Although R is more useful for assessing the raw explanatory power of a fixed collection of variables, adjusted R is often thought of as a better metric for comparing models with varied numbers of predictors