Hey Halt, I know you're still mixed on one-hour compos, but it's one reason why I do it, because it's an hour of my week when I basically KNOW that I should come up with something (anything, even if it's bad, I post it) for the compo because if I don't the opportunity will never present itself again.
So in a way it's the pressure of performing.
Also I find that changing your expectations from the song "sounding like" something to "demonstrating a concept" allows you to return to it more. A lot of us have "visions" of what we want to do but we don't have the skill or practice or experience to do exactly what is in our heads. We try it, it doesn't quite work, and we get frustrated and then do something else. So I find that changing your expectations and instead of saying, "I want to make a sweeping string piece inspired by Mitsuda" you claim something a bit more technical like, "I want to make a song with a I-IV-iii-ii-I progression" or "I'm going to use three guitars and play each part on my acoustic" you can make a song that, even if it's not that good, is at least something that fulfills your goals.
Not everyone can be a Mitsuda in their first few years, and in fact trying to get to grips with the fact that you might not even know how to put the notes together to make a trance song yet (which is often seen to be a simple, repetitive genre) is a bit of an esteem hurdle that is a good one to get over - once you get over it, just practice it and don't worry about whether it sounds "good" - it's much more important to simply "do it" at first and to crap out anything at all.