You'd Better Know Your High Frequency Trading (HFT) Terminology


4 ratings



Elvis Picardo

14 Jul, 2017

in Day Trading & Scalping

With the upsurge of investor interest in high-frequency trading (HFT), it is important for industry professionals to come up to speed with HFT terminology. A number of HFT terms have their origins in the computer networking/systems industry, which is to be expected given that HFT is based on incredibly fast computer architecture and state-of-the-art software. We briefly discuss below 10 key HFT terms that we believe are essential to gain an understanding of the subject.

Locating computers owned by HFT firms and proprietary traders in the same premises where an exchange’s computer servers are housed. This enables HFT firms to access stock prices a split second before the rest of the investing public. Co-location has become a lucrative business for exchanges, which charge HFT firms millions of dollars for the privilege of “low latency access.”

As Michael Lewis explains in his book “Flash Boys,” the huge demand for co-location is a major reason why some stock exchanges have expanded their data centers substantially. While the old New York Stock Exchange building occupied 46,000 square feet, the NYSE Euronext data center in Mahwah, New Jersey is almost nine times larger, at 398,000 square feet.

Flash Trading
A type of HFT trading wherein an exchange will “flash” information about buy and sell orders from market participants to HFT firms for a few fractions of a second before the information is made available to the public. Flash trading is controversial because HFT firms can use this information edge to trade ahead of pending orders, which can be construed as front running.

U.S. Senator Charles Schumer had urged the Securities and Exchange Commission in July 2009 to ban flash trading, saying that it created a two-tiered system where a privileged group received preferential treatment, while retail and institutional investors were put at an unfair disadvantage and deprived of a fair price for their transactions.

The time that elapses from the moment a signal is sent to its receipt. Since lower latency equals faster speed, high-frequency traders spend heavily to obtain the fastest computer hardware, software and data lines so as execute orders as speedily as possible and gain a competitive edge in trading.

The biggest determinant of latency is the distance that the signal has to travel, or the length of the physical cable (usually fiber-optic) that carries data from one point to another. Since light in a vacuum travels at 186,000 miles per second or 186 miles per millisecond, a HFT firm with its servers co-located right within an exchange would have a much lower latency – and hence a trading edge – than a rival firm located miles away.

Interestingly, an exchanges’ co-location clients receive the same amount of cable length regardless of where they are located within the exchange premises, so as to ensure that they have the same latency.

Liquidity Rebates
Most exchanges have adopted a “maker-taker model” for subsidizing the provision of stock liquidity. In this model, investors and traders who put in limit orders typically receive a small rebate from the exchange upon execution of their orders because they are regarded as having contributed to liquidity in the stock, i.e. they are liquidity “makers.”

Conversely, those who put in market orders are regarded as “takers” of liquidity and are charged a modest fee by the exchange for their orders. While the rebates are typically fractions of a cent per share, they can add up to significant amounts over the millions of shares traded daily by high-frequency traders. Many HFT firms employ trading strategies specifically designed to capture as much of the liquidity rebates as possible.

Matching Engine
The software algorithm that forms the nucleus of an exchanges' trading system and continuously matches buy and sell orders, a function previously performed by specialists on the trading floor. Since the matching engine matches buyers and sellers for all stocks, it is of vital importance for ensuring the smooth functioning of an exchange. The matching engine resides in the exchanges' computers and is the primary reason why HFT firms try to be in as close proximity to the exchange servers as they possibly can.

Refers to the tactic of entering small marketable orders – usually for 100 shares – in order to learn about large hidden orders in dark pools or exchanges. While you can think of pinging as being analogous to a ship or submarine sending out sonar signals to detect upcoming obstructions or enemy vessels, in the HFT context, pinging is used to find hidden "prey."

Here's how - buy-side firms use algorithmic trading systems to break up large orders into much smaller ones and feed them steadily into the market so as to reduce the market impact of large orders. In order to detect the presence of such large orders, HFT firms place bids and offers in 100-share lots for every listed stock.

You need to be logged in to post comments or rate this article.

The article is good, the book is a bit bias against the HFTs, I for one think the HFTs are good for the retail trader, because of the liquidity they create and the lowering of the bid/ask spread. For sure the big institution traders are against... The problem for the retail trader are the dark pools (another good book by the same name), because they essentially hide a portion of the market, but the big institutions love them, because they are part of it...

Aug 19, 2017

Member (945 posts)

Very helpful

As a software engineer myself, I find the carefully worded explanations of HFT terms and tactics very enlightening. I have not read the Lewis book, so I come fresh to this article. It is possible, however, that the meaning of the terms and tactics might still escape many readers. I wrote software programs to test the speed and processing accuracy of very large database software. This software is called a Data Warehouse, employed in all large banks, retailers, and manufacturers [and exchanges] to manage and make available all the important data of their businesses. Customers' first priority of these computer/software systems is always speed+accuracy. Software programs written to test "data processing"--the processing of data speeding across the CPU or the little bitty memory chip or the buses [electronic connection lanes between parts on the motherboard] must capture and report on the data's speed in units of a NANOSECOND [1 billionth of a second]. My test programs could report back to the test engineer the time in nanoseconds an action [add number columns] took place. If it took too long, say 15 nanoseconds instead of 8 nanoseconds, the operation could be rewritten to optimize it. After this is done inside the program, the author of this article describes methods to shorten the external path -- or length of the highway -- the data bits travel on. For the test engineer, distance is time added, and that speed is also measured in nanoseconds. There can be an infinite number of slipups along the way. So, I guess, with the author's descriptions of just the mechanical means attempted to optimize the paths [speed] of trading data through computer networks, I have not only become more knowledgeable about HFT setups, but I now feel much less anxious about the effects of it on trading in general. HFT may be good, but as a software engineer myself, I KNOW it can't be perfect.That leaves a LOT of room for the rest of us to PROFIT.

Aug 19, 2017

Member (2 posts)