Another case of "not even wrong". Putting a feather into a 200o oven has nothing to do with lunar surface conditions. But he's even wrong about what will happen to the feather in the oven! So not only is his experiment totally meaningless, he assumed the wrong result.
That's what makes it a Stundie. In order to be a Stundie nominee, it's not enough merely to be wrong. You have to be spectacularly wrong.
On a more serious note, I never know what to do with those arguments that are wrong in multiple ways. In law you attack all the ways in which an argument is wrong, because you never know which one(s) the trier of fact will deem valid, so you present the entire case. But often in informal debate it's better to stick to one line of rebuttal and follow it to the end. The argument is clearer that way and avoids the impression that you're "all over the place."