The average person reacts in 200–250ms. Trained gamers typically land between 150–200ms. Under 150ms is exceptional and may suggest either super human abilities or gaming the system. Elite esports players consistently average around 180ms in controlled testing.

Yes, but less than most expect. A high-polling-rate mouse reduces input latency by a few milliseconds versus a standard office mouse. Monitor refresh rate and your own biology matter more.

Keyboards with analog input and rapid trigger (like the Wooting 80HE) reduce actuation latency versus membrane keyboards. In a browser test the difference is typically under 5ms, within normal human variation between attempts.

Human reaction time varies naturally by 20–40ms due to focus and fatigue. This test averages three rounds for a more reliable result. Browser tests also carry an accuracy margin of ±8–16ms depending on your screen's refresh rate.

Browser-based reaction tests have real limitations. Your result depends on your browser, your screen's refresh cycle, background CPU load, and a dozen other things.

That's fine. This test isn't designed to be a lab instrument. It exists because I was curious whether I could measure a real difference between my own devices, and the only way to find out is to collect enough data from enough people. If you have multiple mice, keyboards, or monitors, try them all and see if you notice any patterns in the data.

At 60Hz a new frame appears every 16.7ms, meaning the green signal can be delayed by up to 16ms from when it was triggered. At 144Hz this drops to 6.9ms. This test detects your screen's refresh rate and displays your accuracy margin in the arena.

Early tests revealed that users were gaming the system with pre-clicks. Cheating their way to the top of the leaderboards (you know who you are). Five rounds averaged together produces a more honest and consistent result, and makes the hardware comparison data significantly more meaningful. Season 1 only required three attempts. Season 2 raised it to five.

If one of your three attempts is significantly faster than the other two, the session is flagged and discarded. This is the signature of a pre-click — pressing the key before the green signal appears, whether intentional or not. This test is a data collection project, not a personal best tracker. A suspiciously fast outlier skews the hardware comparison data for everyone, so it gets thrown out rather than averaged in. Just run the test again with consistent, genuine attempts.

I was surprised by how small the real-world difference is between a 1000Hz and 8000Hz polling rate mouse, actually less than 1 millisecond! That got me curious: would I actually be able to measure a perceptible difference between my own devices? And what better way to find out than collecting data from real users across different hardware? This test exists to answer that question with actual data rather than spec sheets.

When you submit a score, your chosen username, reaction time average, hardware name, and screen refresh rate are stored in a database. A hashed machine and browser identifier is also stored to mitigate leaderboard abuse and database injections. This is a one-way hash — the original value cannot be recovered from it. Your personal best scores and attempt history are stored only in your own browser's localStorage and never sent anywhere. You can clear this at any time by clearing this site's cache in your browser. No cookies, no tracking, no account.

It uses the browser's requestAnimationFrame API to count how many frames are drawn over a fixed period. By measuring the time elapsed across 60 consecutive frames, it calculates your display's frames per second and rounds to the nearest standard refresh rate. This can be influenced by a lot of things — your machine, browser, and other things currently running on your machine. This is no exact science, and it doesn't need to be. It is meant to give a pointer on whether there is a correlation between higher refresh rates and lower reaction scores. No permissions or plugins are required, it just runs silently in the background as soon as the page loads. Neat!

I am testing out this new, neon green on just this page. I kinda like it, but need more time to finally decide if I am to change the whole site's color scheme.