Using R2 with caution

R2 has been in use over the several decades to test the goodness of fit of regression models. However, it has been under criticism over the last three to four decades owing to its diverse limitations as well as its appropriateness for its applicability to nonlinear models. The paper reviews some of the significant work done, especially with respect to its strength and weaknesses, and cautions the users of this widely used statistic in the predictive methodology. The paper also makes a brief mention of the alternative statistics.