depends on the quality of uncertainty i.e. it would be correct, if the beginning number from which  those words were generated was truly random. But that case is ideal. So their entropy in fact is less than it should be ideally.