There's a chance you'll make a good game, but the odds for your first one are pretty low. They get much better over time, though. I took the path of experimenting with small effects (bumpmapping and stuff like that) to learn the underlying parts and I'm glad I did this. I think that may be a good route to go - forget about writing a game for a couple of weeks and just write small projects for each of the things that interest you (e.g. texturing, lighting, and so on if in 3d, drawing bitmaps, pixel plotting and clipping if in 2d, etc.).

Bear in mind that you'll be giving up short-term productivity here. The component sets are pretty slick and you'll end up with your own. As you do more, you'll slowly start to develop classes for images, fonts, etc. This is healthy. Just don't end up with a forum for it at DGDev, coz I'm having a hard time remembering all the component sets as it is! :roll:

I'd say a good compromise would be to take the source code for one of the kits and have a look around. Search msdn for information (via google using the syntax "site:msdn.microsoft.com your_function_here" to see what it's doing if not clear. Harry Hunt's Xcess GDK has quite clean source code, but the rest are good as well.

It's possible to do a good job without extra component sets. Feel free to try! However, there is one thing to watch out for (and which happened to me recently). If you have a tendency to design things "the right way", and have multiple projects, then you might grind yourself to a halt -- each of the projects will have the same work-in-progress framework so all will have to be updated if you find a bug. Just make sure that you're only working on one or two projects at a time to avoid this.