Polygraph Whitepaper (v0.2)

**Canonical markdown mirror** of the Polygraph protocol specification. Section numbers match the full v0.2 document (`Polygraph_Whitepaper_v0.2`); this repository copy is maintained in sync with deployed contracts on Polygon. Rendered at `/whitepaper` on polygraph.news.

8. Protocol economics

Polygraph revenue is **outcome-neutral** wherever it touches credibility adjudication: the protocol treasury must not earn more when content is ruled false than when it is defended. Subscription and API revenue (§8.5–§8.6) use fixed pool splits unrelated to challenge verdicts.

8.5 Verification and reader access

PPOR generation is performed by registered oracle operators (§4). In V1 the flagship reader gates server-side PPOR runs behind an active subscription; there is no on-chain micropayment per score. Product-layer verification fees (where offered) are mostly pass-through to model/API costs with a documented treasury margin — see `docs/FUNDING_CONSTITUTION.md` (Economics).

8.6 Subscription and tip revenue

Reader subscriptions (10 USDC/month) and tips accumulate in `SubscriptionVault`. Each 30-day epoch, `RevenueDistributor` splits undistributed funds **65% authors / 25% treasury / 10% oracle operators** (basis-point constants on-chain). Author shares within the 65% pool are weighted off-chain by engagement and ACS tier, then submitted by the keeper via `finaliseEpochWithWeights`. This split does **not** vary with challenge outcomes.

8.7 Challenge stake settlement

Challenges require a USDC **stake** (minimum tiered by article PPOR — see `FactCheckEngine.minStakeFor`). Authors post a separate **publication bond** (5 USDC) at publish time; disputed challenges may require the author to post a **defense stake** matching the challenger's stake before Kleros arbitration.

When a challenge resolves (author concession, silence after the 7-day response window, or Kleros ruling), stake accounting uses a single **outcome-neutral protocol fee** on the **challenge stake**:

| Symbol | Meaning |

|--------|---------|

| `S` | Challenger's stake for this challenge |

| `protocolFeeBps` | Governance parameter, default **500** (5%), hard cap **1000** (10%) |

| `protocolFee` | `S × protocolFeeBps / 10_000` |

**Treasury:** The protocol treasury (immutable address at `FactCheckEngine` deploy) receives **exactly** `protocolFee` on **every** resolution path — upheld or rejected. There is **no** additional treasury rake on rejected challenges (legacy v0.2 drafts described a 70%/30% author/treasury split on reject; **v0.2.1-on-chain** replaces that with winner-takes-remainder plus the fixed fee only).

**Winner:** The party that wins the challenge receives **`S − protocolFee`** from the challenger's stake (plus separate bond/defense-stake transfers below). The fee rate is identical whether the challenger or author wins.

**Governance:** `protocolFeeBps` is set via `FactCheckEngine.setProtocolFeeBps`, callable only by `GOVERNANCE_ROLE` (TimelockController). Changes do not alter ATS/ACS formulas.

Resolution paths

**Upheld (challenger wins):**

1. `CredibilityLedger.applyChallenge` decrements ATS/ACS per claim type (§7).

2. Challenger receives **`S − protocolFee`** refunded from their stake.

3. Treasury receives **`protocolFee`** (USDC transfer + `ProtocolFeeCollected` event).

4. If the dispute was arbitrated, the author's matching defense stake is transferred to the challenger in full.

5. A proportional share of the publication bond may be slashed to the challenger (50% × Kleros coherence percentage, coherence floored at 50% for penalty math).

**Rejected (author wins):**

1. ATS/ACS unchanged for this challenge.

2. Author receives **`S − protocolFee`** from the challenger's forfeited stake, plus full return of their defense stake if arbitrated.

3. Treasury receives **`protocolFee`** (same amount as on an upheld resolution for the same `S` and `protocolFeeBps`).

**Kleros:** Escalated disputes use `KlerosAdapter`; juror economics follow Kleros Liquid rules. Polygraph's **`protocolFee`** is independent of Kleros juror rewards and is always exactly `protocolFeeBps` of `S`.

Minimum stake tiers (unchanged)

| Article PPOR | Minimum challenge stake |

|--------------|-------------------------|

| ≥ 70 | 20 USDC |

| 40–69 | 15 USDC |

| < 40 | 10 USDC |

Design rationale (outcome neutrality)

Credibility infrastructure must not profit from debunking. A higher treasury take on rejected challenges would create a perverse incentive to uphold challenges regardless of merit. Fixing `protocolFeeBps` across outcomes keeps dispute revenue predictable and constitutionally separate from truth labels (ATS/ACS).

*Other whitepaper sections (§1–§7, §9–§14) remain as in the full v0.2 document; challenge settlement in §8.7 above is authoritative for deployed `FactCheckEngine` behavior.*