|
PureBytes Links
Trading Reference Links
|
Hi Folks,
Recently I've been trying to solve exactly this problem - how to determine
which strategy to use when.
I admit I could use some help.
I have developed 2 strategies (trend following, and breakout) that each seem
to work great for certain markets, until the performance starts to decay
"mysteriously" - then I switch to the other one. Clearly this is not an
efficient system (just like driving blind until you run off the road).
But here's what I *think* I've figured out so far:
I have developed 2 indicators - basically price and volume oscillators.
When the oscillators seem to be "out of phase" the break out strategy seems
to work best.
When they are "in phase" the trend following system seems to work best.
Now here's the hard part.
Both oscillators are really "noisy" so it's hard to tell if they are really
in phase or out of phase (smoothing doesn't seem to help).
Is there a mathematical formula to determine to what degree 2 signals are in
phase?
I have tried just subtracting the 2 signals to get the difference between
them, but since they both oscillate around zero it gets tricky. There must
be a better way - any ideas?
Thanks!
Chris
Chris@xxxxxxxx
----- Original Message -----
From: <fwalaval@xxxxxxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Monday, August 12, 2002 8:39 AM
Subject: trending & non-trending markets
> At 12:45 AM -0700 8/10/02, michel rood wrote:
>
> >The thought is that there are three markets for sytems: trending,
> >non-trending and volatility. But I think in these three markets other
> >subgroups can be found, like e.g. trending, with high volatility,
> >trending with low volatility. For developing a trading system it
> >would give better results if one trades a system only on such a
> >subgroup. The question is if anyone knows more subgroups.
>
> Unfortunately, someone forgot to tell the market that it should be in
> only one mode at a time so it does what it does.
>
> It is easy to make different trading systems that do well in each
> kind of market. Then, this only leaves you with the impossible task
> of deciding which system to use when...
>
> Bob Ful
>
|