A guy named Prospectus put out a free indicator for ThinkorSwim.

https://readtheprospectus.wordpress.com/2009/01/30/volatility-based-trailing-stop/

He designed it as a trailing stop, but I've found that it's very close to Step MA and that it might be helpful as a trend indicator.

Long story short, I'd like to backtest this indicator against my /ES tick data and see how it does. I want to recreate the indicator in NinjaTrader.

The logic is pretty simple.

# Calculate average range for volatility:
#
def atr = expaverage(high-low,length);
#
# Calculate initial short and long volatility stop values:
#
def svs =low+ceil(mult*atr/ticksize())*ticksize();
def lvs =high-ceil(mult*atr/ticksize())*ticksize();

To complete the first part, I basically need the EMA of the (high - low) price.

So what's the best way to calc the EMA for something that's not a price or a dataset? In this case the difference is just a floating point number. Should I set that into a new dataset? Should I use an array?

Any help is appreciated. Thanks in advance.

Create a Data Series and then do the calc for each bar in the Data Series.

HighLowSeries.Set(High[0] - Low[0]);

EMA(HighLowSeries, length)[0];

Curbfeeler 

So what's the best way to calc the EMA for something that's not a price or a dataset? In this case the difference is just a floating point number. Should I set that into a new dataset? Should I use an array?

Since it's just an ema, you can do this (for an object variable 'double ema'):

 

Code

ema = ema + (2.0/(len+1.0))*( (High[0]-Low[0]) - ema );

... but if you want to use ninja's built-in EMA indicator, then define a dataseries 'highlow', and initialize it in Initialize():

 

Code

highlow = new DataSeries(this);

... and use it like so:

 

Code

highlow.Set(High[0] - Low[0]); ema = EMA(highlow, len)[0];

eDanny 

Create a Data Series and then do the calc for each bar in the Data Series. HighLowSeries.Set(High[0] - Low[0]); EMA(HighLowSeries, length)[0];

NinjaTrader has already everything built-in:

EMA(Range(), length)[0];