TODAY marks the start of International Derivatives Week, at which participants from exchanges all over the world will gather in London. Yet traders accounting for more than half those markets’ volume will not be there. They will be at their desks as usual.

These are not the overpaid and overworked dealers in derivatives. They are the algorithmic dealing programs that clinch bargains on exchanges such as the German-Swiss Eurex, Euronext.Liffe in London and the two Chicago exchanges, the Board of Trade and the Mercantile Exchange, without any human input whatsoever.

It is a little-appreciated phenomenon outside the industry that so-called “black box” trading, carried out by computers using sophisticated programs devised by some of the most brilliant academic minds in mathematics, is on the rise.

To understand how the machines took over more than half the market on exchanges, one must go back to 1998 and the closure of the trading floor at Liffe, the London financial futures market and the centre of derivatives dealing in the City. This put out of work or forced an abrupt career change for the “locals” — traders in colourful jackets who worked for the big firms in the market or on their own behalf. Many went off to do other things; others migrated to the electronic screens on which business henceforth would be carried out. A number of firms set up what were effectively mini-dealing rooms on the fringes of the City at which those dealers prepared to work on screen were offered the chance to rent facilities on which to do so.

Peter Green, an old derivatives hand, is chief executive of Kyte Group. This operates a new dealing room at a business park in Islington, North London, on which about 250 people work, all but 30 dealing on their own behalf in return for the payment of fees to Kyte. Mr Green estimates that three years ago about 5 per cent of the business in Islington was algorithmic. A year ago it was about 20 per cent to 25 per cent; now it is approaching 60 per cent.

This is one factor that has contributed to the booming trade in derivatives — others include the growing involvement of hedge funds. At Eurex, for example, volumes last month were up an extraordinary 80 per cent year on year. At Liffe, volumes for May were up 53 per cent over the same period. It is, says Mr Green, the industry’s second revolution.

He says the growth of black box trading has been driven by his customers, who come to him with increasingly developed systems of their own. The programs receive and monitor every piece of trading activity that registers on the screens. The data is analysed for patterns, and the software programmed to carry out a particular trade if a certain pattern is detected.

It may be as simple as instructions to buy a certain derivative at a given price. For example, if the price of a certain derivatives contract set to mature in June reaches a certain level, the programmers may have decided that this is a signal to sell the September contract. The program is in reality likely to be far more complex, with a long skein of “what ifs”, suggested trades and conditions built in.

The point is that this all takes place within the black box, with no human involvement. It is more efficient than its human equivalent, because the computer can map far more data far more quickly. It is also able to get in and out of a bargain without alerting the market and moving the price.

Such trading is more suited to markets that see huge volumes of business but are relatively inefficient, such as the various world stock market indices, but it is already widely used on exchanges elsewhere, even in equities, always seen as the more traditional end of the financial world.

One industry observer says: “The computer doesn’t read the paper, surf the net, take coffee breaks, flirt with the receptionist, eat lunch or get sick. It just sits there day after day scouring its chosen field for trading signals which it activates in less them the blink of an eye.”

Some wonder if there will be a saturation point for algorithmic trading. On one analysis, it is effective only because it exploits tiny inefficiencies in the market. With more and more dealing done by computer, those inefficiencies become less common. Others believe that there will always be an angle for traders to exploit.

“I’m guessing that we are going to grow and grow,” Mr Green says. “There are so many different ways in which you can use software to enhance your trading.” The notion of a battery of machines trading untended at speeds undetectable to the human eye raises obvious concerns. Mr Green insists that there is an array of risk management programmes that can step in if things start to go awry.

Hamish Purdey, chief operating officer at FFastFill, an AIM-quoted company that makes both the systems and the software that is designed to keep any excesses in check, says: “We apply pre- and post-trading risk management to the trades as they come on all our systems. We can be certain if someone has breached a long-short limit or their margin requirements we can immediately restrict them from trading.”

Times Recommends

• Lies and cover-up at MG Rover exposed by report
• High oil prices push manufacturing costs up
• Primark founder steps down as chief executive