Importance of speed in game developement

Imagine a world like this: you have an idea and istantly this idea come into pieces in front of you, on your screen. Mind-reading? Maybe.
But in this blog I want to discuss something important in many ways: how to do not waste time? How to let your time to be important, productive and more more important, get videogames done?
Videogames is just an example, for sure, but I think is something we can understand: you must have one intuition for your gameplay, maybe some images moving on screen and some level of interaction. But many times I've been around with ideas.. but no real example on screen. This kind of activity is named prototyping and in important in many works even in hobby videogame-making.
So when I've read this article, my mind starts to thinking: "find the fun as quick as possible", wow, this is important, really.
Many hours wasted to understand low level engines, when most important in a videogame is gameplay: I can't simply have a professional graphic on my side, so my games can only (after few examples of just copy well established gameplay) innovate, try to find new mechanics and/or fun.
So my idea is: create quick example of little mechanics, try to fix bugs on code and then merge together to make a bigger game. Pretty fun, yeah?
But where to start? From scratch! Don't make so much sense, right?

My idea is to have some little example of strange mechanics: like jump/jump higher and higher, speed, orbit around a point and so on.

Any other ideas?