Skip to main content
DEWS is a forward-looking stress indicator, not a depeg confirmation. The live risk score detects an active depeg; DEWS catches the structural and market conditions that typically precede one — often 6–48h before price breaks.

DEWS vs the Live Risk Score

DimensionLive Risk ScoreDEWS Score
What it detectsActive depeg — price is already movingPre-depeg stress — conditions building before price breaks
Update frequencyEvery 5 minutesDaily (or on-demand)
Primary inputsPrice, slippage, persistence, oracle, supply spikeSupply flows, pool balance, liquidity erosion, whale concentration, blacklist events, yield anomaly
Lead timeConfirms an event in progressWarns 6–48h before price breaks in historical incidents
When to act on itExit / reduce NOWReview position, increase monitoring, prepare contingencies
Score range0–100 (higher = active risk)0–100 (higher = pre-stress)
Tier labelsok / watch / warning / criticalCALM / WATCH / ALERT / DANGER

Threat Bands

Screenshot 2026 04 13 At 11 09 13 AM

The 11 Pre-Depeg Signals

Each signal is normalized to 0–1 before weighting. Missing signals are skipped and all remaining weights are renormalized to sum to 1.0, so data gaps never inflate the score.
SignalWeightWhat It MeasuresMax Threshold
supply_velocity17%Recent mint pressure — minted_24h as a % of total_supply. A sudden supply surge is the leading pre-depeg signal: issuers mint aggressively to meet redemptions, or attackers exploit uncapped minters.≥10% of supply minted in 24h → 1.0
liquidity_erosion12%% drop in liquidity vs the 7-day average. Pool depth quietly shrinks before a depeg as LPs detect risk and exit early. A 50%+ drop in depth signals impending slippage collapse.≥50% liquidity drop → 1.0
pool_balance_drift11%DEX pool imbalance — how far the stablecoin’s share has drifted from 50%. A balanced Curve/Uniswap pool sits at ~50%. Heavy one-sided selling pushes the ratio toward 80–90% before price breaks.100% or 0% pool share → 1.0
price_confidence10%Inverted price source confidence (DeFiLlama). Low confidence = fewer reliable price sources feeding the market. Single-source pricing is a fragility indicator.low → 0.7 · fallback → 0.9 · single_source → 0.4
cross_source_divergence10%Spread across multiple price data sources. Divergence above 5% means market participants are seeing different prices — a precursor to arbitrage breakdown.>10% spread → 1.0 (5%→0, 10%→1.0)
whale_concentration7%Top-10 holder % of total supply. Concentrated holders can trigger a bank-run cascade — when whales exit, redemption pressure overwhelms liquidity and price breaks.<20% = 0.0 · 50% = 0.5 · ≥80% = 1.0
blacklist_activity8%Freeze/blacklist events on this token contract in the last 7 days. Issuers typically freeze large addresses during active investigations or after detecting suspicious behaviour — a leading compliance signal.5+ events in 7d → 1.0 (1 event → 0.3)
mint_burn_flow8%Bank-run gauge from the supply flows tracker (±100). Negative = unusual redemption pressure. When burning significantly outpaces minting, the market is voting to exit.-100 (max burn pressure) → 1.0 (≥0 = 0.0)
mint_role_centralization8%Single EOA holding MINTER_ROLE — the structural vulnerability exploited in the USR March 2026 incident (~$80M minted from ~$200K collateral by one compromised key). An uncapped single-key minter is a systemic risk regardless of current price.critical (single EOA) → 1.0 · warning (2 minters) → 0.5
tvl_supply_divergence5%On-chain supply materially exceeds DeFiLlama per-chain circulating supply. Tokens created on-chain but not reflected in circulating data = potential unbacked minting or bridge mint accounting error.≥20% on-chain excess → 1.0
yield_anomaly4%APY spike above the 30-day rolling average. A sudden APY jump (issuers boost yield to retain depositors during stress) is a counter-intuitive but historically reliable pre-depeg signal.3x baseline APY → 1.0

Case Study: USR March 2026

The mint_role_centralization signal was added directly in response to the Resolv incident. An attacker who gained access to the single EOA holding MINTER_ROLE minted ~$80M USR from ~$200K of collateral in one transaction — the off-chain service was the only constraint, and there was no on-chain cap. DEWS would have flagged this vault with a DANGER pre-stress score based on the mint_role_centralization signal alone, before any price impact.
This is also why the Safety Grade structural pillar includes a mint_cap_status deduction: an uncapped minter gets −10 pts from the structural score regardless of current DEWS readings.