What is a good click latency for a gaming mouse?

Under 1 ms at 1000 Hz is good, and most current gaming mice get there. TechPowerUp measured click latency on 16 mice with NVIDIA's LDAT: wired results ran from 0.11 to 1.5 ms with a median of 0.51 ms, and wireless at 1000 Hz from 0.48 to 2.6 ms with a median of 0.97 ms. The two fastest wired mice use mechanical switches, and the slowest wireless mouse uses an optical one.

Measured click latency, 16 mice

TechPowerUp defines click latency as debounce plus "any potential processing delay", and measures it from "the electrical activation of the left main button" to "the OS receiving the button-down message". Every figure below is at the lowest-latency settings the mouse allows, which on mice with a debounce slider means 0 ms, and without sensor motion.

The sixteenth mouse, the Logitech Pro X2 Superstrike, has no switch at all, so LDAT cannot find a clean edge to time; it is covered under physical latency below. Wired testing "proved to be impossible" on the Model D3 and "was not possible" on the Burst II Pro, per TechPowerUp.

Optical is not automatically faster

Two claims circulate. One says optical switches are faster because light does not bounce, so they need no debounce. The other says a mechanical mouse can match them. At the electrical level, with debounce at 0 ms, the second is what the measurements show: the OP1 8K V2 (0.11 ms) and HTX Ultra (0.14 ms) are the fastest wired results, and the Vaxee NP-01S V3 lands at 0.54 ms, level with the Viper V3 Pro and Superlight 2. Meanwhile three optical mice, the X2N Crazylight, SCYROX V8 and Dash Ultra, sit at 1.4 to 1.5 ms wired.

The first claim persists because it is true of what each switch can safely live with. A mechanical switch needs debounce to stop contact bounce being read as extra clicks, more of it as the contacts wear; an optical switch can run at zero all its life. Our switch and debounce guide covers that trade. What the table shows is that, once debounce is out of the way, the firmware and the polling chain decide the number, not the switch.

On mice with a slider, the 0 ms figure is the floor. TechPowerUp reports that latency scales linearly with the debounce setting on the Dash Ultra, X2N Crazylight, SCYROX V8 and NP-01S V3, and "largely linearly" on the Model D3 and VT2 Max, so any debounce you set sits on top of the table's number. The Maya X is the exception: its debounce "appears to not scale linearly".

Polling rate cuts it on most mice

A click has to wait for the next report like everything else, so a higher polling rate usually lowers click latency. The Burst II Pro goes from 2.6 ms at 1000 Hz to 0.6 ms at 8000 Hz, a 2.0 ms saving, and the X2N Crazylight from 1.4 to 0.40 ms. The wireless median falls from 0.97 ms at 1000 Hz to 0.41 ms at 8000 Hz.

Three mice break the pattern. The OP1 8K V2 and Maya X show no change: "Polling rate has no effect on click latency." On the VT2 Max, TechPowerUp found that only the separate "key scan rate" (button report rate) setting changes it, so raising the report rate alone does nothing for clicks.

Wireless costs a little at 1000 Hz: the Superlight 2 goes from 0.54 ms wired to 1.4 ms wireless, the NP-01S V3 from 0.54 to 1.1. Both gaps are under a millisecond, which fits what our wireless vs wired page says about the radio hop.

The switch travel is the bigger number

LDAT starts the clock when the switch fires electrically. Everything before that, the plunger and leaf travelling until contact, is physical latency, and it is larger. TechPowerUp measured it with a separate rig for its Superstrike review:

On the Superlight 2 the physical part is 89 percent of the total (derived). Switch modes that attack the physical part beat any electrical difference in the main table: Kailh GX Speed mode on the OP1 8K V2 saves 1.4 ms end to end, and SPDT mode on the HTX Ultra 1.76 ms. GX Speed works by registering the click when the top contact opens, instead of waiting for the bottom one to close.

The Superstrike has no switch to travel, which is why it wins this table despite a slower electrical figure. Its ten actuation points scale from 1.8 ms at AP1 to an estimated 9.8 ms at AP10, and TechPowerUp warns that its rig's results depend on the force and speed of the press, so the physical column is a comparison, not a figure you will reproduce with a finger.

Does any of this matter?

Put the electrical spread in a click-to-photon budget. At 240 Hz one frame is 4.17 ms (derived). The gap between the fastest and slowest wired mouse here is 1.39 ms, a third of a frame. Switching the Burst II Pro from 1000 to 8000 Hz saves 2.0 ms, half a frame. An SPDT speed mode saves about as much again. Those are the settings worth changing; debounce at 0 is only worth it if the switch does not start double-clicking.

Where this stops applying

Each figure is one review sample on the firmware TechPowerUp tested, without sensor motion, at the lowest-latency settings. Any debounce above 0 ms adds to these figures, linearly on most of the mice with a slider. The physical figures come from a pen-on-foil rig rather than a finger, and TechPowerUp notes they can be "significantly lower or higher" with different force and speed.

Frequently asked questions

What is a good click latency for a gaming mouse?

Under 1 ms at 1000 Hz, measured from the switch firing to Windows receiving the click. Across 16 current mice tested by TechPowerUp at their lowest settings, the wired median was 0.51 ms and the wireless median at 1000 Hz was 0.97 ms. The slowest was the Turtle Beach Burst II Pro at 2.6 ms wireless at 1000 Hz.

Are optical switches faster than mechanical switches?

Not at the electrical level once debounce is set to 0 ms. The fastest wired results were the Endgame Gear OP1 8K V2 (0.11 ms, Kailh GX mechanical) and the G-Wolves HTX Ultra (0.14 ms, Huano mechanical). The slowest wireless result was an optical mouse at 2.6 ms. What optical switches change is whether 0 ms debounce is safe to leave on.

Does polling rate lower click latency?

On most mice, yes. The Turtle Beach Burst II Pro drops from 2.6 ms at 1000 Hz to 0.6 ms at 8000 Hz. On the OP1 8K V2 and the LAMZU Maya X polling rate had no effect, and on the Rapoo VT2 Max only a separate key scan rate setting changed it.

Is the switch or the button travel the bigger delay?

The travel. In TechPowerUp's end-to-end test the Logitech Pro X Superlight 2 measured 0.9 ms electrical and 7.0 ms physical. Kailh GX Speed mode on the OP1 8K V2 cut 1.4 ms and SPDT mode on the HTX Ultra 1.76 ms, both larger than any gap in electrical click latency between ordinary mice.