Hobby IT: As a system administrator and project managers become the creators of trading robots

After 100 hours of programming, which stretched for three months, Mike Soule was ready to launch his project. At the same time, he did not know what to expect - if everything went as it should, then in the future he could expect financial success. And if not, then he could lose all his savings.

He was not working on a mobile application or another online store. He created a program that should buy and sell stocks 24 hours a day, 5 days a week.
')
Algorithmic trading on the stock exchange is a new dimension of the DIY phenomenon. Driven by their own curiosity, drawing on information from online courses and materials from amateurs like themselves, thousands of day traders write their trading software and enter the market with it.

“This is definitely one of those things when you don’t know if everything will work,” said Sule, a student at the University of Nevada in the city of Reno and network administrator of Tahoe Forest Hospital. “When the robot started trading, it was just wow. I don't know if I expected this, but I did it. ”

Brokerage companies are actively attracting customers developing trade robots - articles about trading collect tens and hundreds of thousands of views, and youtub videos are watched by hundreds of thousands of people. One of these courses was created by a professor at the University of Technology in Georgia, Tucker Balch - a video of his course was watched by more than 170,000 people. And despite the fact that only about 5% of viewers get to the end of it, at an event on algorithmic trading held in April in New York, three people asked for his autograph. “University professors are very rarely asked for an autograph,” says Balch.

Austrian project manager Alexander Sommer watched Balch's video lectures — he wanted to learn the basics of algorithmic trading. Now every morning, before going to work, Sommers wakes up to the sound of an incoming email message, which contains all the upcoming deals of the day. The message is generated by the trading platform created by him, which places orders automatically using algorithms written by him and his three trading partners. The partners have invested $ 200 thousand of their own funds in this “joint venture” - they are used to trade shares of the S & P 500 and Nasdaq Composite list.

Happy Sommers - Project Manager of the European oil company OMV Group. From 9 pm to midnight - he is a member of the team that is working to improve trading algorithms. Sommer has to double-check everything by the time the American stock exchanges open up - he will have at 15:00 by this time in Vienna. After closing the markets, he checks whether all necessary transactions have been executed. Sommer and his partners divide shifts, when you need to follow the work of the robots, in turn.



Multi-billion hedge funds that invest in technology have long attracted the attention of regulators and authorities in different countries. Private traders came to the radar after a trader who worked from his home in West London was arrested and extradited to the USA in the spring of 2015. His actions led to a 1000 points collapse of the Dow Jones Index in May 2010.

There are also failures - when Mike Sule decided in 2013 to go to Iceland and after several days almost without the Internet he returned home and connected to the network, he realized that something was wrong. His score was significantly less than before the trip.

“I realized that not everything is in order. I almost lost everything. ”

Sule updated his trading algorithm shortly before boarding the plane. And while he and five of his friends traveled around the country, his software was losing money due to an error. In total, the robot lost more than $ 6,000 - about 60% of the deposit at that time. It turned out that the problem was a typo, which led to the program buying twice as many shares as it sold.

“When I came back and saw how simple the mistake was, I was very upset. But there was no one to blame. ”

Using algorithms, traders can track the behavior of hundreds of stocks at the same time, which is absolutely impossible to do manually. Strategies can be complex, take into account the emerging news or even discussions in social networks, and can rely on quick response during price movements.

A simple algorithm can work, for example, according to this scheme: if the trading volume of a certain stock goes below the minimum threshold, and the 50-day moving average of this stock crosses the 200-day upward, then you should buy shares for $ 100. If the trading volume reaches the minimum threshold, and the 50-day moving average crosses the 200-day downward direction, then sell the shares for $ 100.

The number of tools for creating trading robots by private traders is constantly increasing. For example, the Quantopian platform allows traders to create their own algorithms for free, and the creators of the most successful receive an award. Another company, Rizm, offers a tool for creating robots in a simple graphic editor - this way even people who are not familiar with programming can do this. In our blog on Habré, we described 11 tools for creating trading robots .

In search of trading ideas, traders read books, blogs, browse Twitter discussions and study academic financial journals.

Some trading systems can even be built using Microsoft Excel, some strategies are quite simple to program. However, it is always necessary to spend a lot of time testing ideas - and here everything is more complicated.

“Some strategies can be programmed with two lines of code,” says Sommer. “And the rest of the work can easily take 5,000 lines.”

After the failure, Sule stopped his algorithm for six months - this time he spent on developing new tools that check the code for errors before the algorithm starts working with real money. Not so long ago, according to him, the trading systems created by him reached a “comfortable profitability”, and the number of profitable months outweighed the number of ended with a loss.

“This is still a hobby for me. It would be great one day to get passive income that goes beyond “active,” says Mike. “But I am not particularly in a hurry in this regard. I still like to have a full day job. ”

Read also : The editors of the ITinvest blog on Habré talked to the developer Andrey Gorkovenko, who told how knowing C # and C ++ helps make money on the stock market.

Links to materials on the topic:

• Stock market: How are exchanges organized and why are they needed?
• Stock Market Toolkit: What are Futures and How Do They Work?
• Stock Market Toolkit: What are Options and How Do They Work?
• What are stock indices and why are they needed?
• Books and educational resources on algorithmic trading
• Algorithmic and automated trading: 13 books on the subject
• ITinvest educational resources