Quick definition High-frequency trading is trading executed at extremely high speeds using sophisticated algorithms and low-latency infrastructure. HFT firms execute millions of trades per day, often holding positions for milliseconds. What it is HFT firms use custom hardware, proprietary data feeds, and co-located servers at exchanges to minimise latency. They exploit tiny price discrepancies that exist for only microseconds. Their algorithms identify and trade on these opportunities faster than any human could. An HFT firm might buy on one exchange at 100.00 and sell on another exchange at 100.01 in the same millisecond, locking in a tiny profit. Across millions of trades, these tiny profits add up. Why it matters HFT has transformed market microstructure. HFT firms provide liquidity, tightening spreads for all traders. But HFT also creates concerns about fairness and market stability. HFT profits come from speed advantages, which some argue are unfair to slower traders. HFT participation also raises concerns about flash crashes and market disruption. The 2010 Flash Crash was partly attributed to HFT algorithms cascading into a market downturn. HFT versus other trading HFT is distinct from regular algorithmic trading. A VWAP algorithm executes gradually over an hour. An HFT algorithm might execute millions of trades in that same hour. HFT also differs from longer-term algorithmic strategies like statistical arbitrage or machine learning models. Latency and infrastructure HFT success depends on latency (the time it takes for an order to reach the market and a response to return). Microseconds matter. HFT firms invest in: - Co-location (servers at exchange sites) - Custom network protocols - Specialised hardware - Proprietary data feeds This infrastructure investment creates barriers to entry and limits HFT to well-funded firms. Practical example An HFT algorithm identifies that AAPL is trading at 150.00 on NASDAQ and 150.02 on NYSE. The algorithm buys 1,000 shares on NASDAQ at 150.00 and sells 1,000 shares on NYSE at 150.02. The transaction completes in 10 milliseconds, locking in 0.02 dollars per share in profit. To a traditional trader, this price difference would not even be visible. HFT algorithms detect and exploit it. Regulation Regulators monitor HFT for market abuse. Algorithms must pass testing before deployment. Exchanges monitor for suspicious order patterns that might indicate manipulation. See also - Algorithmic Trading - Latency - Co-location - Statistical Arbitrage
