Furious bulls: how Wall Street became dependent on "high-speed" trades. Part 5 (and last)
One of the side effects of the growing popularity of algorithmic trading is the explosive growth of the financial data market. One of the most valuable types of information is historical data; quanta require terabytes of such information to test their trading models. Even free information, such as corporate income statements and government statistics, can be sold by converting the information into an easy-to-read format.
[ first part ], [ second part ], [ third part ], [ fourth part ]
This did not escape the attention of Dow Jones, where they realized that the money that the company invests in getting the news first is wasted because of people who need at least a few seconds (!) To assimilate this information. As a result, in 2007, the company created a newsletter directly for its users' trading platforms, all the news in which were adapted for reading by algorithms.
')
Since the Dow Jones newsletters began to include numerical information to display "unstructured" news, it seems that now literally everything can affect the size of profits and stock prices. This practice has spread to the entire "news analytics" industry, led by companies like the Raven Pack, which in a day turns up to 100,000 news reports into data suitable for use in trading.
“The asset manager says,“ I want to model a strategy based on the behavior of the Central Bank, ”and we begin to track every article that mentions the head of the European Central Bank and the word vigilance,” said Rob Passarella, vice president and manager Director of Dow Jones Institutional Markets.
One of the paradoxical rules of analytics is that, all other things being equal, companies about which little is said in the media tend to overestimate the achievements of their more popular competitors.
Image: McKay Brothers uses 20 microwave transponders — specifically this one 215 feet high — to transfer messages between data centers in New Jersey's economic markets and the Chicago Commodity Exchange in just 4.5 milliseconds.
Photo: Steve Aikam
However, not only the information received from the bankers of the Central Bank is important. Virtually any data that is relevant to an economic activity is aggregated and tested to check their potential impact on stock prices. GPS data showing the concentration of mobile phone users in shopping centers and office buildings are used to obtain real-time data on economic activity. Even the most ephemeral snatches of data in a graveyard of digital information, such as old Twitter posts, can have some value if you know how to handle them.
According to his rather sharply titled article, “Twitter Mood predetermines the state of the stock market” [ eng ] Johan Bollen, a computer science researcher at Indiana University, and two of his colleagues collected and analyzed 10 million tweets since 2008, selecting phrases reflecting the authors' emotional state and ranking them in relation to various psychophysical states, such as “calm”, “vigor”, “confidence”, “vigor” and “feeling of happiness”. The researchers then traced the correlations with the stock market and found that an increase in the number of tweets reflecting the “calm” of the authors in the interval from 2 to 6 days was followed by a rise in the Dow Jones Industrial Average.
No one, including Bollen, knows what it means and why it happens, but if quanta had a coat of arms, they would have written on it: “If it works, play on it.”
Moreover, play it right now, this very split second. According to Alexander Wissner-Gross, a Harvardian physicist, from the moment any hedge fund decides that he needs a particle accelerator to generate neutrinos and other organizations decide after him as well, just a few decades is just a matter of time.
And just a few days before the Battle of Quanta, just before the promulgation of disappointing news about the absence of neutrinos moving in space faster than the speed of light, the researchers announced that with the help of neutrinos they were able to transmit a message from the accelerator from the National Accelerator Laboratory. Enrico Fermi in Chicago on a detector device located one kilometer from the laboratory. According to Dan Stencil (Dan Stancil) from the University of North Carolina, the signal was transmitted at a “very close” speed to the speed of light. Unfortunately, the volume of data transmitted in this way is only 0.1 bits per second, so this technology will be useless when transmitting signals larger than a “Yes / No” message.
But we all know the legendary story about Thomas J. Watson from IBM, who predicted that “five computers will be in demand on the global market”. Stensil knows physics, but he seems to have never worked in the financial industry. Any quantum could tell him what to do next.
Buying neutrinos.
PS The II All-Russian Conference on Algorithmic Trading will take place very soon. If you are interested in this topic - then here's an event on Habré , come, it will be interesting.
