Newsgroups: comp.ai.neural-nets
From: Philk@cindex.demon.co.uk (Phil Kellingley)
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!news.sprintlink.net!demon!cindex.demon.co.uk!Philk
Subject: Re: stock prediction :(
References: <CwuBDt.DJy@eng_ser1.erg.cuhk.hk> <36e5bi$6bc@pheidippides.axion.bt.co.uk> <36hejh$q7v@nof.abdn.ac.uk>
Organization: Inspirational Systems
Reply-To: Philk@cindex.demon.co.uk
X-Newsreader: Demon Internet Simple News v1.27
Lines: 19
Date: Tue, 4 Oct 1994 07:08:32 +0000
Message-ID: <781254512snz@cindex.demon.co.uk>
Sender: usenet@demon.co.uk

In article <36hejh$q7v@nof.abdn.ac.uk> t01cjd@nof.abdn.ac.uk "t01cjd" writes:

> .....you can find, as I did, that
> there are small amounts of predictable data in the stock market and no
> predictability from random numbers (if your generator is good enough). 

If that were true - you'd be rich. However, it is comparatively easy to 
retrofit (if that is what you were doing) predictions to data. I find it
hard to believe that there are thousands of professionals - fund managers,
technical and fundamental analysts, etc. - who <haven't> found a predictor
and yet you have. The simple test is to place a bet of one unit (at evens)
for your prediction for the following day to be an up or down movement.
Try it - from now. And let's see how much you make in, say, 20 days. A profit
of more than about 3 units would be considered good.

-- 
-------------------------------------------------------------------------
Phil Kellingley
"it's not the same thing at all"
