Omniousdocs
Market integrityShadow

The economics of honesty

How the versioned Shadow points downside scales. It is a deterrence model, not a profitability or detection guarantee.

Owner
Trust & Safety
Version
s0-2026.07.1
Verified
2026-07-13

The integrity lane is not a court; it is a price system. Every number in it is a published constant (terms version s0-2026.07.1, exposed by the API), and together they are chosen to make review spam costly and an upheld substitution much more expensive than the points earned on one fill. This is a deterrence model, not a guarantee that fraud is impossible or currently unprofitable.

the stake / slash calculatorpublished constants · s0-2026.07.1

Shadow points only · no USDC or token collateral

⚑ challenger locks points
200 GPUs ≈ 2.00 volume basis

max(25 GPUs, 1x base earn). Locked while open.

provider points reduced if upheld
4,650 GPUs ≈ 46.50 volume basis

max(250 GPUs, 25x net earn of 186 GPUs.

points credited to challenger
2,325 GPUs ≈ 23.25 volume basis

Half the points reduction.

✕ credited to no one
2,325 GPUs ≈ 23.25 volume basis

The other half is credited to no one.

An upheld challenge reduces the provider's Shadow points by the equivalent of about 25settled fills of the same size. A rejected challenge forfeits the challenger's 200 GPUs points stake instead. Volume basis does not imply cash value.

The three constants

  • Stake = max(25 GPUs, 1x the receipt's base points earn). Base earn is 100 GPUs per $1 charged, so challenging a $2.00 fill locks 200 GPUs. The floor gives micro-fill spam a non-zero points cost.
  • Slash = max(250 GPUs, 25x the provider's net earn). Net earn already subtracts the router fee. Above the floor, an upheld review reduces provider points by roughly 25 settled fills of the same size.
  • Challenger share = 50%. Half the provider-points reduction is credited to the challenger. The other half is credited to no one.

Why these shapes

Why 25x: a substitution cheat saves the provider some fraction of serving cost per fill, far less than the full charge. For cheating to pay at a 25x penalty, the provider would need the probability of an upheld review to stay very low. Canary operations are Shadow, sampled audits are Proposed, and public bounty access is off, so this remains a sizing rationale rather than a live detection claim.

Why burn half: if the whole slash went to the challenger, a provider could stage challenges against itself through a friendly challenger and convert its own slash into a round trip, cheaply laundering a bad record into bounties. Burning half makes every staged uphold a guaranteed net points loss for the pair.

Why stake scales with the receipt: big receipts have big bounties attached, so the price of being wrong about one scales with the potential points credit. The ratio therefore scales mechanically with fill size.

Shadow terms, not a public rewardAll balances are pure derivations over settlement and challenge rows. The arithmetic is implemented for approved reviewer wallets, but public activation is separate. Points do not promise a token, conversion, value, or date. See The GPU points program.
As a providerRead the calculator as Shadow points exposure under an operator-recorded upheld review. It is not an insurance premium or escrowed collateral.
As a challengerApproved reviewers can see the points downside and potential credit before signing. Public eligibility and rewards are not active.
where this lives in code