# In a random forest, is larger %IncMSE better or worse?

Once I have built a (regression) random forest model in R, the call rf\$importance provides me with two measures for each predictor variable, %IncMSE and IncNodePurity. Is the interpretation that predictor variables with smaller %IncMSE values more important than predictor variables with bigger %IncMSE values?

How about for IncNodePurity?