Back to PJC Framework

Public technical method · Research Preview v0.6

PJC Evidence and Cases

Disclosure date: 25 August 2026
Current conclusion: PJC has an executable method structure, but insufficient evidence that it improves calibration, accuracy, efficiency, or team performance.

1. Status overview

Evaluation targetStatusPublic basis
Method definitionSpecifiedv0.6 documents and field schema
Minimum operabilityUnder evaluationNo complete pilot dataset published
Record completenessNot measuredInternal pilot pending
Settlement rateNot measuredInternal pilot pending
Confidence calibrationNot validatedComparable records required
Error detectionNot validatedNo baseline or control
Work efficiencyNo claimRecording cost unmeasured
Cross-context validityNot validatedNo external cases
External peer reviewNot conductedCriticism invited

2. What current evidence supports

PJC has a consistent field schema and lifecycle; append-only commitment can preserve pre-outcome text; the claim register separates mechanism claims from empirical hypotheses; and the method declares non-use conditions and invalid states.

These are facts about method design and implementation, not behavioural effects.

3. What current evidence does not support

PJC must not currently be described as making people more rational, improving decision accuracy, making agents reliable, lowering costs, fitting every team, or being scientifically validated.

Any future effect statement must link to evidence or remain explicitly untested.

4. Internal pilot plan

The first 30–60 day pilot will cover recurring, low-risk product, engineering, agent-workflow, user-behaviour, and release-risk judgments.

Priority measures:

MeasureDefinition
Completion rateRecords with every required field
Timely commitmentRecords committed before outcome availability
Settlement rateRecords settleable under the original rule
Invalidation rateRecords excluded for leakage, rule switching, or data failure
Continued useParticipants recording across multiple cycles
Recording costTime spent creating, reviewing, and settling

Calibration and decision quality come only after basic operability.

5. Case publication standard

Every published case includes context, roles, agent involvement, duration, record count, original propositions, settlement criteria, actual outcomes, evidence sources, observed and unobserved changes, failed records, data limits, conflicts of interest, and reviewability.

Internal cases carry this notice:

An internal case is not independent validation.

6. Case register

No case currently meets the publication standard.

This is an evidence boundary, not a placeholder. Demonstrations will not be presented as empirical validation.

7. External evaluation

Send criticism, failed cases, or replication proposals to contact@sedesmentis.com. We especially value examples of unsettleable tasks, field-induced bias, cost exceeding benefit, agent contamination, or no improvement over a lighter baseline.

Substantive changes are recorded in the changelog.