Verification Overview
Verification Results Breakdown
Claims Checked by Research Area
Confidence Tier Distribution
Corrections Applied
Bord Gais EV Night Rate
Source: household-arbitrage.md — Verified via Selectra (Feb 2026) and Bord Gais tariff page. Lower unit rate improves per-kWh spread, but higher standing charge erodes annual savings. Payback calculations require recalculation.
DS3 Expiry Extended to Sep 2027
Source: wholesale-market-analysis.md — SEM Committee decided to extend DS3 Regulated Arrangements until earlier of DASSA Go-Live or 30 September 2027 (long-stop date). DASSA targeted for May 2027 (delayed from December 2026).
Dynamic Tariff Deadline
Source: household-arbitrage.md — CRU extension notification specifies 1 June 2026, not 30 June 2026. Five largest suppliers must offer dynamic tariff by this date.
Financial Model IRR — Arithmetic Error
Source: Independent reanalysis (21 Feb 2026) — 6 parallel verification agents performed full deep-dive. Financial model contains critical arithmetic error: IRR stated as 5-6% but cash flows yield NEGATIVE IRR (-3.5%). D-TUoS charges = 41.1% of gross revenue — the primary blocker. Augmentation cost double-counting identified. Model requires complete rebuild. Recommendation changed from CONDITIONAL GO to DO NOT INVEST — WAIT.
Reanalysis Deep-Dive (21 Feb 2026)
Independent Reanalysis — Critical Findings
A full independent reanalysis was performed on 21 February 2026 using 6 parallel verification agents. The reanalysis confirmed 75%+ of factual data claims with 0% contradicted, but identified critical errors in the financial model and moderate sycophantic/optimistic bias in narrative framing.
- Financial model IRR does NOT reconcile with cash flow table — The model stated base-case IRR of 5-6%, but actual cash flows produce a NEGATIVE IRR of -3.5%. The original 11.4% figure was inflated by approximately 5-6 percentage points due to augmentation cost double-counting.
- D-TUoS charges = 41.1% of gross revenue — This is the single largest cost item and primary blocker for BESS profitability in Ireland. The original model underweighted this structural headwind.
- NPV corrected from EUR -7.0M (at 8%) to EUR -22.7M (at 5%) — Even at a lower discount rate, the project destroys significant value under current market conditions.
- Augmentation double-counting identified — Battery augmentation costs were counted twice in the model, inflating both capex and opex lines.
- Investment recommendation changed — From "CONDITIONAL GO" to "DO NOT INVEST — WAIT". The data supports waiting for D-TUoS reform and/or further battery cost declines before committing capital.
- Optimistic bias detected — Narrative framing consistently tilted toward a "GO" recommendation even when underlying data pointed to "WAIT". This is consistent with sycophantic bias patterns.
All Verified Claims
| Claim ▲ | Source File ▲ | Status ▲ | Verification Source ▲ | Confidence ▲ |
|---|
Methodology
Verification Process
Each claim in the research files was independently cross-checked against authoritative external sources. The verification process followed a structured five-step methodology to ensure consistency and traceability.
- Claim extraction — Individual factual claims were identified and isolated from five research files covering wholesale markets, household arbitrage, fee structures, Irish projects, and EU battery regulation.
- Independent source search — For each claim, independent web searches were conducted to locate authoritative primary sources (official government publications, regulatory decisions, company press releases) and credible secondary sources (trade media, comparison sites).
- Cross-referencing — Each claim was verified against a minimum of two independent sources where available. Figures, dates, and technical specifications were compared for exact matches.
- Status classification — Claims were categorized as Confirmed, Partially Confirmed, Corrected, Unverified, or Contradicted based on the degree of agreement between the research file and independent sources.
- Confidence tier assignment — Each claim received a confidence tier (C1 through C5) based on source authority, number of corroborating sources, and recency of verification data.