Feb 14 1994 was a time when Soros lost 600 million dollars on the dollar ‘s
downturn .
Certainly other hedge funds also lost big time . Suppose we can go back in
history and show a machine lots of intelligent spreadsheet data up to Feb 2
1994 , would that machine with help from neural nets , fuzzy logic and fractals
predict a downpattern ? can we then call this a 600 million dollar machine ?
The brainstation could be 1 such machine .
You can download brainstation at http://www.bnn.com for free .
But first down load a training audio and video presentation on running
Brainstation.
This training file is located at ftp.bnn.com under directory pub/ brain .
Look for the file bstrain.zip which takes 5 to 10 minutes to down load . Be
patient , it is good !
As this training video is a lotus screencam file , you would also need lotus
screencam player.exe to view it The latter is downloadable at the same
location under scr.zip .
Having understood the training video you can then go back to the past and
recreate the future .
You can download soros.txt which is a daily spreadsheet model for dollar mark
Take feb.2 as the right learn range go back 109 points , run for 9000
iterations and see if brainstation could predict the big drop on feb 14
onwards .This was the valentine massacre of 1994 and could it have been
avoided by a machine ?
Certainly Feb.14 was a newsy event .The real test is whether brainstation can
generalise fairly well in many other out of sample cases as well . You , the
user can test it yourself in other files like dmtest15.txt (dollar mark )
and brent115.txt (brent crude ) in the same ftp site. Other files to download
are bs.zip which is our fractal converter and pat.xls which describes our
fractal patterns.
Please email me at davidlee@singnet.com.sg for more details . We are looking
for working partners to expand our business globally.