Identity & Verification

Chrono Score is only valuable if it cannot be faked. This page covers how Chronomancy creates verifiable, non-transferable forecasting identity — and defends against the attacks that could undermine it.

On-Chain Attestations via EAS

Chrono Score is attested on-chain using the Ethereum Attestation Service (EAS) — a tokenless, permissionless attestation protocol with two contracts (SchemaRegistry + EAS). As of early 2025, EAS has 420K+ unique attestors in production across Ethereum and L2s. It is natively integrated into Optimism’s Bedrock.

The CS schema encodes per-wallet:

Calibration metrics (Brier score, CS Level)
Glicko-2 Rating + RD
Domain subscores
Attestation timestamp and version

Attestations are updated on-chain after each resolved prediction. The attestation is soulbound — it cannot be transferred or sold.

Soulbound Tokens

Milestone achievements (tier upgrades, seasonal rankings, Wizard Duel wins) are issued as ERC-5192 soulbound tokens — a minimal extension of ERC-721 where the transfer function always reverts.

Account recovery (in case of key loss) is handled via ERC-4337 smart account recovery — users can designate a recovery address that can claim a new soulbound token tied to the same CS record.

Sybil Resistance

A Sybil attacker creates many wallets, gets lucky on predictions, and exploits inflated CS for Fast-Forward premiums. The formal threat model:

Attack cost at κ=30 (minimal Bayesian shrinkage):

Need ~6 accounts to expect 1 “lucky winner”
At $600/account (staking + costs) → $3,600 to get one Premium-tier exploit
Not yet economical to prevent

The defense stack layers from cheapest to most friction:

Layer 1 — Proof of Personhood

Gitcoin Passport (≥20 score recommended) as the baseline identity gate. World ID available as a premium tier for highest-sensitivity contexts.

Passport ≥20 requires: verified Twitter/GitHub/ENS + on-chain history + optional biometric. Attack cost: ~$300/account for credible fake.

Layer 2 — Tiered Staking

Tier	Staking Required	FF Cap
Basic	$0	$1K
Standard	$100 (USDC)	$10K
Premium	$500	$50K

Staking makes attacks unprofitable: at Premium tier, attacking with 33 accounts costs $99K+ — well above exploitable profit from FF margins.

Layer 3 — Temporal Gates

90-day minimum account age before CS access
≥2 predictions/month activity requirement
7-day FF cooldown after large exits

These gates prevent rapid account creation + exploit cycles.

Layer 4 — On-Chain Detection

Automated detection pipeline running continuously:

Signal	What it detects
Funding graph	Multiple CS wallets funded from same source
Position correlation	N wallets making identical predictions
Behavioral clustering	Prediction timing patterns
CS anomaly detection	CS growth rate inconsistent with prediction volume
Anti-correlation	Systematic hedging across wallet cluster

Layer 5 — Difficulty-Weighted CS (Structural Defense)

The hardest-to-attack defense is the CS formula itself. Easy-market farming (predicting 85%+ markets for cheap CS accumulation) is structurally blocked:

difficulty_weight = 1 - 2 × |p - 0.5|

At p=0.85: weight = 0.30. Three certainty-market predictions count as one real prediction.

Response Protocol

Threat Level	Action
Suspicious	Watch: flag wallet, no restriction
Elevated	Restrict: reduce FF caps by 50%
Confirmed	Suspend: freeze CS updates, lock staking
Proven attack	Slash: slash staked collateral, ban wallet

Cross-Platform Import Verification

To prevent importing inflated Metaculus scores without real forecasting skill, imports use TLSNotary proofs — cryptographic proofs that a specific HTTPS response from the source platform (showing your public prediction record) was genuine.

Import credit caps: 50% of face value from Metaculus/GJP, 30% from Polymarket on-chain history. The imported CS is tagged as “imported” and decays faster unless confirmed by subsequent on-platform performance.

Related:

Chrono Score — the reputation system being protected
Precognition — how new users enter without existing identity