Residual variance  

Returns the coefficient of determination (R squared). R2 is a measure of the amount of reduction in the variability of Y obtained by using the regressor variables X. 

If R lies in the [0,1] interval the regression model adequately describes relation between Y and X. However, a large value of R2 does not necessariliy imply that regression model is a good one. Adding variable to the model will always increase R2, regardless of whether or not the additional variable is statistically significant. Thus, models that have large values or R2 may yield poor predictions of new observations or estimates of the mean response.

TRegStatsClass Properties