Working Incentive

This seems to be a big deal for individual game developers who aren't getting paid for what they are doing. You start a project with a grand idea in mind, get it to a playable point, then you go into the dreaded state of level design, and if you don't have a smartly designed level editor at your disposal then you can get yourself into a tedious rut. I'm assuming this is the point where most of us go,
"Aw screw it, I want to go make my awesome amazing super game idea #2 now"
(You know, the one you've been planning all along whilst working on the first game idea you used to think about as you went to sleep but not anymore because you are stuck on level design)
There are a few ways to look at this problem, one of them is acceptance.
Jonathan Mak (Everday Shooter) said something like this. Every time you make a game, you learn something.
Wether it's to never ever try something like that again, or it's to realize that you could have made a script that could have solved that entire problem in the first place. Every time I start a new project I automatically think ahead because I recall the billions of problems that I had come across before. And when you do come across more problems, you are only working forward, knowing that when you come into those problems again, you will know how to handle them... (With a punch to their face!)
I make games for fun, no one is paying me, and I'm lucky to finish something because most of the time my motivation dies with glitch fixing. One solution I found, but it is indeed a hard one to keep up, is to constantly add content. It's tough, because you have to resist from going in and fine tuning things (These sparks are going to look PERFECT)
But when its time to go to bed, you can look back and say:
"Damn. I made a boss that awesome in one afternoon? Sure he has a limp and gets stuck in the wall sometimes, but he goes through 5 different stages of attacks and shoots some pretty awesome complex patterns"
The only games I ever finish quickly are the multiplayer ones. Being able to play with my friends after working on a project is enough incentive to get it finished. So if I had the capability to make an online game, it would be done in a flash.
This is another brain blurb, but I think the following option should be embraced: Procedural Generation. PG (done correctly) offers so much replayability that it even takes the developer a long time to get bored of his/her own project. When I was making Armed Generator Doom Machine, I ended up playing the game when I rested after working on it. Which is weird. I normally detest playtesting for the thousandth time. (Though currently I can't even touch AGDM without going "ugh..." unless aided by the company of a friend)
The game I am currently playing is extremely aided by procedural generation. It is pretty difficult, but it restarts quickly, and the experience is different every time. It is none other than Spelunky by Derek Yu

I've played it more than 135 times, but it is still ever-ready in my games folder. So um.. let me think of a conclusion or something here...
Uhh.
I think it's alright to ditch projects that you don't think are worthy to show the world,because you should never be making a game that you aren't having fun making. I think you should finish projects that you've invested lots of time and thought into, but you should never force yourself into finishing something that you just don't have any ounce of passion for. In the end, you will get a few projects done, but it is nice to know that everything you have up are things that you truly love.
At this moment I have a shooter that I spent a few days on. It has two levels and two bosses, and I think one of them is pretty awesome. I don't really want to keep working on the game, because I got my action fix from making it. I'd like to release it because it is still playable and I think someone might find some enjoyment out of it, but it is not polished enough for release.
What has been happening for a while is that I have been going back to old projects and adding onto them or fixing old bugs that I didn't know how to fix at the time. Slowly (in a yearly process) I think old projects will find their way to the finish line.

Hell, like Grid Shock. At this point I'm not too fond of it because I could have reprogrammed it a lot better, but technically it is finished. To release it after "Armed Generator Doom Machine" which is something I like a lot better, seems a bit silly.

But I will release it, eventually.

Who knows, I could be releasing that said shooter by the time I am part of an actual company, but would the world want it by then?