Glossary index › H

High-Frequency Trading (HFT)

High-frequency trading (HFT) is a form of algorithmic trading in which firms use powerful computers, custom software, and extremely fast data connections to submit and cancel large numbers of orders in fractions of a second, aiming to profit from tiny, fleeting price differences or from providing liquidity many thousands of times per day. Where a typical retail trader might place a handful of trades in a session, an HFT firm's systems can place and cancel thousands of orders in the time it takes a human to blink.

How HFT strategies actually work

Most HFT activity falls into a small number of recognizable strategy types. Market making involves continuously quoting both a buy and a sell price for a stock, capturing the small spread between them across enormous volume - profiting a fraction of a cent per share, but doing so millions of times a day. Latency arbitrage exploits the fact that the same piece of information (a price change on one exchange, say) reaches different trading venues at very slightly different times; a firm whose systems process and act on that information microseconds faster than competitors can trade ahead of the price adjusting everywhere else. Statistical arbitrage uses fast execution to capture small, short-lived pricing anomalies between related securities, such as an ETF and its underlying basket of stocks momentarily drifting out of alignment.

Speed itself is the product being competed for, which is why HFT firms pay substantial sums for "colocation" - physically placing their servers inside the same data center as an exchange's matching engine to shave fractions of a millisecond off data travel time - and, in some well-documented cases, built dedicated microwave transmission towers between financial centers like Chicago and New York because microwave signals travel through air faster than light travels through the fiber-optic cables most data uses. In this world, an advantage measured in single-digit microseconds (millionths of a second) can be worth real, sustained profit.

Origins and the 2010 Flash Crash

HFT grew rapidly through the 2000s as US markets moved to decimal pricing (2001) and regulatory changes, notably Regulation NMS (2005), fragmented trading across many competing electronic exchanges rather than one central floor - conditions that rewarded firms able to trade fastest across all of them simultaneously. The strategy's risks became visible to the public on May 6, 2010, the "Flash Crash," when the Dow Jones Industrial Average plunged about 1,000 points - roughly 9% - in a matter of minutes, before recovering most of the loss within the same session. Regulatory investigations found the crash was triggered by a large automated sell order interacting badly with HFT algorithms that rapidly withdrew liquidity from the market as prices fell, amplifying a routine imbalance into a brief but violent market-wide plunge. The episode led directly to the introduction of market-wide circuit breakers and new rules requiring exchanges to pause trading in individual stocks during extreme, sudden price swings.

The attribution of that day has never fully settled, which is part of its lasting interest. Five years afterwards, a trader working from his parents' house in west London was arrested and later pleaded guilty to spoofing the futures market that morning, placing and cancelling large orders he never intended to fill, and prosecutors described his activity as a contributing factor. Others argued that one participant could not have caused a market-wide collapse and that the episode exposed a structural fragility instead: when the firms supplying most of the quotes withdraw at the same moment, there is nothing underneath the price. Both readings survive, and the rules written afterwards address the second rather than the first.

Cross-market presence and why it matters to other traders

HFT is most dominant in US equities, where various industry estimates have put its share of total trading volume at roughly half of all activity in recent years, but the same speed-based competition exists, in different forms, across other markets. In forex, HFT firms compete on speed in the same way, particularly in exploiting momentary price discrepancies between different currency trading venues around the world. In futures and commodities markets, HFT market makers are active participants providing liquidity in contracts like crude oil and gold futures. Crypto markets have also developed a significant HFT presence on major exchanges, though fragmented liquidity and less mature market-surveillance infrastructure have made crypto more prone to its own version of flash-crash-style events.

For an ordinary retail trader, HFT's practical relevance is mostly indirect: on one hand, HFT market-making has generally tightened bid-ask spreads and lowered the visible cost of trading liquid stocks and ETFs compared to the pre-electronic era; on the other hand, it means retail orders are competing in a market where institutional participants can react to news and price moves far faster than any human, which is part of why "front-running" concerns and payment-for-order-flow debates remain persistent controversies in market-structure regulation.

The part of this that touches an ordinary trader most directly is where their order goes. In the United States a retail order usually never reaches an exchange at all: the broker routes it to a wholesaling firm that fills it from its own book, typically at a price fractionally better than the public quote, and pays the broker for the flow. The arrangement funds commission-free trading and gives the wholesaler a stream of orders it knows to be uninformed, which is exactly the flow a market maker wants. Whether that is a good bargain for the customer has been argued over for years and has survived several regulatory reviews.

The other response has been structural. One American exchange built a deliberate delay of a few hundred microseconds into every incoming order, so that quotes cannot be picked off by a firm reacting to a price change elsewhere fractionally sooner, and several venues abroad replaced continuous trading with frequent auctions for the same reason. Both designs accept a small loss of immediacy to remove the value of being a microsecond faster. They exist because the alternative is an arms race whose cost, in towers, cables and engineering, is ultimately paid out of the spread that everyone else trades against.

What is high-frequency trading? Algorithmic trading that submits and cancels very large numbers of orders in fractions of a second, profiting from fleeting price differences or from quoting continuously across enormous volume.
What are the main strategies? Market making, which quotes both sides and earns the spread millions of times a day; latency arbitrage, which acts on information microseconds before it reaches other venues; and statistical arbitrage between related instruments that drift out of line.
Why is speed worth so much? Because the same trade is only profitable for whoever gets there first. Firms pay to place servers in the exchange's own data centre and have built microwave links between cities, since a signal moves faster through air than through fibre.
How did it grow? Through decimal pricing in 2001 and the American market-structure rules of 2005, which spread trading across many competing electronic venues and rewarded whoever could act on all of them at once.
What was the flash crash? On 6 May 2010 the Dow fell about 1,000 points in minutes and recovered most of it the same session. A large automated sell order met algorithms that pulled their quotes as prices fell, turning an imbalance into a collapse.
What changed afterwards? Market-wide circuit breakers and rules requiring exchanges to pause an individual share during extreme sudden moves.
How big is it? Industry estimates put it at roughly half of US share trading volume. The same competition exists in currencies, in futures including oil and gold, and increasingly on crypto exchanges.
Is it good or bad for a retail trader? Both, and the honest answer is mostly indirect. Quoting competition has tightened spreads and cut the visible cost of trading liquid instruments, while the same participants react to news faster than any person can.
Where does my order actually go? In the US, a retail order usually never reaches an exchange: the broker sells the flow to a wholesaler that fills it internally, often fractionally better than the public price. That arrangement is the subject of the longest-running argument in market structure.
Has anyone tried to slow it down? Yes. One US exchange added a deliberate delay of a few hundred microseconds to every order so that speed advantages cannot be used against resting quotes, and several venues elsewhere use auctions at intervals instead of continuous trading for the same reason.

By InsiderTicker Editorial Team · Last updated

All news on Stocks →