I have a couple of semi-ambitious games in my pipe, but they're not close to completion in any way. This is kind of frustrating, doing this over and over again. And to top it of, I have the urge to start coding on more games. That would leave me with even more uncompleted games in the end, but with a more solid framework. That's the only thing that becomes better with every iteration.
Somewhere in my head there's a voice telling me that I should do a custom engine for every game, but then I wouldn't progress and save time in the future. On the other hand, every custom engine would probably be better. The one I have right now supports Unity, XNA/MonoGame and WPF3D. The WPF3D and XNA/MonoGame versions aren't complete at all. They're mainly used for fast testing when I do my game logic. My primary target is Unity since it covers so mant platforms w/o many compromises.
So, what about back to the roots. I've thought about remodelling my old game, Blow him up!, to run on my new framework in Unity. That should be a pretty small feat and would give me some motivation. So I don't feel like a complete failure when it comes to making games. Of course, I do it for fun, but it never feels good to only have unfinished projects in your history.