Hyperliquid Vault Risk Score:
Exact Formula, Bands, and How to Use It

Stop chasing APR. Start reading risk like an operator.

Most people pick vaults the same way: open leaderboard, sort by APR, ape in.

Works great until it does not.

One regime change later, that “genius” vault is down double digits and everyone acts surprised. This is exactly why we built a deterministic Risk Score: fast triage first, deeper diagnostics second.

TL;DR

Why this exists

APR is a rear-view mirror metric. Risk is what keeps you alive.

A vault can print a beautiful 30-day number while silently loading up leverage, concentration, and bad liquidity conditions. By the time APR rolls over, you are late.

So the workflow is simple:

1. Use Risk Score to narrow the field quickly.
2. Use diagnostics to avoid getting farmed by a pretty chart.

The exact formula

• Range is clamped to 0-100 (higher = riskier)
• Band logic: 0-33 Low, 34-66 Moderate, 67-100 High

Component 1: Performance Risk (35%)

Inputs:

• volatility_30d (std dev of daily returns)
• worst_day_30d (largest daily downside)

Interpretation: this is path quality, not just endpoint return.

Component 2: Drawdown Risk (25%)

Input: max_drawdown_30d (peak-to-trough decline)

APR is the pitch. Drawdown is the bill.

Component 3: Liquidity Risk (25%)

Inputs:

• tvl_usd
• tvl_volatility_30d

Small plus unstable TVL is where exits get painful.

Component 4: Data Confidence Risk (15%)

Inputs:

• quality_label = real | derived | simulated | demo
• data_points_30d

Less reliable data should carry a penalty. Always.

Missing-data defaults (conservative on purpose)

If input is missing, we do not silently pretend everything is fine. Deterministic defaults are applied:

• missing volatility -> 50
• missing worst day -> 50
• missing max drawdown -> 60
• missing TVL -> 75
• missing TVL volatility -> 50
• missing quality label -> 45
• missing data points -> 55

How to use it in practice

Step 1: Triage fast

Filter by Risk Score band first. This takes you from timeline noise to a shortlist.

Step 2: Diagnose deeply

On shortlisted vaults, check diagnostics:

• Sortino / Calmar
• Beta / correlation vs BTC
• Stress scenarios
• Brinson-Fachler attribution

Step 3: Size with intent

• Good score + good diagnostics -> normal or larger size
• Good score + mixed diagnostics -> reduced size
• Mid score + strong diagnostics -> selective bet
• Bad score + bad diagnostics -> easy skip

Score picks candidates. Diagnostics decide size.

Why diagnostics are not useless extras

Sortino / Calmar

Two vaults can share the same risk band while one has much cleaner downside behavior. If Sortino and Calmar are weak, the equity curve is usually more fragile than APR suggests.

Beta / Corr vs BTC

This tells you if you really diversified or just bought disguised beta with extra steps.

Stress test

Shock scenarios expose hidden posture. A vault can look stable in baseline data and still have ugly sensitivity under simple market shocks.

Brinson-Fachler attribution

It splits active return into allocation (sizing), selection (direction), and interaction. If gains come mostly from allocation while selection stays negative, edge is thinner than it looks.

Real examples

Example 1: Same score, different trade

Both vaults: Risk Score 45.

• Vault A: Sortino 2.1, Calmar 1.8, beta 0.28, corr 0.10
• Vault B: Sortino 0.6, Calmar 0.4, beta 1.35, corr 0.72

Same band, different decision. A can be hold/add. B is smaller size or pass.

Example 2: Stable score, ugly stress profile

Vault C: Risk Score 39 looks clean. Stress scenarios still show outsized downside in a broad market move. Translation: hidden positioning risk. Do not oversize it just because the score is green-ish.

Example 3: Attribution shows where edge really comes from

30d active return is positive, but breakdown is allocation-heavy with negative selection. That usually means regime-sensitive edge. Useful, but not something to blindly max-size.

Common mistakes

1. Treating Risk Score like a buy button.
2. Ignoring diagnostics after shortlist.
3. Overweighting APR and underweighting drawdown structure.
4. Reading one 30d window like gospel.
5. Going full degen size on one hot streak.

Final take

Risk Score is simple by design. That is the feature.

• It helps reject weak setups quickly.
• It gives a consistent risk language across vaults.
• It keeps decisions grounded when timeline hype gets loud.

Use Risk Score to find candidates. Use diagnostics to avoid landmines. Then size like you plan to be here next cycle too.

Risk Score live, across bands

Here is how the three bands look on real Hyperliquid vaults right now. Click through to see full composition, drawdown chart, and attribution:

• Low band (0-33): FC Genesis Quantum — typically scores around 18/100. Low volatility, small TVL absorbs liquidity penalty, clean data.
• Moderate (34-66): HLP at ~64/100 (size earns TVL-size score, but tail exposure from market-making drags performance component), Orbit Value Strategies at ~41/100.
• High (67+): BredoStrategy and Long LINK Short XRP sit in the 70s — aggressive volatility + small TVL + directional exposure.

The pattern is consistent: low-risk-band vaults trade path quality for modest APR, high-band vaults flip that trade. Your job is not to avoid one band, it is to pick the right band for your sizing.

Related Reading

• How to Choose a Hyperliquid Vault: 5-Factor Framework
• Top Hyperliquid Vaults Ranked by Risk-Adjusted Return
• How Hyperliquidity Provider (HLP) Works: Mechanics & Real Numbers