Choose one action
Pick an action you can describe with a verb: jump, steer, stack, hide, or trade. The first prototype exists to explore that action. It does not need an inventory system, a character creator, or a branching story.
Write down the moment you hope will be enjoyable. For a jumping game, that might be landing on a moving platform after a narrow miss. Build only enough to reach that moment.
Create a complete tiny loop
A playable loop needs a beginning, an action, feedback, and a way to try again. A single room with three targets can teach you more than a large level with no ending.
Use plain shapes and temporary sounds. Keep a restart button available. The faster you can repeat the loop, the easier it is to understand whether a change helped.
Watch someone play
Give the prototype to a person who has not heard your explanation. Ask them to say what they think they can do. Watch where they hesitate and what they try before helping.
Separate a control problem from an idea problem. If the player cannot understand how to jump, you have not yet learned whether the jumping challenge is interesting.
Finish before you expand
Choose a small definition of done: the player can start, understand the goal, succeed or fail, and restart without your help. Save a working version before adding another feature.
A finished small prototype is a useful decision tool. It can tell you to continue, change direction, or put the idea aside without losing months of work.
Keep exploring. Keep creating.
