Lesson 56: Waiver Renewal Evidence Completeness Scorer for Policy Alignment in RPG Live-Ops

Lesson 55 gave you deterministic sunset checkpoints so waivers cannot quietly expire. The next reliability gap is renewal quality: teams can still extend waivers using partial evidence that does not meet policy expectations.

This lesson adds a renewal evidence completeness scorer so renewal approvals are based on verifiable proof quality, not reviewer memory or urgency pressure.

Random Illustrations artwork for waiver renewal evidence completeness scorer lesson

What you will build

By the end of this lesson, you will have:

  1. A waiver_renewal_evidence_requirements.md policy contract
  2. A waiver_renewal_evidence.csv schema for renewal packets
  3. Deterministic completeness states (complete, partial, insufficient)
  4. A release-routing action map that blocks weak renewal extensions

Step 1 - Define the renewal evidence contract

Create one policy file for all renewal packets:

  • required proof fields per waiver type
  • freshness window for each evidence field (for example, 7 days)
  • accepted source systems for each field
  • reviewer lane ownership and escalation lane
  • minimum score required for approval

Without an explicit contract, "approved renewal" often means "approved by habit."

Step 2 - Build waiver_renewal_evidence.csv

Track one row per renewal request:

column purpose
waiver_id renewal target waiver
release_cycle_id active release cycle
renewal_request_id unique request identifier
policy_version active policy baseline
required_field_count number of required fields
present_field_count fields actually submitted
fresh_field_count submitted fields within freshness window
policy_match_count fields matching policy rule format
completeness_score deterministic score 0-100
completeness_state complete, partial, insufficient
review_owner_lane renewal decision owner
routing_action approve_path, needs_followup, block_extension

Keep this file adjacent to Lesson 55 outputs so sunset timing and renewal quality are reviewed together.

Step 3 - Add deterministic scoring logic

Use a simple weighted model:

  • field presence: 40%
  • field freshness: 30%
  • policy-format match: 30%

State mapping:

  • complete: score >= 90 and no critical-field gaps
  • partial: score 70-89 or one critical-field gap
  • insufficient: score < 70 or multiple critical-field gaps

This gives renewal discussions a stable baseline under patch pressure.

Step 4 - Route scorer outcomes into release controls

Before renewal approval:

  1. attach waiver_renewal_evidence.csv snapshot to packet
  2. block expiry extension if state is insufficient
  3. allow conditional extension for partial only with due timestamp and owner
  4. approve only complete rows with reviewer acknowledgement in UTC

Extension decisions should be evidence-bound, not calendar-bound.

Step 5 - Add one daily renewal quality review

Run a short daily pass:

  • list all renewal requests entering decision window in next 24h
  • recompute completeness state from latest evidence rows
  • route partial and insufficient requests to follow-up owners
  • log unresolved gaps with escalation lane and due time

A short repeatable cadence prevents low-quality approvals.

Common mistakes

Mistake: Counting uploaded files as complete evidence

Fix: score only required policy fields, not file count volume.

Mistake: Accepting stale evidence during urgent release windows

Fix: enforce freshness counts and route stale rows to follow-up lanes.

Mistake: Treating partial packets as automatic approval

Fix: require explicit conditional-extension routing for all partial states.

Pro tips

  • Keep one critical_missing_fields column for fast triage
  • Version renewal policy changes and include policy_version in each row
  • Recompute scores automatically after every packet update

Mini challenge

  1. Create 12 renewal request rows in waiver_renewal_evidence.csv.
  2. Apply the weighted completeness formula.
  3. Label each row as complete, partial, or insufficient.
  4. Produce one lane summary for next release-gate review.

FAQ

Should we allow renewal when one required field is missing

Only as a conditional extension with explicit owner, due timestamp, and blocker note.

How often should completeness scores be recomputed

At least daily during active release cycles and immediately after packet updates.

What if policy rules changed after request submission

Re-score against the current policy_version and log any new gaps before approval.

Lesson recap

You now have a waiver renewal evidence completeness scorer that turns renewal approval from a subjective check into a deterministic policy-aligned decision workflow.

Next lesson teaser

Next, continue with Lesson 57: Waiver Renewal Decision Confidence Tracker for Borderline Approval Escalation in RPG Live-Ops so repeated borderline renewal approvals are escalated before governance debt compounds.

Related learning