Methodology
OmniScore is a wallet reputation signal. It summarizes onchain behavior, risk exposure, wallet maturity, and trust indicators into a single score you can read quickly — and then decompose into pillars for deeper review. Use it as a screening layer, not a final verdict.
Score Bands
Raw scores are normalized to a 0 – 100 scale and mapped to a letter grade.
Strong / Healthy
Few immediate warning signs. Still apply normal diligence.
Lower Risk
Relatively stable behaviour across pillars.
Moderate Risk
Mixed picture — inspect pillars before acting.
High Risk
Meaningful concerns that deserve manual review.
Very High Risk
Deeper validation and caution strongly recommended.
Weights
Pillars sum to 100 %. Each factor is z-scored on its own scale before the weighted sum.
| Factor | Weight | Measures |
|---|---|---|
| Behavior | 30 % | Transaction patterns, counterparty diversity, dApp engagement depth. |
| Risk Exposure | 25 % | Interactions with sanctioned, flagged, or high-risk clusters. |
| Liquidity | 20 % | Balance stability, realized volatility, drawdown behavior. |
| Maturity | 15 % | Wallet age, continuity, and cross-chain footprint. |
| Trust Signals | 10 % | Verified attestations, reputation proofs, protocol endorsements. |
Formula
OmniScore =
0.30 * z(Behavior)
+ 0.25 * z(RiskExposure)
+ 0.20 * z(Liquidity)
+ 0.15 * z(Maturity)
+ 0.10 * z(TrustSignals)
Normalized = clamp(0, 100, 50 + 10 * OmniScore)
Band = mapToGrade(Normalized) // A | B | C | D | F
Confidence = evidenceCoverage * dataRecencyConfidence reflects how much evidence the model had. Low-confidence scores should be treated more cautiously than the letter grade alone suggests.