How We Grade Risk
Hindenrank grades protocols on a scale from A (lowest risk) to F (highest risk) based on economic design analysis. Higher scores mean more risk — not necessarily "bad," but more novel, untested, or complex.
The Six Dimensions
How many of the protocol's economic mechanisms deviate from established patterns? Novel mechanisms have less battle-testing and more unknown failure modes.
How dangerous are the interactions between mechanisms? We flag known dangerous pairs (e.g., vesting cliffs + thin liquidity) and weight by severity: Critical (10), High (5), Medium (2).
How dependent is the protocol on external price data? Custom oracles, single-source feeds, and hardcoded values increase this score.
How well-documented are the protocol's mechanisms? Poor documentation makes it harder to assess risk and may indicate incomplete design thinking.
Has the protocol experienced exploits, bugs, or security incidents? Confirmed vulnerabilities increase this score significantly.
How much capital is at risk? Higher TVL and FDV mean higher dollar impact if something goes wrong.
Grade Scale
| Grade | Score Range | Meaning |
|---|---|---|
| A | 0–10 | Battle-tested, low novelty, strong docs, clean track record |
| B | 11–35 | Some novel mechanisms, moderate interaction risk |
| C | 36–65 | Multiple novel mechanisms, notable interaction risks |
| D | 66–80 | Extreme novelty, critical interactions, unproven at scale |
| F | 81–100 | Multiple critical risks, confirmed vulnerabilities, or systemic concerns |
FAQ
Can a protocol improve its grade?
Yes. Grades change when protocols fix documented risks, improve documentation, survive stress tests, or add audits. We re-scan protocols periodically.
Does a low grade mean the protocol is a scam?
No. A low grade means the design has more novel, untested, or complex elements. Innovation inherently carries risk. Some of the most interesting protocols score poorly because they're genuinely new.
How often are protocols scanned?
New protocols are added weekly. Existing protocols are re-scanned when major events occur (exploits, TGE, mechanism changes) or on a rolling schedule.
Is this financial advice?
No. Hindenrank provides economic design analysis. It is not a recommendation to buy, sell, or deposit into any protocol.
How do you rate Morpho lending vaults?
We score each vault by the risk of its collateral protocols (weakest-link) plus lending market parameter penalties for LLTV, oracle type, liquidity depth, and utilization. LLTV penalties are adjusted for collateral-loan correlation — derivative pairs (wstETH/ETH) are penalized less than standard pairs (wstETH/USDC). Morpho platform risk is shown separately as a contextual callout.
Why do your ratings differ from curator self-ratings?
Vault curators like Steakhouse Financial rate their own products using credit-focused frameworks. Our ratings also capture cryptoeconomic mechanism risk of the underlying protocols — consensus design, oracle dependencies, governance surface, and track record. Both perspectives are complementary.
Vault Risk Ratings
Hindenrank rates Hyperliquid trading vaults by combining the underlying protocol risk of the tokens they trade with structural risk factors: leverage, max position leverage, and directional concentration. The result is a vault risk score that reflects the overall risk of depositing into the vault — not just the protocol risk of the tokens themselves.
Position-Weighted Risk
Each vault holds perpetual futures positions across multiple tokens. We map each token to its underlying protocol risk score and compute a notional-weighted average. Short positions are treated differently depending on the grade of the token being shorted:
| Token Grade | Short Treatment | Rationale |
|---|---|---|
| A or B (score ≤ 35) | Raw protocol score (no penalty) | Pure market bet — direction irrelevant to protocol risk |
| C (score 36-55) | Dampened toward 50 | Mixed — some protocol signal |
| D or F (score > 55) | Full inversion (100 − score) | Betting on protocol failure |
Shorting BTC or ETH is a directional market call, not a bet that the protocol is broken. We use the raw protocol score (no penalty) since the risk is market-directional, not cryptoeconomic. Shorting riskier protocols retains the full score inversion because you're betting on structural weakness — which is exactly what our protocol risk scores measure.
