What the number means

USB HID devices don’t push input — the host asks, on a fixed schedule. The polling rate is how often: 125Hz means the controller is asked every 8ms, 250Hz every 4ms, 500Hz every 2ms, 1000Hz every 1ms. Your button press lands wherever it lands inside that interval; the machine hears about it on the next poll. Worst-case staleness is one full interval, average is half.

That’s the entire mechanism. Double the rate, halve the worst-case freshness gap — nothing more exotic than that.

Stock rates, roughly

• Xbox One / Series pads, wired: ~125Hz (8ms interval) — the most conservative of the majors.
• DualShock 4: ~250Hz wired; community measurements have long shown it behaving faster over Bluetooth on PC — one of the rare cases where wireless wins.
• DualSense: ~250Hz wired.
• Switch Pro / Joy-Cons: ~60–120Hz-class behavior in typical stacks; not the strong point of the platform.
• Third-party “pro” pads: 500–1000Hz wired is now a selling point; check spec sheets, because it varies.

Where polling sits in the lag chain

Controller latency is a pipeline, and polling is one hop: poll interval → USB/Bluetooth transport → OS HID stack → game engine sampling → frame render → display scanout. The 8ms worst case of a 125Hz pad is real — but a 60fps frame is already 16.7ms and displays add their own slice. This is why the honest ranking of interventions is: raise the frame budget first, wire the connection second, chase polling rate third.

Bluetooth deserves its own line: it isn’t just slower, it’s jittery. Retransmission and scheduling turn a nominal interval into a variable one — the average penalty is a few milliseconds but the spikes are what you feel.

Overclocking

The poll interval is set by the host driver, which is why Windows tools like hidusbf can override it — pads get pushed from 125/250Hz to 500/1000Hz and the staleness floor drops proportionally. Caveats: it’s unsigned-driver territory on a per-device basis, some pads glitch or refuse, and gains above ~500Hz shrink into display noise unless you’re also on a high-refresh setup. Verify results with a tool that counts real HID reports, not a web page.

What this site can and can’t see

The update-rate panel measures the layer a browser legitimately owns: how often the browser engine refreshes the gamepad snapshot (gamepad.timestamp changes), the animation-frame rate, and the gap between a press registering in the snapshot and the page detecting it. Those numbers describe the delivery of input, capped at ~16.7ms per sample on a 60Hz display.

What it cannot do is count USB polls — that happens on the wire, below the browser. If a tool on a web page shows you “1000Hz polling”, it’s showing you its own refresh rate. Anything you see here is labeled accordingly; that’s deliberate.

Frequently asked questions

Does controller polling rate affect input lag?

Yes, modestly. The poll interval is a floor under how fresh your input is when the machine reads it: 125Hz means up to 8ms of staleness (~4ms average), 1000Hz means up to 1ms. Real but small — the frame pipeline and display typically contribute more total lag than the difference between 250 and 1000Hz.

Can a web page measure my controller's polling rate?

No — and any page claiming to is measuring something else. A browser can only observe how often the browser engine refreshes the gamepad snapshot, sampled once per animation frame (~16.7ms on a 60Hz display). Our tester's update-rate panel reports exactly that, labeled honestly. True USB polling needs a desktop tool or hardware analyzer.

Is Bluetooth or wired lower latency for controllers?

Wired is consistently lower and steadier. Bluetooth adds a radio hop with retransmission jitter — typically several extra milliseconds and occasional spikes. For latency-sensitive play, wire the pad or use a 2.4GHz dongle where the pad supports one.

Can I overclock my controller's polling rate?

On Windows, community drivers like hidusbf can raise a pad's USB poll rate — it's how many DS4s get pushed toward 1000Hz, and it measurably drops worst-case staleness. It's driver-level surgery: expect some pads to refuse it or behave oddly, and verify gains with a tool that reads actual HID reports.