Arbitration rules

Review, reports, and arbitration can chain: e.g. A’s contribution rejected by B → C reports B → D arbitrates C’s report → E reports D → F arbitrates… Without limits, one matter could be reported and arbitrated indefinitely. So the platform defines: a single contribution or single entry update is one «case»; the report/arbitration chain for that case has a level cap; once the cap is reached the case is closed and no new reports on that case are accepted.

One case = one contribution submit or one entry update, identified by contribution_id; if it’s an update to an existing entry, also by entry_id.

All later reports and arbitrations about that contribution, that first review, or reports/arbitrations on them belong to the same case. Multiple updates to the same entry (same entry_id) create multiple contributions and multiple cases, each with its own level and closure.

Each case allows at most:

• Level 0 — First review: reviewer’s first decision (approve / reject / request changes). Only once.
• Level 1 — One appeal or one report + one arbitration: contributor’s appeal of that review, or anyone’s report on that contribution/review; one arbitration. Only one Level 1 per case (appeal and report share the slot).
• Level 2 — One report on Level 1 arbitration + one arbitration (final): only one report against the Level 1 outcome; a different reviewer/panel does one arbitration. This is final; after Level 2 the case is closed.

After closure, no new reports are accepted for that case (whether the target is the contribution, the first review, Level 1 or Level 2 arbitration). Submitting a report on a closed case is rejected with a clear message.

When Level 2 is done, the case is marked closed. Then:

• New reports on any object in that case (contribution, first review, Level 1 or Level 2 arbitration) are rejected.
• API reports targeting a closed case return 409 Conflict with code CASE_CLOSED, e.g. «Reports and arbitration for this entry/contribution are closed; no new reports accepted.»

This keeps «supervisors can report unfair review» while avoiding infinite report/arbitration chains.

Role recusal in the same case: contributors cannot review their own; reviewers who did Level 0 or Level 1 cannot do Level 2 for that case.

Multiple updates to the same entry (same entry_id) are separate contributions and separate cases. E.g. entry X has contribution 1 (approved), then 2 (update), then 3 (update). Case 1, 2, 3 each have Level 0 → 1 → 2 and close independently. So one entry is not «never closed»; each version’s case closes on its own.