When to Use Each, What It Fixes, and the Retest Probes
Why Recovery Modes Exist
Most systems fail recovery for one reason:
They try to fix everything at once.
That creates chaotic reform, reform fatigue, metric gaming, and political backlash—then the system drifts back.
Civilisation OS uses a different rule:
Pick one dominant drift mode. Run one recovery mode. Retest. Lock it into standards. Then iterate.
Recovery modes are standardised “repair patterns” that work across domains because drift mechanisms repeat.
Start here: What is Civilisation OS: https://edukatesg.com/what-is-civilisation-os/
How to Select the Right Recovery Mode (Fast)
Use DLT:
- If T is lowest → start with Truth Restoration
- If L is highest → start with Load Shedding
- If D is falling → start with Capability Rebuild
If multiple problems exist, choose the one that unlocks the others.
Mode 1 — Truth Restoration Mode
When to use (Triggers)
- bad news can’t travel upward
- reporting becomes performative
- leadership is insulated from ground truth
- errors repeat because nobody can admit them
What it fixes
Restores signal integrity. Without truth, every other reform becomes theatre.
Core actions
- protect reporting channels (no retaliation)
- shorten feedback loops (faster escalation)
- require “reality probes” alongside claims
- remove narrative penalties for admitting mistakes
- separate “truth reporting” from “status games”
Retest probes (Hard to fake)
- Bad-news escalation test: can a frontline issue reach decision-makers and trigger change within a defined time window?
- Error-repeat rate: do the same failures decline within 30–60 days?
- Truth latency: time from incident → acknowledgment → corrective action
Mode 2 — Incentive Realignment Mode
When to use (Triggers)
- people are rewarded for optics, not outcomes
- “doing the right thing” is punished
- compliance beats competence
- corruption/extraction increases even with “good policies”
What it fixes
Stops drift caused by perverse incentives—aligns behaviour to mission.
Core actions
- rewrite reward metrics to include reality probes
- penalise metric gaming explicitly
- link advancement to long-run outcomes, not short-run reports
- remove incentives that profit from failure (the silent killer)
Retest probes
- Behavioural shift test: do people change behaviour within one cycle without new speeches?
- Gaming detection: reduction in audit anomalies / suspiciously perfect reports
- Outcome alignment: do real-world outcomes improve alongside metrics?
Mode 3 — Capability Rebuild Mode (Depth Repair)
When to use (Triggers)
- foundation skills decay (literacy, numeracy, reasoning)
- dependence on a few “heroes” grows
- training volume rises but performance falls
- institutions can’t execute policy even if they want to
What it fixes
Rebuilds Depth—competence density, skill pipelines, adaptation speed.
Core actions
- rebuild foundations before adding complexity
- train for demonstrated competence, not certificates
- standardise mastery thresholds
- create rapid remediation loops (close gaps fast)
Retest probes
- Cold performance tests: can people perform without prompts?
- Transfer tests: can they solve unfamiliar problems?
- Gap closure speed: time to fix known weaknesses month-to-month
- How Civilisation OS Repairs Drift
Mode 4 — Load Shedding Mode (Overload Repair)
When to use (Triggers)
- complexity grows faster than capacity
- backlogs rise steadily
- response times slow
- coordination costs explode (too many approvals, agencies, processes)
What it fixes
Reduces Load so the system can breathe and correct itself again.
Core actions
- kill low-value work and redundant approvals
- cap process growth (no new steps without removing old ones)
- reduce scope temporarily to restore performance
- simplify interfaces between departments
Retest probes
- Cycle time: time from request → action decreases
- Backlog trend: backlog stops growing, then falls
- Coordination count: fewer handoffs per decision
Mode 5 — Standards Reset Mode
When to use (Triggers)
- exceptions become normal
- enforcement becomes selective
- minimum competence declines
- “quality” becomes negotiable
What it fixes
Stops gradual erosion by restoring non-negotiable floors.
Core actions
- define minimum viable standards (few, sharp, enforceable)
- set expiry on exceptions automatically
- enforce consistently (status-independent)
- link standards to retest probes
Retest probes
- Exception count: exceptions decline month-to-month
- Consistency audit: similar cases get similar outcomes
- Minimum performance: weakest decile improves
Mode 6 — Anti-Proxy Mode (Metric Decontamination)
When to use (Triggers)
- metrics improve while reality worsens
- gaming becomes obvious
- audit culture grows but competence shrinks
- “targets achieved” yet problems remain
What it fixes
Restores reality coupling by preventing proxy replacement (Goodhart drift).
