Need Help- Same Output with different DataSeries

[Trendseek]

Hi There,

Below you will find a little Indikator. It Plots the ATR and StdDev, to give a hint for Trend / No Trend.

If you switch on the Smooth-Option, both Lines will smoothed with an EMA, the Code is like this

dstddev = EMA(StdDev(period), smoothperiod)[0];
datr = EMA(ATR(period), smoothperiod)[0];

(---> for more details see the .zip File below):

THIS DID NOT WORK - The Data of the second EMA are identical to first EMA.

How can I solve this problem?

Best
Trendseek

[NinjaTrader_Bertrand]

Trendseek, have you tried just plotting either value alone, does this calculation then hit home as you would expect?

[Trendseek]

Hello Bertrand,
thanks for your fast response!

I just try it out - it is very easy to modify the Code below in that way.
The calculation works as expected if it only plotting one instead of two values.
It also works, if it use two different averages for smoothing the values (i.e. EMA for the StdDev and SMA for the ATR).

I am just a Beginner with C#, but perhaps it is because of the lokal visibility of the used EMA-method (or better: the data.set, using by that method)?
I supose, that the first with EMA smoothed Dataline will be overwritten with the second One.

regards from Germany
Trendseek

PS: I am using the actual Ninjatrader 7 Version.

[NinjaTrader_Bertrand]

Thanks for reporting this in, I saw the same issue on my end and we're looking into.

[NinjaTrader_Bertrand]

Trendseek, thanks for the sharp eye and for reporting, this will be fixed in the next beta round.

[Trendseek]

Hello Bartrand,

thanks again for your quick answer - this are good news, and I will looking forward to the next Update :-)

I made verry bad experiences with serious problems like this in other Trading-Solutions (the german software Tradesignal). This positiv Response will affirmed my decision, to work a little longer with the Ninjatrader.

Greetings from good old Germany
Torsten

[NinjaTrader_Bertrand]

You are welcome Torsten, this will be fixed in the next NT7 update, however your code works with no issues in the current production release of 6.5.

[Ralph]

...perhaps it is because of the lokal visibility of the used EMA-method (or better: the data.set, using by that method)?
I supose, that the first with EMA smoothed Dataline will be overwritten with the second One...

Trendseek,

EMA(...) is a method which is provided by a NT cache-mechanism. That approach checks the signature of EMA() to select the correct indicator instance. You could try something else, if you like:

private EMA ema1, ema2;
private StdDev stddev;
private ATR atr;
...
protected override void Initialize()
{
atr = ATR(period;
stddev = StdDev(period);
ema1 = EMA(atr, smoothperiod);
ema2 = EMA(stddev, smoothperiod);
}
...
datr = ema1[0];
dstddev = ema2[0]

This approach selects the indicator instances explicitely.

Regards
Ralph

[Trendseek]

Hello Ralph,

thanks for your helpfull advice!

beside, that this problem was fixed in the new Beta, I will definitely give them a try - i like that manner of Variable-Declaration ;-)


Regards
Torsten

[Ralph]

Hello Torsten,

and this approach (explicit indicator declaration) executes faster because it skips the cache look-up procedure.

Regards
Ralph