by Julian Richards
In the high-finance world of trading in equities, commodities,
and derivatives, software 'robot' traders working in
their own computer-designed online exchange markets would be better
at making a profit than human traders currently working in international
exchanges, says HP Laboratories researcher Dave Cliff.
Cliff is an invited speaker at the prestigious TradeTech 2004
conference in Paris on April 28. TradeTech is the premier conference
for technology in the international equity trading industry. More
than 400 senior executives from leading asset managers, investment
banks and technology providers are expected to attend the meeting,
with many companies sending their global or regional heads of dealing.
Cliff's keynote is entitled Rise Of the Robots: Increasing
Market Efficiencies By Eliminating Human Traders And Eliminating
Markets Designed For Human Traders. It is based on his
research at HP Labs Europe, in Bristol, UK, which showed that
software traders operating in specialized, computer-designed
online auction exchanges can be significantly more efficient
Potential for huge cost savings
to Cliff, such robot traders are currently typically designed by
humans, and they have to operate in online exchange market-mechanisms
that were also designed by humans for humans.
Recent work by Cliff and his research colleagues has demonstrated
that if these robot traders are set to work in their own virtual
marketplace, automatically designed and optimized by computer to
take advantage of the fact that no human traders are present, the
results are even more promising.
Since the amount of money traded on the world's finance
markets is so huge, even small percentage gains in efficiency made
by the artificial traders in their specialized marketplaces could
translate into millions or even billions of dollars a year in savings
Gains in market speed and efficiency
in the technology of adaptive software agents mean that these sophisticated
robot traders can now readily replace human traders in many real-world
international markets and exchanges. But if you know that none
of the traders in a market is likely to be human, you can alter
the market rules and mechanisms to tailor that market to robots,
rather than to humans, and this can result in the significant gains
in market speed and efficiency.
Cliff leads the Complex Adaptive Systems research group at HP
Labs. Before joining HP Labs in 1998, he was an associate professor
at the MIT Artificial Intelligence Lab and is the author of more
than 70 academic journal or conference papers and some 30 patent
Interest in the finance community
In 2001, researchers
from IBM demonstrated that Cliff's ZIP trading algorithm could
consistently outperform human traders; and in 2002 a favorable
feature article in The
the international finance community to his pioneering work on new,
high-efficiency automated market mechanisms designed by computer
rather than by human professional economists.
In his keynote presentation Cliff will discuss using new automated
techniques to allow computers to search automatically for, and
find, the best designs for exchanges and markets, and for robot
traders to work in those markets.
HP is the only major global IT hardware, solutions, and services
company to be represented in the list of speakers at TradeTech2004.
Other speakers include representatives from Allianz Dresdner; AXA
Investment Managers; Barclays Global Investors; Citigroup; Credit
Lyonnais ; Credit Suisse First Boston; Deutsche Bank; Deutsche
Bourse; Euronext; Goldman Sachs Asset Management; J.P. Morgan;
Liquidnet; Morgan Stanley; and Reuters.