>What always perplexes me about journalistic accounts of bulls vs bears is
>that no mention is ever made about the math and computer wizards on Wall
>Street and London etc. that help make the markets function. Faith in the
>stock market nowadays is faith in "android epistemology" and all those
>expert systems that can compute scenarios of future prices at a level of
>complexity that leaves day traders in the dust. Surely LTCM, Goldman Sach
>etc. must have a flipped out night shift tending all those machines and
>programs. Do investment strategies really have a fractal structure?
Hrrrmmmmm .... I disagree that the automated trading systems that exist out in the world are concocting anything which deserves the name "investment strategy". Typically, the computers are responsible for devising hedging strategies for people holding onto options and other instruments that require dynamic hedging. But this is decision support rather than investment strategy -- the computer is just telling you how the latest market movement affects your plan.
One level up in terms of sophistication would be the automated order-crossing systems of the big (ie; not us) market makers, and some of the neural net sets that people have put on NASDAQ and other markets. Here, you're talking about a computer's ability to monitor more screens, and hit the bid faster than a human being -- it's intelligence of a sort, but the computers are still not really making the big decisions.
Actual, big money making or losing bets put on by prop desks and hedge funds are still made by blokes in suits, talking into telephones. I had this discussion with Rakesh a while back, on whether computers had made much of a difference to the business (Susan Strange argues quite persuasively that I'm wrong). But I still think that computers have just increased the productivity of the people involved -- the nature of the business remains the same.
>If so,
>does the meaning of the "business cycle" become a timeworn metaphor? Do
>Markovian and other types of stochastic analysis improve forecasting
>horizons somewhat, or is there no temporal logic built into those programs
>at all?
I've played around with some of these models in the LBS' Computational Finance & Forecasting Lab. I've never seen one that worked worth a lick, no matter how sophisticated the time structure. The trouble is that, on average, stock prices are a random walk. If you want to make money, there's not that much substitute for looking at the underlying fundamentals, using your experience of market psychology, and working out what the market's going to be thinking the day after tomorrow, quicker than the next guy.
>Evidence indicates that the attacks that triggered the Asian
>meltdown were rather well planned. That takes one hell of a lot of
>computing to accomplish.
Not really. It doesn't take that much planning to make a conference call and say "Let's fuck Thailand". After that, the computing involved is just that necessary to execute the orders.
dd
___________________________________________________________________________
_____
---------------------------------------------------------
This email is confidential to the ordinary user of the
e-mail address to which it was addressed. If you are not
the intended recipient, please notify the sender
immediately on (44) 20 7638 5858 and delete the message
from all locations in your computer. You should not copy
this email or use it for any purpose, or disclose its
contents to any person : to do so may be unlawful.
Email is an informal method of communication and is
subject to possible data corruption, either accidentally
or on purpose. Flemings is unable to exercise control
over the content of information contained in
transmissions made via the Internet. For these reasons
it will normally be inappropriate to rely on information
contained on email without obtaining written confirmation
of it.
----------------------------------------------------------
___________________________________________________________________________ _____
---------------------------------------------------------
This email is confidential to the ordinary user of the
e-mail address to which it was addressed. If you are not
the intended recipient, please notify the sender
immediately on (44) 20 7638 5858 and delete the message
from all locations in your computer. You should not copy
this email or use it for any purpose, or disclose its
contents to any person : to do so may be unlawful.
Email is an informal method of communication and is
subject to possible data corruption, either accidentally
or on purpose. Flemings is unable to exercise control
over the content of information contained in
transmissions made via the Internet. For these reasons
it will normally be inappropriate to rely on information
contained on email without obtaining written confirmation
of it.
----------------------------------------------------------