Front-running how Hyperliquid’s stock perps set their price
Hyperliquid's stock perps don't use a live market price directly, they rebuild it from an oracle and a lagging formula. The idea was to exploit that lag three ways: the weekend catch-up jump, the split-second the perp trails the index during the day, and the gap between two versions of the same stock. We simulated all three from the venue's own published formulas. Even in the best case the edge was smaller than the fee to trade it, so every version lost money after costs. A one-hour simulation settled it, before any weeks of data collection.
Hyperliquid's stock perps don't take a live market price directly, they rebuild it from an oracle and a lagging formula. We tested the three ways that lag could theoretically be exploited, straight from the venue's published formulas. Every one is smaller than the fee to trade it. A one-hour desk simulation settled it, before any capital or weeks of data collection.
Track-0 feasibility gate, external feeds + a pure-mechanics simulator only (no live position, no pre-registration). Three angles, all killed sub-cost: (1) the weekend reopen-snap, gross 4.9 bps realistic / 38 bps at a perfect-direction ceiling, nets −2 to −8 bps; (2) the split-second the perp's mark trails the live index during market hours, 0.7 to 4.4 bps of mechanical lag that isn't actually tradeable, nets −2.6 to −8.6 bps; (3) the basis between two deployers' versions of the same stock, −4 to −12 bps against a ~26 bps two-leg cost. Only an unconfirmed 'growth-mode' fee (~2.2 bps round-trip) flips two of them marginally positive (+2.2 / +2.7 bps), and that is an upper bound that ignores the perp already pricing the move. The cheap test killed the effort at the gate instead of after weeks of forward accrual.
- Kill date
- 2026-06-16
- Sample
- 3 angles, external feeds + mechanics sim
- Method
- Documented kill
- Verdict
- cheap test, killed at the gate
Tested on the record and published in full, with the real numbers, whatever the result.
See all kills