High Frequency Trading : Always Getting Faster?

High Frequency Trading : Always Getting Faster?

High-frequency trading (HFT) is one of the primary algorithmic trading strategies. Nowadays, above half of the market is classified as high-frequency trading.

High-frequency trading was started in the 1930s, and it boomed in 1983 after NASDAQ used electronic trading. Then, the speed of HFT was increased in the 2000s to several seconds. At that time, HFT still just occupied less than 10 percent of the market orders.

The speed of HFT has increased to milliseconds and even microseconds since 2010. Also, 73% of all equity orders volume was HFT in 2009. From 2000 to 2010, we can say that the speed and market share of HFT has increased rapidly.

Even though many institutions took advantage of HFT to eliminate bid-ask spreads and increase the stock price precision, many shortcomings remain as problems in the market.

For example, on May 6th, 2010, Dow Jones dropped 600 points within 5 minutes. The market value lost nearly $1 trillion in value due to algorithms built to trade based on the reactions of other traders.

Right now, the HFT world is still getting faster.

The time of executing an order could be smaller than 600 nanoseconds.

From a physics standpoint, the fastest processor can read high and low voltage in about 0.12 nano-seconds. And this is bounded by the physical barrier.

Computers use ASCII code to store information and one ASCII code takes 8 bits which is a byte.

Then, this goes to about 1 nano-second.

Usually a stock id is around 12 digits.

Including buy and sell information, this time shall go up to 13-14 nano-seconds.

Also, each firm shall have a specific id for trading (12 digits) which makes this time go up to 25 nano-seconds.

This is how much time takes from the end to process the information. If we include the decision end and the wifi speed limits, this time is around 100 nano-seconds.

That is how fast if we just tell the computer “who do action on which stock”.

Python For Trading : The Benefits.

I believe there are security messages within the system, so this time shall go up more. So, if the 600 nano-seconds is true, it is potentially already very close to the most recent physical barrier. 

However, not every trader wants the HFT to be too fast. The faster the order execution, the faster the flash crashes when most traders have negative emotions. IEX founder says too much high-frequency trading can rig the market.

Investors Exchange (IEX) is a stock exchange founded in 2012 to mitigate the effects of high-frequency trading. IEX delays the flow of data from the exchange. Traders cannot co-locate equipment adjacent to IEX’s servers.

The data delay prevents many predatory behaviors.

We asked Mark Fisher, the trading legend, his thoughts on HFT’s future and his response was, “always getting faster”. Hence the title of the article!

Mark Fisher the Trading Legend on Crude’s Price in 2030.

In conclusion, high-frequency trading is a great type of algorithmic trading. It always becomes faster and gives the market liquidity. But we still need the SEC to regulate HFT, since sometimes HFT may negatively affect the market.

Written by Frank Chen

Edited by Ang Li, Avhan Misra, Lika Mikhelashvili & Hakil Haxhiu