committee analysis convention · 2 May

Normalise each period to 2,000 HDD as non-heating gas + (total − non-heating) × 2,000 / actual HDD. Compare P1 with P2, using P0–P1 to inspect pre-trends only. Provisional effect per home = treated P1 normalised use − treated P2 normalised use − matched comparison group's normalised reduction. Use C-A's absolute reduction for T-A, T-C and T-D, and corrected C-B for T-B. This stipulated convention does not prove that all unobserved causes are controlled. An alternative sensitivity may supplement but not replace it.

16-D · Confounders and incomplete records · 2 May

Eight C-B boiler replacements account for 500–900 weather-normalised kWh saved per replaced home. Add their weighted saving back to C-B P2 before estimating the common trend, without scaling this adjustment for weather again.

T-C's independently documented occupancy and care-related temperature requirements would have increased untreated P2 normalised use by 2,000–4,000 kWh/home above P1 less the C-A trend, without refurbishment.

Each T-D home's validated annual P2 total bound is 8,000–14,000 kWh. All eight may share an extreme. These are deterministic bounds; no probabilities, independence assumption or midpoint claim is authorised.

16-E · Billing ledger · Finance lead · 3 May

P1 tariff £0.07/kWh plus £88/home/year; P2 tariff £0.11/kWh plus £112/home/year, for every cohort member. Compare actual P1/P2 bills using only the same fully observed 40 treated homes. Provisional avoided bills use P2 £0.11/kWh and no avoided standing charge. Carbon conversion is fixed at 0.184 kg CO₂e/kWh; use a range for a ranged energy effect.

16-F · Publicity workbook and comfort return · 3 May

Original headline: “48 refurbished homes; gas fell 35%; annual bill savings £15,456.” Stored P1 625,000 kWh, P2 404,200 kWh, tariff £0.07.

Comfort replies: T-A 4, T-B 4, T-C 10, T-D 2. Eighteen of 20 report improved comfort, including ten T-C respondents. Participation was voluntary. No representative indoor-temperature series or nonrespondent assessment exists.

16-G · Monitoring decisions · Ben Orr · 4 May

Available one-off budget £3,000: all T-D mapping repairs plus validated annual readings £1,500; independent representative comfort follow-up £900; individual C-B boiler histories and commissioning dates £600. All can finish in six weeks if authorised now. Ben can own the benefits register and quarterly review; Mei has one analysis day per quarter. Committee may defer payment pending a dated return. Consent covers specified checks and the next year of meter access. No further data acquisition is required.

Recover initial assumptions and interpretation
  • Fixed 48-home P1 normalised total: 625,000 kWh; target 125,000 kWh.
  • Normalisation uses 2,000 HDD and independently measured non-heating gas; P0–P1 checks only pre-trends.
  • Initial sensitivity selects boiler saving 500, T-C increment 2,000, T-D actual total 14,000 kWh/home. This is the least favourable corner, with no point-estimate status.
  • Full bounds also include the opposite corner: boiler 900, T-C 4,000, T-D 8,000. No distribution, probability, midpoint or independence is assumed.
  • All eight T-D values are unresolved observations. Their hypothetical common value only represents a possible cohort mean for calculation.
  • Current-price equivalents use £0.11/kWh; carbon uses 0.184 kg/kWh; standing charges and comfort contribute no avoided value.
  • 7 May 2027 recommendation: hold payment; request all £3,000 specified checks; six-week return 18 June, conditional on authorisation now.
  • Reload/reset restores the initial selection. No state is saved, exported, authorised or transmitted by this page.
Read the complete supplied case · BPB2-16
# BPB2-16 — Housing energy retrofit — test realised benefits

Suite BPB2 v0.3-pilot; Foray 247, third essai.

### Scenario
Linden Homes has refurbished forty-eight homes and wants its final benefits payment released. Sponsor Asha Vale must defend a claim to the investment committee on 7 May 2027. The reported gas reduction mixes populations, billing costs moved differently from consumption, comparison homes received other interventions, and eight meters remain unresolved. Asha needs a benefits decision that is useful despite those limits: an inspectable claim, a defensible range and a proportionate plan for reducing the uncertainty that matters.

### Exact task prompt

Produce one usable in-browser answer page for the named practitioner, using the supplied fictional evidence. Keep observed facts, forecasts, assumptions and proposed decisions distinguishable, and cite source IDs beside material claims. No external research or real-world certification is required.
Prepare Asha's browser benefits-assurance brief. Audit the published claim, estimate weather-normalised project-attributable gas effects under the committee's supplied convention, and test how the specified uncertainties affect the portfolio result. Separate actual bill movement, current-price avoided bills and comfort evidence. Recommend whether the benefits payment is supportable and which monitoring work to authorise. Provide a clear claim suitable for the committee, not just a warning that attribution is difficult. Retain cohort denominators, intermediate calculations and the limits of every estimate; do not invent observations, probabilities or statistical significance.

### Supplied records
All records are fictional. Money is GBP; gas is kWh per dwelling unless marked total. Evaluation date 7 May 2027. Pre-period P0 is April 2023–March 2024; P1 is April 2024–March 2025. Installation occupied April 2025–March 2026 and is excluded. All measures were commissioned 31 March 2026. Post-period P2 is April 2026–March 2027. Heating degree-days are P0=1,800, P1=2,000, P2=1,600.

