What Is High-Frequency Trading (HFT)? Basics, Mechanics and Example
By Forex & Brokers DeskUpdated Nov 10, 2026
Definition
High-frequency trading is automated trading in which algorithms send, modify and cancel huge numbers of orders in fractions of a second, holding positions for seconds or less to capture tiny price differences.
Key takeaways
- HFT is algorithmic trading defined by speed, volume and holding periods of seconds or less.
- It earns tiny amounts per trade from spreads, rebates and brief price gaps, not from forecasting market direction.
- Well-known HFT firms include Tower Research Capital, Citadel and Virtu Financial.
- HFT operates in equities, futures, forex and cryptocurrency markets.
High-frequency trading (HFT) is a form of automated trading in which computer algorithms send, modify and cancel very large numbers of orders within fractions of a second, holding positions for seconds or less. HFT firms try to capture very small price differences many times a day. Their activity shows up as very high turnover and order-to-trade ratios, with far more orders placed and canceled than filled.
How High-Frequency Trading Works
An HFT system has three parts: fast data, fast decisions and fast execution.
- Market data. The firm takes direct data feeds from exchanges instead of slower consolidated feeds, so it sees quote changes as early as possible.
- Algorithms. Programs read the order book, recent trades and prices on related instruments, then decide in microseconds whether to quote, trade or cancel.
- Execution infrastructure. Firms place their servers in or next to exchange data centers (co-location) and use dedicated fiber or microwave links between trading venues. This cuts the time a signal takes to travel.
A high-frequency trade can be executed in about 10 milliseconds, and the fastest systems work well below that. At these speeds no human approves individual trades. People design the strategy, set the risk limits and monitor the system, but the software trades.
Because positions are held so briefly, HFT firms usually end the day with little or no net exposure. Their risk comes from inventory picked up between trades and from software or connectivity failures.
Common HFT Strategies
- Electronic market making. The algorithm posts both a bid and an offer and earns the spread when both sides fill, often collecting an exchange rebate for adding liquidity.
- Arbitrage. The system exploits brief price gaps for the same or linked instruments, such as a stock on two exchanges or an index future against its underlying shares.
- Statistical arbitrage. Models trade short-lived deviations between instruments that historically move together.
- Latency arbitrage. The firm reacts to a price change on one venue before slower participants update their quotes elsewhere.
- Event-driven trading. Algorithms parse economic releases or headlines and trade in the first moments after publication.
Example of High-Frequency Trading
Hypothetical example; the prices, rebate and volume are illustrative.
A stock is quoted at $50.00 bid and $50.01 ask. An HFT market maker posts a bid at $50.00 and an offer at $50.01. A seller hits the bid, and moments later a buyer lifts the offer.
- Spread captured: $0.01 per share
- Assumed liquidity rebate: $0.002 per share on each side, so $0.004 for the round trip
- Gross result per share: $0.014
On 100 shares that is $1.40. If the strategy completes this cycle on 5 million shares in a day, the gross result is $70,000, before exchange fees, data costs, technology spending and losses.
If the price drops to $49.98 after the bid fills but before the offer fills, the firm holds stock bought at $50.00 that is now worth less, and the algorithm must decide within milliseconds whether to exit at a loss. A one-cent loss erases the profit from about two completed round trips ($0.028), so the model must avoid being picked off by better-informed or faster traders.
Why Exchanges Encourage HFT
Exchanges compete for order flow, and liquidity attracts it. Many venues pay rebates to firms that post resting orders, and HFT grew quickly once these incentives spread. The New York Stock Exchange’s Supplemental Liquidity Provider (SLP) program, introduced after the 2008 collapse of Lehman Brothers when liquidity was a serious concern, pays participants for adding competitive quotes. Across millions of daily transactions, these payments can be a meaningful share of an HFT firm’s revenue.
Benefits of High-Frequency Trading
- Liquidity. HFT market makers quote continuously, so other traders more often find a counterparty quickly.
- Narrower spreads. Competition among fast quoters compresses bid-ask spreads. Research on Canadian markets found that after regulators introduced fees that raised HFT costs, market-wide bid-ask spreads widened by 13% and retail spreads by 9%.
- Price efficiency. Arbitrage strategies quickly close price gaps between venues and related instruments.
Limitations and Risks of High-Frequency Trading
- Market disruptions. On May 6, 2010, the Dow Jones Industrial Average fell nearly 1,000 points within minutes and then largely recovered. Regulators found that the “Flash Crash” was triggered by a very large automated sell order. Fast trading amplified the moves as liquidity pulled back.
- Ghost liquidity. Quotes can be canceled in microseconds, so displayed depth may vanish before a slower trader’s order arrives.
- Uneven playing field. Only large firms can afford co-location, direct feeds and private networks. Critics argue this lets them profit at slower participants’ expense.
- Loss of human judgment. A faulty model or bad data can produce large losses or erratic prices before anyone can intervene.
Regulators have responded with rules aimed at automated trading. In the US, SEC Rule 15c3-5 requires brokers with market access to apply pre-trade risk controls. In the EU, MiFID II defines high-frequency algorithmic trading and requires firms that engage in it to be authorized and to keep records of their orders.
HFT vs. Algorithmic Trading vs. Scalping
| Term | What defines it | Typical holding period |
|---|---|---|
| Algorithmic trading | Any trading in which rules coded into software generate or execute orders | Seconds to months |
| High-frequency trading | Algorithmic trading built on ultra-low latency and very high order volume | Microseconds to seconds |
| Scalping | A short-term style aimed at small moves, often done manually | Seconds to minutes |
All HFT is algorithmic, but most algorithmic trading is not HFT. An algorithm that splits a pension fund’s large order over a day is automated but not high-frequency. A scalper clicking through trades on a retail platform is trading frequently, but at human speed.
HFT in Forex and Crypto
In forex, HFT firms trade on electronic interbank venues and ECNs. They make markets in major pairs and arbitrage between platforms. In cryptocurrency, the same methods run on exchanges that trade around the clock, often between spot and derivatives markets or across exchanges quoting pairs like BTC/USD at slightly different prices.
Can Individual Traders Use HFT?
In practice, no. Retail orders pass through a broker’s systems and the public internet, adding latency many times greater than a co-located server’s. Retail spreads and commissions also exceed the fraction of a cent per share that HFT relies on, an edge that only works at institutional cost levels.
The Bottom Line
HFT matters to anyone trading equities, futures, forex or crypto, because fast firms supply much of the quoted liquidity and help keep spreads narrow. Individual traders cannot realistically compete, since broker routing and the public internet add far more latency than co-located servers, and retail costs exceed the fraction of a cent HFT earns. Slower traders should expect displayed depth to vanish quickly in volatile moments.
