One thing that strikes me immediately about his version is that it almost requires one-letter, or at least greatly abbreviated, variable names to keep line lengths manageable. I also tend to find that reducing the number of variables scattered throughout a snippet of code pretties things up and makes them easier to reason through, thus enhancing that ephemeral quality "readability". Further, the complexity of the code in general is greater in the iterative version, as measured in discrete syntactic elements, which strikes me as (all else being equal) a usually useful metric of pleasant succinctness of code.