CHAIN_13 Decision 4 — 6-DOF-only schedule-κ oracle gate (re-run)

Pre-registered: tasks/chain13_oracle_gate_criteria.md (fresh 6-DOF-only gate, two adversarial review passes). Supersedes the pooled 8-run D3 gate (FAILED 3/8), which is preserved unaltered at generated_reports/GNC/Jun12_26/chain13_schedule_oracle.md.

Determinism self-check (pre-registered STOP): max|rel diff| = 0.00e+00 (bar < 1e-6) → PASS.

Stage (a) — calibration (GATE: hits ≥ 4/5 AND f_pred ≤ 0.50, all 3/3; δ_disc reported only)

run	peak hits	n_detected	f_pred (≤0.50)	δ_disc (diag)	err med/max [m]	per-peak phase dist	status
r0_6dof_path	5/5	10	1.000	-0.560	0.0832 / 0.137	[0.0, 0.0, 0.0, 0.0, 0.0]	INVALID
r0_6dof_minus2pct	5/5	10	1.000	-0.130	0.0830 / 0.141	[0.0, np.float64(0.0), np.float64(0.0), 0.0, 0.0]	INVALID
r0_6dof_plus2pct	5/5	10	1.000	+0.160	0.0828 / 0.136	[0.0, 0.0, 0.0, np.float64(0.0), np.float64(0.0)]	INVALID

Stage (a) gate → INVALID. INVALID/STOP — a 6-DOF run is non-discriminating (f_pred > 0.50) or has <5 peaks; the oracle predicts near-singular over most of the orbit at the flown standoff, so it CANNOT localize where to add standoff. Oracle PARKED. A phase-varying R-v2 has no calm phase to exploit → adopt the static rs 0.40 already shown to dominate (Decision 1).

regen: /Users/antoniahoffman/miniforge3/envs/new-pin-env/bin/python validation/schedule_kappa_oracle.py