In statistics, the coefficient of determination, denoted R² or r² and pronounced “R squared”, is the proportion of the variance in the dependent variable that is predictable from the independent variable(s) (wikipedia).

The R-squared indicator gives a correlation coefficient between 0 to 1 (0 = no correlation , 1 = highly correlated) by comparing the injected data to a straight linear regression line.

This indicator gives the same values as the R2 instruction. Some asked me the code to study the embedded instruction of the platform, here it is.

//PRC_R-Squared coefficient | indicator
//19.09.2018
//Nicolas @ www.prorealcode.com
//Sharing ProRealTime knowledge

// --- settings 
period = 20
// --- end of settings 

Data = customclose

SumX  = 0
SumXX = 0
SumXY = 0
SumYY = 0
SumY  = 0

if barindex>Period then

 for k=0 to period-1 do
  tprice = Data[k]
  SumX  = SumX+(k+1)
  SumXX = SumXX+((k+1)*(k+1))
  SumXY = SumXY+((k+1)*tprice)
  SumYY = SumYY+(tprice*tprice)
  SumY  = SumY+tprice
 next
   
 Q1  = SumXY - SumX*SumY/period
 Q2  = SumXX - SumX*SumX/period
 Q3  = SumYY - SumY*SumY/period

 iRsq=((Q1*Q1)/(Q2*Q3))

endif

return iRsq as "R²"