|
PureBytes Links
Trading Reference Links
|
OK, let's not get into a "religious war" about parameters and optimization
!!
I love to use them EARLY in a system's development and also for an UNKNOWN
system to see the cause and effect.
In the case of Demark Sequential, I for instance wanted to see if the SELL
SIDE 9-4 defaults should have been 11-5 for instance.....and just maybe the
9-4 should become 11-5 if a certain change in the market takes place.
Without optimization, you'll never know.
Optimization studies make you keenly aware of your system's "chinks in the
armor". More importantly, how the win/loss ratio and win% relationship
varies by parameter and assumption.
After those studies are done and all of the correlations are made, then the
parms can be reduced to a few.
The "Dual Thrust" system by Mike Chalek takes this approach. What he doesn't
tell you is that his 2 critical parameters kick-off the settings and
formulas for a dozen INTERNAL parameters that you never "see".
They are mandatory for setting stop loss and scalping/quick gain amounts
(expressed in percentage of course)
I believe many developers cannot really use optimization AT ALL in their
systems because they have them built and expressed in POINTS......that is a
big, BIG , B I G "no - no". It makes the system totally non-portable for
one thing.
> -----Original Message-----
> From: Dennis Holverstott [mailto:dennis@xxxxxxxxxx]
> Sent: Friday, July 14, 2000 9:26 AM
> To: Omega List
> Subject: Re: Demark Sequential - a good study....one more time
>
>
> Thanks for the sequential indicator code Robert. I got it several times.
> <grin> However, I think we need to see Simms' system code to be able to
> tell why it might have suddenly quite working in 1998. I've got some
> ideas about trends, volatility and overoptimization but they are just
> guessing without seeing the code.
>
> --
> Dennis
>
>
|