How Risk Assessment Scores and Assessment Grades are built

Risk Assessment Score and Assessment Grade

The Risk Assessment Score (0–100) is the weighted sum of six dimension subscores; higher scores indicate lower assessed risk. Each score maps to an Assessment Grade from A to F using the published ranges below. Scores and grades are our opinions about relative risk, formed by applying this methodology to the information available on the stated as-of date. They are not credit ratings, statements of fact, predictions or recommendations, and no grade means an asset is free of risk.

Hard-stop caps

Some conditions cap the score regardless of other dimensions:

Six dimensions

The full methodology is public. Asset-level subscores, evidence and analysis are available to subscribers. Weights reflect the channels through which a stablecoin has historically failed to hold value or to be redeemable: reserve shortfalls, liquidity runs and peg breaks carry the most weight, with regulatory, operational and concentration risks as amplifiers.

Confidence and coverage

Coverage is the share of a dimension's indicators for which current, sourced data was available at the as-of date. Confidence (high, medium or low) is the analyst's judgement of how reliable the inputs are, considering source quality, recency and consistency. Low confidence or coverage is shown next to every score so readers can see where the evidence is thin. Assets without enough data to assess are designated NR (Not Assessed) and tracked qualitatively.

• Missing submetrics are scored 0 (highest risk) in the reported score. An upper-bound score (missing = 100) is also published so readers see the effect of evidence gaps.
• A submetric older than its max_age_days is capped at 50 and flagged stale.
• If an analyst searched for evidence and confirmed it is not publicly available, the submetric is scored using the explicit 'absent' category (e.g., none_found, undisclosed) and counts as observed.

Process and governance

• Monthly cadence. Data are collected to a fixed as-of date (the last day of the prior month) and published early in the month.
• Event-driven updates. A material event (a significant depeg, reserve or redemption change, enforcement action, or major incident) triggers an interim update.
• Human sign-off. Automated collection and analysis are reviewed, and every issue is signed off by a named analyst before publication.
• Versioning. The methodology is versioned. Changes to weights or definitions are announced in advance and noted in the issue.
• Back-testing. The framework is tested against historical events such as the March 2023 USDC depeg and the 2022 collapse of UST. Back-test results will be published with methodology version 1.0.
• Conflicts. Commercial relationships with any covered issuer are disclosed, and conflicted analysts recuse themselves.

Data sources

Issuer attestations, audits and disclosures; redemption terms; regulatory filings, orders and the Federal Register; market data on prices and order-book depth; on-chain data on supply and holder concentration; and public third-party assessments, cited where used. Every figure in an issue links to its source with the time it was retrieved.

Limitations

• Scores are point-in-time assessments based on information available at the as-of date and can change without notice.
• Issuer disclosures and third-party data are relied on and are not independently audited by the publisher.
• A Risk Assessment Score is an opinion, not a credit rating. It does not predict a depeg and is not a recommendation to hold, buy, sell or use any asset.
• Readers remain responsible for their own research and decisions; institutions remain responsible for their own due diligence and regulatory obligations.