The fourth story.How the investment marketplace grew from the pairs trading

Flash crash

On May 6, 2010, an algorithmic trading settled down in the UnitedTraders office forever.

For most traders, this date became associated with a real “Ideal Storm”. During that session, the fall of the major US indices reached the fantastic 10%, and shares of some of the largest companies fell by 90%. What is more, it happened within a few minutes. Just think of it: about trillion dollars have disappeared from the world economy within a few minutes!

In those years, Ben Lichtenstein, who commented the course of trading futures for S&P 500 Index from the pit of the Chicago Stock Exchange (CME) was very popular in traders circles. Here is the record of his comments that day, it perfectly describes the market tone:

The investigation of the SEC and CFTC, which followed those events, has established the cause of that collapse. It turned out to be the actions of one mutual fund, which started selling large packages of contracts for the S&P500 E-mini index futures. These sales took up high-frequency algorithms that triggered a chain reaction and an avalanche drop in prices. The resonance was so strong that even the White House became interested by this problem.

That event did not pass us by. Many intraday traders could take short positions in time, but it was more like a lucky coincidence. That time, everyone perceived all that’s happening as the second wave of the mortgage crisis which everyone was waiting for eagerly. However, very soon it became clear that we did not understand the nature of the events happened well enough.

Pairs Trading

At that time, we were already engaged in the automation of manual trading. But the events of the 6th of May forced us to pay special attention to whole algorithmic trading.

Since then, we have become especially interested by the algorithmic traders. We began to look for them for involving in the remote work and to develop the internal team.

It is everything clear in respect of the remote work: a trader comes and, in the course of interviews and test launches of trading, the work with him becomes arranged. There are practically no differences with manual trading.

Everything is different in respect of the internal team development. We had to conduct the surveys on our own. This task had been undertaken by the traders, who were interested by so-called “pairs trading”.

Pairs trading is the purchase and sale of 2 instruments similar in the fundamental features. When the instruments’ prices “differ”, more expensive one to be sold and cheaper one to be bought. When the prices “come to one value”, reverse transactions to be made. So, a trader makes profit on the assumption that the prices of similar instruments must behave in a similar way.

As it seemed then, this strategy can be algorithmized easily. Having built the simplest models, we made the most banal mistake usually made by a beginning algotrader. We believed that the incredible results obtained in theory can be obtained in practice. Thus we learned the important algotraders’ wisdom: if the strategy gave incredible results when testing — do not trust it, look for the error

Algorithmic Trading

At that time, we already conducted public activities. A lot of people, who began interesting by trading, became aware of us, especially while American campaigns. There was very few materials on this subject in the Internet and the most part of the resources in Russian have been created by us. It made and now makes it possible for us “to feed” the team with new personnel.

Exactly according such scenario, we got into the eyeshot of Konstantin Ivanov, the Ukrainian programmer interested by trading. He was persistent enough to achieve our involving him in the work of our team which develops the pairs algorithmic trading.

The ideas of pairs trading quickly evolved into the more advanced “statistical arbitration” (simultaneous trading of a large number of pairs — it is a very simplified definition!). It was decided to move Konstantin to Moscow and close him in the room with other traders until the start of the “black box” operation.

Truth be told, the team could not solve the set task directly and had been disbanded after 2 years of experiments. But within 2 years, many surveys have been carried out. In the UnitedTraders, the correct ideas about algorithmic trading, its areas, potential and limitations have been formed. This makes it possible for us to involve the remote algo-teams and to recruit developers of individual strategy easily.

High Frequency Trading

The formed ideas about algo-trading put the HFT (high-frequency trading) on the top of our concept. It was perceived as the coolest thing which is present in the market. Such algorithms make thousands, tens of thousands transactions a day and earn thousands percents of annual interest for their developers. Such algorithms were especially of scalpers’ interest, because they above all others competed with scalpers in the market. For 2–3 years after the mortgage crisis, the algorithms pushed them aside significally in those spheres where they accustomed to make money easily.

The development of HFT line required the competencies, which the team has not yet had. Aleksey Nizev undertook the leader’s role in the development of such competencies. He came to UT as an infrastructure developer for the team of Konstantin Ivanov.

In the beginning, Alexey was engaged in testing the strategies and automating the trading of manual traders, but very soon he completely undertook managing the algorithmic trading in UT and, actually, restarted the team from scratch.

The success has not come immediately. We experimented a lot and were choosing where to start from. Finally, we have chosen the Russian market for launching the high-frequency project. In the Russian Federation, the huge investments in servers, quotes and connections were not required unlike the US stock market, which we knew much better.

Quantitative

When you trade at ultra-high speeds (milliseconds –microseconds-nanoseconds), you stop perceiving the market as something continuous. The time no more be uniform. Each transaction or bid made has a strictly defined owner with certain motives. The better you can determine the bid’s owner and his motives, the higher the probability of success of your transaction.

The analysis of the events of May 6, 2010 from Nanex illustrates such concept well: http://www.nanex.net/FlashCrashFinal/FlashCrashSummary.html

Don’t be upset if you did not understand anything from the material on the link above. With this material we wanted to illustrate the material with which the HFT team has to work. The people who make such analysis are called quantitative analysts, abbreviated to “quants”. They analyze the exchange data, interpret them, search for the opportunities for profit making strategies and, finally, build the mathematical models for the strategies.

About two years ago, the quants began to appear in our office too — the process circle has been closed:

the quants develop ideas, the programmers and engineers develop the infrastructure, the strategies are tested in the test environment and sent for updating, then tested again. They are launched “in field conditions” when all performance indicators become satisfactory.

KvadratBlack

As you know, practically since the foundation of UT, we have made our activity public. This attracted the attention not only of potential student traders, but also of investors.

The appearance of a large number of algorithmic traders in our team made it possible for us to identify a part of the most stable and scalable strategies for managing investors’ money. They were united into the pool called KvadratBlack. This name simultaneously symbolizes the “black box”, as algorithmic strategies usually called, and refers to the picture of Kazimir Malevich who has drawn the portrait of the algorithmic strategy 😉

From 2013 to 2016, the pool of the KvadratBlack strategies have been managing the funds of the hedge fund of the same name. The fund showed stable results, having earned dozens of percents for its investors. In addition, KvadratBlack received two awards, information about which is available on the main page of our website UnitedTraders.com.

Investment Marketplace

We learned a lot while working with investors. The most important thing we understood is the desire to choose. Each investor is individual. Someone wants to undertake more risk and get high profitableness. Someone is not ready to risk under any circumstances. Someone is ready to invest money in the company only because he takes a liking to it and he does not care about the profit. Someone wants to try, but someone consciously conducts his investment portfolio.

We already have several products. We supplemented the KvadratBlack algorithmic strategies with the shares of non-public companies (OTC) and of those who are going to launch IPO, and with crypto-currency strategies. Also we are going to expand the product line.

The appearance of each of our investment products deserves a separate article. But let’s back to the main line of the narration. We will continue to talk about how our business lines emerged from various traders….

Like what you read? Give United Traders a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.