What 10 milliseconds of high-frequency trading look like
Eric Hanseider, the founder of Nanex, a company engaged in analytics and information systems for high-frequency trading, publishes on YouTube very interesting visualizations of high-frequency trading. For example, 10 milliseconds of Merck stock trading:
Circular rectangles are exchanges. Triangles and circles flying between them - changes in quotes and deals. Lower rectangle (6 hours) - the best buying and selling rates among all exchanges. Visualization is about 40,000 times slower than real time.
This is how the first few seconds after the Facebook IPO looked like (due to technical problems, the start of trading was delayed, so at the beginning of the video, about 45 minutes were quickly accelerated):
')
And here is half a second of trading Johnson & Johnson shares:
Here is a recent anomaly in the stock market, when an outbreak of trading programs occurred at 245 milliseconds before the publication of a report on consumer sentiment that affected the price of many stocks:
Such anomalies can be both a sign of information leakage and a consequence of the tactics of some traders who tend to artificially “shake” the market before publishing news important for quotes. This tactic in the jargon of financiers is called “bang the hornet's nest” (Banging the beehive). In February of this year, a similar anomaly occurred with shares of energy companies - active trading began 400 milliseconds before the publication of the report on natural gas.
According to Erik Hansaider, the program that visualizes high-frequency trading was written specifically to visually demonstrate to officials who regulate financial markets what they are dealing with. However, he writes, there are big doubts that he managed to convey to them an idea of ​​modern speeds and how important the technical aspects of networks and stock exchange computers are for the market.
It is not surprising, because high-speed trading, which became possible after the appearance of electronic exchanges, began about 15 years ago, and initially implied the conclusion of transactions within a few seconds. Now the bill goes on milli- and even microseconds. Traders are ready to spend hundreds of millions of dollars on laying cables that reduce ping by 6 milliseconds. Supercomputers are used to model financial markets, the best mathematicians and programmers of the world are engaged in developing algorithms for trading robots, and trading programs run on FPGAs specially created for them , which are much faster than general-purpose hardware. In 2012, in the US, high-frequency traders made about half of all transactions in the stock markets.
Circular rectangles are exchanges. Triangles and circles flying between them - changes in quotes and deals. Lower rectangle (6 hours) - the best buying and selling rates among all exchanges. Visualization is about 40,000 times slower than real time.
This is how the first few seconds after the Facebook IPO looked like (due to technical problems, the start of trading was delayed, so at the beginning of the video, about 45 minutes were quickly accelerated):
')
And here is half a second of trading Johnson & Johnson shares:
Here is a recent anomaly in the stock market, when an outbreak of trading programs occurred at 245 milliseconds before the publication of a report on consumer sentiment that affected the price of many stocks:
Such anomalies can be both a sign of information leakage and a consequence of the tactics of some traders who tend to artificially “shake” the market before publishing news important for quotes. This tactic in the jargon of financiers is called “bang the hornet's nest” (Banging the beehive). In February of this year, a similar anomaly occurred with shares of energy companies - active trading began 400 milliseconds before the publication of the report on natural gas.
According to Erik Hansaider, the program that visualizes high-frequency trading was written specifically to visually demonstrate to officials who regulate financial markets what they are dealing with. However, he writes, there are big doubts that he managed to convey to them an idea of ​​modern speeds and how important the technical aspects of networks and stock exchange computers are for the market.
It is not surprising, because high-speed trading, which became possible after the appearance of electronic exchanges, began about 15 years ago, and initially implied the conclusion of transactions within a few seconds. Now the bill goes on milli- and even microseconds. Traders are ready to spend hundreds of millions of dollars on laying cables that reduce ping by 6 milliseconds. Supercomputers are used to model financial markets, the best mathematicians and programmers of the world are engaged in developing algorithms for trading robots, and trading programs run on FPGAs specially created for them , which are much faster than general-purpose hardware. In 2012, in the US, high-frequency traders made about half of all transactions in the stock markets.
Source: https://habr.com/ru/post/183484/
All Articles