Prototype and Test Improvements

A prototype is a rough first version of a design — built quickly, tested honestly, and improved on purpose. Real engineers never trust a first draft. They build, test, learn, and rebuild. In this lesson you will plan a prototype, test it fairly, record what happens, and iterate. The goal is not a perfect first try; the goal is to fail forward — to learn something useful from every attempt.

Low fidelity vs high fidelity

• Low fidelity: cheap, fast, made of paper, tape, cardboard, or sticks. The job is to test the BIG idea, not the looks. Build it in minutes, not hours.
• High fidelity: closer to the real product. Better materials, more careful build, longer to make. Only worth doing AFTER the low-fidelity test proves the idea is worth it.
• Fidelity = how close the prototype is to the real thing. Start low, raise fidelity only as questions get more specific.

Why we test

A design that has not been tested is just a guess. Testing turns a guess into evidence. We test to find out: does it work, how well does it work, and what breaks first? A failed prototype is not a wasted prototype — every failure points to the next improvement. That is what fail forward means: keep moving, but learn from each fall.

Setting up a fair test

1. Pick one variable to change (the independent variable). Example: ramp height.
2. Pick one variable to measure (the dependent variable). Example: distance the car rolls.
3. Keep everything else the same (controls). Same car, same floor, same release method.
4. Run each test more than once and average the results — one trial can lie.

Baseline first

Before you change anything, record a baseline — how the prototype performs in its starting form. Without a baseline, you cannot tell whether your change was an improvement, no change, or a step backward. The baseline is your "before" picture.

Recording observations

• Use a simple data table: trial number, what you changed, what you measured, what you noticed.
• Write what actually happened, not what you wanted to happen. Honest data is the only data worth having.
• Photos and quick sketches count as observations — capture before, during, and after.

The iteration cycle

1. Build a low-fidelity prototype.
2. Test it and record the baseline.
3. Identify one thing that worked and one thing that failed.
4. Change ONE thing to fix the failure — keep what worked.
5. Test again. Compare to the baseline. Repeat.