[ first part ], [ second part ], [ third part ], [ fourth part ]
This did not escape the attention of Dow Jones, where they realized that the money that the company invests in getting the news first is wasted because of people who need at least a few seconds (!) To assimilate this information. As a result, in 2007, the company created a newsletter directly for its users' trading platforms, all the news in which were adapted for reading by algorithms.')
Since the Dow Jones newsletters began to include numerical information to display "unstructured" news, it seems that now literally everything can affect the size of profits and stock prices. This practice has spread to the entire "news analytics" industry, led by companies like the Raven Pack, which in a day turns up to 100,000 news reports into data suitable for use in trading.
“The asset manager says,“ I want to model a strategy based on the behavior of the Central Bank, ”and we begin to track every article that mentions the head of the European Central Bank and the word vigilance,” said Rob Passarella, vice president and manager Director of Dow Jones Institutional Markets.
One of the paradoxical rules of analytics is that, all other things being equal, companies about which little is said in the media tend to overestimate the achievements of their more popular competitors.
Image: McKay Brothers uses 20 microwave transponders — specifically this one 215 feet high — to transfer messages between data centers in New Jersey's economic markets and the Chicago Commodity Exchange in just 4.5 milliseconds.
Photo: Steve Aikam
However, not only the information received from the bankers of the Central Bank is important. Virtually any data that is relevant to an economic activity is aggregated and tested to check their potential impact on stock prices. GPS data showing the concentration of mobile phone users in shopping centers and office buildings are used to obtain real-time data on economic activity. Even the most ephemeral snatches of data in a graveyard of digital information, such as old Twitter posts, can have some value if you know how to handle them.
According to his rather sharply titled article, “Twitter Mood predetermines the state of the stock market” [ eng ] Johan Bollen, a computer science researcher at Indiana University, and two of his colleagues collected and analyzed 10 million tweets since 2008, selecting phrases reflecting the authors' emotional state and ranking them in relation to various psychophysical states, such as “calm”, “vigor”, “confidence”, “vigor” and “feeling of happiness”. The researchers then traced the correlations with the stock market and found that an increase in the number of tweets reflecting the “calm” of the authors in the interval from 2 to 6 days was followed by a rise in the Dow Jones Industrial Average.
No one, including Bollen, knows what it means and why it happens, but if quanta had a coat of arms, they would have written on it: “If it works, play on it.”
Moreover, play it right now, this very split second. According to Alexander Wissner-Gross, a Harvardian physicist, from the moment any hedge fund decides that he needs a particle accelerator to generate neutrinos and other organizations decide after him as well, just a few decades is just a matter of time.
And just a few days before the Battle of Quanta, just before the promulgation of disappointing news about the absence of neutrinos moving in space faster than the speed of light, the researchers announced that with the help of neutrinos they were able to transmit a message from the accelerator from the National Accelerator Laboratory. Enrico Fermi in Chicago on a detector device located one kilometer from the laboratory. According to Dan Stencil (Dan Stancil) from the University of North Carolina, the signal was transmitted at a “very close” speed to the speed of light. Unfortunately, the volume of data transmitted in this way is only 0.1 bits per second, so this technology will be useless when transmitting signals larger than a “Yes / No” message.
“When using the correct modulation scheme, the amount of data transferred can be increased by one or two orders of magnitude,” says Stensil, adding: “I don’t know what commercial application can be found for this technology.”
But we all know the legendary story about Thomas J. Watson from IBM, who predicted that “five computers will be in demand on the global market”. Stensil knows physics, but he seems to have never worked in the financial industry. Any quantum could tell him what to do next.
Buying neutrinos.
PS The II All-Russian Conference on Algorithmic Trading will take place very soon. If you are interested in this topic - then here's an event on Habré , come, it will be interesting.
Source: https://habr.com/ru/post/204474/
All Articles