Not a Knuth exercise, I know, but no less important

Mr Steve Mayer argues that it is wrong to start with what you want to prove end up with a true statement. Given that I've been doing just that for quite some time now, I have no choice but to disagree.