They were one of the first widespread ensemble methods (based on bagged decision trees), and share with other bagging/boosting methods fairly good empirical performance on a wide range of problems.
But then, some models are pretty much impossible to understand or interpret.
You can, however, calculate importance scores for the features used. Brennan's original paper gives a good algorithm for doing this (in short: for each tree, permuting the data along some feature for an out of bag sample and seeing how much worse it does.)