Stick drift test
See exactly how your analog sticks behave: how big the deadzone is, whether they sit off-centre when you let go, and how close a full roll comes to a true circle. Runs in your browser: no install, no account, no limit.
Left stick
centredLaps 0/3
Right stick
centredLaps 0/3
How does the test work?
Three steps, all from the same live stream of stick positions:
- Rest position. With both sticks released, the tool averages the reported position. Anything clearly above zero is drift, and the measured rest point becomes the real centre for the next steps.
- Range. Push each stick slowly to its left, right, top and bottom edges and back through the centre. The tool records how far the stick travels in each direction, and the smallest movement that registers on the way out is the effective deadzone.
- Circularity. Roll the stick around its rim. The tool records the furthest reach in each of 72 sectors of five degrees and compares it with a circle drawn from your measured range. The plot shades the area you have swept, so you can see which directions you have not covered yet.
The circularity result stays locked until nearly every sector has been passed three times (the tool tolerates a few hard corners). A maximum only settles when you come back to it: after a single lap some sectors always fall short of the rim, and that shortfall would count as error the controller does not have. The progress bar counts laps, not time.
What are deadzone and stick drift?
Drift is a stick that does not return to its centre: released, it still reports a small position, and in a game your character or camera moves on its own. It usually comes from wear or dust in the stick mechanism, explained in controller care and stick drift.
A deadzone is the answer in software: ignore everything inside a small radius, so a slightly off-centre stick reads zero. Games add their own on top of what this test sees. Microsoft's XInput guidance for developers suggests a default of 7,849 out of 32,767 for the left stick and 8,689 for the right, about 24% and 27% of the travel, which is why many games feel unresponsive near the centre.
Read the two numbers together. A large deadzone with a clean rest position is a design choice. A large deadzone hiding a large rest offset is a worn stick, and the travel lost inside the deadzone is travel you cannot use for slow, precise aim.
What is circularity error?
Push a stick fully right and it reports 100%. Push it fully up and to the right and, ideally, it still reports 100%, split across both axes. Trace the whole rim and you should get a circle. Real sticks show two common shapes:
- Squared output. Both axes can reach full scale at once, so diagonals read up to about 141%. It is not a fault, but games that expect a circular range treat those diagonals as faster than intended.
- Clipped diagonals. The stick cannot travel as far diagonally, so diagonals fall short, often at 85 to 95%, and diagonal movement is slower than straight movement.
The average circularity error sums this up as the mean distance between your traced shape and the ideal circle, in percent. Once the Range step is done, the circle is drawn from the travel you measured, and the result is labelled calibrated: it shows the shape of the stick without its drift or its shorter travel. Without that step, the comparison is with a nominal circle of radius 100%, labelled vs nominal. Compare like with like, for example the same controller before and after a repair.
How do I read my results?
The tool colours each value with these thresholds:
| Reading | Healthy | Worth attention |
|---|---|---|
| Rest offset | Under 2% | Above 5% is real drift |
| Deadzone | Under 12% | Above 20% costs noticeable precision |
| Circularity error | Under 10% | Above 18% is uneven |
| Range reached | About 100% in every direction | Below 90%, or more than 10% between two directions |
A difference of more than 10% between directions is typical of wear on one side: the stick feels fine one way and short the other. The calibrated circularity error leaves that difference out on purpose, so read the range next to it.
What can I do about it?
Recalibrate if your controller offers it, then clean around the stick base, and run the test again after each step. If the rest offset stays above 5%, the stick is worn: controller care and stick drift covers cleaning, warranty claims and stick replacement.
To check every button and trigger as well, use the gamepad tester. If the controller does not appear at all, follow the controller detection checklist.
What cannot this test tell me?
The browser reads your stick after the operating system and driver have processed it, through the same Gamepad API a web game uses. That is what software receives, with these limits:
- A deadzone applied by the driver or by software such as Steam Input cannot be told apart from one built into the controller.
- Resolution depends on the controller. A DualSense reports each stick axis as a value from 0 to 255, so the smallest step is about 0.8% of full travel. Xbox controllers report 65,536 steps per axis on Windows.
- It measures the reported geometry of the stick, not the mechanism. It can show that a stick is uneven, not which part is worn.
- Reaching the rim at every angle depends on your hand. Low coverage makes the average less reliable, which is why the tool shows coverage.
Turn your controller into a live, transparent input overlay for OBS, Streamlabs or any app that accepts a browser source, in a couple of minutes.
Build my overlay