Free online reaction time tests and reflex training — measured in milliseconds, no signup
These games measure how fast you respond to something appearing. Reaction time has an unambiguous unit, which makes it one of the few cognitive measures you can genuinely track — but it also means your hardware matters, and a phone and a gaming monitor will give the same person noticeably different numbers.
Simple reaction time has one signal and one response. Reaction Time, Simple Click, Quick Tap and Reflex Test all measure it, and it is the floor — nothing can be faster, because there is no decision to make.
Choice reaction time adds options. Lightning Choice and Multiple Stimuli both measure it, and it is always slower. The relationship was described by Hick in the early 1950s: response time rises roughly with the logarithm of the number of alternatives.
Inhibition tasks are a third thing again. Go/No-Go asks you to sometimes not respond, and it measures restraint rather than speed. A fast score there with several false alarms is a worse result than a slower clean one.
Aim Trainer and Moving Target add a motor component — reacting is only half the job when you also have to arrive somewhere.
A 60 Hz display adds up to 16 milliseconds of frame latency before the signal even reaches your eye. Input lag, browser timing and touchscreen polling add more. Together they can move a result by several tens of milliseconds.
That is a large fraction of the range that separates fast people from slow ones, which is why cross-site and cross-device comparisons are close to meaningless.
Use the same device, at the same time of day, and compare only against your own history. That number is real and it will move with sleep, caffeine and warm-up.
Alertness moves it most. Sleep loss slows reaction time reliably and substantially, and it is one of the most sensitive measures of fatigue there is.
Caffeine shortens it in the short term — one of the more robust findings in the area.
Age slows it gradually and consistently across every large study that has looked. Practice and alertness both shift it back somewhat, but they do not reverse it.
Practice itself does relatively little to the underlying speed. What improves is your consistency and your anticipation, which is why averaging several clean trials matters more than chasing a single best.
Hover, do not rest. A finger already touching the button will press before you decide to, and that is anticipation rather than reaction.
Run at least five clean trials and use the average. Single results are dominated by noise.
Do not count the delay. Every test here randomises it specifically so that anticipating loses.
Anything much under 150 milliseconds on ordinary hardware almost certainly means you anticipated. Physiology plus display latency makes a genuine result that fast very difficult.
Figures around 200-250 milliseconds for visual reaction are widely quoted, but measured values shift by tens of milliseconds with display and input hardware. Track your own average on one device rather than comparing to a headline number.
Yes. Free, instant, no signup and no download.
Touchscreen polling and display latency both add real milliseconds. It is a hardware difference, not a difference in you.
Only modestly. Most of what improves with practice is consistency and anticipation rather than the underlying speed. Sleep and alertness move it more than training does.
A true reflex is a spinal response that does not involve the brain, like a knee jerk. Everything measured here is a voluntary reaction, which is slower and much more trainable.