When ID advocates talk about specified complexity they are starting with an existing outcome and calulating the odds of getting from a past condition to the current condition. Such odds are always astronomical if the conditions involve complex systems.
Evolution isn't about predicting specific winners and losers, except in the sense that the house always wins. You can take that to the bank.
Another fallacy that we're using a made-up name for on these threads: Retrospective Astonishment. PH's idea, I think.
How can one practice science at all without "starting with an existing outcome?"