You may have the best idea on paper, but software will show you if it is fun or not.
So I started a simple challenge a while ago: write one idea of a small game mechanic every day and prototype it. The goal was to push me to learn blueprint in Unreal 4, but this time with fast test scenarios rather than going to wide with complicated projects.
Some ideas were fun. Some were not. Here are two examples of prototypes: one fun, one not.
Air Hockey game
Push the puck and get it to the goal. Simple, but one rule: the more you touch it, the more your score goes down. Why? To push the player to throw it again the walls and make it bounce. It is fun and I started working anew mechanics on top of that (collectables, bouncy walls…).
Runner with slomo
While following the Unreal 4 runner tutorials, I wanted to add a “SuperHot” feel to it: a very fast runner in which you can slomo to make better turns and avoid obstacles. It sounded interesting, but I quickly get the problem that the character was going too fast and it made it too hard, yet the character needed to go fast to make slomo useful.
One way to fix it was to take out the slomo and make it that the character can become invincible, but I thought that was really overdone in so many games.