Re: Time Series Forecast Formula

["Steve Karnish" <kernish@xxxxxxxxxxxx>]

Robert,

I can't answer your question, but here's a nice optimizer that uses TSF:

Enter long:
Cross(opt1,((CLOSE-Ref(TSF(C,opt3),-1))/CLOSE*100)) 

Close long:
Cross(((CLOSE-Ref(TSF(C,opt3),-1))/CLOSE*100),opt2)

Enter short:
Cross(((CLOSE-Ref(TSF(C,opt3),-1))/CLOSE*100),opt2)

Close short:
Cross(opt1,((CLOSE-Ref(TSF(C,opt3),-1))/CLOSE*100)) 

opt 1:  zero to -2 (with .1 step)
opt 2:  zero to +2 (with .1 step)
opt 3:  2 to 8 (with 1 step)

I use this for futures and the above parameters (optimized settings) keep
it in the ballpark.  If you are applying it to equities (or commodities), 
it always makes sense to look at the indicator and understand the outside
parameters for each of it's "steps".  It makes no sense to limit your
outside limits to -2 and +2 if the TSF oscillates between -8 and +8.  So do
a little homework on the "outside" limits of the indicator and then
optimize accordingly.

This might not be what you're looking for, but if it helps you or someone
on the list, it was worth the time of the post.

Good luck,

Steve Karnish
CCT
----------
From: Robert Lambert <lambertb1@xxxxxxxxx>
To: metastock@xxxxxxxxxxxxx
Subject: Time Series Forecast Formula
Date: Tuesday, May 11, 1999 7:07 AM

All:

I would like to know if the following formula (taken from Equis
website) is actually the formula for the Time Series Forecast, or a
modified formula which is simply using the Time Series Forecast as part
of it’s computation.

I’m asking because I’d like to setup a Time Series Forecast of an
indicator as a crossover trigger, rather than use a moving average. So,
if I plug an indicator into the below referenced formula( in place of
the close value), will this particular formula actually give me the
Time Series Forecast of the indicator, or will it give me something
modified?

Thanks in advance for feedback.



The End Point Moving Average was introduced in the October 95 issue of
Technical Analysis of Stocks & Commodities in the article "The End
Point Moving Average", by Patrick E. Lafferty. 
The exact formula for the End Point Moving average is as follows: 

( 14 * Sum( Cum( 1 ) * C,14 ) - Sum( Cum( 1 ),14) * Sum( C,14) ) / (14
* Sum( Pwr( Cum( 1 ),2),14 ) - Pwr( Sum( Cum( 1 ),14 ),2 ) ) * Cum( 1 )
+ (Mov(C,14,S) - Mov( Cum( 1 ),14,S) * (14 * Sum( Cum( 1 ) * C,14) -
Sum( Cum( 1 ),14 ) * Sum( C,14) ) / (14 * Sum( Pwr( Cum( 1 ),2 ),14) -
Pwr( Sum( Cum( 1 ),14 ),2 ) ) )

The above formula plots the last value of a linear regression line of
the previous 14 periods. The Time Series Forecast takes this value and
the slope of the regression line to forecast the next day and then
plots this forecasted price as today's value. 



_________________________________________________________
Do You Yahoo!?
Free instant messaging and more at http://messenger.yahoo.com
----------