Account Leverage Penalty
Leverage amplifies both returns and losses. A vault running 10x leverage can be liquidated by a 10% adverse move. We apply an exponential penalty based on account-level effective leverage (total notional / equity):
| Leverage | Penalty | Wipeout Move |
|---|---|---|
| 1x | +0 | 100% |
| 2x | +3 | 50% |
| 3x | +6 | 33% |
| 5x | +14 | 20% |
| 10x | +25 | 10% |
| 25x | +25 | 4% |
Capped at 25 points. Formula: min(25, 1.5 × (leverage1.5 − 1)).
Max Position Leverage Penalty
A vault with 2x average leverage but one position at 25x can still be liquidated by that single position cascading. We add a separate, gentler penalty for the most-leveraged individual position, capped at 10 points. Formula: min(10, 0.5 × (maxLev1.2 − 1)).
Directional Concentration
We measure the net-to-gross ratio — how one-sided the portfolio is:
|long − short| / (long + short)
A ratio of 0 means market-neutral; 1.0 means fully directional (all long or all short). Above 0.7, we add a penalty up to 8 points. A portfolio that looks diversified across 8 tokens but is 100% short is still making one bet: “crypto goes down.”
Performance Metrics (TWR)
Vault returns are computed using Time-Weighted Return (Modified Dietz method), which strips out the effect of deposits and withdrawals. Raw AUM growth is misleading — a vault that grows from $10k to $10M mostly through new deposits has not returned 100,000%. We fetch the full deposit/withdrawal history and adjust each period's return to reflect actual investment performance.
Sharpe ratio, max drawdown, and volatility are derived from the flow-adjusted NAV series, not raw AUM. Due to the weekly granularity of Hyperliquid's account value snapshots, these metrics are approximations — directionally correct but not precise to the decimal.
Morpho Lending Vault Risk Ratings
Hindenrank rates Morpho lending vaults by combining the risk profiles of their collateral protocols with lending market parameter penalties. Morpho platform risk is shown as a contextual callout, not baked into individual vault grades.
Collateral Protocol Exposure
Each vault allocates deposits across multiple Morpho lending markets. We map each market's collateral asset to its underlying protocol risk score. The vault's base risk is the worst (highest) score across all collateral protocols — the weakest-link model. A vault lending against both wstETH (Lido, B) and sUSDe (Ethena, C-) inherits Ethena's risk as its floor.
Market Parameter Penalties
Lending market configuration adds risk beyond the underlying protocol. We score four parameters as additive penalties — aggressive settings increase the score, conservative settings are neutral:
| Parameter | Penalty | What it measures |
|---|---|---|
| LLTV | 0–8 | Liquidation buffer (higher LLTV = less buffer) |
| Oracle | 0–5 | Price feed reliability |
| Liquidity | 0–5 | Can collateral be liquidated quickly |
| Utilization | 0–4 | Market stretch / withdrawal risk |
Collateral-Loan Correlation
Not all high-LLTV markets carry the same risk. A 96% LLTV on wstETH/ETH (a derivative of the same underlying) is structurally safer than 86% LLTV on wstETH/USDC. We adjust the LLTV penalty by pair type:
| Pair Type | Multiplier | Example |
|---|---|---|
| Derivative | ×0.25 | wstETH/ETH, cbBTC/BTC |
| Stablecoin | ×0.5 | sUSDS/USDT, USDC/DAI |
| Standard | ×1.0 | wstETH/USDC, WBTC/USDT |
Synthetic stablecoins like sUSDe (Ethena) are classified as “standard” despite being USD-denominated, because their mechanism (delta-neutral perp positions) introduces depeg risk that traditional stablecoins do not carry.
Independent Ratings vs. Curator Self-Assessments
Vault curators like Steakhouse Financial publish their own risk ratings for the products they manage. These ratings are valuable, but they face a structural limitation: the curator is rating a product they profit from.
Hindenrank provides independent, third-party risk ratings. Our methodology differs from curator self-assessments in two ways: we measure the full cryptoeconomic mechanism risk of underlying protocols (consensus design, oracle dependencies, governance surface, track record), and we score each vault's market configuration (LLTV, oracle type, liquidity, utilization) as additive penalties.
Both perspectives are complementary. A curator's “A” rating reflects their view that credit risk is low. Our “B+” rating reflects that the underlying mechanisms carry complexity a credit-only framework does not capture. Together, they give a more complete picture.