Methodology
Health Score v1 is rules-based, descriptive, and auditable.
Module A avoids predictive modeling. Each component maps directly to observable data or explicit profile metadata, with fixed thresholds and published weights.
Scoring formula
The final score is the weighted sum of five deterministic subscores. No hidden overrides or qualitative analyst adjustments are applied in Module A.
Health Score = 0.30 x Peg + 0.20 x Volatility + 0.20 x Market Cap Stability + 0.15 x Volume Trend + 0.15 x DisclosurePeg stability
30%Uses both the 30d count of daily deviations beyond 0.5% and the maximum absolute deviation across the same window.
Volatility
20%Uses the standard deviation of daily returns over 30 days. Lower realized price volatility receives a higher subscore.
Market cap stability
20%Uses the absolute 30d percentage change in market capitalization. Larger moves are treated as less stable.
Volume trend
15%Uses the ratio of average 7d volume to average 30d volume. Large spikes or drops away from 1.0 are penalized.
Disclosure score
15%A deterministic transparency proxy based on profile metadata: attestation cadence plus availability of reserve, attestation, and redemption links.
Disclosure Score v0
Disclosure Score v0 is a transparency proxy based on publicly stated attestation cadence and the existence of reserve, attestation, and redemption documentation links.
- +30 monthly attestation frequency
- +15 quarterly attestation frequency
- +20 reserve breakdown link
- +20 attestation link
- +20 redemption terms link
It is not a legal opinion, reserve audit, or solvency guarantee. It only measures the presence of basic public disclosure artifacts.
Limitations
- Partial CoinGecko market-chart windows can reduce available metrics; those cases are marked as partial data in score components.
- Disclosure Score v0 is a proxy for transparency and documentation, not a legal opinion, reserve audit, or guarantee of redeemability.
- Module A remains descriptive rather than predictive and does not model regional flows or counterparty dependencies.
Risk Event Feed Note
Event feed items are generated by deterministic rules comparing day-over-day changes in score, peg stress, supply, and rank context. The feed is an auditable change log, not a predictive alerting model.