I was not aware until yesterday that Windows (ALT-PRINTSCRN) would capture a
bitmap of the active window into the "clipboard" . I then activated the
Windows accessory "Paint" program and pasted the clipboard followed by
saving the result as a GIF file, serendipity!

So now I have a way of showing my graphical results obtained by processing
historical SP index data into a market prediction plot and have done so at:
http://www.mypivots.com/forum/ making my first ever gif file viewable in
the CHARTS forum under topic "market map" posted by me, davegemini.

My stock-market intraday pattern predictor program (100%% Fortran) really
nailed the last day of 2007 (monday dec 31) pattern. Here is my first-ever
gif posted in that topic (my forum postings may continue thru January if
anyone is interested in following my success or lack of success) predicting
what most think is a "random walk" down wall street.

http://home.earthlink.net/~dave_gemini/test.gif

and today's pattern predicted is:
http://home.earthlink.net/~dave_gemini/jan02.gif

"David Frank" hotmail.com> wrote in message
news:13nmad6cpljdq5d@corp.supernews.com...

I'm amazed at how much this looks like my output, for 1987. I didn't think
to save the source; maybe you could post yours?
--

Gerry Ford

"Und es begab sich"
~ BoM, german version

3 Republican Square
251 Michigan Av.
GR, MI 49503

David Frank wrote:
...

in need of more inputs it appears... :)

--

On Jan 2, 1:05 am, "David Frank" hotmail.com> wrote:

Unless you are willing to sell or give away your Fortran code or
describe your forecasting algorithm in detail, I don't understand what
we should do with your message. Sure, a mathematical algorithm for
predicting financial markets can be coded in Fortran. The question is
whether such an algorithm exists.

I think there is a tendency for the stock market to bounce back after
sharp declines, such as the one occurring today (S%%P 500 down 1.5%% as
of 3:09 EST), and a statistical model such as a Smooth Transition
Autoregression can capture this. One needs a Fortran code for non-
linear least squares, and such codes are publicly available.