|
PureBytes Links
Trading Reference Links
|
Steve,
I agree that MetaStock DOES NOT SUPPORT Multi-Period Market Indicators. :-(
Multi-Period Market Indicators CANNOT be created accurately without direct
access to Multi-Period data. i.e. without access to Multi-Period
TimeSeries.
Equis, PLEASE, add a new feature to version 7.0 which allows creation of
new TimeSeries based on a user-selected period. This should be EASILY
capable of being added. (I've done it in other spreadsheets. e.g. Wave
Wi$e).
While you're at it, PLEASE ensure that all indicators can be applied to
OTHER INDICATORS, rather than to an implied C(lose) or P(rice) of the
Chart's underlying TimeSeries. This way we can create Oscillators of
Oscillators, etc, ad infinitum without reinventing the wheel in the
Indicator Builder.
All this not withstanding, here's my offering.
It is NOT PERFECT, because of the aforementioned limitations in MetaStock.
The biggest problem is that the resulting Weekly Stochastic on Daily Data
produces INCORRECT values at the hard-right-edge of your chart. i.e.
exactly where you need it to be correct. :-(
(Equis: I realise this is because I'm "looking into the future with Ref(P,
4) etc).
I reproduce it here to illustrate the point to EQUIS. (Are you listening?)
For those readers who are equally frustrated by this short-coming, may I
refer you to the Fibonacci Trader charting package. Download a demo from
http://www.fibonaccitrader.com/. It costs more than MetaStock but does the
job properly. It also has some other really neat indicators which can't be
reproduced in MetaStock. ;-)
Anyway...
(Steve's not going to like this because it involves the creation of a swag
of supporting indicators :-).
{ Weekly Close }
If( DayOfWeek() = 1 , Ref(P, 4),
If( DayOfWeek() = 2 , Ref(P, 3),
If( DayOfWeek() = 3 , Ref(P, 2),
If( DayOfWeek() = 4 , Ref(P, 1),
If( DayOfWeek() = 5 , P ,
LastValue( C)
)))));
{ Weekly High }
Ind:= HHV(H, 5) ;
If( DayOfWeek() = 1 , Ref(Ind, 4),
If( DayOfWeek() = 2 , Ref(Ind, 3),
If( DayOfWeek() = 3 , Ref(Ind, 2),
If( DayOfWeek() = 4 , Ref(Ind, 1),
If( DayOfWeek() = 5 , Ind ,
LastValue( Ind)
)))));
{ Weekly Low }
Ind:= LLV(L, 5) ;
If( DayOfWeek() = 1 , Ref(Ind, 4),
If( DayOfWeek() = 2 , Ref(Ind, 3),
If( DayOfWeek() = 3 , Ref(Ind, 2),
If( DayOfWeek() = 4 , Ref(Ind, 1),
If( DayOfWeek() = 5 , Ind ,
LastValue( Ind)
)))));
{ Weekly Stochastic }
WClose:= Fml( "Weekly Close" ) ;
WHigh := Fml( "Weekly High" ) ;
WLow := Fml( "Weekly Low" ) ;
KPer:=Input("%K Time Periods", 1, 200, 5)*5;
KSmooth:=Input("%K Slowing", 1, 200, 3)*5;
DPer:=Input("%D Time Periods", 1, 200, 3)*5;
{ Reinvent the wheel here... }
PercentK:=( Sum( WClose
- LLV(Fml( "Weekly Low" ),
KPer),
KSmooth ) /
Sum( HHV(Fml( "Weekly High" ),
KPer)
- LLV(Fml( "Weekly Low" ),
KPer),
KSmooth) ) * 100;
PercentD:= Mov(PercentK, DPer, S) ;
{ Plot the %K & %D lines }
PercentK;
PercentD;
{ You'll need to separate PercentK & PercentD out into }
{ separate indicators if you want different colors :-( }
Heh! I didn't say it was neat! <g>
Drag and drop { Weekly Stochastic } onto a Daily chart and voila!
Not EXACTLY the same as on a Weekly chart, but very, very, very close.
PLEASE NOTE:
This only works for TRADING DAY charts.
i.e. if "Application Properties/Chart Options/Dates"
has "Show Holidays" and "Show Weekends" unchecked.
Of course, these indicators have been written for MS6.5
but should be easily adapted back to at least MS6.0.
Paul Chivers
Western Australia
----------
> From: Steven Buss <sbuss@xxxxxxxxxxx>
> To: metastock-list@xxxxxxxxxxxxx
> Subject: Re: Multi-period Market Indicator Values
> Date: Tuesday, January 13, 1998 12:09 PM
<snip>
> In my intra-day activities, I'm learning to be very clear about where
each
> of these indicators are for 5-6 time periods at once. (1 minute, 5
minute,
> 10 minute, 30 minute, daily, and weekly.)
Paul> You're on the right track!
<snip>
> I should note that it's clear that one can substitute larger parameter
> values for some indicators at the smaller periods to provide an
> approximation of the indicators at larger intervals:
Paul> No. Substituing a larger parameter DOES NOT reproduce the desired
result.
<snip>
> Dinner on me in San Francisco for the
> list member who can successfully do this!
Paul> Tempting... Shall we say 7:30pm? :-) <grin>
<snip>
> Equis should be interested in this too ...
Paul> I should bl**dy well think so too.
<snip>
> ==>> The point in the paragraphs above is that if, in fact, one can't
> closely approximate indicator values for a larger period by adjusting the
> parameters for the smaller period, then, Elder has discovered or, at the
> least, clearly written about an issue that not a whole lot of other
people
> discuss. I can't figure out why this is so. Is Elder really on to
> something?
Paul> Yes, he's onto something, but so are a lot of other people. :-)
<snip>
> Anyone tired of my tirades about this multi-period issue yet? <G>
Paul> Who, me?
> Steven Buss
> Walnut Creek, CA
> sbuss@xxxxxxxxxxx
<snip>
Cheer up, Steve. Either Equis gets it's act together (and learns to trade)
or we go somewhere else. I've dropped enough hints already!
Paul Chivers
pchivers@xxxxxxxxxxxxxx
Western Australia
+8:00 GMT
P.S. I think I'll go lie down now. (I really don't think MetaStock's that
bad).
|