Profile
last few months that there is no such thing as a technology to ask the kind of questions and
normal distribution. One of the most important do the kind of investigations we are doing
Dietmar maringer cV
things I learned at Cambridge was the extent today. Now it is regarded as a very useful education
to which we are making assumptions and tool. We can rely on standard hardware
1993 ma business Information Systems,
simplifying the real world. and standard software packages and still
university of Vienna and university of
‘If we assume everything is normally investigate the non-linearities.’
Technology of Vienna
distributed and linear we are probably missing Part of Maringer’s research has involved
1996 PhD university of Vienna
1997 m.Phil. university of cambridge, uK
most of what happens in the real world. If you agent-based modelling of financial markets.
2004 PD (habilitation) at the university of
look at the FTSE (London Stock Exchange The behaviour of markets has always been
erfurt, Germany
Index) over the past 25 years there are difficult to model if you assume that people are
employment
17 events that have been more than four making decisions for good reasons. In financial
1993-2002 university of Vienna,
standard deviations from the mean, 10 events markets there are events that create herd
Department of finance, lecturer
with five standard deviations and one event behaviour where investors flock to or from a 2002-2005 university of erfurt, Department
at eight standard deviations.’ particular investment in the same way that of econometrics, lecturer
Maringer returned to Vienna to complete wildebeest move between waterholes on the
2005-2008 co-Deputy director, centre
his PhD and started taking an interest in Serengeti, and usually for similar reasons.
for computational finance and economic
using artificial intelligence methods to solve Maringer’s approach is to create a number
agents, university of essex
problems in finance. In particular he used of independent agents and give half of them a
2008 to date university of basel, Professor
heuristics and evolutionary methods applied ‘trend following’ behaviour and the other half
of computational management Science
and visiting Professor at the econometrics
to portfolio management. He started attending ‘fundamentals’. The first group makes decisions
Department at the university of Geneva
computational finance meetings and, while at based entirely on what happened yesterday and
one of these, he met Peter Winker, who had the second calculates a price for a security based 1980s was that the trading programs were not
just been appointed professor of economics on knowledge. The behaviour of the market very sophisticated and the sell orders from one
at the University of Erfurt. He was also alters wildly according to the proportion of machine trigger sell orders from others.
interested in the same type of research and trend followers and fundamentalists. ‘A lot of the hedge funds now have trading
he invited Maringer to join him. Winker is a
‘As the world sits on
algorithms that are their own and so take the
mathematician by background and they have place of the human trading decisions. We need
collaborated on many published papers.
the edge of a financial
to know what effect this algorithmic trading
After three years at Erfurt he moved
precipice it may be a good
has on the markets and modelling is very useful
to the University of Essex, where there
time to reflect on the fact
for this. We can set up a model of different
was an established research programme in algorithms trading against each other. We can
computational finance and using artificial
that little fundamental
then run some experiments without ruining
intelligence to solve business-related
research has been done on
our stock exchanges. It is not enough to just
problems. In 2008 the University of Basel
how financial markets work’
suspend machine trading when certain events
decided to set up a new chair in computational happen, because that in itself completely
management science and Maringer landed He says: ‘We can come up with reasons changes the composition of the market.’
the position. It was the chance to set up a why certain things happen. We have found He says that one of the complications of
new research group, which he hopes will be that extreme events usually happen when financial markets is that the most rewarding
international in composition, in a country the composition of the market changes. If outcome is often obtained by those whose strat-
famous for its banking industry. you have a bull market or a bear market, egy is adopted by the minority rather than the
He says: ‘To my knowledge it is the first things can suddenly switch when there is majority. For example, those who are buying a
chair in computational management science. an outside influence that changes people’s security while the majority are selling stand to
There used to be a chair of computer science behaviour. This can often be events taking gain the most from the restoration of an equi-
within the economic department, but that was place in another market. What we are doing librium or fair value. He says that this is very
around database management and business is modelling the individuals in the market difficult to analyse using conventional game
systems. They wanted to get involved in the rather than the market as a whole; and those theory, but it does lend itself to modelling.
newer fields like simulation and optimisation. individuals can change its behaviour. What One of the most surprising aspects of
I am now setting up the team and determining people tend to do is assume that everyone this work is that it is not besieged by hedge
the research agenda here, which will be behaves the same and averages the market, fund managers offering grants for research.
computational economics and finance.’ but in fact everyone is different and the Maringer believes that people are waiting
The whole field of computational finance average person does not really exist. to see if this work is consistent with actual
was very small when Maringer began his ‘To understand the markets, you also have recorded market data first. There have been
research and there were many people who to understand the machine trading, which plenty of false dawns in the past. Secondly
believed that this kind of research would never accounts for about 50 per cent of the volume there is a self-belief among traders that their
really get anywhere. Maringer says: ‘Five or of trades on the London Stock Exchange. One experience is far superior to knowledge gained
10 years ago we simply did not have the of the explanations for the crash in the late from computer modelling.
www.scientific-computing.com SCieNTifiC CoMPUTiNG WorlD february/march 2009
9
SCWfeb09 pp08-11 Profile.indd 9 4/2/09 10:41:09
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44