One fast click is not a baseline
Running the same simple reaction test twice rarely produces matching scores. That variation is normal. Reaction time is a noisy biological signal, not a fixed constant.
A fair baseline is the median of several clean trials under stable conditions, with false starts removed. This guide covers common sources of bounce, what to discount, and how to use the Dashboard so history stays comparable.
Common causes of score bounce
Most session-to-session swings come from a short list of factors:
• Warm-up: The first 1 to 2 rounds are often slower while attention settles. Treat those rounds as practice, then use the remaining trials.
• Anticipation and false starts: Responding before the cue, or guessing the rhythm, can produce unrealistically low times. Visual scores under about 120ms are almost always anticipation, not a true reflex. See how to interpret results.
• Attention lapses: A blink, notification, or brief distraction can add hundreds of milliseconds to a single trial without changing underlying speed.
• Fatigue and sleep: Tiredness slows both speed and consistency. A rested session and a late-night session are not comparable.
• Hardware and display: A 60Hz screen, Bluetooth mouse, or heavy browser load adds delay unrelated to neural speed. Prefer a wired mouse or trackpad for comparable numbers. See the hardware latency guide when comparing setups.
How to run a fair baseline session
Use this checklist for scores that can be compared week to week:
1. Choose one protocol and keep it fixed (usually Simple Reaction). Do not mix test types when tracking a baseline.
2. Close extra tabs, use the highest available display refresh rate, and prefer a wired input device.
3. Complete 5 to 10 rounds. Ignore the single best click. Record the session median.
4. Discard sessions with many false starts or large outliers. One messy session should not redefine the baseline.
5. Keep history on the Dashboard. Export JSON or CSV when a training log is needed.
Median versus best
Best time reflects peak potential under ideal conditions. Median reflects performance across a short block of trials. Prefer the median for training and comparison. It is less sensitive to one early click or one distracted miss.
Consistency matters as well. Two medians can look alike while stability differs: one session may stay within about 15ms, another may swing about 60ms. Track both speed and consistency when reviewing progress.
A simple weekly check-in
Once or twice a week, run the same test under similar conditions and record the median. Over several weeks, that series is more informative than a single personal-best screenshot.
For structured drills after a baseline is set, see how to improve reaction time. For testing methods, see how to test reaction time.
