Stick Drift Test
Release both sticks and observe their browser-reported offset for ten seconds.
Controller input
Up to four controller slots. Labels describe the standard browser mapping; they do not identify your hardware model.
Press a button to connect
Your inputs will appear here when your browser exposes the controller.
Controller not detected? Follow the connection checks →10-second stillness observation
Observe stick drift
Put the controller on a flat surface and release both sticks. Keep this tab visible for the full measurement.
Default 0.05. This line marks the result; it does not filter or correct your inputs.
Connect a controller with a standard mapping and four valid axes to begin.
Local report
Keep your test results
Download the current readings and measurement conditions, or print a summary. The JSON includes the controller name your browser reports.
Your controller input and report stay in this browser. Refreshing clears the session; files you download stay on your device.
Using the test
Prepare a repeatable stick drift test
Place the controller on a stable, flat surface. Let go of both sticks before starting, and keep this browser tab visible throughout the sample. Use the same connection and reference value for later comparisons.
Select USB or Bluetooth in the test conditions if you know which connection is active. The tool records your selection; it does not infer the physical connection from the device name.
Understand mean, P95, and maximum offset
Each foreground sample uses the current browser snapshot, including values that have not changed. Stick offset is the distance from the center: √(X² + Y²). A diagonal can therefore exceed 1.
Mean describes the average observed distance. P95 is the value at or below which at least 95% of the samples fall, using the nearest-rank method. Maximum records the largest distance observed. Each axis also keeps its minimum and maximum in the local report.
The default 0.05 reference is a comparison aid, not a pass/fail standard. Changing it does not filter the readings or calibrate the controller. A result above the line alone cannot determine whether a game will behave differently.
When a measurement cannot complete
Switching controllers, disconnecting, hiding this tab, or losing valid standard axes interrupts the current sample. Start again when the connection is stable. A previous completed result remains labeled with its own time.
The controller drift test needs a standard mapping and at least four valid axes to label left and right sticks. With a different mapping, inspect the raw axes in the general tester instead of assuming a stick assignment.
Related tools