What Is High-Frequency Trading (HFT)?
High-frequency trading uses powerful computers to execute thousands of trades per second, exploiting tiny price discrepancies across markets in microseconds.
What Is High-Frequency Trading (HFT)?
High-frequency trading (HFT) is a subset of algorithmic trading that uses powerful computers and low-latency infrastructure to execute thousands of trades per second. HFT firms compete to be fastest to act on tiny price discrepancies, holding positions for fractions of a second and closing flat by the close.
| Feature | HFT |
|---|---|
| Hold time | Microseconds to seconds |
| Trades per day | Thousands to millions |
| Profit per trade | A fraction of a cent |
| Infrastructure | Co-located servers, FPGA hardware, microwave links |
| Overnight risk | None (always flat at close) |
| Participants | Institutional firms, not retail |
What HFT firms do
- Market making — Continuously quote both sides, earning the spread thousands of times per second.
- Arbitrage — Exploit tiny gaps across exchanges, futures vs. spot, or ETFs vs. baskets.
- Statistical arbitrage — Bet on rapid reversion of correlated instruments.
- Latency arbitrage — React to price changes on one venue before slower participants can on another (controversial).
- Event arbitrage — Ingest news in milliseconds and trade before humans react.
Why speed matters
At HFT's scale, being 1 microsecond faster can capture an entire opportunity. Firms invest in co-location (servers inside the exchange), microwave/laser links (faster than fiber), FPGAs (hardware-coded logic), and specialized data feeds. The race is now measured in nanoseconds.
The 2010 flash crash
On May 6, 2010, the Dow dropped nearly 1,000 points in minutes before recovering. HFT algorithms weren't the sole cause, but they amplified the cascade: as liquidity providers withdrew, prices fell further, triggering more withdrawal — exposing how automated systems can interact to create instability.
Pros and cons for the market
| Pros | Cons |
|---|---|
| Adds liquidity | Latency arbitrage taxes slower participants |
| Tighter spreads | Algorithms can amplify flash crashes |
| Faster price efficiency | "Phantom liquidity" vanishes in chop |
| Lower retail trading costs | Favors those who can pay for infrastructure |
Why retail can't compete
- Cost. Co-location, fast data, and custom hardware cost millions.
- Latency. A home connection is hundreds of milliseconds too slow.
- Access. Retail brokers don't offer DMA or low enough fees.
- Capital. Profit per trade is tiny; meaningful returns require huge notional.
How HFT affects everyday traders
HFT tightens your spreads and speeds your fills — both good. But it also prices in news before you can read it and exploits predictable retail patterns. The lesson: avoid predictable patterns. Obvious stop clusters, time-of-day trades, and momentum-chasing at the open all get picked off by faster participants.
Bottom line
HFT is an infrastructure-driven corner of the market where firms compete on microseconds for fractions of a cent. Retail traders cannot and should not try to win on speed. Compete on strategy and discipline instead — the speed race is one you cannot win.