Goal:
Make a fun game.
Make something you want to make.

That's it.

Nope!

Keep your priorities straight. It should be:

1) Make something you want to make.
2) Make something you will have fun making.
3) Make something you will learn from.
...
1058) Make something that other people will like playing.

Trust me. You really should not worry about making your game marketable (because that's what you're worrying about right now). If you make a game that you love then it will be fun and therefore it will be marketable. But if you make a game because "that would be something people would love to play" or "that would be something that would make me a lot of money" then you're totally in the wrong ballpark.

Facebook was made because someone wanted a tool for easy communication.
Google was made because someone wanted a better search engine.
Dell was made because someone wanted to construct cheap and affordable computers.

Make sense?

Naturally there are some companies started because they want to make money. But I think that's the rare exception. They might hope to make money, but they're not saying "this is the best idea to make money." That logic is inherently flawed.

Hmmm, I think you should follow Demonpants advice, REALLY!!!

Why? Well, no matter how hard you try if you are just a beginner that have no experience in making game, you first game will probably not be fun. So just make what you want to make.

Making a game is fun for about two weeks.

Then it becomes rewarding hard work for about two months.

Then pain begins. Burnout sets in, and you'll start to hate the game and wonder if there are better ways to spend your time.

My advice is: make a game you will finish.  If you don't think you can survive the burnout phase, build something you can finish in two months.  Once you've completed it, the satisfaction justifies all the suffering, but when you're on the road you'll have to keep reminding yourself that believe it.