Use a stable setup
Connect the controller, close heavy background tabs and keep the same cable, USB port and browser for every comparison.
Controller polling rate test
Controller polling rate test for browser-observed Hz.
Move a stick continuously for 10–15 seconds. Gamepad Bench estimates report frequency from the median interval between Gamepad API timestamp changes and separates that comparison from true end-to-end latency.
No downloadNo accountLocal processing
Live drift0.0%Awaiting signal
Observed rate— HzBrowser estimate
Pressed now016 inputs exposed
Mapping—0 samples
Left stick0.0% from center
CenteredX 0.000Y 0.000Peak 0.0%
Right stick0.0% from center
CenteredX 0.000Y 0.000Peak 0.0%
ToleranceDeadzone threshold
10%1%Recommended 8–12%30%
Ready when you are Connect a controller or open the demo to see live diagnostics.
Polling rate is inferred from Gamepad API timestamp changes. It is useful for comparison, not proof of end-to-end hardware latency.
Repeatable diagnosis
Controller Health Report
Measure ten seconds of untouched stick input, compare a before and after result, then export a local report.
Sampling and exports run locally. Shared links contain only aggregate values in the URL fragment; they omit the controller name and raw input history.
01
Three-minute check
How to get a reliable result
Keep input moving
Rotate one stick continuously for at least ten seconds so the controller produces a useful stream of changing reports.
Repeat before comparing
Run the same motion three times. Compare the browser-observed median rate only after the reading settles and keep every other variable unchanged.
Read the signal
Numbers that actually help
Browser tests are best for repeatable comparisons. Use the same cable, port, browser, and controller state before judging a change.
02
Interpret the test
Read the number without confusing frequency with latency.
A polling-rate result is useful when its measurement path and limits are clear. Use it to compare controlled setups, not to promise a specific reaction time inside a game.
What controller polling rate means
Polling rate describes how frequently a controller produces or exposes fresh input reports. A 250 Hz stream has a theoretical interval of about 4 ms, while 1000 Hz corresponds to about 1 ms. That interval is only one part of responsiveness: controller scanning, firmware processing, transport, operating-system scheduling, the game engine, rendering and the display all add their own behavior.
How this browser estimate is calculated
The tester watches the timestamp exposed through the Gamepad API while an input is moving. It stores recent positive timestamp differences, uses their median to reduce the influence of occasional long browser frames, and converts the interval to Hz. The result represents updates visible to this browser session; it is not a direct capture of every USB or Bluetooth packet.
Why the value may be lower than the box
A controller advertised at 500 or 1000 Hz may show a lower or less stable browser-observed value. The Gamepad API implementation, operating system, firmware mode, Bluetooth transport, USB hub, power management and background load can affect visible timestamp changes. Some devices also report at different rates depending on whether a stick, trigger or button is changing.
USB versus Bluetooth comparisons
Test the connection you actually use, then compare USB and Bluetooth as separate sessions. Keep the browser, computer, controller firmware and input motion the same. A data-capable cable connected directly to the computer provides a useful reference, while Bluetooth results can vary with battery level, distance, interference and the host radio.
How to make a repeatable test
Close demanding tabs, disable the tester's demo mode, reset the session and rotate one analog stick smoothly for 10–15 seconds. Wait for the observed rate to settle, record it, and repeat twice. Use the middle of the three settled results for comparison. Changing one variable at a time makes it possible to attribute a difference to a cable, port, connection or firmware setting.
When the reading stays blank or unstable
Press a controller button after opening the page, confirm the operating system detects the device and keep a stick moving so timestamps change. If the value remains blank, try current Chrome or Edge, reconnect the controller and test a direct USB port. Large swings do not automatically mean faulty hardware; repeat the test before drawing a conclusion.
03
Straight answers
Frequently asked questions
Is polling rate the same as input lag?
No. Polling rate describes report frequency. End-to-end input lag also includes controller processing, USB or Bluetooth transport, the operating system, game engine, rendering and display latency.
Why do I see a lower rate in the browser?
The Gamepad API, browser scheduling, power saving, connection type, controller firmware and whether the input is actively changing can all affect observed timestamp updates. Compare results inside the same browser instead of treating different tools as interchangeable.
Should I use USB or Bluetooth?
Use the connection you actually play with. For controlled comparisons, test both separately without changing the browser, computer, controller mode or input motion. USB is a useful reference, but it does not make every controller poll faster.
Can a browser prove a controller is 1000 Hz?
Not reliably on every platform. This page reports the update frequency visible through browser timestamps and intentionally avoids claiming hardware-lab accuracy. A native USB capture tool can observe a different layer of the input path.
Why must I move a stick during the test?
Some browsers or controllers update the exposed timestamp only when the gamepad state changes. Continuous stick movement produces a stream of changing reports and gives the estimator enough intervals to calculate a stable result.
Does a higher polling rate always feel faster?
No. Higher report frequency can reduce one source of waiting, but it does not guarantee lower end-to-end latency. Firmware, wireless transport, game processing, frame rate and display response may have a larger effect.