PMFA
← Return to Index
005

Audit Without Replay Is Fiction

If past decisions cannot be replayed under audit, audit reports are narratives, not evidence.
Version 1.0.0 — Ratified

Audit is not observation.
Audit is verification.

Verification requires replay.
Without replay, audit is fiction.

I. The Purpose of Audit

Audit exists to answer one question:

Did this actually happen as claimed?

This question cannot be answered by logs.
It cannot be answered by reports.
It can only be answered by replay.

Replay means: given the same inputs, execute the same logic, observe the same result.

If the result differs, the claim is false.

II. The Current State of Audit

Modern enterprise systems produce audit artifacts:

These artifacts describe what happened.
They do not prove what happened.

Description is not proof.

III. The Replay Requirement

A system is auditable only if:

  1. All inputs are recorded
  2. All inputs are immutable
  3. The execution path is deterministic
  4. Re-execution produces identical results

If any condition fails, audit is impossible.

Most enterprise systems fail all four conditions.

IV. Why Replay Fails

Replay fails because systems are designed without audit in mind.

Common failures:

Each failure makes replay impossible.
Each impossibility makes audit fictional.

V. The Cost of Fictional Audit

When audit is fictional:

The cost is not immediately visible.
It appears during crisis.

Crises expose fiction.

VI. The Architectural Solution

Audit must be built into the execution layer.

Requirements:

These are not features to add later.
They are foundational constraints.

VII. Final Conclusion

Audit without replay is not audit.
It is theater.

A system that cannot prove its past cannot be trusted with its future.

The ability to replay is not optional.
It is the minimum requirement for a governed system.

Without replay, there is no audit.
Without audit, there is no governance.
Without governance, there is no system.

Only stories.

Canonical text. Interpretations are invalid.
SHA-256: bce7ca2fa529b05462e7165de5bd93f9d0dd3e2e34ae70e4cbcde024395342cc