High-Frequency Trading
A class of latency-sensitive automated trading strategies that profit from microstructure phenomena — passive market-making, statistical arbitrage at sub-second horizons, and liquidity-taking signals — typically holding inventory for milliseconds to minutes.
Details
- aliases: HFT
- category: quantitative
- investment thesis: Speed, inventory management, and adverse-selection control let market-makers earn the bid-ask spread net of expected adverse selection.
- edge source: Co-location, optimized execution, microstructure modeling, and queue-position management.
- typical holding period: Microseconds to minutes
- drawdown profile: Very small drawdowns, but capacity-limited and prone to flash-crash episodes (May 2010).
- correlation to equities: 0.0
- correlation to bonds: 0.0
- liquidity profile: daily
- instruments used: equities, futures, options, FX, fixed income
- asset classes: multi-asset
- primary risks: technology failure, regulatory intervention (e.g., FSB, SEC Reg AT), queue-position decay, predator/prey dynamics with informed traders
- key metrics: Sharpe ratio, P&L per share, cancel ratio, queue-fill rate, adverse selection
- notable practitioners: Citadel Securities, Virtu Financial, Jane Street, Jump Trading, DRW
- exemplar funds: citadel-securities, virtu, jane-street, jump-trading
- related terms: market-making, latency, co-location, adverse-selection, tick-size
- related strategies: statistical-arbitrage, market-making
- academic foundations: Glosten-Milgrom (1985), Kyle (1985) — Continuous Auctions and Insider Trading, Ho-Stoll (1981)