The Gold Standard · Archaeology
Make the decision path visible
A suspended bridge-model control does not return to service simply because the immediate defect was fixed. Reviewers may agree that the site-code filter works again yet disagree about whether the test window, case mix and operating evidence are sufficient to restore release authority.
The gold standard turns that disagreement into an explicit evidence decision. It separates necessary proof from desirable reassurance, preserves dissent and restores only the scope supported by current validation and enforceable monitoring.
Choose the route closest to your decision
Chapter contents
Chapters 1–3
Chapters 4–6
Chapters 7–9
Chapters 10–12
CHAPTER 1 OF 12 · FRAME RESTORATION
1. Define the restoration decision
State the suspended model version, affected release scope, breached obligation and authority that can restore use. Separate restoration from retrospective repair of cases in the breach window.
A technically repaired component does not itself decide whether the complete archaeological workflow is safe to resume.
CHAPTER 2 OF 12 · FRAME RESTORATION
2. Set the burden of evidence before reviewing results
Identify necessary evidence for each restoration claim: cause controlled, relevant cases represented, workflow enforced, provenance retained and residual risk bounded. Mark desirable evidence that can be collected during conditional monitoring.
Predeclared burdens reduce the temptation to move the goalposts after seeing favourable or unfavourable results.
Evidence check
Check whether the scope, owner and decision consequence are explicit before continuing.
CHAPTER 3 OF 12 · FRAME RESTORATION
3. Build a claim-to-evidence map
Connect each proposed use to test sets, operational runs, independent review and unresolved limits. Record which evidence is direct, inferred or missing.
A large evidence package can still be insufficient when none of it tests the context that triggered suspension.
| Claim | Necessary evidence |
|---|---|
| Cause controlled | Mechanism test |
| Workflow works | End-to-end run |
| Scope represented | Relevant cases |
| Residual risk bounded | Explicit limit |
CHAPTER 4 OF 12 · TEST SUFFICIENCY
4. Verify the repair and its enforcement path
Test the code, input mapping, queue filters, reviewer assignment and release gate. Confirm that a detected condition changes the actual decision state, not only a dashboard.
Evidence must cover the path from archaeological record to enforced review; component success alone can hide a broken handoff.
CHAPTER 5 OF 12 · TEST SUFFICIENCY
5. Use independent challenge for contested claims
Ask a reviewer outside the redesign team to reproduce the test, inspect exclusions and search for plausible boundary cases. Give the team a chance to answer with traceable evidence.
Independence improves challenge, but it does not replace subject knowledge. Record reviewer competence and any limits.
Evidence check
Ask a second reviewer to reconstruct the example from the preserved evidence and stated assumptions.
CHAPTER 6 OF 12 · TEST SUFFICIENCY
6. Worked example: the fix works, but the window is narrow
A renamed site-code filter is repaired and every test join enters review. The restoration team presents two weeks of ordinary cases, while the independent reviewer notes that no high-complexity cross-trench joins occurred during that period.
Governance restores ordinary joins but keeps high-complexity joins under manual release until a specified challenge set and four operating cycles succeed. Changed-context check: if archived high-complexity cases can be replayed faithfully through the current workflow, they can strengthen evidence but should not be presented as live workload proof.
| Question | Record |
|---|---|
| Direct or inferred | Evidence strength |
| Excluded cases | Reason |
| Independent challenge | Outcome |
| Missing item | Decision effect |
CHAPTER 7 OF 12 · RESTORE PROPORTIONATELY
7. Resolve disputes by the consequence of being wrong
For each disputed item, ask whether the missing evidence could plausibly change release integrity, provenance or reversibility. Require stronger proof where archaeological evidence cannot be recreated after disturbance.
Not every uncertainty deserves equal delay. Low-consequence process detail can become a monitoring obligation rather than a barrier to all restoration.
CHAPTER 8 OF 12 · RESTORE PROPORTIONATELY
8. Choose the narrowest defensible restoration scope
Options include full restoration, phased restoration, restricted site types, manual review overlays or continued suspension. Record why excluded uses remain outside authority.
Scope should follow evidence, not organisational pressure to return to the old operating envelope.
Evidence check
Verify that the control still reaches the intended decision under a changed but plausible context.
CHAPTER 9 OF 12 · RESTORE PROPORTIONATELY
9. Renew conditional monitoring with enforceable actions
Define metrics, populations, thresholds, review cadence, evidence owners and the immediate state change for warning, restriction and suspension. Set an expiry for renewed conditions.
Monitoring is part of approval only when its triggers can alter release authority.
| Choice | Use when |
|---|---|
| Full | All claims supported |
| Phased | Operating proof pending |
| Restricted | Boundary unresolved |
| Suspend | Protection insufficient |
CHAPTER 10 OF 12 · WATCH THE RETURN
10. Reconcile the breach window and early restored cases
Complete targeted re-review of material cases from the plausible breach period, then sample early restored releases for the same failure mechanism and new workarounds.
Restoration should not erase unfinished corrective work or treat the first quiet weeks as proof of permanence.
CHAPTER 11 OF 12 · WATCH THE RETURN
11. State limits and misconceptions
No validation set represents every site formation, taphonomic complexity or workflow adaptation. More evidence can also repeat the same blind spot.
Record untested strata, workload conditions, reviewer dependencies and the circumstances that require fresh scope revalidation.
Evidence check
Confirm that versions, notices and unresolved limits remain visible in the final record.
CHAPTER 12 OF 12 · WATCH THE RETURN
12. Audit restoration as a new authority decision
Track claims, evidence, exclusions, dissent, scope, conditions, thresholds, early cases and later breaches. Preserve the decision version so future reviewers know what was actually approved.
The gold standard restores useful capability without converting uncertainty into silence. Approval returns in proportion to evidence and remains answerable to monitoring.
| Field | Action |
|---|---|
| Metric | What changes |
| Threshold | State transition |
| Owner | Who responds |
| Expiry | When reassessed |
Frequently asked questions
Does fixing the trigger breach restore approval automatically?
No. Restoration is a new authority decision based on the repair, relevant validation and enforceable operating evidence.
Must every reviewer agree?
No, but material dissent and its evidence should be visible to the accountable authority.
Can approval return in stages?
Yes. Phased or restricted restoration is appropriate when evidence supports some uses but not the full previous scope.
What makes renewed monitoring meaningful?
Named evidence owners, defined populations and thresholds, and a trigger that actually restricts or suspends release authority.
Continue through verified routes
Core routes
Archaeology subject root · Gold Standard of Education · Gold Standard directory · How X Works hub
Specialist routes
- Post-acceptance monitoring trigger breaches, conditional approval suspension and scope revalidation
- Redesign acceptance, conditional approval, residual obligations and post-acceptance monitoring
- Control redesign validation disputes, competing success criteria and governance acceptance
- Regression root-cause disputes, independent review and control redesign