Why your reaction time changes every time you measure it

Measure your reaction time five times and you get five different numbers. One day it says 240ms; a few days later, same chair, same desk, it says 275ms. Most people read that as “bad day today.” But part of those 35ms may have come from the measuring side, not from you.

This guide is about separating the two. Not how to get faster — how to work out what actually moved when your number moved.

What is packed inside that one number

Between the screen turning green and your finger pressing the button, at minimum all of this happens:

1. The display physically emits green light
2. That light reaches your eye and becomes a signal
3. Your brain decides “now”
4. The command travels to the muscles in your finger
5. Your finger actually closes the switch
6. The input arrives at the browser

Only steps 2 through 5 are what people mean by “my reaction time.” Steps 1 and 6 are equipment. But the measurement hands you all six added together as a single number. From that one number you cannot tell which part moved.

What moves on the human side

Arousal. Just-woke-up and fully-engaged are clearly different, and this moves continuously through the day. Which factor matters most, though, varies by person and by day, so ranking them is not something this guide can do.

Fatigue and sleep. Long observed in reaction-time work. But there is no conversion table that turns “hours slept” into “milliseconds” for an individual.

Age. Pooled across many people, the mid-twenties tend to be fastest, with a gradual slowing after. But individual differences within one age group are far larger than the differences between age groups. So explaining your own score by your age is usually wrong.

Practice and familiarity. Repeat the same test and the first few runs get noticeably faster. That is less “better reflexes” and more getting used to this screen and this button. There is research on action-game experience and processing speed (Dye et al., 2009), but it should not be read as “play games, gain N milliseconds in a week.” The populations and conditions those studies used are not our screen.

What moves on the measurement side

Refresh interval. A 60Hz display only changes its picture about every 16.7ms. If the moment that should turn green lands just after a refresh, the actual green light waits up to one more frame. That wait differs every round, so it enters your score as random noise.

Input path. Mice, keyboards and touchscreens each go through different processing. Phone touches in particular need their own processing before a contact is judged to be a “tap.”

The browser. Our measurement uses the clock the browser reports. It does not use a photosensor to catch the instant the panel actually emits light. With many tabs open, or something heavy running alongside, those clock readings can slip.

So which one moved?

There is exactly one method: change one thing at a time.

• Freeze your equipment completely and measure repeatedly across several days. Whatever variation remains is your record’s ordinary spread. It is usually wider than people expect.
• Not all of that spread is you. Freezing the equipment does not remove random error like frame waiting, so read it as a spread with both you and the hardware in it.
• It is still useful. A new value inside that spread has moved within your ordinary range, which on its own is thin evidence that anything changed. Go looking for causes when a value clearly leaves the spread.
• If you changed equipment, treat before and after as two separate records. An improvement that appears right after moving from 60Hz to 144Hz has the display’s share mixed into it.

A single measurement tells you nothing. Comparison only starts once you have a spread of several.

What this test cannot measure

The FlowScore reaction test measures simple reaction time: one signal, one thing to do. So it says nothing about how fast you are when you have to choose what to do. The moment a choice appears, times grow substantially — and that is a different ability, the one the Hard Brake test measures.

This number is also not a grade for your driving or your reflexes in general. It is the time one finger took to answer one signal in front of one screen. It means something only inside that range.

In short

• Your score mixes you and your hardware. One number cannot separate them.
• Freeze the equipment and measure repeatedly to learn your own spread first.
• A value that moved within that spread is thin evidence on its own that anything changed.
• Split your history at the point where you changed equipment.