The lag your device adds
Between your finger and the number on the card sits a second nervous system, made of glass and radio. It has reflexes too, and they are billed to you.
Every timed score on this site is the sum of two reaction times: yours, and your equipment’s. The site is named after the first one, but the second one is real, measurable, and larger than most people guess. This guide itemizes it.
The screen owes you milliseconds before you see anything
A monitor does not show change continuously; it repaints on a schedule. At 60 Hz, a new frame arrives every 16.7 milliseconds, so a green light that “turns on” does so, on average, 8 milliseconds after the software asked, and in the worst case a full frame late. A 144 Hz panel cuts that average wait to 3.5 milliseconds; 240 Hz to 2. The panel then needs time to physically change its pixels, and many displays add internal processing on top: a TV in its default picture mode can sit on a frame for 50 to 100 milliseconds, which is why consoles have a game mode and why a reaction test on a living-room TV reads like a bad day.
The tests here start their clock on the frame the browser actually paints, tied to the display’s own schedule, so the software’s share of the wait is accounted for. The panel’s physical share cannot be seen from a web page, and stays in your number.
The input device bills you on the way back
Your press also travels. A wired gaming mouse polls at 1000 Hz and delivers a click within about a millisecond; an ordinary office mouse at 125 Hz can sit on it for up to 8. Keyboards add scanning and debounce delays of a few milliseconds. Bluetooth adds radio scheduling on top, commonly 10 to 20 milliseconds and occasionally worse, which is why a wireless keyboard in a quiet room can still lose a duel with a cheap wired one.
Touchscreens are the expensive instrument. The screen samples touch on its own clock, the operating system filters the stream (partly to guess whether a touch is the start of a scroll), and the result is that a tap commonly reaches the page tens of milliseconds after the finger physically arrived. Phones have improved, and the browsers’ handling has improved, but a phone still reads slower than a wired mouse on the same human.
What it sums to
Stack the pieces and the spread between setups is large: a fast desktop (high-refresh panel, wired mouse) might add 15 to 30 milliseconds around your nervous system; an average laptop on its touchpad, 40 to 60; a phone, or a TV out of game mode, more. The spread between the best and worst common setups, 30 to 50 milliseconds, is bigger than the gap between an average human and a quick one. On the aim trainer, where thirty pointer movements are timed, the instrument shows most of all: a mouse, a trackpad and a thumb produce three different distributions from one and the same person.
Why the site does not subtract it
Because it cannot, honestly. A web page cannot see your panel’s response time or your mouse’s polling rate; any correction would be a guess dressed as a measurement, and this site does not do that. Instead the position is declared openly on the Method page: the device’s lag is part of the number, every test’s pool mixes phones and desktops, and the comparison that survives all of this is you against yourself on the same setup.
One number here is nearly immune, and it is worth knowing which: the Stroop test’s interference lag is a difference of two averages measured through the same hardware, so the machine’s share cancels in the subtraction. The choice test’s decision lag enjoys the same protection. Differences travel between devices; raw times do not.
If you want your fastest honest number
Use the fastest screen in the house at its highest refresh rate, plug the mouse in, close the app that is mining something in the background, and skip the TV. On a phone, a screen in high-refresh mode and a firm tap with a finger already resting close to the glass beat a hover-and-stab. None of this changes you; all of it changes the number, which is exactly the point of this guide. When you compare with a friend, agree on the instrument first, and then let the test settle the rest.