Write down what you are trying to learn
A prototype should be built around a decision. Can a player judge the jetpack boost? Does a single board position create competing choices? Would a short run still feel satisfying without a reward screen? A concrete question prevents unrelated work from taking over the test.
Keep the mechanics that affect the answer
An early racing test needs input, track shape and a readable loss of traction. It does not need a garage, seasonal progression or a collection of liveries. Remove what makes the build harder to judge; keep the signals and timings that directly shape the player’s response.
Look at the second attempt
The first attempt shows whether the rules are understandable. The second often reveals whether the player noticed a choice and adjusted their behaviour. Instead of only counting successes, record whether the player could describe what they intended to change.
Make a decision while the lesson is fresh
After the test, identify the observation that matters and choose an action: keep, revise or stop. A one-page note with a specific example is more useful than a vague verdict about fun. Even a discarded prototype leaves behind knowledge that can improve the next one.
