Message:
If we or other come up with good ideas, does that help you to improve
the models that produce the ANO? Or, have you done enough research that we figure
that improvement is not likely to occur? And, we will need to create our own
more private models?
If it were to me to look into some more ideas, the 2 next things I would
do would be:
1) to create a value which is the difference between the % change of the
DJIA and the % change in copper prices that day, and see how that affects or
helps predict the next day's djia change;
2) to create the same value for the djia and the dax.
If a person uses JJC for copper, the results are fairly strong, but it
doesn't always work--as with everything!
but jjc only goes back to October 07 . . . so it might not be so easy
to use that to test things prior to that time.
dz
|
Response:
I'm always open to new ideas. In fact, I hardly ever stop testing new
models to see if any improvements in accuracy can be gained. I do ask that you
send you ideas via email as
this board may become too busy.
Please note that adding
new variables does create a problem. I need to have daily data going
back to January of 1982. The data types currently in my database
are shown on the About page.
|