Khalim Conn-Kowlessar 8520a52ee9 Slice 25c: 000477 §4/§5/§6 closure — Table 3c (p.162) M+L lower bound ... Fixed a single-character spec adherence bug: SAP10.2 Table 3c (p.162) specifies the M+L profile's DVF lower bound as `V_d,m < 100.2`, not `< 100.0`. The 0.2 L/day window matters when V_d,m sits between 100.0 and 100.2 — exactly where 000477's May lodgement lands (100.16 L/day). For V_d,m = 100.16: Spec: DVF = 0 → (61) = E × r1 × fu = 134.84 × 0.015 × 1.0 = 2.0225 ✓ Buggy: DVF = 100.2 - 100.16 = 0.04 → (61) = 2.0233 (off by 0.0008) The cascade through the missing 0.0008 W on May LINE_61 propagated to LINE_62/64/65 and then §5 LINE_72/73 + §6 LINE_84 — clearing one constant unblocks the entire 000477 §4-§6 cluster. Per-fixture cluster status (was/now): §3 24/24 → 24/24 §4 46/54 → 53/54 (only 000487 LINE_65 remains) §5 50/54 → 52/54 (only 000487 LINE_72/73) §6 10/12 → 11/12 (only 000487 LINE_84) All remaining cascade failures cluster on 000487 (slice 25d — derive LINE_64A electric-shower kWh from cert per Appendix J step 8) plus §7 LINE_92/93 marginal residuals on 4 fixtures (precision artefact). Scoreboard: section_cascade_pins: 286 → 293 PASS (+7) e2e SapResult: 32 → 32 PASS (still cascade-blocked by 000487 LINE_65 + downstream §8-§12 pins not yet asserted) Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>		2026-05-23 23:04:23 +00:00
..
adr	docs: handover for Table 3c two-profile combi loss → close 4 Elmhurst fixtures	2026-05-22 12:14:00 +00:00
sap-spec	Slice 25c: 000477 §4/§5/§6 closure — Table 3c (p.162) M+L lower bound	2026-05-23 23:04:23 +00:00