**16-A — Benefits agreement, investment committee, 1 March 2025.** Portfolio is the same 48 treated dwellings throughout. Target is at least 20% reduction in weather-normalised total gas attributable to the refurbishment, with comfort not worsened. For this target, divide the aggregate attributable kWh effect by the same 48 dwellings’ P1 normalised total gas; do not change the percentage denominator between scenarios. A £50,000 final benefits payment is held until the committee accepts the evidence; no automatic capital clawback follows if the target is not demonstrated. Asha recommends; the committee releases the payment. Work cost £360,000. No monetary value was assigned to comfort. No homes disconnected from gas.

**16-B — Validated meter/cohort extract, analyst Mei Tan, 30 April 2027.** Figures are arithmetic means within each row. Rows are distinct dwellings, not samples from overlapping populations. Non-heating gas was independently measured except where stated; it is included in total gas and must not be weather-scaled.

| Cohort | Homes | P0 total | P1 total | P2 total | Non-heating gas P0/P1/P2 | Status |
|---|---:|---:|---:|---:|---|---|
| T-A treated | 18 | 13,000 | 13,600 | 9,800 | 2,000/2,000/2,000 | Stable households; matched to C-A |
| T-B treated | 12 | 14,000 | 14,600 | 10,400 | 2,100/2,100/2,100 | Stable households; matched to C-B |
| T-C treated | 10 | 10,000 | 10,500 | 10,300 | 1,800/1,800/2,600 | Occupancy and care-related heating requirements changed |
| T-D treated | 8 | 12,000 | 12,500 | Unresolved | 2,000/2,000/2,000 | See meter bounds in 16-D; no point estimate validated |
| C-A comparison | 20 | 12,800 | 13,400 | 10,400 | 2,000/2,000/2,000 | No refurbishment or boiler change; stable households |
| C-B comparison | 12 | 14,200 | 14,800 | 11,000 | 2,100/2,100/2,100 | Eight received another programme's boiler replacement |

**16-C — Committee analysis convention, technical review, 2 May.** Normalise each period to 2,000 HDD: non-heating gas + (total minus non-heating gas) × 2,000/actual HDD. Compare P1 with P2, using P0–P1 only to inspect pre-trends. Provisional project effect per home = treated P1 normalised use − treated P2 normalised use − the matched comparison group's normalised reduction. Use C-A's absolute per-home reduction for T-A, T-C and T-D; use corrected C-B for T-B. This is a stipulated provisional convention, not proof that all unobserved causes are controlled. Explain whether the pre-trends and other records strengthen or weaken the interpretation. A clearly labelled alternative sensitivity may supplement it, not replace it.

**16-D — Confounders and incomplete records, Mei and the building-services reviewer, 2 May.** The eight C-B boiler replacements account for between 500 and 900 kWh of **weather-normalised** saving per replaced home. Add their weighted saving back to C-B's P2 mean before estimating the common trend; do not weather-scale that adjustment again. For T-C, independently documented additional occupancy and care-related temperature requirements would have increased its untreated P2 normalised consumption by 2,000–4,000 kWh/home above P1 less the C-A trend. These changes would have happened without refurbishment. For each T-D home, validated meter-bound evidence permits annual P2 total gas between 8,000 and 14,000 kWh; all eight can be at the same end of the range. These are deterministic bounds, not probability distributions. No independence assumption or midpoint claim is authorised.

**16-E — Billing ledger, finance lead, 3 May.** P1 tariff £0.07/kWh plus £88 per home/year; P2 tariff £0.11/kWh plus £112. These rates apply to every cohort member. Use only the same forty fully observed treated homes when comparing actual P1/P2 bills. For provisional avoided bills, use P2's £0.11/kWh and do not claim any avoided standing charge. Carbon conversion for the exercise is 0.184 kg CO2e/kWh, fixed; show a range if the energy effect is a range.

**16-F — Publicity workbook and comfort return, 3 May.** Workbook headline: “48 refurbished homes; gas fell 35%; annual bill savings £15,456.” The sheet stores P1 total 625,000 kWh, P2 total 404,200 kWh and tariff £0.07. Comfort returns: T-A four replies, T-B four, T-C ten, T-D two. Eighteen of twenty respondents report improved comfort; ten of those eighteen are T-C. Participation was voluntary. No representative indoor-temperature series or nonrespondent assessment exists.

**16-G — Monitoring decisions, operations manager Ben Orr, 4 May.** A £3,000 one-off monitoring budget is available. Repairing all T-D meter mappings and obtaining validated annual readings costs £1,500; an independent representative comfort follow-up costs £900; obtaining individual C-B boiler histories and commissioning dates costs £600. All can complete in six weeks if authorised now. Ben can own the benefits register and quarterly review; Mei has one analysis day per quarter. The committee can defer its payment decision pending a dated return. Consent covers these specified checks and the next year of meter access; no further data acquisition needs to be invented.

### Browser deliverable and completion conditions
Provide an evidence-linked claim audit, cohort calculation bridge, portfolio bounds, billing comparison and an owned monitoring/payment decision. Controls for the boiler correction, T-C counterfactual increment and T-D consumption must update cohort/portfolio energy, current-price bill equivalents, carbon and target status together. Keep unknown observations visibly unknown. Show the two pre-periods without pretending they establish causality. The reader must be able to recover the original reported claim and all initial assumptions.

The page must work in the supplied browser/runtime; a download button or file preview alone is insufficient. Include a clear “Return to gallery” link using the shared shell route, or ../index.html if no route is supplied. Required controls must change the dependent results, not just their appearance. Original evidence and the initial dated position must remain recoverable; hypothetical changes do not create approvals or alter source records. Optional exports are secondary.