Oleg Nikishenkov

The battle between man and machine has started on the Russian stock market - but for now, at least, the humans have the upper hand.

Mikhail Shlyachkov is a part-time player on the forex market who used to leave the decision-making to a "robot", or high-frequency trading machine. But, after the machine lost his deposit, he decided to terminate their relationship.

The problem occurred when the robot mistook a bunch of "buy" deals as a  new trend and jumped on the bandwagon - only for it to start going rapidly downhill.

"Any algorithm generates a history when the market changes chaotically," said Shlyachkov, with a touch of bitterness.

High-frequency traders, or HFTs, make trades in microseconds, or one millionth of a second, to profit from short-term market gyrations, using a strategy chosen by programmers.

In Russia, they are used mainly on the MICEX, the more liquid of Russia's two stock exchanges, by small and medium-size brokerages.

On Western markets, their widespread use has become a controversial topic, with critics worrying that the robots may be injecting new risks into the financial system - particularly after the recession was brought about in part by Wall Street banks' widespread use of risky financial instruments. HFT now accounts for 42 per cent of all US trades, according to one recent survey.

For Shlyachkov, the robot's lack of a wider perception of trends and the ability to only execute decisions based on numerical indicators is a significant chink in its armour.

"If you find me an algorithmic machine able to think globally, I'll buy it from you for any price," said Shlyachkov, who now works in real estate.

These days, he only uses "advisers" - machines that quickly analyse huge volumes of historical data on previous deals, but cannot take any decisions. The final say remains strictly in human hands.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Adding human intuition to the machine's analysis is still not guaranteed to produce positive results because robots available for general retail are too simple to beat contemporary markets.

"When you apply a plain, straightforward strategy, even if it's the right one, you may lose your money anyway, as it can be easily repeated by 100,000 people," said Arsen Yakovlev, head of algorithmic trading at Moscow firm Zerich Capital. "So you could share your profit between these 100,000 people - and your income doesn't exceed your commission."

Despite the problems Shlyachkov experienced, robots are gaining popularity with small investors, with several hundred attending a seminar in Moscow on the subject last month.

To compete with the banking cyber-sharks with more powerful and sophisticated programs, students were advised to vary their approach and get several robots executing different strategies at the same time.

"When you invent a new fruitful strategy, keep in mind that someone else sooner or later will start to make money exploiting your idea," said Vasily Strizhanov, an analyst and trading methods teacher at online trading firm Alpari, which hosted the event.

The machines' share of the Russian market has shot up to 70 per cent in terms of volume said Yakovlev, but despite their popularity, they're not able to beat the banker just yet.

"It's still a fight for profit between humans, not computers," said Yakovlev. "But only talented and highly-skilled managers will be able to beat smart programs."

Even older investors have started siding with the machines, taking up programming classes to compete with a generation that is taught IT skills in school.

"Last year the sales of algorithmic programs doubled, as many people look to robots as a remedy against the crisis turmoil," said Pavel Efremov, deputy head of Alor, a Moscow brokerage.

Russia's financial authorities have been positive about the rise of the machines as they add liquidity to the system, but the European Commission and US Securities and Exchange Commision have raised worries that they could lead to lack of transparency.

Svetlana Shvetsova, a spokeswoman for the Federal Financial Monitoring Service, said it was up to stock exchanges to regulate the machines.

But last year the RTS complained that some robots were hyperactive and congesting the floodgates with multiple inquiries. The bourse applied an extra charge and a threshold for the maximum number of mechanical inquiries.

"Every additional 20 inquiries will cost 1 rouble, which decreased the number of empty calls by 15 percent," said Sergei Zamolodskikh, who heads the RTS's infrastructure department.

 

The breakdown

How robot trading works

High-frequency trading uses supercomputers to catch very short-term fluctuations in stocks, currencies and commodities. They seek out temporary market "inefficiencies" and trade in ways that allow them to profit before the distortions disappear.

The robots aim to make money on the spread between how much investors are prepared to pay to buy and sell a stock, and buy and sell on both sides of the trade.

While many brokers use high-frequency traders, not all have the highly sophisticated strategies employed by some of the dedicated HFT firms or Wall Street banks.