Introduction to High Frequency Trading – What Individual Traders Really Need to Know

Michael J. Carr, CMT

After being blamed for the flash crash of May 6, 2010, high frequency trading (HFT) took on a legendary status with the investing public. Some had found yet another reason why the little guy can’t get a break on Wall Street, while others saw another step in the evolution of trading. As always, the truth lies between two extremes, and HFT does impact small investors, in both good and bad ways, so we all need to understand what it is.

There’s no easy way to explain HFT so we will start with a technical definition and then dig a little deeper. HFT involves using algorithmic trading strategies to generate a large number of orders using very low latency data, seeking a very small profit on the few orders that are actually executed.

Algorithmic trading strategies are computer generated buy and sell orders. They’ve been around for years, but as processing speed has increased, the number of orders they can generate has risen dramatically. Today, computers can generate thousands of orders a second based upon the buy and sell orders being created by individuals and institutional traders. Most of these algorithm generated orders expire unfilled very quickly. Buys and sells take place so quickly, in fact, that the timeframes involved in HFT are difficult for the individual trader to comprehend.

Very low latency data means knowing exactly what’s going on at the exchange in real-time. Latency describes the amount of delay between the exchange computer and the HFT system. It’s technically impossible to have zero delay, but the delay that HFT firms experience is measured in milliseconds. It takes 300 milliseconds for you to blink your eye. Hundreds of orders are placed and either filled or cancelled while you’re blinking your eye, and the process continues throughout the trading day. Most web sites offer individual traders quotes with a 20 minute delay, making it impossible for individuals to compete in this ultra short-term game.

Firms fight to place their computers as close to the exchange as possible, and vie for the most talented programmers. One example highlights the process. Futures are traded in Chicago, while stock trading is centered in New York. One company laid a fibre optic cable in a straight line between the exchanges to gain a small advantage over its competitors. That meant drilling through mountains, going under rivers and highways, and overcoming a variety of physical obstacles. In the end, it covers 825 miles and data moves between the two cities in 13.3 milliseconds. The next fastest cable runs 100 miles longer and data takes 3 milliseconds more to make the trip. The cost was $3 million to run the cable; the payoff is that traders pay $1.2 million a year for access to the line.

The goal is to make a fraction of a cent on each share they buy or sell. This is a very low margin business, and an academic paper recently estimated that the total available profits total only $21 million. Obviously the academic community is missing something, or no one would have built that $3 million cable to connect all those multi-million dollar computer systems. However, the reality is that there is a limited amount of money to be made in HFT, and only highly specialized firms will be able to make that money.

While individuals aren’t directly involved in HFT, they are participating in the process. You may want to buy a stock that’s trading at 10.00 a share. That’s what the last trade was at. Stocks trade in an auction market, so there are always two prices for a stock. The highest amount someone is willing pay to buy the stock is the bid, and the ask price is the lowest amount someone is willing to sell the stock for. In this example, the bid may be 9.99 and the ask is now 10.01. You want to buy, and if you place a market order, you’ll pay 10.01 and own your stock in a matter of seconds. Sellers at the market will receive 9.99 a share. The difference between the bid and ask is known as the spread and spreads have always existed in the stock market. When stocks traded in factions, the spreads totaled 12.5 to 25 cents per share, and are now down to a penny or less in some cases.

During the milliseconds that lapse between the time you enter your order and the time it’s executed, an HFT firm can come in and enter a buy order at 10 even, knowing you’ll take it off their hands for a price that’s a penny higher. If their bid of 10 is accepted, they’ll immediately offer it to you, at perhaps 10.005, slightly below the previous ask price and low enough to ensure they don’t hold the stock longer than a second. This is a riskless transaction from their perspective since they know they have a buyer lined up at a higher price to take the stock off their hands.

This process plays out millions of times a day, and the HFT firms are competing with each other to buy the shares to sell to you. Competition forces the difference between the bid and ask price to become smaller and smaller, and it really is often a penny or less on the most actively traded stocks. HFT firms can profit from these transactions, but in many ways the individual traders are also winners since the spreads on stocks have narrowed significantly because HFT firms are competing with each other for these orders.

HFT definitely has an upside since it provides liquidity in the markets and helps lower transaction costs for traders. But it also has a downside for individuals who don’t understand the impact HFT has on the market. For years, many individuals have been taught to place a stop loss order on their holdings to ensure they don’t suffer very large losses. Some traders will place a stop 8-10% below their buy price, long-term investors may use 25% or more as the point where they just sell out at the market price.

Much of the wealth lost in the May 6 flash crash was likely due to these standing stop orders. The S&P 500 lost a little more than 8.5% at its low that day, before quickly bouncing back. Many individuals use exchange traded funds (ETFs) to invest in stock indexes. ETFs come pretty close to tracking the market over the long-term, but can differ from their index stocks for brief periods of time. One of the most active ETFs that tracks the S&P 500 is the SPDR, which trades with the symbol SPY. At its low, SPY fell by 10.1% on May 6, but closed in line with the S&P 500. The larger decline was due to market inefficiences, and these opportunities are rapidly seized by HFT firms and corrected. That’s why the difference didn’t last more than a few moments and the ETF matched the index again by the end of the day.

Many individuals use the iShares S&P 500 ETF, which trades under the symbol IVV, as an alternative to SPY or an index mutual fund. The annual expense ratio is slightly lower for IVV compared to SPY or a mutual fund, and individuals profit from that cost difference over the long term. Individuals like to place stop loss orders, as countless magazine articles have told them to do, and we saw that on May 6. The crash sent IVV about 24.75% below than its previous close, significantly different than the underlying S&P 500 or even other ETFs that track the index.

HFT firms saw the stop orders in IVV, and their goal is to execute any customer order. The investor sought protection with the stop loss order, while a high frequency trader saw it as a request to trade from the client. Computers drove the price lower, until as many orders as possible were cleared from the market.

That is what individual investors need to know about HFT. The specifics shouldn’t matter because individuals can’t afford the data access. You need to know that market structure has changed, and that stop orders are now likely to be executed in a steep decline. Check the close, and enter a sell order for the next morning if you need to. But, leaving a standing stop order in the market may lead to losses for you and a tiny profit for an HFT firm in today’s fast moving markets.


Trading involves substantial risk of loss and is not suitable for all individuals. Past Performance is not indicative of future results.