Iteration

While prototyping, you essentially iterate on the concept, stripping it down to its core.

For example, you may find out that the essence of the game you're trying to make can be described as your unit charging the enemy in a medieval battle. That's your next direction for prototyping, just focus on the charge element, on the enemy's reaction. They can flee if you're charging from one side, or react by firing arrows or standing their ground if you're charging from the front.

All you will see on your screen is going to be just colored cubes chasing each other. But that's already a great start. The sooner you have something playable, the sooner your vision will become more tangible and you will have real feedback not only on its playability but also on its feasibility.

Remember, there's no need to prototype something that's already been proven. There's no point in prototyping running and jumping for a 2D platformer if you already have a strong reference to them. Your prototypes should focus on the USP! For example, the portal gun and its mechanics in Portal are more important to prototype than the behavior of the automated turrets.