It's certainly complicated and my answer probably does not apply to such a trivial example code that is using mostly stdlib functions. But the potential gain fo

It's certainly complicated and my answer probably does not apply to such a trivial example code that is using mostly stdlib functions. But the potential gain for Haskell is with loop fusion. The user could implement all of those functions themselves (sum, take, etc) and then loop fusion will bind them into a single performant loop that the lower optimizers will turn into a straight-line code, whereas in procedural languages only stdlib functions and certain patterns of general functions get the fusion treatment. Worse, throw in an unfortunate alias into the procedure and the whole thing collapses into sequential loops.