Base commit: 6139de6173ea7e538db6097170a1ccfe93f311cb. The dates below record the research day, not reconstructed wall-clock times. The retained scripts, numerical outputs, and full prover receipts specify the inputs and checks. Cloud commands incur another run when repeated.
id: cycle_moments-2026-09-05-spectral-and-mixed
hunt: cycle_moments
started: 2026-09-05
finished: 2026-09-05
ran:
- .venv/bin/python hunts/cycle_moments/probe.py
- .venv/bin/python hunts/cycle_moments/mixed_moments.py --self-check
- .venv/bin/python hunts/cycle_moments/mixed_moments.py --grid-points 4801 --self-check
- .venv/bin/python -m modal run hunts/cycle_moments/modal_probe.py
- .venv/bin/python -m modal run hunts/cycle_moments/modal_mixed_probe.py
outcome: Finite spectral inequalities passed their checks; the tested narrow mixed observers had negative numerators and supplied no asymptotic gain.
artifacts:
- hunts/cycle_moments/results.json
- hunts/cycle_moments/mixed_integral_results.json
- hunts/cycle_moments/mixed_integral_refined.json
- hunts/cycle_moments/cue_results.json
- hunts/cycle_moments/mixed_cue_results.jsonThe completed Modal app identifiers are ap-kczpMD2GBPTaRhTYnB80L2 and ap-yxAuPSoMKkEjFs6ZeLYQS2. The null mixed-observer results are retained in full. Their interpretation and Fourier support costs are in MIXED-MOMENTS.md.
id: cycle_moments-2026-09-05-repeated-indices-and-counts
hunt: cycle_moments
started: 2026-09-05
finished: 2026-09-05
ran:
- .venv/bin/python hunts/cycle_moments/distinct_cycles.py
- .venv/bin/python hunts/cycle_moments/same_kernel.py
- .venv/bin/python hunts/cycle_moments/counting_overlap.py
- .venv/bin/python hunts/cycle_moments/overlap_bound.py
outcome: Exact repeated-index corrections agreed with enumeration; one fixed Fourier kernel gave equal full spectra but different simple counts, and the overlap bounds passed the retained finite checks.
artifacts:
- hunts/cycle_moments/distinct_cycle_results.json
- hunts/cycle_moments/same_kernel_results.json
- hunts/cycle_moments/counting_overlap_results.json
- hunts/cycle_moments/overlap_bound_results.jsonThe first same-kernel example had equal simple counts and therefore did not establish a counting improvement. The subsequent five-occurrence example does distinguish the counts. Both examples remain in the record.
id: cycle_moments-2026-09-05-lean
hunt: cycle_moments
started: 2026-09-05
finished: 2026-09-05
ran:
- AXLE check of SpectralJensen.lean in lean-4.33.0
- AXLE check of QuarticScore.lean in lean-4.33.0
- AXLE check of CycleMomentAssembly.lean in lean-4.33.0
- AXLE check of ThirdMomentObstruction.lean in lean-4.33.0
- AXLE check of DistinctCycles.lean in lean-4.33.0
- AXLE check of IsospectralCycles.lean in lean-4.33.0
- AXLE check of CountingOverlap.lean in lean-4.33.0
- AXLE check of OverlapBand.lean in lean-4.33.0
outcome: All eight exact source files have successful Lean receipts with no admitted proofs; the printed or guarded theorem audits use only standard axioms.
artifacts:
- hunts/cycle_moments/FORMAL-CHECKS.json
- hunts/cycle_moments/receipts/SpectralJensen-axle-receipt.json
- hunts/cycle_moments/receipts/QuarticScore-axle-receipt.json
- hunts/cycle_moments/receipts/CycleMomentAssembly-axle-receipt.json
- hunts/cycle_moments/receipts/ThirdMomentObstruction-axle-receipt.json
- hunts/cycle_moments/receipts/DistinctCycles-axle-receipt.json
- hunts/cycle_moments/receipts/IsospectralCycles-axle-receipt.json
- hunts/cycle_moments/receipts/CountingOverlap-axle-receipt.json
- hunts/cycle_moments/receipts/OverlapBand-axle-receipt.jsonEach receipt retains its request identifier, exact input, and toolchain. The source hashes are indexed in FORMAL-CHECKS.json (FORMAL-CHECKS.json). The formal scope is limited to the declarations and their explicit hypotheses; the Fourier, inertia, and asymptotic connections are separated in the notes.