Make the next attempt answer a question

Name the question before moving a piece

Suppose an object stops before its target. You could steepen the ramp, remove a surface and replace the object. If the next run reaches the target, you have improved the result—but you cannot tell which change mattered. Several possible explanations changed together.

Start with a baseline: the arrangement and outcome you are comparing against. State one question, choose one variable to change, and keep the other relevant conditions the same. Decide what you will observe before starting. “It looked better” is harder to compare than “it reached the marker” or “it travelled farther after the same interval.”

A small investigation log

Imagine a browser circuit with a connected battery, a bulb and an open switch. The bulb is off. Your question is whether that switch is interrupting the useful path. Predict that closing it will light the bulb, close only the switch, and run again.

If it lights, reopen the switch and repeat the baseline. Off, on, then off again is evidence for the switch’s role under these conditions. It does not prove that every dark bulb in every circuit has the same cause. A future board might contain a disconnected terminal or a bypass. Keep the conclusion as specific as the test.

Choose a test of ramp angle

Baseline: the same object starts from the same point on a 20° ramp, with grip on. The question is whether changing angle affects the travel result. These are planning examples, not a duplicate of the game engine.

Which trial would you run next?

Choose a plan, then check what the comparison can tell you.

A useful outcome measure might be distance travelled at a fixed time. Use the same measure and reset conditions for both settings.

Read the comparison without the exercise

The 30° trial with grip still on changes only the angle. Changing angle and grip together confounds their effects. Repeating the 20° baseline checks consistency but does not test a new angle. Both repetition and controlled comparison have useful, different jobs.

Reset, repeat and notice hidden changes

A fair comparison needs a consistent starting state. Reset positions, stored charge, switch states or inventory when those conditions affect the question. A second run beginning where the first ended may not be testing the same situation.

Repetition can reveal inconsistency. If an outcome varies between apparently identical runs, inspect timing, collisions, initial placement or another uncontrolled factor. Do not hide the inconvenient run. Record what differed and adjust the question. “One variable” is a useful starting method, not a claim that complex systems never need more advanced experiments.

Use one question across the gaming world

In Oops! Lab, test a ramp angle while holding the object, release position and grip constant. In Spark Island, test a switch or one suspect wire before rearranging the whole circuit. In Monster Market, compare a price choice while keeping the relevant costs and quantity assumptions visible.

Detective Zero uses a related habit: change your explanation when a new observation demands it, rather than inventing a missing fact to protect your first guess. Across the games, write a short sentence after each attempt: “I changed X, observed Y, and now think Z under these conditions.” That record turns a sequence of clicks into an investigation you can explain.

Source & scope

NASA JPL Education: How to Do a Science Fair Project. Reference for planning fair tests and controlling variables. The diagnostic log and ramp choices are original browser-game teaching examples.

This is an Aurism learning article, with AI-assisted development as described in our editorial approach. It does not claim an independent expert review or that a simplified model reproduces every game outcome. Report a correction.

TRY THE IDEA IN A GAME

Oops! Lab

Use the mission’s actual instructions and feedback to test your next prediction.

Play Oops! Lab