Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!uknet!gdt!bsmail!danno!wildgoos
From: wildgoos@danno (E.A. Wildgoose)
Subject: Making a killing on the stock market
Message-ID: <D5u81n.Lv9@info.bris.ac.uk>
Sender: usenet@info.bris.ac.uk (Usenet news owner)
Nntp-Posting-Host: danno.cs.bris.ac.uk
Organization: University of Bristol, England
X-Newsreader: TIN [version 1.2 PL0]
Date: Wed, 22 Mar 1995 10:34:34 GMT
Lines: 39

Suppose that the stockmarket is greatly influenced by the way people invest in
it.(reasonable)

Then because everyone goes to the same business schools they all learn the 
same basics of business finance, ie stocks go up, then at this point will go
down, but the reason that the market behaves like this is because the poeple
investing on it expect it to, and invest in such a way as to perpetuate it.

Now, everyone does actually invest slightly different for various reasons and
so most of the predictability must be lost by the sheer volume of different
investing strategies tugging the market around.

BUT, if someone started selling a neural net (say) which was able to gain a 
few percent advantage over standard techniques then everyone would presumably
start to clamour for a copy.  Although the selling would start very slowly
it would not be long before demand increased, at which point enough people 
may be using your program that Mark 2 has to take into account of the influencesof mark 1!  Mark 3 could be built before Mark 2 had even started selling becausewe could try to predict the number of people who would buy Mark 2 (and 
competitiors products) and start to assume that these would be the dominant 
market effects!

Perhaps it would be possible to turn the stockmarket into a huge race to get
the next AI program on to the market which predicts the effects of all the
others!

Thinking about this further though, this would eventually lead full circle
back to the starting point where we would have so many different market 
predicting computers that predicatability would be once more lost in the huge
tug that each different computer would exert on the market...

Still there is the potential for a killing to be made while this transition is 
taking place?


Thoughts on a postcard to:

		E. Wildgoose

(who knows nothing about stockmarket trading what so ever)

