High Frequency Trading Explained

How exactly does high-frequency trading work, and who wins the battle when HFT firms and the best AI stock trading software solutions compete?

Viktor Tachev

By Viktor Tachev | Updated November 26, 2018

We may receive a commission if you click on links in this article. Learn more

Financial markets have notably transformed in the last few decades. The shift towards technology brought speed, efficiency, transparency and comfort for every market participant. High-Frequency trading was born, and we reached a stage where thinking about trading without a computer is impossible. What are the best HFT companies, and how to get started with HFT trading?

high frequency trading

Trading Technology and Infrastructure

High-Frequency Trading is a subset of algorithmic trading based on high-speed trade execution. Or in other words – orders are opened and closed in fractions of a second.

High-frequency and day traders must utilize algo trading software to ensure lightning-speed trade executions. Although based on the same principles, High-Frequency Trading is different from algorithmic trading in that it requires significant investments in infrastructure, co-location rights and data feed products to ensure a lightning-fast trade execution process that provides the given company with a competitive advantage.

While high-frequency trading doesn’t require owning a supercomputer, your hardware might need an update. As an HFT trader, you need one of the best laptops for stock trading with a fast integrated LAN, fast CPU, GPU and sufficient RAM. Alternatively, you build your own trading computer.

While your setup is important, the more significant distinction is on the server level since you need your servers and order routes as close as possible to the relevant exchanges.

The point is that traders with fast trade executions are more profitable, and to have the most rapid trade executions, and you need to be as close to the exchange as possible.

Co-location rights are primarily used by institutions with server farms and mainframes, e.g., near the New York Stock Exchange when HFT trading stocks.

Retail investors typically go with direct-access brokers with server farms near the stock exchange or forex brokers executing client orders within their pools and exchanges.

Prime Broker Connection and Data Feeds

You will also need someone to execute your trades. There are a handful of large prime brokerages in the foreign exchange market, and usually, they are large investment banks. Prime brokers typically require you to have a minimum capital to trade besides the commissions they may charge for their services.

Also, remember that high-frequency trading requires tick data (level 3). Unlike historical data, live data costs, especially for the stock market, can be higher due to its highly fragmented nature. Besides, you often need more than a single data provider, especially if you plan on running match-making strategies.

High-Frequency Trading Nowadays

Today, the industry has reached its maturity. In the last few years, it has gone through a major consolidation due to ever-increasing competition, rising trading costs and constant regulatory developments.

A few bigger players run the industry, like Virtu Financials, Citadel Securities, Flow Traders, Hudson River Trading, Jump Trading, Optiver, Quantlab, TradeBot Systems, etc. High-Frequency Trading and quant trading companies vary in size, trading strategies, and type, as some are public, while most are prop firms.

Institutions and venues use HFT algorithms to route the PFOF orders, and brokers like TD Ameritrade and Robinhood receive payment for order flow in exchange for routing orders through HFT algos to venues like Citadel.

Whether it is a propriety or public company, all high-frequency trading shops are similar regarding their main goal – to outmuscle their competitors and execute as many trades as possible, even if the profit per trade is tiny.

To achieve that, high-speed traders focus on investments in new infrastructure to speed up the trade execution process. The speed at which trades are executed is measured in milli- and even microseconds.

The process will become even faster with microwave networks and high-performance chips.

  • 1 millisecond = 0.001 (10−3) of a second
  • 1 microsecond = 0.000001 (10−6) of a second

High-Frequency Forex Trading

So far, most analysis research papers have focused on high-frequency trading in the equities and commodities markets. Still, high-frequency forex trading is a viable option with markets open 24 hours, 5 days per week, enormous liquidity, OTC business and without any short-selling restrictions.

High-Frequency Trading Strategies

Generally, high-frequency traders’ strategies focus on capturing small profits from many executed trades. Their speed and technological advantage allow them to place many orders and front-run other market participants.

That way, they can buy a particular instrument and sell it back to the next one in the queue at a higher price, thus pocketing the difference.

