What is a primary advantage of using random forests in predictive modeling?

Question $1$ options:

They prevent overfitting by averaging multiple decision trees.

They require less data preprocessing than linear models.

They always produce the most accurate predictions.

They are easy to interpret and understand.