When I did a transformation for my linear regression (1/x), the coefficient of determination (r^2 value) only increased by 0.1. The original regression was already linear.

Would I say that the transformed one is more effective method than the original linear form or not considering the r^2 value increased but not by much?

Thanks!