ULD Repair converts diagnosis into change.
It answers one question:
Once the limiting capability is identified, how do you repair it without creating new instability?
ULD Repair is not a collection of techniques.
It is a rule-based repair loop that prevents overtraining, false confidence, and drift.
The ULD Repair Loop (Fixed)
ULD Repair always follows this closed loop:
- Diagnose the limiting axis (Depth, Load, or Transfer)
- Apply targeted repair to that axis only
- Verify using retests that match the diagnosis
- Lock gains, then move to the next axis if needed
Skipping steps breaks the loop.
Rule 1: Repair One Axis at a Time
Most humans fail because repairs are applied everywhere at once.
ULD forbids that.
- Repairing Load when Depth is weak creates brittle performance
- Repairing Transfer without Load creates collapse under pressure
- Repairing everything simultaneously creates noise, not progress
Always repair in this order:
Depth → Load → Transfer
Repairing Depth (Understanding Integrity)
Depth repair strengthens structure.
Depth repair focuses on:
- explanation in the human’s own words
- reconstruction without templates
- identifying and correcting errors
- connecting ideas into a coherent model
Depth repair avoids:
- speed training
- time pressure
- volume repetition
- pattern drilling
Depth is repaired when the human can rebuild understanding independently.
Repairing Load (Performance Stability)
Load repair strengthens stability under constraint.
Load repair focuses on:
- controlled time pressure
- multi-step execution
- gradual increase of constraints
- maintaining accuracy under stress
Load repair avoids:
- endless untimed practice
- removing pressure entirely
- increasing difficulty without constraint
- confusing stress with motivation
Load is repaired when performance remains stable as constraints tighten.
Repairing Transfer (Generalisation Power)
Transfer repair strengthens adaptation.
Transfer repair focuses on:
- variation of format
- changing contexts
- recombination of known elements
- removal of explicit cues
Transfer repair avoids:
- repeating the same format
- drilling a single template
- increasing difficulty without variation
- teaching “tricks” that do not generalise
Transfer is repaired when the human succeeds in unfamiliar situations.
Rule 2: Repair Must Match the Diagnosis
ULD Repair is invalid if the repair does not match the diagnosed axis.
Examples:
- Low Depth repaired with repetition → invalid
- Low Load repaired with explanation only → insufficient
- Low Transfer repaired with harder versions of the same task → invalid
Mismatch creates false improvement and delayed failure.
Rule 3: Verification Is Mandatory
Repair without verification is guesswork.
After repair, ULD requires axis-specific retests:
- Simplify and explain → verify Depth
- Reapply constraints → verify Load
- Introduce new formats → verify Transfer
If performance does not change as predicted, the repair was wrong or incomplete.
Locking Gains (Preventing Regression)
Once repair is verified:
- reduce intervention intensity
- reintroduce normal conditions
- observe stability over time
If regression occurs, rerun diagnosis.
Do not escalate repair blindly.
When Multiple Axes Are Weak
It is common for more than one axis to be weak.
ULD handles this by sequencing:
- Repair Depth until stable
- Repair Load on top of Depth
- Repair Transfer last
Never reverse the sequence.
Reversal creates the illusion of progress followed by collapse.
What ULD Repair Is Not
ULD Repair is not:
- cramming
- brute-force practice
- motivation coaching
- discipline enforcement
- content dumping
Those may temporarily raise scores but do not repair capability.
How ULD Repair Fits in the Full System
ULD Overview:
https://edukatesg.com/uld/
ULD Protocol (run before repair):
https://edukatesg.com/uld-protocol/
Next layer (measurement after repair):
ULD Scoring & Bands:
https://edukatesg.com/uld-scoring/
If You Are Using ULD Correctly
You repair deliberately.
You verify relentlessly.
You change only what the diagnosis demands.
That is how ULD converts insight into reliable improvement.
Next Page
To make ULD comparable across time and humans, you need scoring bands:
https://edukatesg.com/uld-scoring

