Do stuff -> Realize errors in design -> Refactor -> Do stuff...
I'd say that most design flaws aren't really apparent until you want to do something with a piece of code that hasn't been considered in its design. With that in mind, you can train your ability to foresee what systems could be used in ways you didn't thought about.
As with most things, it's matter of practice.
That's why people spend 4+ years studying it in school before they become software engineers. Have you tried obtaining formal education?
You're biting off more than you can chew.
Start with Tetris, not a platformer.
Make sure you FINISH it. Starting and giving up halfway through to start working on something else doesn't count.
Don't give up. Stay on track. Don't develop a pattern of unfinished projects. When you finally finish one you'll be so happy for yourself for sticking through it. I've been working on a game off and on for 4 years now. I devoted myself to complete it. When I started I definitely bit off more than I could chew... but that didn't stop me. I was determined. I don't work on it every single day, and I'm an Indie working completely alone... but the more challenges I face the more adamant I become. I now have a working game after many years. It's so rewarding. Maybe when I'm finally done I'll share with the community here. My point is... attitude and intention is everything. Don't let a few little obstacles get in your way. Overcome them. You can do it. You know you want to.. otherwise you wouldn't be here and part of this community if you didn't.
We believe in you. Stick to your guns. If you have to take a huge step back and hit some books.. then do it. Do whatever it takes. 
Do stuff -> Realize errors in design -> Refactor -> Do stuff...
I'd say that most design flaws aren't really apparent until you want to do something with a piece of code that hasn't been considered in its design. With that in mind, you can train your ability to foresee what systems could be used in ways you didn't thought about.
As with most things, it's matter of practice.
You are correct. When I designed my code from the beginning I never actually thought about things as items(collectible or usable by the character), more than a single entity(i.e enemies or passive creatures) and even how these entities would behave simply because I've never actually done anything like that.
After about 2 refactorizations of a project, then not working on it for 5 months, then completely starting it from zero, I was able to make a meaningful and useful design up-front about that project (knowing all the flaws in the previous attempt), and that project is a pretty basic stuff (an automatic test bench with basic pneumatic control, very basic data acquisition, basic user interface). It is simple but needs multi-threading, has to be very reliable and the deadline is close, that's why I needed to design it up-front at all.
It's a learning process that never ends. That's not to say that you won't ever get there. Just relax and work on that next roadblock to progress. As long as it's not a primary source of income, your life shouldn't fall apart as you look for the answers. I've learned / am learning that it's a "slow and steady" road to learn how to develop games. You'll only fail if you quit. Until then, you're still in the game.
The point is for you to love it soo much that you don't give a shit how hard it is.
If you doubt yourself, maybe it's not for you?
I think the important part is learning from what's gone wrong. Even big studios with hundreds of man-years of experience will look back at a project and find plenty of things to not like about how it wen't. The fact that you're having difficulties isn't a sign that you're too stupid for this, or that you're not cut out for it, it just means you've taken a journey into unknown territory and discovered a dead-end. Just back out and try another path. It can be frustrating, sure, but how many hard things in life can you expect to get from A to Z without making some mistakes along the way? This applies to life in general, not just programming. The trick, I think, is to find something you like doing enough that you're willing to put up with all the temporary setbacks. And it feels good to defeat those problems you've been having for hours, days, or weeks, when you finally get to claim victory, flip your computer screen two birds, and do your happy dance.
Besides, when you think about it, you often learn less from straight-line success than from failures. Experience is not all about knowing the happy-path, but also about knowing which paths lead to unhappyness. For example, if I'm sitting in a design meeting and some good-sounding idea is proposed, the person who's been down that path and can say with authority why its a bad idea is more valuable than the guy who's never ventured far from the happy-path and goes along with it because it sounds like a good idea. Success in most things is not about knowing where you're going, but knowing--or finding out--where not to go.
