Linear Regression of 2 identical DataSeries different results ?

[GKonheiser]

Hi,

I am trying to figure out why a Linear Regression of the Closes[0] , in this example 10min bar closes , and a custom Data-series which holds Closes of 10min bars apart from index 0 which holds the current 1 tick close, Closes[1][0], gives me a different result even when both data Series are the same ?

So for example this is the output that shows the output using a print statement. You can see that both Data Sets Close and fast Close are identical but the Linear Regression results, Average and FAverage, are different

anyone know why this would be ?

22/02/2016 11:00:00 10min Close= 165.14 165.17 165.2 165.22 165.22 165.25 165.24 165.25 165.22 165.13 165.02 164.95 165.08 165.08 165.1 165.09 165.01 164.92 165.04 165.05
22/02/2016 11:00:00 fast Close  = 165.14 165.17 165.2 165.22 165.22 165.25 165.24 165.25 165.22 165.13 165.02 164.95 165.08 165.08 165.1 165.09 165.01 164.92 165.04 165.05
Average  = 165.237
FAverage = 165.192428571428

[NinjaTrader_Jesse]

Hello,

Thank you for the question.

I would be unsure of the reason with the data you provided, do you instead have syntax that can show this result?

Without further testing and seeing the syntax used to see what is different it would be impossible to answer.

Please let me know if I may be of further assistance.

[GKonheiser]

Well the first data series is just Closes[0].


For the second DataSeries, fastClose in my primary BarsInProgress I set fastClose to the primary Bar series on the first bar and then for subsequent bars I am setting 1 bar back to the close of the primary bar 1 bar back:-

if(BarsInProgress == 0)
				{
					if(CurrentBars[0] == 0)
						fastClose.Set(0, Closes[0][0]);
					else
						fastClose.Set(1, Closes[0][1]);


				}

Then in my secondary bar object I set the 0 index of the fastClose to allways be the most recent tick:-

if(BarsInProgress == 1 && CurrentBars[baseBars] > 1)
				{

fastClose.Set(0, Closes[1][0]);


// linea regresion of my cutom series
avg =	Average(fastClose, averageBase, 20); 

// linea regression of Closes[0]
double avg2 = Average(Closes[0],  averageBase, 20);

}

and my average method in UserMethods is:

public double Average(IDataSeries input, AverageBase averageBase, int sDPeriod)
		{
			
			double avg = 0;
				
			switch(averageBase)
			{
				case AverageBase.LRC:
					avg = LinReg(input,sDPeriod)[0];
					break;
					
				case AverageBase.EMA:
					avg = EMA(input,sDPeriod)[0];
					break;
					
				case AverageBase.WMA:
					avg = WMA(input,sDPeriod)[0];
					break;
					
				case AverageBase.SMA:
					avg = SMA(input,sDPeriod)[0];
					break;
			}
			
			return avg;
		}

So you can see from the output in the previous post , even when both series are identical I am getting different results.

[NinjaTrader_PatrickH]

Hello GKonheiser,

Thank you for your response.

Can you provide the full Output from the first bar it calculates to the last? If possible can you attach as a .txt file after copying and pasting the information from the Output to the .txt file?

I look forward to your response.

[GKonheiser]

Here is the output file from a complete run

[NinjaTrader_PatrickH]

Hello GKonheiser,

Thank you for your response.

May I test the full script on my end?
If you do not want to attach it here you can send it to platformsupport[at]ninjatrader[dot]com with 'ATTN: Patrick H - 1469781' in the subject line.