Data Sources & Methodologies
How we collect, process, and score DeFi vault data — from raw on-chain reads to risk analytics.
Coverage
Data Sources
On-Chain RPC
PrimaryDirect reads from 16 EVM chains via multicall-batched RPC. 20 vault protocol adapters (ERC-4626 share prices, oracle dependencies, governance structures, holder concentration, deposit/withdrawal flows, deployer metadata) plus 12 lending protocol readers (supply/borrow rates, utilization, LTV, market depth) — all read straight from smart contracts.
DIA Oracles
PrimaryPrimary token price source via the DIA API. Fetches real-time USD quotations, yesterday prices, and 24h volume for every token in the pipeline. Covers all 16 chains with per-asset granularity at contract-address level.
DeFiLlama
AggregatorProtocol-level TVL, pool yields (base vs reward APY), 365-day APY history, protocol fee revenue, and security incident data. Used for ecosystem-wide coverage and cross-protocol comparisons where on-chain reads alone would be impractical.
+ 4 additional sources
Curator and vault metadata from the Morpho protocol API — entity names, vault deployment info, and cross-references for the risk curator ecosystem.
Fallback token prices, market cap, circulating supply, and price change data for peg deviation monitoring on stablecoins and LSTs.
Implied APY, underlying APY, swap fee APY, PENDLE rewards, expiry dates, PT/YT/SY tokens, and trading volume for Pendle markets across 11 chains.
Trailing 7d/30d APY, collateral breakdown, strategy allocations, and protocol revenue for OETH, SuperOETH, OUSD, and OS.
Methodologies
Realized APY
((share_price_now / share_price_then) - 1) × (365 / days) × 100Computed from on-chain ERC-4626 share price snapshots. Annualized over 30-day, 90-day, and 365-day windows. Caps at ±10,000% to reject anomalies. Requires at least 2 snapshots with ≥1 day gap.
Max Drawdown
(peak − trough) / peak × 100%Peak-to-trough decline from daily share price snapshots. Measures the worst historical loss a depositor could have experienced.
Sharpe Ratio
(mean_apy − 5%) / std_devRisk-adjusted return using 90-day APY history. The 5% risk-free rate proxies US Treasury bills. Requires a minimum of 7 daily observations.
Credit Spread
entity_tvl_weighted_apy − benchmark_rateMeasures how much excess yield an entity offers above the DeFi risk-free rate. The benchmark is the median Aave v3 USDC supply rate across 6 major chains. A 30-day rolling z-score flags outliers.
Jump-to-Default (JTD)
Σ vault_tvl where primary_asset = XTotal TVL at risk if a specific asset goes to zero. Assets are classified into stablecoins, LSTs, LRTs, wrapped BTC, and native tokens. Concentration is measured via HHI (Herfindahl-Hirschman Index).
Yield Sustainability
apyBase / (apyBase + apyReward) × 100%Percentage of yield from protocol mechanics (lending interest, staking rewards) vs token emissions. Vaults with >80% reward-driven yield are flagged as emission-dependent.
Peg Monitoring
market_price − peg_targetTracks stablecoin prices against their $1 peg and LST exchange rates against their underlying (e.g., stETH:ETH). Data from CoinGecko and on-chain oracle reads.
Governance Scoring
Composite of owner type + timelock + multisig + proxy statusEvaluates vault admin structure: multisig vs EOA ownership, timelock delay presence, guardian roles, and proxy upgradeability. Higher scores indicate stronger governance safeguards.
Data Pipeline
A 30-phase cron pipeline runs every 6 hours, with a separate daily discovery pass that scans protocol factory contracts for new vaults. Data flows through three caching tiers: in-memory (5 min), database (24 hours), and incremental static regeneration (60 seconds).
Overview stats, pool metadata, lending markets, on-chain vault reads, token prices (DIA + CoinGecko)
APY history, entity discovery (DeFiLlama + Morpho), protocol profiles, lending-pool bridge, data validation
Oracle deps, holder stats, governance, oracle freshness, capital flows, deployers, backing analysis
Drawdown & Sharpe, entity performance, vault fees & capacity, asset exposure (JTD), credit spreads
Asset correlations, layer summaries, asset intelligence, origin enrichment, pendle enrichment