> We then gathered the most efficient (i.e. fastest) version of the source code in each of the remaining 10 benchmark problems, for all the 27 considered progra

> We then gathered the most efficient (i.e. fastest) version of the source code in each of the remaining 10 benchmark problems, for all the 27 considered programming languages. I wonder if anyone has attempted to rate how idiomatic or typical CLBG entries are, and whether choosing the most idiomatic implementation for each language would have obtained different results?