Problem: losing the MA of a custom indicator.

[pj909]

Is there a problem creating 2 series in a called function?

I am creating Series1 and using that to make Series2 then getting an MA of Series 2. The calculation of Series1 uses an ATR. No problem at this point. But when I added a volume series to fCalcBarVal, I lost the MA. If I don't use the volume series I can retrieve the MA of the Series2 without problem.

if (!bInit) {
     xSeries1 = efsInternal("fCalcBarVal", var1, var2);
     xSeries2 = efsInternal("fCalcS2", var3, xSeries1);
     xMA = ema(nLen, xSeries2);           // losing this value
     bInit=true;
}

...


function fCalcBarVal(v1, v2) {
if (!bInit2) {
     xATR=atr(v1);
     xMAV=sma(v2, volume() );             // when I add this
     bInit2=true;
}
nATR=xATR.getValue(0);
nMAV=xMAV.getValue(0);
// do stuff with these values 
... 
return(BarVal);
} 

[waynecd]

Just a guess since I can't test out your code, but try:

    if (!bInit2) {
        xATR=atr(v1);
        xMAV=sma(v2, volume() );             
        xMAV=getSeries(xMAV);             //ADD THIS LINE OF CODE
        bInit2=true;
    } 

[pj909]

Thanks Wayne.

I tried that in both bInit sections: main() and the UDF. The problem is with the MA created in main(). Raw bar values are created in fCalcBarVal. That series is used by a 2nd UDF (xTrendy) to create the indicator. Then I'm plotting a moving average in main() of my new improved whizbang indicator.

The UDF, fCalcBarVal, computes a value for each bar based on the price data of that bar. When I modify the calculation in the UDF to also use the volume of the bar the MA (in main) of the indicator returns a value of zero.

The MA is fine if I don't add the volume data to the calculation of fCalcBarVal.

(I didn't post the code because it is over 400 lines and would be laborious to go thru. It is functional w/o the volume calc. But, will do so if that would help)

if (!bInit) {
     xCalcBarVal = efsInternal("fCalcBarVal", var1, var2 );
     xTrendy= efsInternal("fCalcS2", var3, xCalcBarVal );
     xMA = ema(nLen, xTrendy );               // the MA of series 2 is lost 
     xMA=getSeries(xMA);                      // with or w/o the get series command
     bInit=true;
}
var nBarVal = xBarVal.getValue(0);
var nTrendy  = xTrendy.getValue(0);
var nMA = xMA.getValue(0);                    //  this value is returning 0


...
return new Array(nBarVal, nTrendy, nMA);
}  // end of main

function fCalcBarVal(v1, v2) {
if (!bInit2) {
     xATR=atr(v1);
     xMAV=sma(v2, volume() );             // when I add this
     bInit2=true;
}
nATR=xATR.getValue(0);
nMAV=xMAV.getValue(0);
// do stuff with these values 
... 
return(BarVal);
} 

[waynecd]

The code would help so it can be tested.

Wayne

[pj909]

Started to upload the efs 2 wks ago and got sidetracked -- here it is.

[waynecd]

See if these changes do what you intended.

See attached study. Changes I made follow:

1- commented out line 75.

2- uncommented menu item for "fpVol"


3- changed line 177 to:

xCalcBarVal = efsInternal("fCalcBarVal", fpATR, fpLen, fpVol);  // returns 2 series 

4- on line 384 added:

if(!isFinite(nVWT)) nVWT = 1;//@@ ADDED (problem was that "nVWT" sometimes evaluated to infinity) 

under:

(nAvgVol != 0) ? nVWT = ( volume(0) / xMAV.getValue(0) ): nVWT = 1; 

5- changed line 328 to:

function fCalcBarVal(_nLA, _nLV, bVolAdj) {//@@ ADDED "bVolAdj" 

Wayne

[pj909]

Thanks Wayne.
That fixed it. Apparently just checking for a zero value wasn't sufficient. Didn't think of using the isFinite test. Really appreciate your feedback.

(now i need to find out why email notfiy isn't working - I subscribed to this thread but never got an email)

PJ

[waynecd]

Glad it worked.

Wayne