There are several good ways to guarantee a project will fail:
1. Announce It
If you announce a project before you actually have a working version, it will fail. The lil' boxing game I mentioned previously isn't really a project per se, more of a means to create more tutorials - so hopefully I haven't cursed it. It may never see the light of day, aside from whatever knowledge I glean from it and put inside of my articles.
I can't tell you how many people I've seen announcing "I need developers to make a totally huge project of amazing proportions." And everyone ignores them. Why? No one cares. You can't pretend to be the manager of some huge project unless you know what you're doing. And you don't know what you're doing unless you've done it.
If I made two ugly stick figures headbutt each other until you hit ^C I'd get way more interest than by claiming I'm creating a virtual universe unlike one that has ever been seen by Sony, Lucas Arts or Blizzard.
2. Name It
If you think of a name for your project before you actually have a working version, you're in it for the wrong reasons. A project doesn't need a name until you want to communicate regularly about it, and it doesn't need to come up in that many conversations unless you've announced it. And if you've announced it, you've violated point 1.
3. Promise It
This is the biggest one. If you ever, ever promise something will be released it never will be. Have you ever noticed that the best game development companies keep everything under tight wraps until they go into beta testing?
Hype kills a project. You have a split second to get people interested in your project, and if you don't have enough stuff to fuel there interests then they'll move on. Projects promised on concept art and a five-person development team might generate interest for five seconds, but
unless you have a demo, a beta or at least a smattering of screen shots everything will quickly turn to vapor.
Low expectations! That's what I'm all about!
Damn. Do I sound cynical or what?