Arbitrage strategies are the most frequently used HFT strategies. Whenever there is a way to make free money instantly, without putting the invested money at risk, HFT systems are there to capture the profit.

There are separate types of arbitrage-based strategies. Yet, all of them have the same idea – to exploit existing gaps in the pricing of specific instruments.

Thanks to the extreme speed at which high-frequency traders can place and execute orders, an instrument can be bought and sold simultaneously on different exchanges, where statistical arbitrage is among the most popular and widely applied High-Frequency Trading strategies.

Let’s say Apple stock trades at a lower price on NYSE. High-frequency traders can buy it from there and sell it on another exchange with higher prices, e.g., via cross-exchange orders on European or Asia exchanges or even alternative ECNs in the U.S.

Liquidity

Do HFT systems increase liquidity, or are most orders fake without the aim of getting ever executed? Many research papers, such as Hendershott (2011), point out that high-frequency traders help narrow the spreads for specific stocks by increasing liquidity, while some experts suggest this is the case only when markets are calm.

When things go wrong, high-frequency traders are often accused of the opposite – consuming liquidity to avoid losses. Because HFTs liquidity provision is unstable, it is also called ghost liquidity.

Market transparency and the other investors

Some of the most popular High-Frequency Trading strategies are built around the idea of submitting and canceling large portions of orders so that a certain instrument’s price can be manipulated. Such actions result in fake illusions about the buying and selling interest.

This can confuse other investors and have a particular effect on the large players in the industry, as they cannot tailor their strategies to the real supply and demand of the market.

Apart from that, High-Frequency Trading is known to minimize the spreads and decrease trading costs. While the first part is true, the second one is controversial.

When market participants are front-run by high-speed traders, they are basically unaware of the hidden trading costs charged as they buy instruments at higher prices and sell at lower prices.

Market stability

One of the most prominent examples of the harmful effect of high-frequency trading on the market’s stability is the so-called Flash Crash. On May 6th, 2010, for just 36 minutes, the DJIA lost almost 1,000 and regained approximately 700 basis points.

Later, a CFTC report confirmed that the high-speed trading activity was responsible for the market crash. During these 36 minutes, numerous large-cap companies traded down to pennies while other stock prices exploded.

Numerous research papers, such as the one from the Joint CFTC – SEC Advisory Committee on Emerging Regulatory Issues (2010) and many others, confirmed that high-frequency trading activity could cause market instabilities and even periodical flash crashes.

How Much Should High-Frequency Traders Have to Start

This is as complex as a question can get. Yet, we can make some basic estimations. While real-time data for retail traders with a brokerage account is free, more reliable, to-the-tick data is needed for HFT systems. It might sound surprising, but live trading data can cost between $5,000 and $50,000 per provider per month.

Co-location rights and exchange cross-connections can vary between $2,000 and $8,000 monthly.

The software is the most expensive part. While off-the-shelf solutions are on the cheaper side, usually up to $20,000 per month, they have many drawbacks (slow, hardly customizable, you don’t have full ownership, etc.). On the other hand, designing a high-frequency trading system from scratch can cost you up to $1m, depending on its complexity.

The cost of prime brokerage services and the hardware setup vary case-by-case, so they are hard to estimate. However, the rest of the figures can indicate whether you can afford high-frequency trading.

Conclusion

High-frequency trading is a phenomenon that transformed financial markets completely. Like every other disruptive technology, it has its supporters and critics. Some traders are specialized in day trading using HFT algorithms and make profits. Overall, high-speed traders are taking advantage of the gaps in the existing market structure.

Private investors come closest to the idea of high-frequency trading by using a fast trading computer, ensuring access to a high-speed network, and using day trading software, automated trading platforms and artificial intelligence-based trading strategy development algorithms for strategy development and automatic trade execution purposes.

See Also:

Alexander Voigt
Alexander Voigt is the founder of DAYTRADINGz, was a regular contributor to Benzinga and has been featured and quoted on leading financial websites such as Business Insider, Investors, Capital and Forbes.