Why Your Reaction Time Test Score Is Wrong (and How to Test Fairly)
Every reaction score is really your reaction time plus your device latency chain. The same brain can look 80 ms slower on a different screen.
Take the same reaction test on a gaming monitor and on a three-year-old budget phone and you can get two different numbers from the same brain. Every reaction score you have ever recorded is really your reaction time plus your device's latency chain. Here is what is inside that chain, how big each piece is, and how to get a number you can compare.
The latency chain
| Stage | What it is | Typical added latency |
|---|---|---|
| Display refresh | A 60 Hz screen updates every 16.7 ms; the stimulus appears somewhere inside that window. | 0-17 ms at 60 Hz, 0-8 ms at 120 Hz |
| Display processing | Panel response and image processing before pixels change. | 5-30 ms |
| Input sampling | Touchscreens and mice poll at fixed rates; your tap waits for the next poll. | 5-20 ms |
| Software | Browser event handling, OS scheduling, and background load. | 5-50+ ms |
Total: roughly 20 ms on a well-configured gaming setup, and 50-100+ ms on an older phone with background apps running. That is why cross-device comparisons are noisier than they look.
Phone vs. PC: which is right?
Neither. A test cannot remove device latency; it can only keep it constant. Modern flagship phones with 120 Hz screens are respectable test instruments. Budget devices at 60 Hz with slow touch sampling are the worst case.
- Same device, same conditions for any score you track over time.
- Same test design so you measure reaction, not anticipation.
- Median of multiple attempts, never a single personal best.
How to run a fair test
- Charge the device and close background apps.
- Set maximum screen brightness.
- Enable the highest refresh rate your device supports.
- Sit still and use your dominant hand's index finger.
- Run five to ten attempts; record the median and spread, not the best.
- Retest under the same conditions and time of day.
What a competitive test adds
A one-off test tells you your number; a competitive format tells you whether you can repeat it under pressure. Racing starts, not lab buzzers, are the gold standard: a randomized hold punishes anticipation, a jump start punishes gambling, and averaging across a whole race punishes inconsistency.
ReactionRacing's leaderboard uses exactly that design: every ranked millisecond is an average across real race starts on the player's own device. In-game rankings reward the tight, repeatable reactor over the one-tap hero.