Core actions
- demote easy-to-game metrics
- replace with hard-to-fake probes
- rotate probes so gaming is costly
- reward “truthful bad results” over fake good ones
Retest probes
- Probe-metric alignment gap: gap narrows
- Gaming indicators: fewer suspicious patterns, fewer anomalies
- Reality outcomes: improvements show up outside reporting channels
Mode 7 — Resilience Build Mode (Fragility Repair)
When to use (Triggers)
- system works until shocks, then collapses
- single points of failure multiply
- maintenance debt grows
- recovery time increases
What it fixes
Turns brittle growth into durable performance.
Core actions
- invest in redundancy where failure is existential
- reduce concentration risk (suppliers, energy, infrastructure)
- pay maintenance debt first
- run stress tests and drills monthly
Retest probes
- MTTR: mean time to repair decreases
- Stress test pass rate: improves over cycles
- Maintenance backlog: declines steadily
Mode 8 — Coordination Simplification Mode (Institutional Drag Repair)
When to use (Triggers)
- too many agencies own one problem
- decisions stall in committees
- frontlines can’t act without permission
- the system becomes “administration-first”
What it fixes
Reduces coordination friction and restores speed of execution.
Core actions
- clarify ownership (one accountable owner per outcome)
- reduce handoffs and interfaces
- empower frontlines within clear standards
- compress decision loops
Retest probes
- Ownership clarity: fewer “not my job” outcomes
- Decision latency: falls
- Frontline autonomy test: more issues solved without escalation
Mode 9 — Constraint Coupling Mode (Reality Rebinding)
When to use (Triggers)
- reality is denied until crisis hits
- stability maintained by debt, subsidies, or narrative
- resource constraints ignored (energy, water, demographics, ecology)
- repeated “surprises” that were predictable
What it fixes
Re-couples civilisation to physical reality—so constraints are seen early.
Core actions
- publish constraint audits (where we’re pretending)
- create hard thresholds and early warnings
- align production plans to constraints
- stop postponing pain with hidden liabilities
Retest probes
- Shock simulation: improved response speed and preparedness
- Constraint visibility: fewer “surprise” crises
- Future-borrowing trend: reliance on debt/masking declines
The Recovery Mode Ladder (If You’re Unsure)
When systems are truly messy, run modes in this typical order:
- Truth Restoration (so you can see)
- Anti-Proxy (so metrics stop lying)
- Incentive Realignment (so behaviour shifts)
- Load Shedding (so the system can breathe)
- Capability Rebuild (so it can execute)
- Standards Reset (so gains stick)
- Resilience Build (so shocks don’t kill it)
- Coordination Simplification (so it moves faster)
- Constraint Coupling (so reality stays visible)
Not always, but often.
The Civilisation OS Recovery Rule
A recovery mode is “real” only if:
- retest probes improve
- gaming decreases
- standards lock in the gain
- the improvement persists across cycles
If the system needs constant speeches to behave, recovery is fake.
Real recovery changes incentives, standards, and feedback loops.
Q&A: Recovery Modes
Can I run multiple modes at once?
You can, but you shouldn’t. Civilisation OS prefers one dominant mode per cycle to avoid chaos reform.
Which mode is most urgent?
Truth Restoration. If truth cannot travel, nothing else can be trusted.
How do I prevent relapse?
Standards Reset + retest probes. Recovery must be structural, not campaign-based.
How do I know the mode worked?
Retest probes improve within 30–60 days and the e/t slope stays positive.
Companion Article to this series
Part 1 — What is Civilisation OS: https://edukatesg.com/what-is-civilisation-os/
Part 2 — How it works: https://edukatesg.com/how-civilisation-os-works-why-these-layers-govern-human-reality/
Part 3 — Academic foundations: https://edukatesg.com/civilisation-os-what-are-the-academic-foundation-of-civilisation-os/
Part 4 — Detect + repair trajectories: https://edukatesg.com/how-civilisations-os-detect-rise-stagnation-regression-and-collapse-and-how-to-repair-trajectory-with-limited-prediction/
Part 5 — This Field Manual (execution method, recovery modes, probes) https://edukatesg.com/civilisation-os-field-manual/