I think that IainPeregrine's Out with the Old, In with the New post has a good message, but here's my take on the reason why projects tend to not be completed and what developers can do about it.
I suspect the biggest reasons that projects don't get finished are:
1. You just lose interest. It happens. There's not much we can do about this, even the best ideas won't keep your interest forever.
2. You're afraid to fail so you set yourself up to fail so it's less disappointing when you do, indeed, fail.
This is like the kid who takes half-court shots when forced to play basketball because he's expected to miss, so he doesn't really look like a failure - if he tried to make a layup and missed he'd look like a loser, but missing a half-court shot is expected. The half-court shot has the added benefit that if he comes close to hitting the backboard or rim he feels successful, even though it'd still be worth zero points if we were keeping score. These are the people who set out to make epic games that never quite come together.
The problem is that these people don't understand all it takes to make a game. They might work for months, give up, and think "I was almost there, the game was almost playable". But that's not how it works. If you split a game into engine and content, the epicness comes from the amount of content. BYOND hands you most of the engine, so even the most epic of undertakings should be playable early in development. It might lack content and isn't as epic as you'd like, but it's playable.
3. The game is better as an idea than it is as a game.
These are the people who have the environment and story planned out but never seem to make progress on the game. What I suspect happens is that the game is perfect in the developer's mind, so anything they create won't do their imagination justice. You have a vision of what the game will look like, but once you start coding it up the game falls short of your vision so you turn to the less concrete aspects of development. You end up planning what the whole world will be like, but you've still made no progress towards creating the world.
This is similar to problem #2. Instead of being afraid of failure (in other words, being afraid of disappointing other people) you're afraid of disappointing yourself.
For problem #2, think about what features you want the game to have and split them into "engine" and "content". You might want to have a complex crafting system in your game, but what creates the complexity? Is the system (the engine part) complex? Or is there just a lot of content? Focus on getting things to work and worry about content later. The system either works or it doesn't, you have to pass some threshold to get it to work. There is no such threshold for content - it doesn't matter if you can craft 10 items or 1,000 items. If you can craft 10 items and the system works, it's playable. If you could craft 1,000 items but the system doesn't work, it's not playable.
For problem #3, understand that your game will evolve over time. Each step you take doesn't have to be perfect. Even if you make the perfect grass icon, your game still doesn't live up to your vision of it because your vision is of more than just a grass icon. You can't let yourself get bogged down with every little detail. Get things working so that the game becomes playable and the details will fall into place over time.
I don't like articles that say "here's what I did, it worked for me, so that's how you should do it." But this approach is what I did and it worked for me, so take my advice if you'd like. A Miner Adventure started out with very simple graphics (screenshot) but these placeholers let me work on more important things. I could add the basic features (digging, lighting, bombs, flowing water) and improve the graphics later (screenshot). The first screenshot wasn't how I envisioned the game, but I knew that it would eventually take shape.
These problems stem from a fear of failure and a lack of understanding of what it takes to make a game. The problem with these problems is that when you fail you don't learn much about the process. You're not likely to improve. To go back to the basketball analogy, you could spend months practicing half-court shots and never see improvement. Your chances of making the half-court shot are as slim as when you started.
You don't have to make small, simple games, but understand that even epic games should be playable early in development. Don't use the epicness of the game as an excuse for why things aren't coming together. Your game doesn't have to be complete in a few weeks, but you should have something playable by then. If you've been making blog posts about your game for a month and don't have screenshots to show, something is wrong with your development process and unless you make some significant changes, it's likely that your game will never be finished.
What has stopped you from completing games? Why do you think that some people never seem to make progress on their games? What you do think can be done about it?
Mar 26 2011, 2:43 am