[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Setting colour on a multi-plot indicator



PureBytes Links

Trading Reference Links

Thanks for the excellent information Dave.  Now to see if I understand it.

> There is a faster way.  You'll need to use a very structured approach.

Being a programmer I should be able to handle this... 8^)

> Start with a custom formula with three lines which are
> something like this:
>
> LineA:=<time period one>;
> LineB:=<time period two>;
> LineC:=<time period three>;
> Mov(C,LineA,S); Mov(C,LineB,S);Mov(C,LineC,S)

I'm using something very similar to this in my System.  The Enter Long is:

Short:=Mov(C,opt1,S);
Medium:=Mov(C,opt2,S);
Long:=Mov(C,opt3,S);
Long > Medium AND
Short > Medium AND
Ref(Short,-1) < Ref(Medium,-1)

Are you talking about this being a custom Indicator or a System?  I think
you mean System because of your reference to OPT1, OPT2, etc.  I have an
Indicator called Triple Moving Average, which is:

Short  := Input("Short Moving Average", 1, 1000, 5);
Medium := Input("Medium Moving Average", 1, 1000, 15);
Long   := Input("Long Moving Average", 1, 1000, 75);
Mov(CLOSE, Short, S);
Mov(CLOSE, Medium, S);
Mov(CLOSE, Long, S);

> You can later change these to OPT1, OPT2, and OPT3 in your
> system test.
>
> Build your next formula based on the first, and give signals
> for enter, short, exit, etc. as 1,2,3,4 or something like that.

This is where you lost me.  Build the 2nd formula where?  Base it on the
first, how?  Do you mean to use fml("Triple Moving Average")?  Are these the
column formulas in an Exploration?

> Then a third formula(Indicator3), based on the second, would
> use Barssince(<your long signal>)<Barssince(<your short signal>) AND
> Barssince(<your long signal>)<Barssince(<your exit signal>) =
> 1 for long or =-1 for short and =0 for out.

Here you're saying if the amount of time since I went long is shorter than
the last time I went Short or Out, then I'm in a Long position and return
+1, right?

> Then you can create a fourth custom indicator based on this,
> the equity line.   This one says, Cum(If(Indicator3=1,Open-Ref(Open,-
> 1),If(Indicator3=-1,-1*(Open-Ref(Open,-1),0).

If I'm Long subtract the prior day's Open from the current day's Open and
add this difference into the accumulator. Otherwise, if I'm Short subtract
this difference from the accumulator, else add zero.  This will keep a
running total of the change in stock price, with an increase when the price
increases and I'm Long, and a decrease when the price drops.  When I'm Short
and increase on a drop in price and a decrease on a rise?  Am I on track?

> Now you've got four indicators based on those three variable
> definitions which are easy to change.  You can run an
> exploration with
> this equity line to guage the performance across a portfolio
> of stocks.

Am I creating four custom Indicators here and using the name of the first
one in the second, and the name of the 2nd in the 3rd, etc.?

> You can write an expert that is based on the formulas and a
> template.  Your arrows for exits and entries as well as the lines
> themselves will adjust whenever you go back to adjust those
> three first lines of the first indicator.

It seems that you're telling me to build custom Indicators, but I've read
that when a custom formula is referenced, which has multiple plots, only the
last one is returned.  I'm confused about how I'd be able to use the first
Indicator to have the second one based on it.

> You can also run a new exploration once you've adjusted those
> three values in the formula.  You can do this as many times as you'd
> like.
>
> I hope this explanation makes sense.  I'm happy to clarify.
> It is quicker, but still has some manual steps.

I'm glad you're willing to clarify because, as you can see, I'm very green
with MS and still need the obvious things included in any explanation.
Sorry if this is more than you contemplated when you replied.  You have me
very intrigued about how this would work - especially for screening through
large numbers of stocks to glean those that are cycling nicely so I can test
different values to optimize.  I suspect that I'm going to find that the
optimal values for one stock are not the optimal values for another stock.
What I think I'm looking for is a set of values that does reasonably well on
many stocks, but not all, but I could be wrong on this.

I hope there's some value in this thread to other readers.  Thanks again
Dave.

Bill

>
> Dave Nadeau
> Fort Collins, CO
>
> 12/20/00 10:53:21 AM, "Bill Irwin" <Bill-Irwin@xxxxxxxx> wrote:
>
> >Thanks Frans & Steve for the suggestion.  Maybe if I'd more
> fully explained
> >what I'm doing perhaps another method is available, and
> would be better.
> >I'm working on a System that's based on this triple plot.
> I'm going to be
> >optimizing the values of this triple MA and, after the test,
> observe the
> >most profitable scenario.
> >
> >What I'd really like is for the System tester to plot the
> triple MA (in 3
> >colours) along with the Buy/Sell/Exit signs so that I can see the MAs
> >crossing at these triggers.  As it is, it appears I'll have
> to apply the
> >Triple MA Coloured Template, then adjust the lengths of the
> MAs to match the
> >winning System (after digging into the winning System to see
> the values of
> >OPT1/2/3 in a window that can't be sized to avoid the extra
> step and time of
> >scrolling).  This doesn't make for a speedy review of a
> given System across
> >many charts.  Of course, having to run the System test on
> each individual
> >chart (rather than on a selection, as in an Exploration) is
> what really
> >takes the time.
> >
> >If there's a faster way of doing this, please advise, and
> thanks again.