The one place UTF-16 massively wins is text that would be two bytes as UTF-16, but three bytes as UTF-8. That's mainly Chinese, Japanese, Korean, etc...