Trading in the Sunshine or in the Shade: Market Impact and Adverse Selection on Hyperliquid

Battle Count: 74

Highly relevant to quantitative trading. The paper provides direct empirical evidence on execution costs under different transparency regimes, quantifies the announcer discount (~9 bps lower temporary impact for TWAPs), measures permanent impact differences (~5 bps), and documents how adverse selection shifts to non-announcers. The findings on execution schedules (U-shaped for hidden vs. uniform for TWAP), impact trajectories, and order-book responses are directly actionable for algorithmic execution strategy design. The DeFi/Hyperliquid setting is increasingly relevant as crypto perpetual futures grow in institutional adoption.

The empirical methodology involves multiple layers: metaorder reconstruction from raw fills, square-root impact law estimation, separable power-law surface fitting, event-time placebo regressions with fixed effects, side-decomposed dominance measures, and propagator model benchmarks. The data processing pipeline (641M fills, address-level attribution, 30-minute splitting hypothesis) is computationally intensive. The econometric specifications are standard but numerous. Replication requires access to Hyperliquid on-chain data and careful implementation of the metaorder identification algorithm.

The paper uses publicly available Hyperliquid data via Hydromancer's Reservoir. The methodology is well-documented with explicit parameter definitions, sample construction rules, and regression specifications. However, no GitHub repository is mentioned, and the statistical metaorder reconstruction algorithm (splitting hypothesis with 30-minute inter-arrival threshold) would need careful replication. The order-book data availability window (Dec 15, 2025 onward) is specified. The paper provides detailed descriptive statistics and fitted parameters.