[TylerE]

Because what you’re really measuring is something like how well the block size the interpreter uses for IO matches the underlying os/hardware

[cutler]

If they are both doing similar things and one is faster that's still significant. So Python's i/o is faster than Ruby's.