Reaction time test world record, and why results under 100 ms are doubtful
· 6 min read · Tyamka team
No body keeps a world record for voluntary reaction time. What the fastest measured times are, and why sprint rules and fair tests doubt anything under 100 ms.
A reaction time test world record does not exist in any official form. We found no organisation that measures and certifies the fastest voluntary human reaction the way records are kept for running or swimming. What careful research does give is averages. A large lab study measured a mean of 231 ms for a light and a mouse button, and Francis Galton recorded times below 190 ms in young people in the 19th century. The very small numbers on the leaderboards of online tests, under 100 ms, are best read with doubt. Sprint rules treat a start under 0.100 seconds as a possible false start, and a browser adds its own delay on top of the real reaction. The same applies to our own reaction time test.
The one listed record is not a human reaction
Guinness World Records does list the fastest artificially-assisted reaction times. The record is 50 ms from stimulus to action, set in Chicago in 2019 by the researchers Pedro Lopes, Jun Nishida and Shunichi Kasahara. Their volunteers wore electrical muscle stimulators, and in the assisted runs a small current moved the muscles. So it is a record for a machine moving a hand, not for a person reacting.
Where the 100 ms line comes from
In sprinting, the starting blocks measure the moment an athlete pushes off after the gun. World Athletics Technical Rules (edition in force from 1 November 2019, rule 16.6) say that when a certified Start Information System is used, it signals a possible false start when "the reaction time is less than 0.100 second". The starter then checks the reaction times and decides who is responsible.
The rule treats a start that fast as a sign that the athlete may have anticipated the gun rather than reacted to it. It is a judging rule, and the starter still makes the call.
Is 100 ms a real human limit?
Not exactly. In 2007 Pain and Hibbs at Loughborough University measured sprint starts with starting blocks that had force sensors in each footplate. Five of their nine athletes had a mean reaction time under 100 ms in at least one test condition, and 20% of starts in the first two conditions were under 100 ms. The authors concluded that the nerve and muscle part of a simple reaction to sound can be under 85 ms.
So the 100 ms threshold is a practical rule for judging races, not proof that nobody can go faster. It also concerns sound and a push against a sensor in the blocks. Online tests use a light on a screen and a press on a button, a different sense and a different measurement, and the numbers do not carry over.
Fast visual reactions in research
For a light and a button, one large and careful measurement is the 2015 study by Woods and colleagues. They tested 1,469 adults with a calibrated screen and a gaming mouse and found a mean of 231 ms, or 213 ms after taking off the 17.8 ms of hardware delay they had measured. They also note that Galton may have kept each person's best of three tries, which would make his numbers look faster.
These are averages, and single tries spread around them. In the same study the trial-to-trial standard deviation within one person averaged 40 ms, so a try well below your own average happens. The study itself treated any response faster than 110 ms after the target as a false alarm rather than a reaction.
Why very small numbers on test sites are not records
A browser test cannot see the real moment the light appears on your screen or the real moment your finger touches the glass. It sees its own events, and there is a gap. In a lab study of web timing, Pronk and colleagues pressed phones and laptop keys with a machine at precisely known times. The web page recorded every interval as longer than it was, on average by 57.6 to 132.9 ms depending on device and browser, and on no single press by less than 44.9 ms.
As an illustration, and only as arithmetic, take a result of 90 ms on one of those devices. Subtract even the smallest delay they saw, and the real press would have come about 45 ms after the light, faster than anything in the research above. More likely explanations are these:
- A guess. The player pressed at the moment they expected the signal and happened to be lucky.
- A wait that can be learned. A test that picks from only a few waits, or rounds them, lets people learn the rhythm.
- A glitch or a script. A leaderboard of a free online test usually cannot tell how a result was made.
None of this means the person cheated on purpose. A lucky early press feels exactly like a very fast reaction.
How a fair test deals with it
A test that wants reactions rather than guesses needs a new random wait every time, so there is nothing to learn. It also needs to reject presses that come too fast and make them cost something, so guessing does not pay.
In our reaction time test the circle appears after a pause of one to four seconds, drawn at random to the millisecond. A tap before the circle, or less than 0.12 s after it, is a false start and costs one of your three lives. From level 10 a square sometimes appears instead of the circle, and tapping it also costs a life. We treat a tap under 120 ms as a guess. That line is our rule for this game, set above the sprint threshold, and not a measured limit.
Formula 1 draws the line differently. A jump start there means moving before the lights go out, judged by a transponder on the car, with a tolerance the FIA does not publish. The article on the F1 reaction test explains that start.
What to do with your own best time
Your best single time is the one where everything went your way, possibly including a well-timed guess. The number worth watching is the middle of several tries on the same device. The article on the reaction time test average explains why your times move around and how to compare them fairly.
In our game the record is not a single time anyway. The results screen shows points for the whole level, 10 for each correct answer plus up to 10 more for speed, and your previous best for that level. Even the first level has 8 rounds, so one freak tap cannot carry it, and a false start takes a life instead of adding points.