Draw.RegressionChannel upper and lower channel values.

[JeffCO7]

HI,

I am using Draw.RegressionChannel to draw a regression channel from bars A to B. I would like to get the value of the regression channel upper and lower bounds. If I use the indicator, I can get these values (RegressionChannel(A,B)[0].upper, etc. Can I get the same information when I use the RegressionChannel drawing tool? I would like to draw multiple RCs on the screen and so using the indicator is not my solution.

I see I can return the number of SDevs to the upper or lower channel bounds. But I don't see how to then return what the standard dev is at that point.

Thanks for any help.

JeffCO

[NinjaTrader_ChelseaB]

Hello JeffCO,

I don't think the values you are wanting are available with the Draw.RegressionChannel() drawing tool.

Below is a link to the help guide with the properties available.

https://ninjatrader.com/support/helpGuides/nt8/regressionchannel.htm

The standard deviation to the high and low are available. The start and end anchors are available.

[JeffCO7]

Chelsea,

Thank you. I went ahead and wrote my own method to do the calculation. It uses start and end Bar# instead of BarsAgo which suits my use case. I will still use the draw.RegressionChannel to do the drawings. This method will give me the standard deviations. I left some debug Print statements in the code for others to use if needed. I am also creating an update method to add additional bars to the end of the LinReg calc without redoing the entire calc. It is a simple modification of the init method. That is the reason the sum?? variables are global variables and not defined in the method.

I checked this method against the RegressionChannel to verify the results.

[JeffCO7]

		public double mInitLinRegChannel(	int i_start,  // start bar #
											int i_end )   // end bar #   (should already include Lag)
		{
			double start = i_start;
			double end	 = i_end;
			double mPrice;
			int Idx;

			// NOTE: these parameters are defined in the global variables section of the intidicator.
			nBars	= 0;
			sumX	= 0;
			sumY	= 0;
			sumXY	= 0;
			sumXSqr	= 0;
			sumYSqr	= 0;

			for (int i = i_start; i <= i_end; i++)
			{
				Idx 	= CurrentBar - i;
				nBars	+= 1;
				mPrice	= Input[Idx];
				sumX	+= i;
				sumY	+= mPrice;
				sumXY	+= i*mPrice;
				sumXSqr	+= i*i;
				sumYSqr	+= mPrice*mPrice;

				Print(string.Format("CalcLR C1  Bar:{0}  i:{1}  Idx:{2}   Price:{3}  sumX:{4}   sumY:{5}   sumXY:{6}   sumXSqr:{7}   xumYSqr:{8}   nBars:{9}",
									CurrentBar, i, Idx, Input[Idx], sumX, sumY, sumXY, sumXSqr, sumYSqr, nBars));

			}

			divisor		= ( nBars * sumXSqr	- sumX * sumX );
			slope		= ( nBars * sumXY	- sumX * sumY ) / divisor;
			intercept	= ( sumY * sumXSqr	- sumX * sumXY ) / divisor;

			slopeSeries[0]		= slope;
			interceptSeries[0]	= intercept;

			// Next we calculate the standard deviation of the
			// residuals (vertical distances to the regression line).

			double sumResiduals = 0;
			for (int i = i_start; i <= i_end; i++)
			{
				Idx 	= CurrentBar - i;
				double regressionValue = intercept + slope * i;
				double residual = Math.Abs(Input[Idx] - regressionValue);
				sumResiduals += residual;

//				Print(string.Format("LRC SR1 - Bar:{0}  slope:{1}  intercept:{2}  RegVal:{3}  Y:{4}  Residual:{5}   Sum:{6}",
//									CurrentBar, slope, intercept, regressionValue, Input[Idx], residual, sumResiduals));
			}

			double avgResiduals = sumResiduals / Math.Min(CurrentBar - 1, nBars);

//			Print(string.Format("LRC SR2 - Bar:{0}  AvgResidual:{1}",CurrentBar, avgResiduals));

			sumResiduals = 0;
			for (int i = i_start; i <= i_end; i++)
			{
				Idx 	= CurrentBar - i;
				double regressionValue = intercept + slope * i;
				double residual = Math.Abs(Input[Idx] - regressionValue);
				sumResiduals += (residual - avgResiduals) * (residual - avgResiduals);
//				Print(string.Format("LRC SR3 - Bar:{0}  slope:{1}  intercept:{2}  RegVal:{3}  Y:{4}  Residual:{5}   Sum:{6}",
//									CurrentBar, slope, intercept, regressionValue, Input[Idx], residual, sumResiduals));
			}

			double stdDeviation = Math.Sqrt(sumResiduals / Math.Min(CurrentBar + 1, nBars));
			stdDeviationSeries[0] = stdDeviation;
//			Print(string.Format("LRC SR2 - Bar:{0}  sDevWidth:{1}  StdDev:{2}  2SDev:{3}   3SDev:{4}   4SDev:{5}",
//				CurrentBar, sDevWidth, stdDeviation, stdDeviation*2, stdDeviation*3, stdDeviation*4));

			double middle = intercept + slope * i_end;
			Middle[0] = CurrentBar == 0 ? Input[0] : middle;
			Upper[0] = stdDeviation.ApproxCompare(0) == 0 || Double.IsInfinity(stdDeviation) ? Input[0] : middle + stdDeviation * sDevWidth;
			Lower[0] = stdDeviation.ApproxCompare(0) == 0 || Double.IsInfinity(stdDeviation) ? Input[0] : middle - stdDeviation * sDevWidth;

			double temp = intercept + slope * i_start;
//			Print(string.Format("LRC - Bar:{0}  Divisor:{1}   Slope:{2}   Intercept:{3}   sumX:{4}  sumY{5}, start-Bar:{6}  val:{7}  end-Bar:{8}  val{9}  Upper{10}   Lower{11}",
//								CurrentBar, divisor, slope, intercept, sumX, sumY, i_start, temp, i_end, middle, Upper[0], Lower[0]));

			return (Upper[0] - Lower[0]);
		}