Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!news.moneng.mei.com!uwm.edu!msuinfo!harbinger.cc.monash.edu.au!news.cs.su.oz.au!metro!news
From: peter@magna.com.au (peter)
Subject: Re: Newbie question on stock prediction
Message-ID: <CyA51C.I0o@ucc.su.OZ.AU>
Sender: news@ucc.su.OZ.AU
Nntp-Posting-Host: enterprise.magna.com.au
Organization: PJB
X-Newsreader: WinVN 0.90.3
References: <38j865$2jn@news.ycc.yale.edu> <38k36n$omp@newsbf01.news.aol.com>
Date: Wed, 26 Oct 1994 12:22:24 GMT
Lines: 21

In article <38k36n$omp@newsbf01.news.aol.com>, tiednbound@aol.com (TiedNBound) says:
>
>In article <38j865$2jn@news.ycc.yale.edu>, jwang@salmon.zoo2.cs.yale.edu
>(Gang Wang (Jonathan)) writes:
>
><I was just
>wondering, are there any proofs that using neural nets in stock prediction
>is better/worse than random prediction?>
>
>Witness:   Most of the volume in the stock market today is the result of
>programmed trading . . . .most of which is driven by a series of very
>top-secret redundant AI systems on the part of large NY-based
>broker/dealers!  Case closed . . . . 

Outsider: Are all of the NY based financial institutions using ANNs, and if they
what about the rest of the world. NY is an important financial centre but
there are anothere sixteen hours in the day.

Is the use of ANNs just like technical analysis, ie feed-back based (every body 
follows the same rules)?
Case unresolved......
