1. Define the symptom, not the shopping list
Write down where the problem happens: corner entry, middle or exit; left, right or both; on power or off power. ‘Rear steps out when I accelerate from the hairpin’ is a useful observation. ‘Needs more upgrades’ is not a testable diagnosis. Inspect tyres and mechanical condition before changing a setting.
2. Establish an A run
Use a known battery condition and the same warm-up routine. Run enough clean laps to settle into a rhythm. Record the full times, collisions and obvious mistakes rather than deleting inconvenient laps. Keep clean-lap pace and incident count as separate measures. Your fastest lap shows potential; your average and variation show how repeatable it is. The median is the middle time when laps are sorted, or the mean of the two middle times for an even-sized sample.
3. Change one variable for the B run
Choose one change with a clear reason: a tyre pair, an allowed steering adjustment or a smoother driving input. Keep the other settings, driver and test duration as similar as possible. Record the changed value, not just ‘softer’. Gearing tests also need temperature and run-time observations; a ratio that feels quick for one lap may be unsuitable for a full race. Check the manual’s limits rather than adopting another racer’s pinion blindly.
4. Return to A before drawing a conclusion
A repeat of the original configuration helps reveal changes in grip, battery performance or your own driving. If both A runs differ greatly, the session may be too variable for a confident verdict. The range between fastest and slowest clean laps is a simple consistency check, but one outlier can distort it. Compare several runs and the car’s behaviour, not only one summary number.
5. Keep what you can repeat
For example, clean laps of 10.2, 10.3, 10.1, 10.4 and 10.2 seconds have a 10.24-second mean, 10.2-second median and 0.3-second range. These are illustrative values, not a Rick race result. A small improvement that disappears in the next run is weak evidence. Prefer a stable car that you can race for the whole heat, and save the test conditions with the result.
One change, comparable conditions, repeated evidence. Let the notes guide the next experiment.