|
PureBytes Links
Trading Reference Links
|
All:
Recently there has been some talk about Standard Error Bands.
I have confirmed that the Standard Error Bands produced within AB do NOT
match the Standard Error Bands produced within MS. I do not care which
one is "theoretically correct." I have a system idea that needs to
generate a signal when the Close crosses above the upper Standard Error
Band, or below the lower Standard Error Band.
It does not work in AB but does work using the MS version of the
Standard Error Band (I really do not care what the name is, I just want
the values to be correct)....
Here is the AB version of the error band:
Pr=Param("Pr",8,2,21,1);
Lr = LinearReg( Close, Pr);
Se = StdErr( Close, Pr );
width = 2;
Uband = Lr + width * Se;
Lband = Lr - width * Se;
Plot(C,"Close",4,64);
Plot(Uband,"Upper",2,1);
Plot(Lband,"Lower",5,1);
Here is the more complex (understatement) expression for the standard
error band from Metastock:
21 period Upper Band (smoothed):
Mov((21 * Sum(Cum(1) * C,21) - Sum(Cum(1),21) * Sum(C,21)) / (21 *
Sum(Pwr(Cum(1),2),21) - Pwr(Sum(Cum(1),21),2)) * Cum(1) + (Mov(C,21,S) -
Mov(Cum(1),21,S) * (21 * Sum(Cum(1) * C,21) - Sum(Cum(1),21) *
Sum(C,21))/ (21 * Sum(Pwr(Cum(1),2),21) - Pwr(Sum(Cum(1),21),2)))
+2*(Sqrt(((Sum(Power(C,2),21)-(Power(Sum(C,21),2)/21))
-((Sum(Cum(1)*C,21))-((Sum(Cum(1),21)*Sum(C,21)/21)))/
((Sum(Power(Cum(1),2),21))-(Power(Sum(Cum(1),21),2)/21))
*((Sum(Cum(1)*C,21))-((Sum(Cum(1),21)*Sum(C,21)/21)))) /19)),3,S)
21 period Lower Band (smoothed):
Mov((21 * Sum(Cum(1) * C,21) - Sum(Cum(1),21) * Sum(C,21)) / (21 *
Sum(Pwr(Cum(1),2),21) - Pwr(Sum(Cum(1),21),2)) * Cum(1) +(Mov(C,21,S) -
Mov(Cum(1),21,S) * (21 * Sum(Cum(1) * C,21) - Sum(Cum(1),21) *
Sum(C,21))/ (21 * Sum(Pwr(Cum(1),2),21) - Pwr(Sum(Cum(1),21),2))) -
2*(Sqrt(((Sum(Power(C,2),21)-(Power(Sum(C,21),2)/21))
-((Sum(Cum(1)*C,21))- ((Sum(Cum(1),21) * Sum(C,21)/21))) /
((Sum(Power(Cum
(1),2),21))-(Power(Sum(Cum(1),21),2)/21))*((Sum(Cum(1)*C,21))-
((Sum(Cum(1),21)*Sum(C,21)/21)))) /19)),3,S)
Can someone suggest how I might recreate the equation (whatever it is
called) so that I can get the same signal out of this "error band"
parameter?
For example, what is the equivalent in AB for "POWER"? I will for now
assume that all of the other functions - Cum(), Sum(), etc are the same.
I know how to convert the MS Mov() formula to the AB MA(). Any help on
Power or Pwr??
If the above is too imposing, can you just translate this into AFL
Power(Sum(C,21),2)/21) ===> ???? in AFL ?????
Thanks for any help,
Ken
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|