Campaign session

01KWZ6DN5FWFP0T0CMY4D6KG6S

finished 2026-07-07 21:05:54.351563+00:00 → 2026-07-07 21:08:03.772149+00:00 · 0 runs · supervisor: react-agent

“Prerequisite-blocked: PostGIS unreachable from runner (connection refused :5432); zero runs launchable since exp-csi-static stages scenes via gis export-floor. No new evidence; C2/C3 stand resolved from prior session 01KWYBEJQK, C1 remains under-powered/Phase-2-blocked.”

Archive snapshot, as of 23 h ago — the run corpus is rebuilt once a day, so this page is not a live reading. The fleet panel is the live one; it refreshes every 30 s.

Success criteria

Criterion	Resolved
Primary: bootstrap-over-floors 95% CI on LOFO inflation ratio lower bound > 1.5x	no
Variance split: sigma2_floor vs sigma2_seed; geometry-dominates iff between-floor SD of per-floor CV(N) > within-floor between-seed SD	no
Comparator honesty: LOFO denominator is held-out within-floor leave-one-occupancy-out, never in-sample residual	no

Synthesis

c-csi-cross-geometry-scaleout — session 01KWZ6DN5FWFP0T0CMY4D6KG6S (2026-07-07)

Verdict: prerequisite-blocked — zero runs launched. PostGIS is unreachable from this runner environment (connection refused, 127.0.0.1:5432 via gis_experiments). Every exp-csi-static run stages its scene through gis export-floor, so no floor could be simulated. This is exactly the Prerequisite #1 risk the brief flagged (sibling session 01KWWJ0REM hit the same refusal on 2026-07-06). No new corpus, no new figures — fanning out sim-runners would only have produced guaranteed staging failures.

Inherited state (NOT re-established by this session). The immediately-prior same-day session 01KWYBEJQK ran the full affordable Phase-1 pilot: 90 gate-passed runs over the 6 staged floors × 3 seeds × 5 occupancy levels {0,2,4,6,8} at 2.4 GHz. Its verdict stands:

C1 (primary inflation-ratio CI, lower bound > 1.5×): FAIL, under-powered by design — floor-cluster 95% CI [1.04, 9.85] at only 6 clusters; pooled ×3.37 masks extreme per-floor heterogeneity (0.70×–49.8×). Powered test needs Phase 2 (≥40 floors, n₀ ≈ z²σ²/d² ≈ 40 at σ=4.82, d=1.5).
C2 (variance split): PASS — σ²_floor 0.0109 vs σ²_seed 0.000165 (66×); between-floor SD 0.105 vs between-seed 0.0116 (9×). Geometry dominates seed decisively on CV(N).
C3 (comparator honesty): PASS — honest leave-one-occupancy-out within-floor baseline (0.026) vs held-out (0.088); the original in-sample-residual bug is fixed.

This session adds no evidence to any criterion — all three are booked unresolved here. C2/C3 remain resolved on the prior session's record; C1 remains under-powered and Phase-2-blocked.

Recommended next actions. (1) Restore PostGIS reachability on the runner (brief names :5435; gis surface hit :5432 refused), then re-probe with gis_experiments. (2) Stage the 4 still-missing Phase-1 floors (10156/10425/6090/8266) to complete the 10-floor pilot. (3) Run the ResPlan → gis_rebuild_nrg → gis_validate import batch to reach the ≥40 floors Phase 2 requires for a powered C1. Until (1), no exp-csi-static run can execute from this environment.

Honest scope

Zero-run, prerequisite-blocked session. All quantitative results cited above are inherited from 01KWYBEJQK and are not independently re-derived here.

Attached runs

The attached runs are not in the current atlas snapshot — rebuild via web --build-sim-atlas.