Scratch

You guys might think differently about this, but there I believe the path to a good game is this.
Knowledge:
There are many blocks that need more attention. They're used by a few people and have great results. Color effect seems to be underused in some cases. It may seem surprising because so many projects use it, but only for artsy things where the color effect is changed by 1 or something like that. You'll be astounded at what it can do in other cases. Pen and stamp are used in many projects such as 1s1s, but when you learn everything about it, you can come up with something amazing and foresee problems before running into them.
Ideas:
Think about the many uses for certain blocks. If you think some amazing use for the stamp block and the color block in unison you've never seen used before, test it out. You can even make a game around it. I once heard on another site, that "it's alright to think backwards. Think about what you want to happen and about how to make it happen." In this case it means that if you come up with something cool, you can post it online, or you can wait and think about what kind of game it can go into. An awesome textbox can fit into almost anything as it's good enough to suit the needs of the game.
Planning:
Another good idea is forethought. Plan everything out. A few simple questions are: How do you make the program start? What style will you use*1? What forms of gameplay will you have*2? How will you make each form of gameplay? Are you using an engine, or are you going to make a set of scripts that run their own loops? If so, how will you incorporate the different forms of gameplay into the engine? *1 You can have a certain style for your programming. Some use different scripts with their own loops. I prefer to use one loop that runs the whole game with broadcasts. Thus, there is only one game loop. *2 There can be multiple forms of gameplay in a game. Gameplay is how you play a game. If you have minigames, each minigame will have a different gameplay.
I have had several projects where I only had a couple ideas about it. I tried to code them, but gave up very early on. But the program I'm working on right now is thought out. I've spent 10 hours tops coding it and it's already halfway done (the coding, not the actual game). How much time have I spent thinking about it? At least than 4 times that (I got the full idea/outline last Sunday). Not to mention the several projects I had made and things I studied.
Experience:
Make several attempts to make the thing that you want. It might change a little. I've recently been into visual novels. I've made at least 4 attempts at a visual novel that even got remotely anywhere. Each one had a different approach, and they keep getting better and better. I suspect that I might even drop the project I'm working on and make something even better.
Experimenting:
As soon as anything peaks your interest, test it out. Open Scratch in another window and make a small program that tests the result. at least 90% of the time, your program will work. Then you're ready to add it into your program, so close the test window and rewrite the code and add a couple of extra features.
Testing:
Make it easy to test and run the program from the beginning. I had a project that I didn't test because it wasn't possible. It was full of bugs and I took a while to get rid of everything. The project I'm working on now can be tested, except for one feature. Thus, there are no bugs except for any that could be in the untestable part.