Overfitting can be reduced by using Pruning in Decision tree , as random forest is an ensemble , more tress capture more variance And we could say it results in better performance on test set