Methodology
A rating you cannot inspect is a vibe. Every Assay grade decomposes into five pillars, each with its inputs shown on the asset page. Same inputs, same score — deterministic and reproducible.
Pillars & weights
Peg integrity — 30%Measured on-chain price vs the underlying's last trade, sampled every 5 minutes (DEX price vs last real trade from a US market-data feed). Penalties scale with the 7-day p95 deviation and, separately, with weekend deviation — the window where arbitrage is weakest and holders are most exposed. Outside US market hours the reference is the last trade before the close, not a live quote — so off-hours figures measure divergence from that last print, which is the only real price available while the market is shut.
Backing & reserves — 25%Public proof-of-reserves, attestation cadence, named custodian, explicit 1:1 model. Each verified fact adds points; each unverified one is stated as unverified.
Issuer & legal — 20%Named legal entity, regulatory framework, jurisdiction, and the incident record we maintain.
Liquidity — 15%Live pool depth, log-scaled: $10k→25, $100k→50, $1M→75, $10M→100.
Redemption rights — 10%Mechanism, window, and who is eligible to redeem for the real asset.
Conservative by construction
Unknown facts are never guessed. A pillar with unverified inputs keeps a low base score and its reasons say "unverified". Hard caps apply on top: fully unverified backing caps the overall grade at B; an unresolved high-severity incident caps it at C; assets without live peg data receive a conservative peg score of 40 until data accrues.
Data sources
On-chain prices and liquidity from public DEX data; underlying reference prices from a licensed US market-data feed including extended hours; issuer and backing facts from primary sources linked on every page. The registry is hand-maintained; issuers can submit documentation to complete or correct their record.
What Assay is not
Not investment advice, not an audit, not a guarantee. Assay measures structural soundness signals — it cannot see inside custodians or predict conduct. Grades change when facts change.
Assay publishes methodology-based opinions on the structural soundness of tokenized assets, computed from public data. Nothing here is investment advice, an offer, or a recommendation to buy or sell any asset. Unverified facts are treated conservatively — issuers can submit documentation to complete their record.