Modeling Perpetual Futures Autodeleveraging as Online Learning

Summary

This paper models autodeleveraging, a last-resort mechanism used by perpetual futures venues when liquidations and insurance funds cannot cover a solvency shortfall. At each round, the venue chooses a solvency budget and profitable accounts whose gains can be reduced to help restore solvency. In the proposed framework, haircuts apply to positive unrealized profit rather than posted collateral.

The authors frame these choices as online learning over possible haircut allocations and derive robustness guarantees and upper bounds on regret. For the October 2025 Hyperliquid stress episode, they compare the production queue with algorithms inspired by the framework; the reported counterfactuals show lower overshoot from the best alternative. These results illustrate potential improvements for this event, but do not establish performance across other stress episodes or venues. The excerpt provides no details on the data or assumptions used to calibrate the bounds.

Key ideas

• Autodeleveraging is used when liquidation and insurance buffers cannot restore venue solvency.
• The model treats haircut allocation across profitable accounts as a sequential online learning problem.
• In the framework, haircuts reduce positive unrealized profit rather than posted collateral principal.
• The paper derives robustness results and regret bounds for solvency recovery mechanisms.
• A case study of the October 2025 Hyperliquid stress episode compares the production queue with counterfactual algorithms.

Tags

# Autodeleveraging as Online Learning


# Autodeleveraging as Online Learning









Autodeleveraging (ADL) is a last-resort loss socialization mechanism used by perpetual futures venues when liquidation and insurance buffers are insufficient to restore solvency. Despite the scale of perpetual futures markets, ADL has received limited formal treatment as a sequential control problem. This paper provides a concise formalization of ADL as online learning on a PNL-haircut domain: at each round, the venue selects a solvency budget and a set of profitable trader accounts. The profitable accounts are liquidated to cover shortfalls up to the solvency budget, with the aim of recovering exchange-wide solvency. In this model, ADL haircuts apply to positive PNL (unrealized gains), not to posted collateral principal. Using our online learning model, we provide robustness results and theoretical upper bounds on how poorly a mechanism can perform at recovering solvency. We apply our model to the October 10, 2025 Hyperliquid stress episode. The regret caused by Hyperliquid's production ADL queue is about 50\% of an upper bound on regret, calibrated to this event, while our optimized algorithm achieves about 2.6\% of the same bound. In dollar terms, the production ADL model over liquidates trader profits by up to \$51.7M. We also counterfactually evaluated algorithms inspired by our online learning framework that perform better and found that the best algorithm reduces overshoot to \$3M. Our results provide simple, implementable mechanisms for improving ADL in live perpetuals exchanges.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.