Slippage-at-Risk (SaR): A Forward-Looking Liquidity Risk Framework for Perpetual Futures Exchanges

Imagine a massive, high-speed casino where people are betting on the future price of digital assets like Bitcoin. This casino is called a Perpetual Futures Exchange. Unlike a normal casino where you bet on a single hand of cards, here, people can hold their bets open forever, using borrowed money (leverage) to multiply their potential wins—or losses.

The paper you provided introduces a new way to measure how safe this casino is, called Slippage-at-Risk (SaR).

Here is the breakdown in simple terms, using everyday analogies.

1. The Problem: The "Thin Ice" Problem

In this casino, if a player loses too much money, the house has to force them to sell their bet immediately to stop the bleeding. This is called a liquidation.

The Old Way (Looking in the Rearview Mirror): Traditional risk managers look at history. They ask, "How many people lost money last year?" It's like driving a car while only looking in the rearview mirror. If you see a crash behind you, it's too late to avoid the cliff ahead.
The New Way (Looking Through the Windshield): The author, Otar Sepper, says we need to look at the Order Book. Think of the order book as the "depth of the pool" where players place their bets.
- If the pool is deep (lots of buyers and sellers), a big fish can swim through without splashing.
- If the pool is shallow (thin liquidity), a big fish swimming through creates a massive wave that crashes everything.

SaR measures how shallow that pool is right now, before anyone actually falls in.

2. The Three Key Metrics (The Dashboard)

The paper proposes three specific numbers to put on the casino manager's dashboard:

SaR (Slippage-at-Risk): Imagine you have to sell a giant bag of gold coins instantly.
- Normal day: You sell them, and the price drops a tiny bit (1%).
- Stress day: You try to sell, but there are no buyers, so you have to slash the price to get rid of them (10% drop).
- SaR tells you: "95% of the time, the price drop will be under 3%. But 5% of the time, it could be a disaster." It's a warning light for the worst-case scenarios.
ESaR (Expected Slippage-at-Risk): This answers the question: "If we do hit that disaster scenario, how bad will it actually be?" It's not just the warning light; it's the estimate of the damage.
TSaR (Total Dollar Slippage-at-Risk): This puts a price tag on the disaster. "If everything goes wrong, we will lose $127 million in value." This helps the casino know exactly how much cash they need to keep in the safe.

3. The "Fragile Liquidity" Trap (The Concentration Adjustment)

This is a crucial part of the paper. Imagine two swimming pools that both hold 1,000 gallons of water.

Pool A: 1,000 people are holding buckets, each contributing 1 gallon.
Pool B: One guy is holding a giant hose, contributing all 1,000 gallons.

Both pools look deep. But if the guy with the hose leaves (or gets sick), Pool B instantly dries up. Pool A is safe because no single person matters much.

The paper introduces a "Concentration Haircut." It penalizes pools where one or two people hold all the water. If the casino sees that one "whale" is providing 50% of the liquidity, the SaR number goes up, warning that the pool is actually very fragile, even if it looks deep.

4. The "Insurance Fund" Connection

When a player gets liquidated and the price crashes so hard that the house loses money, the casino uses its Insurance Fund (a safety piggy bank) to cover the loss.

The Mistake: Many casinos keep a tiny piggy bank because they haven't had a big crash yet.
The SaR Solution: The paper says, "Don't wait for a crash. Look at the pool depth now."
- If the pool is shallow and concentrated, the SaR calculation says: "You need a piggy bank of $300 million, not $25 million."
- In the paper's case study (the October 10, 2025 event), the casino had $25 million in the bank. The SaR metric predicted they needed $312 million. Because they didn't listen, they had to force other winning players to pay the difference (a process called "socializing losses"), which caused panic.

5. The Real-World Test: The October 10, 2025 Cascade

The paper analyzes a real (simulated future) event where the market crashed.

Before the crash: The SaR metric started rising 12–24 hours before the crash. It saw the "water level" dropping and the "whales" getting nervous.
During the crash: The actual losses matched the SaR prediction almost perfectly.
The Lesson: If the casino managers had watched the SaR dashboard, they would have seen the warning signs, increased their safety fund, and lowered the betting limits before the disaster happened.

Summary: Why This Matters

Think of Slippage-at-Risk as a weather radar for financial markets.

Old Risk Management: "It hasn't rained in 10 years, so we don't need an umbrella."
SaR: "Look at the clouds right now. The pressure is dropping, the wind is changing, and the ground is getting dry. A storm is coming in 6 hours. Put up the umbrella now."

It shifts the focus from "What happened?" to "What is about to happen?" By measuring the actual structure of the market (who is providing the money and how deep the pool is), it prevents the casino from running out of money when the storm hits.

Here is a detailed technical summary of the paper "Slippage-at-Risk (SaR): A Forward-Looking Liquidity Risk Framework for Perpetual Futures Exchanges" by Otar Sepper.

1. Problem Statement

Perpetual futures exchanges face systemic risks arising from the interaction between high leverage, liquidation mechanisms, and order book liquidity. When leveraged positions are liquidated during market stress, they consume available order book depth. If the order book is too thin, execution prices deviate significantly from the theoretical bankruptcy price, generating bad debt (deficits).

Existing risk management frameworks suffer from critical limitations:

Backward-Looking: Traditional metrics like Value-at-Risk (VaR) rely on historical returns or realized deficit distributions, failing to capture current market capacity.
Blind Spots for New Assets: Historical metrics cannot assess risk for newly listed tokens lacking crisis history.
Static Calibration: Leverage caps and position limits are rarely adjusted dynamically based on real-time liquidity conditions.
Ignored Microstructure: The actual resource determining execution quality—the order book—is rarely incorporated into formal risk models.

