Worked example · 05
Correct the duration without doubling the event
Should the old and corrected records be added together?
Evening · record correction
The recorded duration of the synthetic learning session is corrected from four hours to three. The earlier revision remains visible but is no longer active.
This is a synthetic variation on the paper’s redacted public worked day. It teaches a branch of the model; it is not a private record, observed personal event or empirical finding.
World 03 / 03
Record Integrity Lab
Performance · superseded revision · active correction
Short answer
No. Only the active revision contributes; the superseded revision remains in the audit chain.
01 · Moment
Observed time slice
A bounded event with context and provenance.
02 · Economic question
Vcom 1500
Typed VC output: 1050
03 · Human question
+0.021 WellbeingScore movement
Wbefore 700.000 → Wafter 700.021
Show the calculation
Four questions, in order.
- 01
Start with the prior state
The event reads Wbefore = 700.000.
- 02
Ask the economic question
Validated rate × duration produces Vcom = 1500 hypothetical credits.
- 03
Ask the human-state question
The reviewed input produces deltaW = 0.021, so Wafter = 700.021.
- 04
Keep the units honest
VCperformance = (Wbefore / 1000) × Vcom = 1050 hypothetical credits.
What this means
Replay uses the corrected duration for valuation while preserving one occurrence and one transition.
What it does not mean
Correction is not deletion and does not permit silent editing of the historical record.
Scenario structure
Follow the integrity rule.
- 01
Earlier revision remains
- 02
Corrected revision becomes active
- 03
Only the active revision contributes
Technical layerAudit the calculationInspect statuses, intermediate values, provenance, uncertainty and raw scenario structure.
- Source settlement
- final
- Impact review
- applied
- Replay eligibility
- eligible
- Valuation
- valued
- Projection
- canonical
- Wbefore
- 700.000
- Component delta
- 0.07
- deltaW
- 0.021
- Wafter
- 700.021
- Vcom_event
- 1500
- VCperformance_event
- 1050
Pinned implementation calculation
Complete component-transition trace
Every input and intermediate value is carried by the stable synthetic fixture or derived from its pinned issuer-policy snapshot.
- Components before
- { "Physiology": 700, "Emotion": 700, "Feeling": 700, "Thought": 700, "Habit": 700, "Performance": 700 }
- calculated_Wbefore
- 700
- Affected component
- Performance
- component_before
- 700
- Band
- (600,700]
- band_rate
- 0.03
- adjustment_factor
- 0.005
- Direction
- positive
- Intensity
- medium
- intensity_multiplier
- 2/3
- maximum_move
- 0.105
- calculated_component_delta
- 0.07
- component_after
- 700.07
- component_weight
- 0.3
- calculated_deltaW
- 0.021
- calculated_Wafter
- 700.021
- Typed Vcom_event
- 1500
- Typed VCperformance_event
- 1050
maximum_move = 700 × 0.03 × 0.005 = 0.105
component_delta = 0.105 × 2/3 × 1 = 0.07
component_after = 700 + (0.07) = 700.07
deltaW = 0.07 × 0.3 = 0.021
Wafter = 700 + (0.021) = 700.021
Vcom_event = validated rate × validated duration = 1500
VCperformance_event = (700 / 1000) × 1500 = 1050
Required case note
Impact justification
- Assumed synthetic evidence
- A synthetic active revision corrects validated duration from four hours to three and explicitly supersedes the earlier revision; the impact decision input is unchanged.
- Reviewed decision input
- positive · medium on the active occurrence
- Calculation demonstrated
- The calculator applies the pinned HouseOfBros Founder Policy v0.1 transition matrix to the declared component state. It does not infer direction or intensity from the evidence narrative.
- Decision rationale
- Only the active revision contributes to valuation and one occurrence still produces at most one transition.
- Alternative interpretations
- A new occurrence would not be a correction.
- Uncertainty
- This synthetic case has not been calibrated against observed longitudinal evidence or independent reviewer agreement.
- Evidence-to-intensity rubric
- unvalidated. Direction and intensity are reviewed inputs; the calculator does not infer them from prose.
- Validation research required
- Append-only supersession and replay conformance tests.
Raw scenario structure
Active correction selection
The superseded revision remains in the audit chain. Only the active revision supplies duration and valuation to replay; its values are not added to the superseded output.
[
{
"revision_id": "rev-1",
"status": "superseded",
"duration_hours": 4,
"old_output": {
"Vcom_event": 2000,
"VCperformance_event": 1400
}
},
{
"revision_id": "rev-2",
"supersedes": "rev-1",
"status": "active",
"duration_hours": 3,
"replayed_output": {
"Vcom_event": 1500,
"VCperformance_event": 1050
}
}
]