TheStumbler: The problem with the common designer argument, as I'm sure you are aware, is that any conceivable reality is consistent with design so the idea is not testable.
Thats the beauty of it. you dont even have to go over falsifiability, the analogy that is relevant is this:
Suppose you have two (supposed!) oracles Alice and Bob. So you ask them for the lottery numbers. Bob give you 3 numbers, alice says: "Ooooh. i sence it is a number greater than 10!"
It turns out one of the 3 numbers bob gave was the right one.
So you come back the next week and ask again, again bob give you 3 numbers and alice just says its a number greater than 10. Again it turns out one of bobs numbers is the right one.
So after a couple of times, everyone would think Bob was really psychic, even though alice was correct and could have been falsified. it has nothing to do with falsifiability, but with the relative narrowness of the hypothesis.
Ofcourse thats not how creation science work. Any creationists worth his salt would say that bob gave you 3 numbers not one, so he would be wrong once in a while, and that alice infact told you the number was greater than 10, and was right most of the time. Therefore, the creationist conclude, Alice is really the psychic one.
Its the same way with this type of evidence, and thats why its impossible to dismiss: even though creation is compatible with the observations, the evolutionary assumption is much narrower and fit the data relatively well. In the example, statistics offer a way to quantify exactly how much narrower Bobs assumption is, ie. to tell us how many bits of evidence evolution gain over creation.
Thats also why its a total red herring that some retroviral insertions are modified and used by the organism... it does not affect the narrowness of the evolutionary hypothesis that the sequence is used.