Mouse DPI test
A browser cannot read a mouse’s hardware DPI, so this runs the sum backwards: you push along a ruler for a known physical distance, the page measures how many pixels the cursor travelled on screen, and dividing one by the other gives counts per inch (CPI/DPI). This test needs an actual ruler, and pointer acceleration has to be switched off first — otherwise there is no point reading the numbers.
Before you start
- Turn off pointer acceleration. Control Panel → Mouse → Pointer Options → uncheck “Enhance pointer precision”. Skip this and there is no point reading anything below.
- Drag the pointer speed slider to the middle (notch 6). That is the only notch with 1:1 mapping; one notch either way and the result is out by a chunk.
- Press Ctrl+0 to reset browser zoom, lay the ruler across the mousepad, and line its 0 mark up with the dashed line on screen.
- Rest the mouse at the ruler’s 0 mark and click the strip to the left of the dashed line, then push straight along the ruler for the chosen distance and click once more to finish.
Line the ruler’s 0 mark up with the dashed line on the right, rest the mouse there, and click the strip on the left to begin;
push straight along the ruler to the target mark, then click once more to finish.
Push 1 cm straight along the ruler, then click once more to finishThe cursor has hit the edge of the screen — pushing further will not move it. Click once to finish, then lower the target distance and measure again.
0 pxIs D really D? It is actually CPI
Strictly, this quantity is CPI (counts per inch), while DPI (dots per inch) was originally a printer specification. But the mouse industry has muddled the two together for decades, and makers, drivers and player communities all write DPI today, so this site does too — what you need to know is that here they mean the same thing.
Reference tiers
These are common nominal tiers, and all they give you is a rough “about the same as”. They are not a verdict: a 5% gap between nominal and measured is normal, and this page measures by hand-pushing a short distance, so a few percent of error is built in.
| Tier | Counts / cm | Typical use |
|---|---|---|
| 400 DPI | 157 | Office work, multi-monitor setups, low-sensitivity players |
| 800 DPI | 315 | Office work, multi-monitor setups, low-sensitivity players |
| 1200 DPI | 472 | The most common gaming tier |
| 1600 DPI | 630 | The most common gaming tier |
| 2000 DPI | 787 | High-resolution screens or high-sensitivity settings |
| 2400 DPI | 945 | High-resolution screens or high-sensitivity settings |
| 3200 DPI | 1260 | High-resolution screens or high-sensitivity settings |
| 4000 DPI | 1575 | Usually paired with a sensitivity-reduction multiplier in the driver |
| 6400 DPI | 2520 | Usually paired with a sensitivity-reduction multiplier in the driver |
| 12800 DPI | 5039 | Usually paired with a sensitivity-reduction multiplier in the driver |
| 25600 DPI | 10079 | Usually paired with a sensitivity-reduction multiplier in the driver |
How far you can push is up to your mouse
This looks like a test where the longer the push the better, and it is not: the cursor has edges on screen. Every count the mouse produces per inch of movement becomes one pixel of cursor travel, so for the same physical distance a higher DPI sends the cursor further. Push 10 cm at 800 DPI and the cursor has to cover 3150 pixels; at 3200 DPI the same 10 cm is 12600 pixels — long past the edge of the screen. Once the cursor reaches that edge, pushing the mouse further does not update its coordinates at all; the distance you measured is cut short, the result comes out low, and nothing on the page looks wrong.
So this page starts from 1 cm by default. The point of that run is not to hand you an accurate number but to find out how far this mouse can actually be pushed — once it has a rough DPI, the page works that length out and tells you, and you measure again at that distance. That second run is the most accurate result this page can give you.
Where the error comes from
- A hand-pushed distance is imprecise. This is the biggest one, and the shorter the distance the worse it bites: 1 mm off over 1 cm is 10%, while the same 1 mm over 10 cm is only 1%. That is exactly why the section above tells you to push further on the second run. Starting and stopping on the ruler’s marks makes it noticeably better.
- The pointer speed slider is not in the middle. Off notch 6, the system applies a whole-number scale first and what you measure is the product. This one can skew the result badly, so put it back where it belongs first.
- Pushing off-axis. This page uses net displacement: it measures the straight line between the start and the end point. Push in an arc and what you measure is less than the path you actually travelled, so the result comes out low. The “Off-axis” reading above is there to show you exactly this.
- The system is still accelerating. Some mouse drivers ignore the system’s “Enhance pointer precision” and apply acceleration of their own. If two runs — one pushed slowly, one pushed fast — come out far apart, something is still accelerating; go and confirm with the acceleration test page.
Common questions
Because that is pointer acceleration. With it on, the system scales your movement up according to how fast you move — push the same 10 cm quickly and the cursor travels further. So the conversion “how many pixels per centimetre” simply does not exist, and the number you measure is tied to how fast your hand moved and cannot be reproduced. Turning it off is a precondition for this test, not an option. Where it lives: Control Panel → Mouse → Pointer Options → uncheck “Enhance pointer precision”.
It is, and the longer the better — but you cannot push that far on screen. Every count the mouse produces per inch of movement becomes one pixel of cursor travel, so an 800 DPI mouse pushed 10 cm has to send the cursor 3150 pixels, while a 1920-wide window gives you only somewhere around 1800 from the left edge to the right. Once the cursor hits the screen edge it stops for good: push the mouse further and its coordinates never update, the measured distance is quietly cut short and the DPI comes out low. What makes it worse is that this ceiling depends on your DPI — and DPI is the very thing this page is here to measure. So the page starts from 1 cm: get a rough figure first, then use it to tell you how far this mouse can actually be pushed, and measure again at that distance for a much better result. If the cursor reaches the screen edge mid-run, the toast changes its wording on the spot.
Because the coordinates the browser reports are CSS pixels, not mouse counts. At 150% scaling on the desktop, one mouse count is only 1/1.5 CSS pixels, so the CSS pixels you measured have to be multiplied back by 1.5 to give real counts. The page reads the system scale by itself, but if your browser zoom is not 100% (Ctrl+0 resets it) the value it reads is polluted by that zoom, and you have to set it by hand.
Yes, and there will almost always be a gap. The error has three sources: the distance you push by hand can never be exact (1 mm off over 1 cm is 10%), a pointer speed slider away from the middle (notch 6) adds a fixed multiplier, and a mouse’s own rated DPI is commonly out by around 5%. So this page presents its result as roughly equal to some tier, and shows you the gap between the DPI you entered and the one it measured rather than passing or failing it — below 5% that gap is the same order of magnitude as this page’s own error, so it says nothing about whether your mouse is off.
Windows’ pointer speed slider has 11 notches, and only the middle one (notch 6) is a 1:1 mapping — one mouse count to one screen pixel. To the left it scales down, to the right it scales up, and it does so in whole-number steps, not smoothly. With the slider off centre, what this page measures is the product of the slider’s scale and the hardware DPI.
