QEC v170.1.1 — Replication Receipts and Report-Claim Validation

QEC v170.1.1 hardens the evidence layer introduced by the Exact Ququart FER Oracle and Harmonic Fault Battery.

This release does not alter the packed [[5,1,3]]₄ code, bounded decoder or harmonic-receiver mathematics. Instead, it makes benchmark results independently reproducible, hash-bound and substantially harder to misreport.

Merged through PR #584 with merge commit 73afc2a.

Highlights
Exact FER oracles for every declared ququart error channel
SHA-256-bound lane-exchange symmetry certification
Canonical receipts for independent replication runs
Fail-closed validation of generated report claims
Correctly separated harmonic-receiver telemetry
Improved dependency-free GitHub Pages evidence lab
Full repository validation under the QEC Validation Law
Exact FER Oracles for Every Channel

The finite-code oracle now provides exact weight enumerators and FER curves for:

full_packed_depolarizing
lane0_only
lane1_only
same_pauli_correlated

The complete packed channel retains exhaustive classification across all:

16⁵ = 1,048,576

packed Pauli patterns.

Each restricted channel is evaluated over its exact 4⁵ alphabet rather than being compared against the full packed-channel curve.

The Pages visualisation now overlays each selected channel’s Monte Carlo samples against the matching exact channel oracle, eliminating misleading cross-channel comparisons.

Lane-Exchange Symmetry Certificate

A new SHA-256-bound certificate verifies that the exact weight enumerators for:

lane0_only
lane1_only

are identical under encoded-lane exchange.

This certificate is included in the generated methodology, browser report and benchmark manifest. A lane-symmetry claim can no longer pass report validation without the corresponding exact certificate.

Independent Replication Receipts

This release introduces the canonical receipt schema:

qec.ququart-fer-replication.v1

Replication artifacts are classified using explicit verification states:

full_hash_match
prefix_consistent
parameter_variant_expected
unverified
mismatch

A mismatch fails closed.

Receipt identity fields such as the target release and commit must now be present and non-empty. Missing, empty or whitespace-only identities are rejected rather than silently becoming values such as "None".

qBraid Replication Record

The independent qBraid execution of v170.1.0 is preserved as a parameter-bound replication:

docs/replications/QBRAID_V170_1_0_REPLICATION.md
docs/replications/qbraid_v170_1_0_receipt.json

The original report exposed only eight-character artifact hash prefixes. Exact deterministic artifacts are therefore recorded as:

prefix_consistent

rather than being overstated as complete SHA-256 matches.

Sampled artifacts are marked:

parameter_variant_expected

because the qBraid execution used:

1,000 Monte Carlo trials per cell
500 harmonic trials per cell

while the canonical release evidence uses:

10,000 Monte Carlo trials per cell
4,000 harmonic trials per cell
Report-Claim Validation

QEC now emits two additional proof objects:

report_claims.json
claim_validation.json

Claims can be independently validated with:

qec-ququart-validate-report
--claims report_claims.json
--evidence /path/to/evidence

The validator rejects:

numeric claims that disagree with generated evidence;
conflation of expected-accept and adversarial-reject populations;
zero-false-trust claims when false trust occurred;
lane-symmetry claims without the exact certificate;
artifact equality claims based only on truncated hashes;
“all tests passed” claims without a separate successful test receipt;
threshold claims derived from a single finite-code study;
hardware claims lacking device, circuit, timing, leakage and readout declarations.

Controlled curve language is restricted to finite-code descriptions such as:

low_error_quadratic_regime
finite_code_crossover
Harmonic Telemetry Layering

Receiver and decoder outcomes are now represented as distinct evidence layers:

receiver rejection;
decoder rejection;
trusted correct syndrome;
receiver false trust;
accepted incorrect syndrome;
accepted logical residual.

This prevents an accepted logical residual outside the guaranteed correction radius from being incorrectly classified as harmonic-receiver false trust.

The methodology now explicitly declares every emitted telemetry layer, including:

accepted_incorrect_syndrome

A new operating-curve artifact reports receiver behaviour across physical error rate and harmonic-noise sigma:

receiver_operating_curve.csv
GitHub Pages Evidence Lab

The dependency-free browser report now displays:

selected-channel exact FER and Monte Carlo evidence;
exact per-channel weight enumerators;
lane-symmetry certification state;
report-claim validation state;
qBraid replication status;
layered receiver and decoder telemetry;
receiver operating data;
canonical methodology and proof hashes.

Threshold reporting has also been clarified. The interface now states:

Threshold claim made: no

or, for a rejected payload:

Threshold claim made: yes — invalid

It no longer presents threshold-claim presence as a permission state.

New Generated Artifacts
exact_channel_weight_enumerator.csv
exact_channel_fer.csv
receiver_operating_curve.csv
lane_symmetry_certificate.json
report_claims.json
claim_validation.json
qbraid_replication_receipt.json

These are added alongside the existing FER, Monte Carlo, harmonic-fault, methodology and manifest artifacts.

CI and Validation

The dedicated Ququart FER workflow now:

checks out complete Git history;
installs QEC v170.1.1;
runs the full repository test suite;
generates the canonical release evidence;
independently reloads and validates the emitted report claims;
publishes the upgraded Pages report after changes reach main.

Complete Git history is required because existing repository-boundary tests inspect previous commits through HEAD^.

Authoritative validation:

19,036 passed
7 skipped
4 warnings

v170.1.1 evidence generation: SUCCESS
generated report-claim validation: PASS
Canonical Proof Identities
Benchmark manifest:
1daec95984d2c914ede2b207ce8f6e2402ddbd5889c8095969838ce8700d66af

Claim validation:
7211dba1b71ddca9bcbd47e9599344270882f4851dbd4d507603924573005822

Methodology:
b8d9d3ffec3e9c443818683751fa9d48b3ca7407b72965985977fab1e20e63eb
Scientific Boundary

QEC v170.1.1 remains an exact finite code-capacity Pauli analysis combined with deterministic classical harmonic-readout simulation.

It does not claim:

a hardware or circuit-level threshold;
break-even quantum error correction;
physical pulse fidelity;
leakage or SPAM performance;
hardware timing performance;
universal quantum advantage.
Summary

QEC v170.1.1 turns benchmark reporting into part of the proof system itself.

Exact channel results, independent replications, telemetry classifications and public-facing claims now carry explicit evidence contracts. Truncated hashes cannot masquerade as complete matches, parameter differences cannot be hidden, threshold language cannot escape the finite-code boundary, and generated claims are revalidated directly against their underlying artifacts.