The paper cites the October 10, 2025 Hyperliquid event as a case study where $2.1 billion in liquidations caused $304.5 million in deficits, overwhelming an insurance fund that was undercapitalized by an order of magnitude.

2. Methodology: The SaR Framework

The authors propose Slippage-at-Risk (SaR), a forward-looking framework that quantifies liquidation execution risk directly from the current order book microstructure.

A. Core Metrics

The framework defines three complementary metrics based on a calibrated stress notional $Q_i$ (typically a percentage of Open Interest):

SaR( $\alpha$ ): The cross-sectional $\alpha$ -quantile of slippage across all tokens. It answers: "What is the maximum slippage 95% of tokens will experience during a stress event?"
ESaR( $\alpha$ ): The Expected Slippage at Risk. The conditional expected slippage for tokens in the tail (those exceeding SaR). This measures the severity of the worst-case scenarios.
TSaR( $\alpha$ ): Total Dollar Slippage at Risk. The aggregate dollar-denominated slippage for all tail tokens. This maps directly to potential deficit exposure.

B. Concentration Adjustment

Raw slippage calculations often ignore the structure of liquidity. Two order books with identical depth can have vastly different fragility if one is dominated by a single market maker.

Metric: The framework uses the Herfindahl-Hirschman Index (HHI) and the Effective Number of Providers ( $N_{eff}$ ) to measure concentration.
Haircut Formula: A concentration haircut ( $h_{conc}$ ) is applied to raw slippage. It penalizes books with low $N_{eff}$ or a single dominant provider (high CR1), simulating the impact of a liquidity provider withdrawing.
Adjusted Slippage: $S_{adj} = S_{raw} \times (1 + h_{conc})$ . All SaR metrics are computed using these adjusted values.

C. Theoretical Connections

The paper establishes a causal chain: Liquidity Withdrawal $\to$ Thin Order Books $\to$ High Slippage $\to$ Bad Debt $\to$ Autodeleveraging (ADL).

Insurance Fund Sizing: By linking slippage to deficits, the authors derive a formula for the optimal insurance fund ( $IF^*$ ):
$IF^* \approx c \cdot TSaR_{adj}^{\$}(\frac{r}{\kappa})$
Where $c$ is a multiplier (empirically 2.0–2.5) accounting for cascade dynamics, and $r/\kappa$ represents the ratio of opportunity cost to reputational cost. This allows exchanges to size capital reserves based on current liquidity rather than past deficits.

3. Key Contributions

Forward-Looking Risk Metrics: Introduced SaR, ESaR, and TSaR, which are computable for any asset with an order book, regardless of historical crisis data.
Concentration-Aware Modeling: Formalized the fragility of concentrated liquidity using HHI-based haircuts, addressing the risk of "ghost liquidity" (spoofing) and correlated withdrawals.
Dynamic Capital Requirements: Derived a direct mapping from SaR metrics to optimal insurance fund sizing, enabling proactive capital management.
Empirical Validation: Validated the framework using Hyperliquid data, demonstrating that SaR acts as a leading indicator for systemic stress events.

4. Empirical Results

The authors analyzed 26 days of order book data from Hyperliquid (Oct 9 – Nov 3, 2025), including the October 10 liquidation cascade.

Predictive Validity:
- Lead-Lag Correlation: TSaR showed a correlation of 0.61 with realized deficits 12 hours in advance and 0.42 at 24 hours.
- Granger Causality: Statistical tests confirmed that TSaR Granger-causes deficits ( $p < 0.001$ ), proving it is a genuine leading indicator, not just a coincident measure.
Case Study (Oct 10, 2025):
- Pre-Event Warning: 24 hours before the cascade, SaR metrics rose by 30%, and total exchange depth collapsed by 75%.
- Prediction Accuracy: A regression of pre-event adjusted slippage against realized slippage yielded an $R^2$ of 0.78.
- Insurance Fund Gap: The SaR-implied insurance fund (calculated pre-event) was $312.6M, closely matching the realized deficit of $304.5M. The actual exchange fund was only ~$25M, highlighting a massive undercapitalization detectable via SaR.
Concentration Dynamics: During the cascade peak, 73% of tokens had an effective provider count ( $N_{eff}$ ) of less than 3, confirming that concentration amplified slippage far beyond raw depth metrics.

5. Significance and Implications

Proactive Risk Management: SaR shifts the paradigm from reacting to realized losses to anticipating liquidity failures. It allows exchanges to tighten leverage limits or halt new positions before a cascade occurs.
New Asset Listing: Provides a quantitative risk assessment for newly listed tokens that lack historical deficit data, preventing the listing of inherently fragile assets.
Decentralized Finance (DeFi) Governance: In systems where "code is law" and manual intervention is difficult, SaR offers an objective, algorithmic basis for adjusting parameters (e.g., insurance fund requirements, leverage caps) dynamically.
Systemic Stability: By identifying "tail tokens" and concentrated liquidity structures, regulators and exchange operators can target specific vulnerabilities before they trigger system-wide failures.

Conclusion

The paper argues that liquidity is the "lifeblood" of perpetual futures markets. By measuring the cause (order book depth and structure) rather than the consequence (realized deficits), the SaR framework provides a robust, forward-looking tool for preventing systemic collapses. The empirical evidence from the October 2025 event suggests that implementing SaR could have prevented the massive undercapitalization and socialization of losses seen in that crisis.