High-Frequency Trading (HFT) is the extreme form of algorithmic trading: trading frequency measured in milliseconds or microseconds, extremely short holding periods (typically no overnight positions), relying on ultra-low latency hardware and network infrastructure (co-location servers, fiber/microwave transmission) to receive price information or execute trades before other participants.
## Major HFT Strategy Types
**Electronic Market Making**: HFT firms simultaneously quote bid and ask prices, earning profits through the spread while providing market liquidity. Market maker risk: Inventory Risk (holding positions that move against them). Modern HFT market makers dynamically manage inventory risk through extremely high order placement/cancellation rates (tens of thousands per second). Academic research generally finds electronic market making has compressed bid-ask spreads, benefiting ordinary investors.
**Latency Arbitrage**: exploiting price desynchronization between different exchanges — detecting a price change at one exchange before other participants and “front-running” the trade at another. This is the core of “Flash Boys” criticism — IEX Exchange introduced an artificial 350-microsecond delay (Speed Bump) to eliminate latency arbitrage opportunities.
**Statistical Arbitrage at HFT Speed**: identifying and trading short-term price deviations between correlated assets within millisecond timeframes.
## Regulation and Controversy
China’s CSRC has relatively strict HFT regulation: the A-share T+1 system (shares bought today can only be sold tomorrow), price limit bands (10%/20%), and quant fund registration requirements limit certain HFT strategies’ effectiveness in A-shares — making China’s HFT landscape distinctly different from US markets.
See [Algorithmic Trading Basics](https://sunqi.org/algorithmic-trading-basics-en/) and [Flash Boys — Michael Lewis](https://en.wikipedia.org/wiki/Flash_Boys).
